PAC report (Papermaking) March 2000

REPORTS STATUS to the PAPERMAKING PROJECT ADVISORY COMMITTEE March 8 - 9,200O IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) INSTITUTE OF PAPER SCIENCE AND TECHNOLOGY Atlanta, Georgia ANNUAL PROGRAM REVIEW PAPERMAKING March 8-9, 2000 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) .February lo,2000 TO: MEMBERS OF THE PAPERMAKING PROJECT ADVISORY COMMITTEE Attached for your review are the Status Reports for the projects to be discussed at the Paper-making Project Advisory Committee meeting being held at the Institute of Paper Science and Technology. The Program Review is scheduled for Wednesday, March 8, 2000, from 8:00 a.m. to 6:00 p.m. and the PAC Committee Meeting will meet on Thursday, March 9, from 8:00 a.m. to 12:30 p.m. We look forward to seeing you at this time. Sincerely, Frederick W. Ahrens, Ph.D. Professor of Engineering & Unit Leader Water Removal Research FWA/map Attachments IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Institute of Paper Science and Technology, Inc. 500 10th Street, N.W. Atlanta, GA 303186794 Telephone 404-894-5700 FAX# 404294~IPST (4778) PAPERMAKING PROJECT ADVISORY COMMITTEE . IPST Liaison: RAC Liaison: Chairman: Dr. David Orloff (404) 894-6649; FAX (404) 894-1496 Mr. John Bergin (715) 422-2239; FAX (715) 422-2227 Dr. Jay Shands (608) 364-8501; FAX (608) 364-8600 Dr. Dwight E. Anderson *(I200 1) Vice Chairman Manufacturing Manager Weyerhaeuser Company WTC 2F39 Post Office Box 2999 Tacoma, WA 98477-2999 (253) 924-6466 (253) 9246541 FAX [email protected] Dr. Slava Babinsky *(200 1) Fellow Mead Central Research Eighth & Hickory Streets Post Office Box 1700 Chillicothe, OH 4560 I-5700 (740) 772-3056 (740) 772-3595 FAX [email protected] Mr. John Bergin *@AC Liaison) Director of Research and Development Consolidated Papers, Inc. Post Office Box 8050 300 N. Biron Drive Wisconsin Rapids, WI 54495-8050 (7 15) 422-2239 (7 15) 422-2227 FAX [email protected] Dr. G. Ronald Brown *(2001) Director of Research Westvaco Corporation 11101 Johns Hopkins Road Laurel, MD 20723-6006 (301) 497-1301 (301) 497-1275 FAX [email protected] Mr. Jack Burke *( 1999) Principle Project Manager Radian International LLC 1979 Lakeside Parkway Suite 800 Tucker, GA 30084 (770) 4 14-4522 (770) 414-4919 FAX Dr. Partha S. Chaudhuri *( 1999) Senior Scientist, Papermaking Champion International Corporation Technical Center West Nyack Road West Nyack, NY 10994 (914) 578-7123 (914) 578-7474 FAX [email protected] Mr. Frank Cur-mane *( 1999) Vice President, Technology AstenJohnson 4399 Corporate Road Post Oftice Box 11800 1 Charleston, SC 29423 (843) 747-7800 (843) 747-3856 FAX fi-ank.cunnaneQasten.com Mr. Richard Daniel1 *(Alternate) R&D Laboratory Manager EHV Weidmann One Gordon Mills Way St. Johnsbury, VT 058 19 (802) 75 l-3328 (802) 751-3373 FAX [email protected] Dr. Christopher P. Devlin *(2001) Process Engineering Manager Inland Eastex A Temple-Inland Company Post Office Box 816 Silsbee, TX 77656 (409) 276-3 190 (409) 276-3419 FAX [email protected] Mr. R. Doug Estridge *(Alternate) Staff Engineer BE&K Engineering Company 2000 International Park Drive Post Office Box 12607 Birmingham, AL 35202-6607 (205) 972-64 16 (205) 9726300 FAX [email protected] Mr. Thomas M. Hailer *(2001) Paper Applications Supervisor Specialty Minerals Inc. 9 Highland Avenue Bethlehem, PA 180 17 (6 10) 882-8756 (610) 861-3412 FAX Mr. Ed R. Hendrickson *(2001) Research Engineer Potlatch Corporation Post Office Box 503 Cloquet, MN 55720-0503 (2 18) 879-0626 (2 18) 879-2375 FAX [email protected] * The dates in ( ) indicate the final year of the appointment. Revised 12/08/99:at PapermakingPAC (cont.) Mr. Kenneth Kaufman *(200 1) Senior Research Manager Kimberly-Clark Corporation 1400 Holcomb Bridge Road Building 40012 Roswell, GA 30076 (770) 587-7493 (770) 587-7709 FAX [email protected] Mr. Markku Korpela *(Alternate) Kymni Paper R&D UPM-Kymmene Fine Paper FIN-45700 Kuusankoski, FINLAND +358-204-15-3541 +358-204-l 5-2552 [email protected] Dr. Alexander A. Koukoulas “(2002) Manager, Papermaking Process Research 1ntemational Paper Company Corporate Research Center 1422 Long Meadow Road Tuxedo Park, NY 10987 (9 14) 577-7275 (9 14) 577-7507 FAX [email protected] Dr. Charles Kramer “(2001) Director Albany International Research Company Post Office Box 9114 Mansfield, MA 02048-9114 (508) 337-9541 (508) 339-4996 FAX [email protected] Mr. David J. Lacz *(2OOl) Technical Associate Eastman Kodak Company Paper Support Division B-3 19 1669 Lake Avenue Rochester, NY 14652-3622 (716) 477-6301 (7 16) 588-2680 FAX [email protected] Dr. Jeffrey D. Lindsay *(Alternate) Associate Research Fellow Kimberly-Clark Corporation 2 100 Winchester Road Post Office Box 999 Neenah, WI 54957-0999 (920) 72 l-3990 (920) 721-7748 FAX Mr. Greg Maule “(2001) Production Manager Consol idated Papers, Inc. 1101 Main Street Niagara, W 1 54 15 1 (7 15) 25 l-8326 (7 15) 25 l-1540 FAX [email protected] Mr. Vie Nisita *(200 1) Operations Superintendent Wisconsin Tissue Mills Chicago Operations 13101 S. Polaski Road Alsip, IL 60658 (708) 824-4397 (708) 389-4901 FAX [email protected] Dr. Franc0 Palumbo *( 1997) Riverwood International Corporation Post Ofice Box 35800 West Monroe, LA 7 1294-5800 (3 18) 362-2000 (3 18) 362-2441 FAX Dr. Paul R. Proxmire “(2001) Research Associate Appleton Papers Inc. Post Office Box 359 Appleton, WI 549 12-0359 (920) 730-7254 (920) 991-7243 FAX [email protected] Mr. Jeffrey R. Reese *(2001) Consultant, Paper Mill Georgia-Pacific Corporation 133 Peachtree Street, NE 18th Floor Atlanta, GA 30303 (404) 652-4880 (404) 584-1466 FAX [email protected] Mr. Thomas E. Rodencal *(Alternate) Sr. Paper Mill Staff Engineer Georgia-Pacific Corporation 133 Peachtree Street, NE 18th Floor Atlanta, GA 30303 (404) 652-45 14 (404) 584-1466 FAX [email protected] * The dates in ( ) indicate the final year of the appointment. Revised 12/08/99:at Papermaking PAC (cont.) Mr. Nickey J. Rudd *(2002) Vice President, Special Projects BE&K Engineering Company Post Office Box 12607 Birmingham, AL 35202-2607 (205)969-3600 (205) 972-6300 FAX Dr. Jay A. Shands “(2001) (Chairman) Manager, Forming Systems Beloit Corporation Rockton Research Center 1165 Prairie Hill Road Rockton, IL 6 1072- 1595 (608) 364-850 1 (608) 364-8600 FAX [email protected] Mr. Ralf Sieberth *(200 1) Sr. Vice President, Market & Business Development Voith Sulzer Paper Technology North America inc. Post OffIce Box 2337 2200 North Roemer Road Appleton, WI 549 13-2337 (920) 73 l-0769 (920) 731-1391 FAX [email protected] Mr. Frank J. Sutman *(2002) Sr. Research Scientist Hercules Incorporated Pulp and Paper Division 75 10 Baymeadows Way Jacksonville, FL 32256 (904) 733-7110 (904) 448-4995 FAX frank.j.sutmanQbetzdearborn.com Mr. David G. Thurman *(Alternate) Project Leader, Board Eka Chemicals Inc. 1775 West Oak Commons Court Marietta, GA 30062-2254 (770) 321-4138 (770) 321-5880 FAX Mr. Markku Tuderman *(2002) Vice President, Research and Development UPM-Kymmene Eteltiesplanadi 2 P.O. Box 380 FIN-00 10 1 Helsinki, FINLAND Mr. James R. Watson *(2001) Segment Manager - Printing & Writing Grades Eka Chemicals Inc. 1775 West Oak Commons Court Marietta, GA 30062 (770) 321-4142 (770) 32 l-645 1 FAX 358-204 15 111 358-2041 50 507 FAX [email protected] Mr. Lloyd 0. Westling *( 1999) Vice President, Production Planning Longview Fibre Company Post OffIce Box 639 Longview, WA 98632 (360) 575-5259 (360) 575-5926 FAX [email protected] Dr. Gary L. Worry *(200 1) Research Fellow Fort James Corporation 19 15 Marathon Avenue Post Office Box 899 Neenah, WI 54957-0899 (920) 729-8470 (920) 729-8023 FAX [email protected] * The dates in ( ) indicate the final year of the appointment. Revised 12/08/99:at PROJECT PAPERMAKING ADVISORY COMMITTEE MEETING March 8-9, 2000 Institute of Paper Science and Technology Atlanta, Georgia Location: .March Se m in a r R o o m 8,.2&)0 - COMNflTTEE’,DlSCU~SlONS AGENDA .: ,’ 7:30 a.m. Coffee/Danish 8:00 a.m. - 8:lO a.m. Opening Remarks, Review of Antitrust Statement and Confidentiality Statement Chairman 8:lO a.m. - 8:40 a.m. Portfolio Management System and Project Scoring Gary Baum, David Orloff 8:40 a.m. - 8:45 a.m. Welcome Fred Ahrens, Cyrus Aidun 8:45 a.m. - 9:30 a.m. Project F048 Approach Flow Systems Ted Heindel 9:30 a.m. - IO:15 a.m. Project F022 Flow Through Porous Media Seppo Karrila IO:15 a.m. -IO:30 a.m. Break IO:30 a.m. - 1I:00 a.m. Project FO03 Fluid Dynamics of Suspensions Cyrus Aidun 1I:00 a.m. - 12:00 a.m. Project FO05 Headbox and Fluid Hydrodynamics Cyrus Aidun 12:00 p.m. - 12:45 p.m. Lunch and IPST Update Gary Baum 12:45 p.m. - I:30 p.m. Project F039 Water Removal Limits Tim Patterson I:30 p.m. - 2:15 p.m. Project FO40 Press Dwell Time Limits Paul Phelan 2:15 p.m. - 3:00 p.m. Project F04 1 High Intensity Water Removal Fred Ahrens 3:00 p.m. - 3:15 p.m. Break 3:15 p.m. -4:15 p.m. Project F021 & 4253 4:15 p.m. - 6:00 p.m. Subcommittee Discussions of Projects & Preparation of Summaries Drying Productivity Fred Ahrens Tim Patterson NOTE: IO minutes of project discussioti time is included at the end of each presentation. + Dinner provided at 6:00 p.m. PAPERMAKING PROJECT ADVISORY COMMITTEE MEETING March 8-9, 2000 Institute of Paper Science and Technology Atlanta, Georgia Location: R o o m 1 7 3 nhh 9; 2000 .-C~MMI?TEE’D~SCUSS~ONS AGENDA ” 7:30 a.m. Coffee/Danish 8:00 - 8:10 a.m. Convene - Antitrust Statement - Confidentiality Statement - New Members - Acceptance of Fall, 1999 minutes - Review of Agenda Chairman 8:lO - 8:30 a.m. F022 Flow Through Porous Media (Karrila) Chaudhuri, Shands, Sieberth F048 Approach Flow Systems (Heindel) Hendrickson, Westling, Bergin, Rudd 850 - 930 a.m. F021 Drying (Ahrens/Patterson) Reese, Worry, Beck 9:10 - 9:30 a.m. FO05 Headbox and Forming (Aidun) Anderson, Devlin, Koukoulas 9:30 - 950 a.m. F041 High Intensity Water Removal (Ahrens) Babinsky, Watson, Kaufman 950 - 10: 10 a.m. FO03 Fluid Dynamics of Suspensions (Aidun) Brown, Maule, Proxmire 8:30 - 850 lO:lO-IO:30 a.m. a.m. Break IO:30 - IO:50 a.m. F039 Water Removal (Patterson) Cunnane, Lacz, Haller IO:50 - II:10 a.m. FO40 Press Dwell (Phelan) Kramer, Palumbo ll:lO-II:30 a.m. Proposed Projects IPST PI’s 11:30-12:15 p.m. PAC Discussion of Potential Projects All Final Discussion/Adjourn All 12:15- 12:30 p.m. NOTE: Subcommittee chairs are indicated in underlined bold characters above, + Lunch provided. TABLE OF CONTENTS Page 1 Project F048 Approach Flow Systems ..................................................................... Project F022 Flow Thru Porous Media ................................................................... 43 Project FO03 Fluid Dynamics of Suspensions ........................................................ 61 Project FO05 Fundamentals of Head box and Forming Hydrodynamics ............... .73 Project F039 Water Removal Limits ....................................................................... 85 Project FO40 Press Dwell Time Limits .................................................................. 127 Project F041 High Intensity Water Removal ........................................................ 139 Project F02 1 Drying Productivity . .... ... ... ... ... ... .... ... ... .... .. ... .... ... .. ... .... .. ... ... ... ... ... .. 159 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) APPROACH FLOW SYSTEMS STATUS REPORT FOR PROJECT F048 Ted Heindel (PI) Aklilu Giorges March 8 - 9,200O Institute of Paper Science and Technology 500 10th Street, N.W. Atlanta, Georgia 30318 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 3 F048 DUES-FUNDED Project Project PAC: Title: Number: Project Staff Principal Research PROJECT SUMMARY Approach Flow Systems F048 Papermaking Investigator: Support Staff: Ted Heindel Aklilu Giorges FY 99-00 Budget: $75,007 Time Allocation: Principal Research Investigator: Support Staff: I LINE/ROADMAP: RESEARCH Status Report January 2000 20% 50% LINE #II Improve the ratio of product performance to cost for pulp and paper products by 25% by developing break-through papermaking and coating processes which can produce the innovative webs with greater uniformity than that achieved with current processes. 2 PROJECT OBJECTIVE: The objective of this project is to provide recommendations to improve the spatial and temporal consistency and chemical uniformity of the stock leaving the approach flow area. 3 PROJECT BACKGROUND: The PI in this project began the project on July 1, 1999. The initial focus was to review the work completed in a related project conducted by Dr. Xiaodong Wang (Project FO04). Appendix A of this status report contains brief summaries of the various reports completed by Dr. Wang. Familiarization of other literature related to approach flow systems and pipe mixing was also completed. From this background, it was decided to focus this research on improving the spatial and temporal consistency and chemical uniformity in the approach flow area. The ultimate goal of this project is to be able to provide paper producers and suppliers with recommendations on approach flow piping configurations. To accomplish the goal of providing approach flow recommendations to IPST Member Companies, this project will focus on developing and validating a model of the fiber mixing process. Once a model is validated, it can be used throughout the pulp and paper process where one constituent is mixed with another. The model will also provide a tool to paper producers to allow them to design approach flow configurations to optimize mixing performance given economic and space limitations. This effort will follow two parallel research paths, schematically represented in Fig. 1. The first path stresses the development of a new fiber mixing model, and the second addresses experiments to IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) -.. Project 4 F048 study the mixing process and mixing model for a concentric system were completed. This stages of a water/fiber model Section 10. 4 Status Report January 2000 to validate the model. Since July 1999, a water/water pipe mixer was developed and experiments in a similar work will be detailed in Section 10. Additionally, the initial and the accompanying experiments will be outlined in MILESTONES: This project was initiated by this PI on July I, 1999. In the Fall PAC meeting it was agreed that by June 2000, high speed video analysis of the concentric mixing process with velocity ratios of up to R, = 5 will be completed for a water/water system and a water/l% fiber system. It was further agreed that by June 2000, a model of turbulent fiber mixing in a concentric mixer will be developed. As of January 2000, water/water mixing experiments and numerical predictions have been completed for velocity ratios from I to 6, and a model of the fiber mixing process has been outlined. Details of this work are provided in Section 10. 5 DELIVERABLES: The overall deliverable concentric mixing process. 6 STATUS OF GOALS of this project is a validated model of the water/fiber FOR FY 99-00: The status of the goals for the current fiscal year, as of January summarized as follows: 2000, are 1. Review previous FO04 work. Status: Completed 2. Become familiar with the approach flow and mixing literature. Status: Completed 3. Modify the current experimental facility. Status: Completed 4. Review CFD software options and purchase a new CFD software Status: Completed 5. Water/water concentric mixing: a. Experiments Status: Completed b. Computations Status: Completed c. Comparisons between predictions and experiments Status: In progress 6. Water/l% hardwood fiber concentric mixing: a. Model development Status: In progress b. Experiments Status: Not yet begun c. Computations Status: Not yet begun d. Comparison between predictions and experiments Status: Not yet begun IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) license. Project 7. 7 F048 Status Report January 2000 Member Company Report Status: Not yet begun SCHEDULE: The project schedule for the current fiscal year is summarized in Table numbers correspond to the goal numbers listed in Section 6 (above). Table 1: Project Task Descriptions 1 8 SUMMARY F048 schedule 1999 July - Sept 0-I X 1. The task for FY99-00. 1999 Ott - Dee 2000 Jan - Mar 2000 Apr - Jun OF RESULTS: This project began with the current PI on July 1, 1999. Since that time, the current concentric mixing experimental facility has been modified to achieve velocity ratios of up to 6 without cavitation occurring at the trailing edge of the inner pipe. Using this facility, water/water mixing experiments have been completed. By increasing the velocity ratio from 1 to 6, the mixing distance was reduced by approximately 65%. Additionally, qualitative observations revealed that by increasing the velocity ratio, the inner jet spreading angle was found to be larger (i.e., the jet spread faster) and the downstream mixing region was more uniform. New CFD software was also acquired for this project and turbulent water/water concentric mixing was modeled with the standard k-E and realizable k-E turbulence models. Numerical studies have shown that the closure constants in the turbulence models can influence the predicted mixing effects. Constant values of CE = 1.7 and CZE = 2.4 for the standard k-E model and C2 = 2.2 for the realizable k-c model have been used to obtain reasonable qualitative agreement with the experimental data. It was further shown that qualitative differences between the two k-E models were very small. We are in the process of extending the mixing experiments to a water/fiber system, as well as developing a new model of water/fiber mixing. In the model, we propose to model the fiber mixing process by using a variable local shear stress (i.e., viscosity) which is a function of local fiber consistency and shear rate. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 9 BULLETED l l l l 6 F048 Status Report January 2000 KEY ACCOMPLISHMENTS: Water/water mixing experiments have been completed velocity ratios between 1 and 6. in a concentric mixer for Turbulence model closure constants affect the predicted mixing results using the experimental results, the closure constants were determined. Reasonable predictions Thus, qualitative agreement between experiments and numerical have been obtained with water/water concentric mixing. A new method to model water/fiber mixing has been proposed. 10 DISCUSSION: Details from work completed since July 1, 1999 are summarized below. addresses the experimental work. Section 10.2 and 10.3 focus on modeling mixing and water/fiber mixing, respectively, in a concentric pipe mixer. 10.1 Concentric Mixing Section 10.1 water/water Experiments Concentric mixing is a simple and effective method to mix one constituent with another. It is a common process in many industries, including the pulp and paper industry. A schematic of a concentric mixin’g process is shown in Fig. 2. This process involves mixing a fluid from an inner pipe with diameter d, volumetric flow rate q, and mean fluid velocity u, with a fluid in an outer pipe of diameter D, volumetric flow rate Q, and mean fluid velocity v. Typically, u > v and the inner pipe fluid is referred to as the “primary fluid”, while the outer pipe fluid is called the “secondary fluid”. In many cases, the primary fluid has a specified species concentration C, while the secondary fluid has a species concentration C,. In this case, the purpose of the mixing operation is to provide a uniform species concentration downstream of the primary fluid inlet. As shown in Fig. 2, the jet issuing from the center pipe may be divided into two regions, the potential core region and the entrainment or mixing region. The characteristics of the potential core are identical to those of the primary fluid stream (e.g., u, C,), while the characteristics of the mixing region vary from those of the primary fluid to those of the secondary fluid. This mixing process appears to be simple, but very complex flow phenomena occur to thoroughly mix the two fluid streams. When the two fluid streams enter the mixing region at different velocities, a high shearing region forms at the interface between the two fluid streams. Instabilities at this interface cause vortices to intertwine from each stream, enabling mixing. Depending on the mixing behavior, the two fluids may, or may not, mix completely or uniformly. The degree of mixing in a concentric mixer depends on the following [I]: the ratio of inner-to-outer pipe diameter; the ratio of inner-to-outer pipe flow rates or velocities; the ratio of specific gravities between the two fluid streams; the inner and outer pipe Reynolds numbers; the pipe surface roughness; and any secondary pipe flows. When one of the constituents is a fiber suspension, additional parameters related to the fiber characteristics (e.g., fiber length, coarseness, flexibility, etc.) also affect the mixing process [2-51. One of the important parameters in the mixing process = the primary and secondary fluid, R, u/v. In this research, is the velocity ratio between initial experiments were IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 7 F048 Status Report January 2000 conducted in a water/water system in which 1 < R, < 6 to determine the effect of R, on mixing. In addition, these data will then be used in the model validation phase of this project. 10.1.1 Experimental Equipment A schematic of the former mixing system used by Dr. Wang is shown in Fig. 3. This system was composed of two transparent concentric pipes in the test section, a large mixing tank to hold the primary fluid,~a smaller water tank to hold the secondary fluid, a large discharge tank, a pump, and the associated piping, valves, and flow meters. This system was designed to provide a constant head input into the test section and eliminate any pressure pulsations that may result from the system pump. With this setup, however, a maximum velocity ratio of R, = 1.4 was realized before cavitation occurred at the trailing edge of the inner pipe. This mean velocity ratio is much smaller that those in industry, as well as those recommended by others [5, 61. To increase the possible velocity ratios, the flow loop was modified for this research program to allow the primary fluid to enter the test section directly from the pump (Fig. 4). This increased the maximum possible velocity ratio, but also introduced potential pressure pulsations from the pump, which were assumed to be acceptable if they were present. With this current configuration, a maximum water/water velocity ratio of R, = 6 was achieved. The test section, pictorially represented in Fig. 5, consisted of a transparent inner pipe with inside diameter d = 2.54 cm (1 in) and a pipe wall thickness of 0.3175 cm (0.125 in). The outer pipe was also transparent and had an inside pipe diameter of D = 6.35 cm (2.5 in). As shown in Fig. 5, the inner pipe protruded into the outer pipe approximately 4’ = 39.4 cm (15.5 in) after the 90° bend. The outer pipe extended approximately L = 58 cm (22.8 in) beyond the inner pipe trailing edge before exiting into the discharge tank. Although L = 58 cm, the actual mixing region captured by high speed video was approximately 35 cm (13.8 in) downstream of the inner pipe trailing edge, corresponding to a mixing region of approximately 5.5D. The high speed video equipment used to capture the mixing process Kodak Ektapro Motion Analyzer with a frame rate of 1000 frames/set. 10.1.2 Water/Water Mixing consisted of a Experiments Initial concentric mixing experiments were completed in a water/water system. To identify the primary fluid, red dye (rhodamine) was added to city water in the large mixing tank. This colored water was pumped through the inner pipe, while the outer pipe was supplied with clear city water from the elevated secondary fluid tank. High speed video images at 1000 frames/set were then taken at the test section where the two fluid streams were allowed to interact. These images encompassed a mixing region of approximately 5.5D. During the experiments, the mean secondary fluid velocity was held constant at approximately v = 1.30 m/s, while the mean primary fluid velocity was varied between 1.35 m/s 5 u 5 7.76 m/s. This corresponded to a mean velocity ratio range of 1 2 R, < 6. Table 2 summarizes the actual flow conditions for each experiment. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project Table 2: I 8 F048 Flow conditions used in the water/water Status Report January 2000 concentric mixing experiments. Primary Volumetric Flow Rate, 9 @Pm) Secondary Volumetric Flow Rate, Q @pm) Mean Primary Velocity, u (m/s) Mean Secondary Velocity, v (m/s) Mean Velocity Ratio, R, = u/v 10.8 50.8 1.35 1.30 1.04 20.7 50.0 2.58 1.28 2.02 30.6 50.2 3.81 1.29 2.95 40.4 51.3 62.3 I 50.6 50.3 50.2 I 5.03 6.39 7.76 I 1.30 1.29 3.87 4.95 1.29 6.02 7 I Figure 6 illustrates the flow structure of concentric water jets mixing when R, = 1.04. The outer pipe boundary is clearly identifiable, and the tip of the inner pipe is captured on the left-hand side of the image. Each dark “+” mark on the outside of the outer pipe represents a distance of 1 D. The darker fluid is the colored primary fluid. The mixing process can be visually observed from the dispersion of the dye that was introduced in the center jet. As expected, the mixing region increases in the radial direction as the fluids evolve downstream. The actual mixing process can be attributed to flow and geometric factors that promote the interaction between the two fluid streams. A close visual inspection directly downstream of the trailing edge of the inner pipe in Fig. 6 points to a radial increase in the jet, even though the inner and outer jet mean velocities are nearly identical. This agrees with the flow character reported by Dahm et al. [7]. In their case, they concluded that the boundary layer on both sides of the inner pipe introduced a wake. This resulted in a velocity defect and caused the two fluid streams to intertwine at their interface, creating a vortex ring. In the current experimental geometry, boundary layers are present on both sides of the inner pipe, creating a velocity defect. Additionally, the inner pipe has a finite thickness, which results in wake formation at the pipe trailing edge. This wake also contributes to the interaction between the two fluid streams. Finally, through inspection of multiple images of the R, = 1.04 mixing process, the location at which the colored fluid stream contacts the outer pipe wall (L, in Fig. 2) can be estimated. This will be further discussed below. A representative image from the R, = 2.02 experiment is shown in Fig. 7. The location at which the colored fluid contacts the outer pipe wall has moved upstream compared to that at R, = 1.04. For the relatively small velocity ratios depicted in Figs. 6 and 7, a large weave-like coherent structure is observed along the interface between the two fluids. This structure becomes unstable as it moves downstream, indicating that hydrodynamic instabilities, as well as small scale turbulent interactions, enhance the concentric pipe mixing process. Figures 8 through 11 display representative images of the mixing process for R,= 2.95, 3.87, 4.95, and 6.02, respectively. As the mean velocity ratio increases, the inner jet angle increases. This increase in R, also results in the decrease in the distance between the inner pipe exit and the location at which the colored fluid contacts the outer wall. The increase in jet spread also enhances the entrainment of the outer fluid and its mixing with the inner fluid. The inner jet and downstream mixing region also appear IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) I Project 9 F048 Status Report January 2000 darker and more uniform as the mean velocity ratio increases. This is the result of two factors: (i) the increase in center jet flow rate increases the dye concentration of the mixture causing the image to appear darker; and (ii) the large velocity difference between the inner and outer fluid enhance small scale turbulent mixing creating a more uniform mixture. We define the dark region downstream of the inner pipe as the mixing region, and the distance between the inner pipe exit and the location at which the darker region contacts the outer pipe walls (L, in Fig. 2) as the mixing distance. This distance has been estimated from Figs. 6-l 1, and has been nondimensionalized with respect to the outer pipe diameter, D. This nondimensional distance, Lw/D, has been plotted as a function of mean velocity ratio in Fig. 12. Increasing the velocity ratio from 1 to 6 reduces L,/D by approximately 65%. This decline can also be used to obtain qualitative comparisons between experimental observations and numerical predictions and will be completed in Section 10.25 Further analysis of multiple images from each test condition may also allow for qualitative identification of the local dye concentration through variations in the image intensity. This is currently under development and may be used for model comparisons if encouraging results are obtained. 10.1.3 Water/Fiber Mixing Plans The next phase of experiments will be completed in the same flow loop and test section, but with 1% fiber suspension as the primary fluid. These experiments are currently planned for mid-March (after the Spring PAC meeting). It is hoped that a laboratory repulper and dewatering screw press that will assist with experimental preparation and cleanup will be approved and in place at this time. This equipment has been requested through IPST capital funds and is currently under review by IPST management. If this equipment is not in place by mid-March, the 1% fiber mixing experiments will still be completed, they will just be more time-intensive. 10.2 Numerically Modeling Water/Water Mixing Water/water mixing was numerically modeled in an attempt to simulate the fluid flow in the water/water mixing experiments. Hence, the mixing process was simulated as two turbulent miscible fluids with the same density and viscosity, but with different concentrations of an inert tracer (e.g., a red dye). For this model, the governing equations and two turbulence models are summarized below. Selected numerical results and comparisons to experiments are also provided. 10.2.1 Governing Equations The governing equations for conservation of mass, momentum, and tracer concentration for steady, incompressible, turbulent viscous fluid flow with constant properties are dU -= i dx i fluid 0 i =-- dP P d PU .’ ax*I aXi + dxj dU IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) (2) Project 10 F048 Status Report January 2000 (3) All quantities have been time averaged in the above equations, and ui mean local velocity components in the axial and radial direction. Also, density, p is the time averaged pressure, p is the dynamic viscosity, h (turbulent) viscosity, c is the time averaged local tracer concentration, number, and oc is the turbulent Schmidt number (specified as ~~ = 0.7 calculations). (i = 1, 2) are the p is the fluid is the eddy SC is the Schmidt in our The governing equations are discretized and solved using Computational Fluid Dynamics (CFD) software. The particular software used in this study is called FLUENT, which uses a finite volume method to discretize the governing equations [8]. It was selected because it can be used to model the conservation equations of multiple fluid streams [9], which will be discussed in Section 10.3. The eddy viscosity (h) is specified through various turbulent models. Several models have been developed, but it is beyond the scope to this report to explain the advantages and disadvantages of each model. However, we present two turbulence models that are available in FLUENT v5 and were used to simulate the turbulent mixing process in this study. 10.2.2 The Standard k-E Model The standard k-E model [IO] is widely used due to its robustness, economy, and reasonable accuracy for a wide range of engineering of the model is that the eddy viscosity is defined by Pt =Pc,- computa problems. tiona The basis k* & (4) where CCL is an empirical constant and k and E are the turbulent kinetic energy and dissipation rates, respectively. These parameters are determined from the following transport equations dk UE u =PU .’ ax-1 ax*1 d d& Pt P+ -+ Cl, ;Gk % 1 dxj ii where C,, and C,, are empirical constants, numbers for k and E, respectively. The Gk term represents 0 the production E2 and ok and oE are the turbulent of turbulent kinetic energy PI where S is the modulus (6) -C,,pyy of the mean rate-of-strain tensor defined by IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Prandtl and is modeled by Project 11 F048 Status Report January 2000 (8) S=p& with the mean strain rate given by 1 S ij =2l dU i+J ax 1 dU . (9) ax i 1 In the standard k-E model, the following constant values are used as defaults FLUENT v5: Cp = 0.09, C,, = 1.44, C,, = 1.92, uk = 1 .O, and G& = 1.3. Comments applicability of these values will be given below. 10.2.3 The Realizable in on the k-E Model One variant of the standard k-E model is called the realizable k-8 model [I I]. This model differs from the standard k-E model in two ways [9]: (i) the realizable k-E model contains a new formulation of the eddy viscosity, k; and (ii) a new transport equation for the dissipation rate, E, is used. One class of problems where the realizable k-E model has been shown to be more accurate is the modeling of turbulent planar and round jets PI. For the realizable k-E model, the eddy viscosity given by Eq. (4) no longer has a constant value for CP. Rather, CCL is now a complex function of the mean strain rate (Eq. (9)). This formulation can be found in FLUENT [9]. The turbulent kinetic energy equation for the realizable k-E model is identical to that used in the standard k-E model (i.e., Eq. (5)). The dissipation equation is given by P”j~=~[(Pi~)~]+PCISE-Pc* where S is given by Eq. (8), v is the kinematic constant and ,:j, viscosity SklE C, = max 0.43, (Sk/&)+ [ The default o&= 12. . 10.2.4 values Numerical for the realizable (= p/p), C2 is an empirical 5 k-E model in FLUENT 1 (11) v5 are C2 = 1.9 , (T)(= 1 .O, and Results Numerical simulations of the water/water concentric mixing phenomena have been completed using both the standard and realizable k-E turbulence models available in FLUENT v5 [9]. The flow conditions were assumed to be axisymmetric to reduce the computational domain from three dimensions to two dimensions. The actual computation domain (Fig. 13) encompassed a radial distance of 3.175 cm and an axial distance of 44.45 cm, corresponding to a D/2 by 7D region. Note that the computations encompassed a 1 D length upstream of the trailing edge of the inner pipe and a 6D length downstream. This region was discretized into a numerical computational grid of 36 x 300 nodes, with a slightly higher node density near the inner pipe trailing edge. It is well known that turbulence enhances the mixing process. Thus, the turbulence model used to simulate mixing plays a major role in determining realistic predictions. It IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 12 F048 has also been shown that the values of the standard k-E model constants, affect the relative concentration of the mixing streams [12, 131. Status Report January 2000 CE and C,,, To identify the appropriate values for CE and CPE, selected numerical experiments were completed at the velocity ratio extremes of this study (R, = 1 .O or 6.0). Figure 14 shows the concentric mixer concentration profiles predicted with the standard k-E model when R, = 1 .O. Recall from Fig. 13 that each image shows the upper half of the concentric mixer, with the primary fluid entering from the lower left region and the secondary fluid entering from the upper left region. The isopleths in each figure represent lines of constant concentration. Therefore, at the trailing edge of the inner pipe wall, the closely packed lines of constant concentration represent a region of high concentration gradient. Downstream of the inner pipe, the concentration gradients are much less severe. The results with C,, = 1.44 and CZE = 1.92 (the FLUENT v5 default values) are shown in Fig. 14a. As CZE increases (Figs. 14b, c), the length of the potential core of the inner jet (the region from the inner pipe to the first isopleth) decreases and the jet spread increases. Figures 14d-f represent calculations that were completed to identify the best values for C,, and CZEl and were completed in conjunction with the calculations at R, = 6.0. Figure 15 shows similar numerical experiments using the standard k-E model and R, = 6.0. An analogous trend is observed when CZE is varied. By decreasing C,,, the length of the potential core of the inner jet increases and the jet spread decreases. After that C,, = 1.7 and several numerical experiments at R, = 1 .O and 6.0, it was determined C 2E = 2.4 provide reasonable qualitative agreement to the experimentally observed jet shape. These constant values were used for all standard k-E model calculations at the intermediate mean velocity ratios. The realizable k-E model was also tested to determine the effect of C2 variations on the mixing process. Figure 16b shows the predicted tracer concentration for R, = 1 .O when C2 is set to the FLUENT v5 default value of 1.9. Very little mixing between the inner and outer fluid streams is observed, even after six pipe diameters. By increasing C2, the mixing is enhanced as shown by the increased jet spreading and the shortened potential core of the inner jet (Figs. 16~f), while decreasing C, results in less mixing and a larger potential core (Fig. 16a). Figure 17 shows similar results for R, = 6.0. By comparing the experimental jet shape to the predictions in Figs. 16 and 17, a value of C, = 2.2 was selected to yield the best qualitative agreement between the experiments and predictions. This value was used for all realizable k-E model calculations of this study. 10.25 Flow Visualization Comparisons Calculations were completed using the standard k-E model and the realizable k-E model to predict the experimental jet shape at various R, values. For the standard k-E For the realizable k-E model, C,, = 1.7 and C,, - 2.4 were used for all calculations. model, C, = 2.2. Figures 18-23 illustrate the numerical predictions with a representative high speed video image from the various R, values for this study (Table 2). Each figure shows that there are only small qualitative variations between the two turbulence models. Further analysis is required to identify quantitative differences between the two numerical models, if any. In general, both models predict the qualitative shape of the mixing region between the colored and clear water streams. Figure 24 shows that the dimensionless mixing distance (L,/D) predicted by both turbulence models are similar, and they follow the experimental trend. However, exact lengths at which the inner fluid IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 13 F048 Status Report January 2000 stream contacts the outer pipe wall is a subjective measurement, and Fig. 24 shows only that the correct trend is predicted by each turbulence model. The information determined in the water/water mixing experiments and computational simulations will be useful when extending the mixing system to water and fiber. However, the modeling of this mixing process will be much more complicated. Initial ideas in this area are presented next. 10.3 Modeling Water/Fiber Mixing The mixing of a relatively thick fiber stream with a dilute fiber stream or white water is a very common process in the pulp and paper industry. This occurs in the approach flow area where thick stock is mixed with thin stock to dilute the fiber suspension to the proper headbox consistency. This also occurs when chemical additives are introduced into a pulp suspension moving through the mill. Modeling this mixing process is very complicated because the fiber suspension can behave as a Newtonian or a nonNewtonian fluid, depending on the fiber type, fiber consistency, flow conditions, and local shear rate. This section of the status report will summarize our progress in modeling this complex suspension flow. First, suspension flows without fibers will be briefly discussed. Then, fiber suspension flows will be addressed. Finally, a proposed new method to model the concentric pipe mixing process of two streams, one with fiber and one without, will be outlined. 10.3.1 Suspension Flows Without Fibers Suspensions of solid particles in fluids are commonly encountered in many engineering applications. For sufficiently dilute suspensions in a Newtonian fluid, the suspension behaves as a Newtonian fluid. As the solids loading increases, the viscous behavior of the suspension becomes non-Newtonian [14]. The point where this deviation occurs cannot be reliably predicted because it depends on the characteristics (i.e., size, shape, roughness, etc.) and concentration of the solid material, as well as the solid-solid and solid-fluid interactions [I 4, 151. The rheological behavior of suspensions (and for that matter, all liquids) can be described by the deformation or shear rate (y ) as a result of a shear stress (z>. Several models that describe this viscous behavior for various fluids and suspensions are available [14-l 81. The simplest case is for a Newtonian fluid where the shear stress is proportional to the shear rate. The proportionality constant is commonly called the dynamic viscosity, p. This relationship is described by z = pj (12) For suspensions of dilute spherical particles, the interactions between particles can be neglected and the fluid will behave as a Newtonian fluid. Citing Frisch and Simka [19], Stein [18] describes the viscosity of such a system by the Einstein equation I-L= i-9 0 + 2-w) where $ is the volume fraction of the dispersed solids and h is the viscosity of the suspending medium. Additional models have been proposed to describe the viscous behavior of suspensions. Three of which are described below, which are summarized from Darby [14]. A power law fluid can be described by IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) (13) Project F048 Status Report January 2000 where n is the flow index and m is the consistency coefficient. A Newtonian fluid is described by n = 1, a shear-thinning fluid when n < I, and a shear-thickening fluid when n > 1. Most fluids exhibit a power law region over a specified shear rate range, but the model fails at very high and/or low shear rates. When a suspension is at rest, interparticle contact resistance may develop. Fluid motion will develop only after an apparent yield stress (TV) is reached, It is common to describe these fluid types as Bingham fluids with where pGo is a limiting viscosity. Combining the power model, described by law model with a yield stress produces the Herschel-Bulkley z= zy +mj” This model is more flexible than the Bingham fluid model because adjustable parameters (xY, m, and n), but it still provides unrealistic shear rates. 10.3.2 Suspension (16) it has three behavior at high Flows With Fibers The flow of cellulose fiber suspensions is different from other solid-liquid suspensions and conventional Newtonian or non-Newtonian fluids. The fibers inherently entangle and form structures, even at low fiber concentrations. According to Stenuf and Unbehend [3], the fibers rotate freely in a dilute suspension of fibers and transfer momentum from regions of high velocity to regions of low velocity. This momentum transfer produces additional drag forces which are proportional to the shear rate. The resulting system viscosity increases with increasing fiber concentration and length until a critical concentration is reached. Above this concentration, floes form and may rotate freely. Further increases in fiber concentration produce a more continuous network which must yield or break down when flow begins. This produces a yield stress followed by a pseudo-plastic flow behavior. Because of this behavior, fiber suspensions have been described as Bingham fluids [17]. Previous investigators have studied the pipe flow behavior of fiber suspensions and it has been summarized in many articles [2, 3, 20-291. From this information, a theoretical model of the effective viscosity can be identified [22]. The theoretical shear stress behavior as a function of shear rate is shown in Fig. 25. Location A in Fig. 25 represents the yield stress that must be overcome to produce fluid movement. Between A and B, the flow rheology may be described as a Bingham fluid. From B to C to D, the fluid becomes shear-thinning and a drag reduction region is identified, with the maximum drag reduction occurring at C. At location D, the shear rate is high enough that the flow is highly turbulent and the fiber suspension may be approximated as a Newtonian fluid. The theoretical behavior is shown in Fig. 25, but the actual locations of A, B, C, and D are dependent upon many factors including the fiber type, fiber consistency, and pulping process. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 15 F048 Status Report January 2000 A successful model of the fiber suspension rheology must contain a majority of the characteristics found in Fig. 25. Wikstrom and Rasmuson [30] have numerically modeled a jet nozzle agitator in a large fiber suspension tank. A Bingham fluid model and a Hershel-Bulkley model of the fiber rheology was used in the numerical calculations. They concluded that there were no significant differences between the two rheology models because the dominating parameter was the yield stress. Furthermore, comparisons to experiments showed that the difference between the predicted and experimental velocities increased as the distance from the jet increased. Therefore, they concluded that a Bingham fluid model does not completely describe the fiber suspension rheology, and this was attributed to the shear-thinning behavior of the fiber suspension. We believe that at least four significant behaviors must be incorporated into any model to accurately describe the fiber suspension rheology. These behaviors include: (I) A nearly rigid region where the shear stress is below the yield stress. (II) A low shearing rate region where the fiber suspension rheology once the yield stress of the fiber suspension is overcome. (Ill) A region where the shear stress is independent of shear rate (i.e., line BD in Fig. 25). This will provide a conservative estimate of the fluid viscosity in the shear-thinning region. (IV) A high shear rate region where the suspension Newtonian fluid (i.e., turbulent water flow). behaves is nearly linear as a turbulent Region I and II can be successfully illustrated by modeling the suspension as a Bingham fluid. For this case, the yield stress is required for the fiber suspension. This value has been correlated for various fiber suspensions and has the form [28]: "Y = aCP (17) where C is the fiber mass concentration and a and p are empirical depend on the fiber type. Hence, in modeling the fiber suspension the local shear stress (zij) can be modeled as where peff is the effective suspension constants that mixing in this region, viscosity. In region III, the shear stress is described by the maximum shear stress obtained in region II. This relationship will be valid between locations B and D in Fig. 25, corresponding to critical shear rates jcrl and jcr2. At the second critical shear rate value, jcr2, the shear stress once again becomes a linear function of shear rate. In this region (region IV), the flow may be approximated as highly turbulent Newtonian flow, In summary, the shear stress for a fiber suspension Region I and II: 3j = a@ Region Ill: %j = CCC + peffj P + Peffh flow may be modeled r < hrl = zmax as Wa) Ycrl < ? < L2 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) (2W Project 16 F048 Region IV: zij ++I-& <r Ycr2 The challenging for &I and Ycr2 and consistency. Status Report January 2000 aspect of this simplified 9 which are likely to depend WC) theory is identifying the appropriate values upon various parameters such as fiber type Additional characteristics of fiber suspensions, such as floe formation, are not included in the above discussion, but may be added at a later date. For example, the crowding factor, defined as [4] N =- S&L2 03 (21) where C, is the fiber mass fraction in percent, L is the fiber length in m, and CQis the fiber coarseness in kg/m. Kerekes and Schell [4] have shown that the crowding factor is a measure of the level of fiber-fiber interaction, and hence, the tendency to flocculate. The effect of N on the local shear stress is currently unknown. However, Steen [31] has concluded that the crowding factor and pipe Reynolds number were key variables that determine if the pipe flow conditions are either plug, turbulent, or mixed. This will ultimately affect momentum transport in the fiber suspension and possibly the mixing of a fiber suspension with a dilute fluid. 10.3.3 Modeling Concentric Pipe Mixing of Fiber Streams In this study, we will consider the concentric pipe mixing process with the inner fluid consisting of a fiber suspension and the outer fluid corresponding to water (see Fig. 2). As previously discussed, the two fluids will respond differently to fluid shear. Water behaves a Newtonian fluid (Eq. (12)), while the fiber suspension is non-Newtonian (possibly described by Eq. (20)). As the two streams mix, the local viscous behavior of the suspension will change because it is a function of the local fiber consistency and the local shear rate, both of which vary throughout the mixing region. It is the goal of this research to identify the appropriate relationship that describes this variation. Consider the simplest concentric mixing flow conditions, laminar flow in an axisymmetric (two-dimensional) mixer. The conservation of mass and momentum steady, incompressible fluid with constant properties is dU -= i ax i dU (22) 0 i ax ij dp =-PU .’ ax-1 &i + ac (23) dxj where zij is given either by Eq. (12) for water or Eq. (20) for the fiber suspension. conservation of fiber concentration can also be written for this process; however, assume that fiber transport is governed by advection only and fiber diffusion is negligible. Therefore, PU .-= ‘ax for a 0 i where c is the fiber mass fraction. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) The we (24) Project 17 F048 Status Report January 2000 The solution to these equations is complicated becausethe fiber concentration equation (Eq. (24)) is coupled to the momentum equation through Zij, where Zij is a function of fiber consistency (Eq. (20)). The success of this model rests on the proper definition of zij as described in Section 10.3.2. It is hoped process. 10.4 Goals that this simple model can be extended to the turbulent concentric mixing for FY 00-01 Goals for the next fiscal year are based on a budget request of $115,000 for this project. The goals include experimental concentric mixing studies with the primary fluid composed of 1 and 2% fiber suspensions. Two different fiber types will be studied in these experiments. These experiments will assist with model validation. Modeling of the concentric mixing process will also be continued. The model will be used to complete parametric studies of various factors that affect the mixing process. These studies will allow us to make recommendations for concentric mixing modifications and optimization. The tasks for FYOO-01 are given below. The numbers continue from those is Section 6. 8. 9. 10. 1 I. 12. 13. 14. 10.5 Water/l% softwood fiber concentric mixing a. Experiments b. Computations c. Comparisons Water/2% hardwood fiber concentric mixing a. Experiments b. Computations c. Comparisons Water/2% softwood fiber concentric mixing a. Experiments b. Computations c. Comparisons March 2001 Status Report Member Company Report on the experimental work Numerical parametric studies Member Company Report on the parametric studies Deliverables for FY 00-01 Two Member Company Reports will be completed during the next fiscal year. The first will address the experimental work and comparisons to the model predictions. The second will focus on numerical parametric studies of the mixing process. 10.6 Schedule for FY 00-01 The proposed schedule for Project F048 for FYOO-01 is provided in Table 3. The last three months of FY99-00 are also included. The numbers in the task column correspond to those in Section 6 and 10.4. The marks during any given time quarter represent some work will be performed during that time period, but they do not imply a full-time commitment. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 18 F048 Table 3: Project Task Descriptions 16a F048 schedule for FYOO-01. 2000 Apr - June l 6b I Status Report January 2000 2000 Julv - Sept I 2000 Ott - Dee 2001 Jan - Mar -- 2001 Apr-Jun I I 1 IOa 1 IOb 1IOC II REFERENCES [II Forney, L.J., “Jet Injection for Optimum Pipeline Mixing,” Encyclopedia of Fluid Mechanics Volume 2 - Dynamics of Single-Fluid Flows and Mixing, N.P. Cheremisinoff, Ed., Gulf Publishing Company, Houston, 660-690 (1986). PI Duffy, G.G., and Lee, P.F.W., “Drag Reduction Suspensions,” APPITA Journal, 31(4): 280-286 PI Stenuf, T.J., and Unbehend, J.E., “Hydrodynamics of Fiber Suspensions,” Encyclopedia of Fluid Mechanics: Vol. 5 - Slurry Flow Technology, N. P. Cheremisinoff, Ed., Gulf Publishing Company, Houston, 291-308 (1986). PI Kerekes, R.J., and Schell, C.J., “Characterization of Fibre Flocculation Regimes a Crowding Factor,” Journal of Pulp and Paper Science, 18( 1): J32-J38 (1992). PI Helmer, Mixing,” PI Norman, Control,” VI Dahm, W.N., Frieler, C.E., and Tryggvason, G., “Vortex Structure and Dynamics in the Near Field of a Coaxial Jet,” Journal of Fluid Mechanics, 241: 371-402 (1992). PI Patankar, S.V., /Vumerica/ Corp., New York, 1980. PI Fluent Incorporated, in the Turbulent (1978). R.J.N., Covey, G.H., and Lai, L.C.-Y., “Laboratory APPITA Journal, 52(3): 197-201 (1999). B., and Tegengren, Paper Technology, A., “Stock Preparation 42-43 (January 1990). Heaf Transfer “FLUENT Flow of Wood Study of Co-axial by Stock - A Key to Grammage and Fluid Flow, Hemisphere 5 Users Guide,” Pulp Lebanon, NH, Fluent, IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Publishing Inc. 1998. Proiect J WI F048 19 Status Report January 2000 Jones, W.P., and Launder, B.E., “The Prediction of Laminarization with a TwoEquation Model of Turbulence,” international Journal of Heat and Mass Transfer, 15: 310-314 (1972). [111Shih, T.-H., Liou, W.W., Shabbir, A., and Zhu, J., “A New k-E Eddy Viscosity Model for High Reynolds Number Turbulent Flows: Model Development and Validation,” Computers in Fluids, 24(3): 227-238 (I 995). cw Monclova, L.A., and Forney, L.J., “Numerical Simulation of a Pipeline Tee Mixer,” Industrial and Engineering Chemistry Research, 34(4): 1488-I 493 (1995). WI Giorges, A.T., Forney, L.J., and Wang, X., “Numerical Study of Multi-Jet Industrial and Engineering Chemistry Research, In review. PI Darby, R., “Hydrodynamics of Slurries and Suspensions,” Encyclopedia of Fluid Mechanics.= Vol. 5 - Slurry Flow Technology, N.P. Cheremisinoff, Ed., Gulf Publishing Company, Houston, 49-91 (1986). WI Ganani, E., and Powell, R.L., “Suspensions of Rodlike Particles: Literature Review and Data Correlations,” Journal of Composite Materials, 19: 194-215 (1985). Mixing,” [I 61 Hanks, R.W., “Principles of Slurry Pipeline Hydraulics,” Encyclopedia of Fluid Mechanics: Vol. 5 - Slurry Flow Technology, N.P. Cheremisinoff, Ed., Gulf Publishing Company, Houston, 213-276 (1986). Cl71Hubbard, D.W., “Diffusion and Mixing in Non-Newtonian Fluids,” Encyclopedia Fluid Mechanics: Vol. 6 - Complex Flow Phenomena and Modeling, N. P. Cheremisinoff, Ed., Gulf Publishing Company, Houston, 35-l 09 (1986). WI Stein, H.N., “Rheological Behavior of Suspensions,” Encyclopedia Mechanics: Vol. 5 - Slurry Flow Technology, N.P. Cheremisinoff, Publishing Company, Houston, 3-48 (1986). Cl91Frisch, H.L., and Simka, R., “Viscosity of Colloidal Suspensions Macromolecular Solutions,” Rheology, Theory and Applications, Academic Press Inc., New York, 535-613 (1936). of of Fluid Ed., Gulf and F.R. Eirick, Ed., WI Lee, P.F.W., and Duffy, G.G., “An Analysis of the Drag Reducing Suspension Flow,” TAPPI Journal, 59(8): 119-I 22 (I 976). Regime WI Sharma, R.S., Seshadri, V., and Malhotra, Suspensions: Some Mechanistic Aspects,” 703-713 (1979). in Dilute Fibre Science, 34: P21 Gullichsen, Fundamental R.C., “Drag Reduction Chemical Engineering J., and Harkonen, E., “Medium Consistency Technology Data,” TAPPI Journal, 64(6): 69-72 (1981). E231Duffy, G.G., “The Optimum Design of Pipelines Suspensions,” APPITA Journal, 42(5): 358-361 Pu for Transporting (1989). - I. Wood Duffy, G.G., “Flow of Medium Consistency Wool Pulp Fibre Suspensions,” APPITA Annual Genera/ Conference Proceedings, 507-514 (1993). P51 Powell, R.L., Weldon, M., Ramaswamy, of Pulp Suspensions,” 1996 Engineering 533 (September 16-l 9, 1996). of Pulp Pulp Fibre 47th S., and McCarthy, M.J., “Characterization Conference, Chicago, TAPPI Press, 525. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 20 F048 WI Kerekes, Chicago, m Bennington, Suspensions 69: 251-258 R.J., “Characterizing TAPPI Press, 21-28 Fibre Suspensions,” (September 16-19, Status Report January 2000 7996 Engineering 1996). Conference, C.P.J., Kerekes, R.J., and Grace, J.R., “Motion of Pulp Fibre in Rotary Devices,” The Canadian Journal of Chemical Engineering, (1991). P81 Bennington, C.P.J., Azevedo, G., John, D.A., Birt, S.M., and Wolgast, B.H., “The Yield Stress of Medium- and High-Consistency Mechanical Pulp Fibre Suspensions at High Gas Contents,” Journal of Pulp and Paper Science, 21(4): Jll l-J1 18 (1995). WI Pande, H., Rao, N.J., Kapoor, Nonwood Fiber Suspensions,” WI Wikstrbm, T., and Rasmuson, A., “The Agitation of Pulp Suspensions with a Jet Nozzle Agitator,” Nordic Pulp and Paper Research Journal, 13(2): 88-94 (1998). WI Steen, M., “On Turbulence Structure in Vertical Pipe Flow of Fiber Suspensions,” Nordic Pulp and Paper Research Journal, 4: 244-252 (1989). S.K., and Roy, D.N., “Hydrodynamic Behavior TAPPI Journal, 82(6): 140-I 45 (1999). IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) of 21 Project F048 Status Report January 2000 12 FIGURES Modeling I i / Approach Flow Systems Experiments . Validate mixing model for concentric mixer Concentric mixer wl fiber Parametric studies of concentric mixers Transverse mixer wlo fiber Modeling other geometries Transverse mixer WI fiber b\\ Figure 1: Research Approach Flow Recommendations road map for Project F048. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 22 F048 Status Report January 2000 interface between the inner and outer fluid secondary flow (Q, v, c S ) 1 primary flow (4, u, q - I LW Figure 2: Schematic representation of the concentric mixing process. constant head stuff box city water water overflow sewer thick stock overflow return to flow rate control valves t J 0 flow meters discharge tank Figure 3: Former experimental mixing facility. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 23 F048 Status Report January 2000 constant head stuff box city water water overflow sewer to I thick stock flow rate 1 control va Ives flow meters test section Figure 4: Modified experimental discharge tank mixing facility. secondary fluid (water) Q V 1 prima ry fluid (thick stock) mixing Figure 5: Concentric pipe mixer test section. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) region Project 24 F048 Status Report January 2000 Figure 6: The mixing process at a mean velocity ratio of R, = 1.04. Figure 7: The mixing at a mean velocity ratio of R, = 2.02. process I[PST Confidential Information - Not for Public Disclosure (For IPST Member Company’s lntcrnal Use Only) Project 25 F048 Status Report January 2000 Figure 8: The mixing process at a mean velocity ratio of R, = 2.95. Figure 9: The mixing process at a mean velocity ratio of R, = 3.87. lPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 26 Project F048 Figure 10: The mixing process at a mean velocity ratio of R, = 4.95. Figure 11: The mixing process at a mean velocity ratio of R, = 6.02. Status Report January 2000 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 27 F048 Status Report January 2000 n - 3 3 I 0 1 2 3 4 5 Figure 12: The effect of R, on L,/D. outer Figure 13: The axisymmetric computation domain. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 6 7 Project 28 F048 Downstream Status Report January 2000 distance D 2D 3D 4D 5D (a) Cl, = 1.44 and C,, =I.92 cc) GE = 1.44 and C,, =2.5 w c,, = 1.3 and C,, =2.5 (0 c 1E = 1.7 and C,, =2.3 Figure 14: Concentric mixing predictions various C,, and C,, values. for R, = I .O by the standard IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) k-E model for Project 29 F048 Downstream distance D (a>c,, = 1.44 (WGE= Cd)Cl, Status Report January 2000 2D 3D 4D 5D and C,, = I .92 1.44 and C,, = 2.3 = 1.5 and C,, = 2.3 (e> GE = 1.5 and CzE = 2.5 (9) GE = 1.7 and C,, = 2.3 Figure 15: Concentric mixing predictions various C,, and C,, values. for R, = 6.0 by the standard IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) k-E model for Project 30 F048 Downstream Status Report January 2000 distance D 2D 3D 4D 5D (a) C2 = 1.2 (b) C2 = 1.9 (d) C2 = 2.2 (e) C2 = 2.3 Figure 16: Concentric mixing various C2 values. predictions for R, = 1 .O by the realizable IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) k-E model for Project 31 F048 Downstream Status Report January 2000 distance D 2D 3D 4D 5D (a) C, = 1.9 (b) C, = 2.1 (d) C2 = 2.3 (e) C2 = 2.5 Figure 17: Concentric mixing various C2 values. predictions for R, = 6.0 by the realizable IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) k-E model for 32 Project F048 D 2D Status Report January 2000 3D 4D 5D (C > Downstream distance D 2D 3D 4D 5D Figure 18: Comparisons between the experimental and predicted mixing regions for R, = 1.04. (a) Representative experimental image, (b) standard k-E model predictions, and (c) realizable k-E model predictions. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 33 Project F048 I I 2D Downstream distance D Status Report January 2000 3D 2D 3D I 4D 4D 5D Figure 19: Comparisons between the experimental and predicted mixing regions for R, = 2.02. (a) Representative experimental image, (b) standard k-E model predictions, and (c) realizable k-E model predictions. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 34 Project F048 I I I D Status Report January 2000 I 2D 3D I 4D I 5D (a1 Downstream distance D 2D 3D 4D 5D Figure 20: Comparisons between the experimental and predicted mixing regions for R, = 2.95. (a) Representative experimental image, (b) standard k-E model predictions, and (c) realizable k-E model predictions. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 35 Project F048 D 2D Status Report January 2000 3D 4D 5D (C ) Downstream distance D 2D 3D 4D 5D Figure 21: Comparisons between the experimental and predicted mixing regions for R, = 3.87. (a) Representative experimental image, (b) standard k-E model predictions, and (c) realizable k-E model predictions. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 36 Project F048 D Downstream distance 2D D Status Report January 2000 3D 2D 3D 4D 4D 5D 5D Figure 22: Comparisons between the experimental and predicted mixing regions for R, = 4.95. (a) Representative experimental image, (b) standard k-E model predictions, and (c) realizable k-8 model predictions. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 37 Project F048 D Downstream distance 2D D Status Report January 2000 3D 2D 3D 4D 4D 5D 5D Figure 23: Comparisons between the experimental and predicted mixing regions for R, = 6.02. (a) Representative experimental image, (b) standard k-E model predictions, and (c) realizable k-E model predictions. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 38 F048 Status Report January 2000 I I 5 l I I I I 4 I I b I 17 - 3 3 -I L I I I I I I 2 I I I Experimental Estimates I 1 -I Standard k-e Model I - ------------ Realizable k-e Model - I 0 0 I 2 3 4 5 6 7 RV Figure 24: Qualitative values. comparisons between the experimental and predicted IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) L,/D Project 39 F048 Status Report January 2000 Fiber A 0 / I I / / :/ /I / ’I 0 Y Figure 25: Theoretical cr 1 model to the effective Newtonian Fluid I I I I I I I l Y cr 2 Shear Rate, f fiber viscosity. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 40 Project F048 13 APPENDIX A This appendix contains part of Project FO04. 13.1 Report Status Report January 2000 summaries of the various reports completed by Dr. Wang as I Wang, X., “On the Research Areas Company Report, Project FO04 Report of Approach Flow Systems”, 1, March 1996. IPST Member Report 1 provided an outline of potential research directions related to approach flow systems. It contained a brief summary of a telephone survey of various PAC members on approach flow problems. This survey was conducted in the Fall of 1995. The report also included current design recommendations from Voith, Beloit, Valmet, and TAPPI Technical Information Sheets. This report concluded by providing a research plan on approach flow systems. 13.2 Report 2 Wang, X., Anderson, T., Snyder, C., and Marziale, M., “On the Causes of Basis Weight Variability”, IPST Member Company Report, Project FO04 Report 2, September 1998. This report analyzed various signals from a paper machine to try to determine precursors to basis weight nonuniformity. The recorded signals included thick stock flow rate, consistency at eight different stock locations, local inlet and exit pressure of the pressure pulsation attenuator, beta-gauge scanner position, and beta-gauge basis weight. The signals were sampled at a high frequency (100 Hz), and various signal analysis techniques were used to identify signal variances. Numerous peaks in the power spectrum were identified, but the associated causes were not discussed. 13.3 Report 3 Wang, X., Feng, Z., and Forney, L.J., “Computational Simulation of Turbulent Mixing with Mass Transfer”, IPST Member Company Report, Project FO04 Report 3, March 1998. The use of ADINA software, and a k-E turbulence model to predict the threedimensional mixing with mass transfer in various geometries was presented in this report. All calculations assumed that the rheology of a low consistency fiber suspension was similar to water. Therefore, the mixing process was modeled as two turbulent miscible fluids with the same density and viscosity, but with different concentrations of an inert tracer. Modeling transverse mixing with various injection angles from 30” to 150”, in 30” increments, revealed a 90” pipe mixer provided the best mixing. However, these mixers, as well as those with injection angles greater than 90”, produced large scale vortices where floes could form in actual approach flow systems. These mixer types may also impact the opposite pipe wall and create pressure pulsations in approach flow systems. Although jets with injection angles less than 90” are less efficient mixers, they are less likely to impact the opposite wall. This type of jet mixer was recommended for the pulp and paper industry, but a longer mixing length would be required to compensate for the loss in mixing efficiency. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 41 F048 Status Report January 2000 Compared with a single 90” transverse jet mixer, the mixing efficiency of transverse multijet mixers, each with a 90” jet inlet, was much better. Multijet mixers with an even number of jets equally spaced around the pipe performed better than an odd number of jets. No explanation for this phenomena was provided. Concentric pipe mixers were shown to be more effective mixer. Mixing was also shown to be unaffected by concentric than a single transverse mixer nozzle shape. F.inally, mixing in geometries that would represent pulp and paper silo operations entering a fan pump were completed. It was shown that increasing the outlet pipe length will improve the mixing efficiency. A corner cut in the silo was shown to reduce, rather than enhance, inert tracer concentration uniformity. 13.4 Report 4 Feng, Z., Wang, X., and Forney, L.J., “Single Jet Mixing at Arbitrary Angle in Turbulent Tube Flow”, IPST Member Company Report, Project FO04 Report 4, March 1998. This report considered a transverse jet with injection angle &,, with 0” < e0 < 180”, and derived asymptotic solutions for both jet trajectory and tracer concentration profiles in the region near the injection location. Both fluids were similar and the mixing occurred between different concentrations of an inert tracer. The asymptotic solutions matched the existing experimental data obtained from the literature. 13.5 Report 5 Wang, X., and Bloom, F., “Flow Induced Vibration of Submerged and Inclined Concentric Pipes with Different Lengths”, IPST Member Company Report, Project Report 5, May 1998. FO04 This report presented a mathematical model for a submerged concentric pipe system‘with both confined and unconfined external flows (similar to silo flow right before the fan pump). The natural frequencies of the pipe system were determined and it was shown that the system contains low frequency oscillations around 1 Hz. This may cause low frequency vibrations in inner pipes containing thick stock. It was suggested that to minimize these oscillations, (i) the pipe mass per unit length should be reduced, (ii) the pipe flexural rigidity should be increased (or tapered pipes could be used), and/or (iii) various structural supports should be added to the piping system. The criteria where buckling or flutter occurred was not detailed. However, additional conclusions in this report included: (i) the longer the outer pipe was, the more susceptible the inner pipe was to buckling or flutter; (ii) the pipe inclination angle and the depth of the submerged pipe system did not significantly influence the characteristic vibration behaviors; (iii) for a fixed volumetric flow rate, there was an optimal inner pipe radius to minimize buckling or flutter; and (iv) increasing the outer pipe diameter for a fixed volumetric flow rate reduced the inner pipe buckling or flutter. 13.6 Report 6 Wang, X., and Feng, Z., “A Note on Helmholtz Attenuators with Air Cavity and Membrane”, IPST Member Company Report, Project FO04 Report 6, September 1998. This report presented a derivation of an analytical expression for the resonance frequency of Helmholtz resonators with an air cavity and membrane. This type of IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 42 Project F048 Status Report January 2000 resonator can be used to absorb pressure waves with frequencies similar to the membrane resonant frequency. Suggested design guidelines for these devices included (i) increasing the neck length will decrease the natural frequency, (ii) the membrane natural frequency should be as low as possible to minimize the resonance frequency, and (iii) the effect of changing the neck cross-sectional area depends on the membrane stiffness. 13.7 Report 7 Wang, X., and Bloom, F., “Stability Issues of Concentric Pipes Conveying Steady and Pulsatile Fluids”, IPST Member Company Report, Project FO04 Report 7, April 1999. modeling of a This report extended the work of Report 5 and the mathematical submerged concentric pipe system with both confined and unconfined external flows. The criteria when buckling and flutter occurs was also specified in this report. It was shown that frictional forces suppressed flutter instability, and the longer the outer pipe, the less likely flutter was to occur. Frictional forces also delayed buckling instability. However, it was shown that the longer the outer pipe length, the more susceptible the inner pipe was to buckling. It was also noted that for current pipe system designs in the pulp and paper industry with reasonable flow rates, the concentric pipe system was stable. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 43 FLOW THRU POROUS MEDIA STATUS REPORT FOR PROJECT F022 Seppo Karrila (PI) March 8 - 9,200O Institute of Paper Science and Technology 500 10th Street, N.W. Atlanta, Georgia IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 45 Project F022 Status Report DUES-FUNDED PROJECT SUMMARY Project Title: Project Number: PAC: FLOW THRU POROUS MEDIA F022 PAPERMAKING Project Staff Principal Investigator: Seppo Karrila PAC Subcommittee P. Chaudhuri, J. Shands, R. Sieberth FY 99-00 Budget: $104,000 Time Allocation: Principal Investigator: 25% Supporting Research: Special Students: Andres Navia (MS) RESEARCH LINE/ROADMAP: Line #ll - Improve ratio of product performance to cost for pulp and paper products by 25% by developing breakthrough paper making and coating processes which can produce the innovative webs with greater uniformity than achieved by current processes. PROJECT OBJECTIVE: Improve control of the layered structure in thickness direction, by clarifying the mechanisms affecting formation and retention on the wire section. Strive to decouple formation and retention. PROJECT BACKGROUND: The SOTA was reviewed for formation-improtiing wire section elements as well as for splitting methods enabling layered analysis (tools needed for quantification of structure). Experiments with saturated in-plane flow on nip compression were reported. Drawings (in Finnish, European standards and parts list, with some needs for “debugging”) for a pulsating handsheet former were acquired, with device construction dependent on budget allocation. MILESTONES: Implement pulsating handsheet forming capability at IPST. Acquiring an MBDT (Moving Belt Drainage Tester) device is under way and will be accomplished by the end of this fiscal year (June 2000). Quantify dominant benefits of research into wire-section phenomena. Survey opinions from select DFRC members by Spring 2000 PAC. lnitia te numerical modeling of parficle migration during forming. The model is used to identil the dominant phenomena determining retention (layetwise), to design MBDT experiments and to interpret results. Blocks of the model will be presented in Spring 2000 PAC meeting. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 46 Project F022 Status Report DELIVERABLES: Improved diagnostic laboratorv methods for analvzing the structure of paper samples. Enabling characterization techniques necessary for quantitative experimental results are also coupled to troubleshooting printability and curl. L&orator-v procedures for maximizin_asolids content off couch and tunina the stock composition. Development and validation of Mf3DT device and laboratory procedures for retention and drainage studies will enable fast laboratory studies improving dewatering and machine efficiency at a low cost and with a wide range of parameters, in comparison with pilot studies. Retrofit technoloav using next generation forming elements to decouple formation and retention. Opportunities are identified through mechanism studies and applied through innovations. STATUS OF GOALS FOR FY 99-00: March 2000 PAC qoals. Quantify dominant benefits of research into wire-section phenomena. Identify key gaps in forming and dewatering on the wire section and estimate the economic implications. The survey will be reported in the Spring 2000 PAC meeting. Initiate numerical modeling of particle migration during forming. Phenomena and models related to particle migration have been reviewed to assess where improvement over published models is needed for papermaking application, and to enable composing an improved numerical model based on phenomenological equations describing the essential phenomena. The review is included in this report. Current fiscal vear qoals. (Status will be reported in March 2000 PAC meeting.) Implement pulsating handsheet forming capability at IPST. IPST has made a capital allocation that enables construction of the MBDT device. The drawings have been reviewed in detail, improved where necessary, and organized into several subassemblies so that machining work can be subcontracted to external machine shops as separate subprojects. Develop experimental plan for application of the MBDT. The direction taken is decided in the March 2000 PAC meeting, largely based on results of the survey mentioned above. Adopt and improve dry splitting and image analysis techniques that are used as a tool to inspect wire-section effects on layered structure. A set of pilot-run samples in which the jet/wire speed difference has been varied will be inspected. A flatbed scanner with through lighting capability has been acquired and is connected to an existing image analysis system at IPST. Intermediate status with these samples and techniques is reported in the March 2000 PAC meeting. A supporting MS-project has been initiated, with focus on the curl of fine paper. In this project Andres Navia will apply layerwise analysis of paper samples from a collaborating mill to assess the root causes of curl. Layerwise orientation and possibly also fines/filler distributions will be inspected. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 47 Project F022 Status Report SCHEDULE: The project schedule should tabulate the major tasks of the project that were scheduled to be worked on from the last spring PAC until the end of this fiscal year. The schedule should indicate the approximate dates when each task was to begin and when each task was scheduled to be completed. The completion of a major task is referred to as a project milestone. Time is usually expressed in quarters. Task Descriptions 1999 (example) Apr - Jun 1. MBDT quotes, getting funding allocation 2. Design “debugging” 3. Build Eauipment j&Instrumentation I I 5. Initial testina I 6. Initiate migration modelina numerical 8. PAC report 1999 July - Sept 1999 Ott - Dee mm-X 2000 Jan - Mar I a----X 2000 Apr-Jun I SUMMARY OF RESULTS: Dry splitting of sheet samples was demonstrated in a troubleshooting case. Analysis of fiber orientation from thin peeled layers provided information not available from SEM microphotographs or from ultrasonic scans, and was helpful in solving the problem. The construction of MBDT pulsating drainage device resumed after Fall 1999 PAC meeting; it was dropped in Spring 1999 due to budget constraints. Device is expected to be available for initial testing before the end of this fiscal year. To enhance understanding of the mechanisms determining retention, and its layerwise profile in the paper web, numerical simulation of the particle migration phenomena is being initiated. Fundamental aspects related to building an appropriate model from the paper forming perspective are discussed in some detail within this reporf. The key element absent from earlier models with flow through formed web layers is the detachment of migrating particles. This appears to be the phenomenon responsible for filler depletion at the water exit surface of a formed web, possibly in combination with layerwise compression during a suction pulse. SUMMARY OF KEY CONCLUSIONS:. A clear need for improvement in paper related parficle retention modeling has been identified and will be pursued, to quantify the mechanisms determining thicknessdirection filler profiles. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 48 F022 Status Report Particle migration during forming, review of fundamentals SIGNIFICANCE APPROACH C. RESULTS Published models re Iated to paper-making Deposition Turbulent agitation without mat formation Retentionduring drainagewith mat formation Filtration through preformedweb Steadymodelsand surveydata Hydrodynamic particle release in general Entrapment of particles Interaction of migration with permeability Connections with industrypractice D. CONCLUSIONS ScientiJic Conclusions Economics Deliverables D. REFERENCES A. B. A. Significance Dues-funded Project F022 at IPST seeks to improve the balance formation on the wire section. To achieve this goal, a mechanistic of these is necessary to suggest developments in the equipment between retention and understanding of each or operation practices. A laboratory forming capability with pulsating dewatering-a so-called MBDT device-is under development at IPST. In practice, realistic z-direction profiles of pulp fines and fillers have only been observed when a web is formed under pulsating dewatering with somewhat realistic consistency; such capability is essential for inspecting the mechanisms affecting retention. Conventional handsheet forming operates with smooth water removal and very low fiber consistency. This leads to an “inverted z-profile” of filler content, with enrichment on the wire side. Published modeling work appears capable of qualitatively matching this process, but unable to explain retention phenomena in a realistic situation. Still, models are the main tools needed to interpret experimental data, so an improved model is a necessity. The purpose order to of this review is to inspect the fundamental phenomena and mechanisms in guide the design of experiments with the MBDT aid in interpreting results from experiments l develop both qualitative and quantitative understanding of the active mechanisms, and facilitate numerical simulation of particle migration A task agreed to with the subcommittee steering this project, for the March 2000 review, is to “initiate numerical simulation of particle migration during forming.” This review is part of the task mentioned, surveying available qualitative and quantitative information to be used in numerical simulation. l 0 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 49 Status Report B. Approach Published literature is reviewed to collect submodels related to particle migration. These models will support numerical simulations for either qualitative trends to inspect the relative importance of various phenomena, or quantitative simulations and parameter identification based on experiments. Even a qualitative model can provide insights for modifying the process-a well-known example is Wahlstrom’s wet pressing theory and transversal flow presses. A quantitative model should be able to predict the effects of changes in the dewatering profile on the thickness direction profile of migrating components in the web being formed, and on total retention. Calibrating the prediction could be based on current operating conditions on machine or on laboratory experiments that enable parameter identification. C. Results To gain an understanding of what is missing from papermaking related models, first these are reviewed, and then more generic modeling of particle migration is inspected. The conclusions will emphasize the improvements considered necessary in modeling retention during paper forming. Published models related to papermakinq The literature related to papermaking concentrates on the deposition of fillers. As noted, these models are good for steady forming at low fiber consistency, while pulsating forming at realistic consistency is still not well understood. The reader will observe that particle detachment in the paper-forming related models is only considered in cases where no web is formed and turbulent agitation is applied. Deposition The majority of papermaking related experiments have dealt with turbulent agitation of a suspension of fiber and filler and examination of filler content in samples devoid of fibers to assess the filler concentration in suspension as opposed to filler adsorbed on fibers. In a case where web is formed, the options are to inspect retention in a realistic situation with simultaneous forming, or to assess the more well-defined, simpler situation with filtration through a preformed web. The simpler situation is naturally more amenable to modeling and interpretation of results. . Turbulent aqitation without mat formation These experiments are pertinent to stock handling prior to and within the headbox. Alince et al. provide a detailed derivation of an adaptation of Langmuir kinetics to model the deposition/detachment balance of filler on pulp fibers. He notes that aggregates of fillers will have both faster deposition rate and higher maximum coverage than welldispersed filler [Alince 911.Causing aggregation of fillers prior to deposition would then seem kinetically advantageous, but avoiding aggregation maximizes the light-scattering effect of filler [Alince 961. The observations of Middleton et al. support the Langmuir-type deposition/detachment balance for turbulent suspensions; a static model under equilibrium is inspected IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 50 Status Report [Middleton 911. The authors note that fiber consistency has a significant effect on the apparent retention, since increasing the amount of fluid with filler concentration in balance with some adsorption level on fibers means increasing the amount of filler not adsorbed but suspended in fluid. This matches the practical experience that forming at high consistency improves retention. Kamiti et al. continued along the same lines, inspecting the effects of PEI with calcium carbonate, using the model provided by Alince. They once more confirmed the Langmuir kinetics for turbulent agitation without mat formation [Kamiti 941. Retention durinq drainaqe with mat formation van de Ven discussed the retention of small particles, such as dispersed filler, during forming. He concluded that particle attachment during forming is not significant, instead, most of the particles retained are deposited on the fibers earlier on in the process D/en 841. Note the assumption of fillers being in dispersed state; fillers attached to migrating fines or in agglomerated state may behave quite differently. Wei et al. modeled the forming process in a handsheet mold, with analytical solutions reached using simplifying assumptions. The model predicts maximum deposition on the wire side, due to these layers “filtering” the suspension for a longer period of time than the top layers of the web. No effect of retention level on drainage through permeability changes is included. The mat is assumed incompressible in this work. Simulation results were compared with experimental observations published by others, and a satisfactory match was found [wei 961. Wildfong et al. developed a laboratory drainage tester operating at realistic consistency and drainage rate for forming low basis weight mats-pilot trial data is presented and was used to guide the design. A constant vacuum level drives the drainage, so the superficial flow rate is not constant. The flow resistance of the wire was accounted for on calculating the permeability of the formed mat, and a decreasing trend in permeability was observed with accumulation of formed web [Wildfong 981. Further work with the same device showed that the fines content is the dominating factor determining permeability of formed webs. Observations were presented for furnishes with intentionally varied fines contents to examine this aspect. The authors conclude that compressibility effects are secondary in determining permeability, at least in comparison with fines retention [Wildfong 991. Possibly for the case without fines, with a low basis weight, the compressibility effects on permeability are low-however, the experiments did not show that at a fixed fines content, compressibility effects would be absent. Wildfong et al. also showed that the thickness direction retention profiles were almost uniform for both pulp fines and fillers, in spite of the rapid drainage rate affecting overall retention. Fines migration during forming was demonstrated experimentally by application of layerwise analysis of formed sheets [Wildfong 99b]. Sutman improved a pulsating drainage tester, which uses a rotating foil under the forming wire, by increasing the stock consistency to realistic level and adjusting the operating conditions to get realistic retention levels (coarser wire and moderate vacuum application during drainage). He developed various indicators that correlate with mill observations on the effects of additives, as regards drainage rate and formation [Sutman 991.An interesting observation was the effect of peak vacuum level applied during drainage and dewatering on the final air permeability of the partially saturated sheet; this indicates that the compression level does cause a structure change affecting IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 51 Status Report permeability. The basis weights examined were above 50 gpsm, in reasonable alignment for comparison with Wildfong’s results. Compression effects on permeability may well exist for webs formed with realistic fines content, under ordinary drainage conditions. Filtration throuqh preformed web Ramarao developed a relatively complicated micro-model for particle deposition on fibers. Collection efficiency would describe the rate of capture onto a fiber from the set of bypassing particles [Ramarao 931. Al-Jabari et al. presented a model in which the (in the plane of the web) nonuniform permeability of a preformed web was mathematically treated as axial dispersion. The mechanism by which this affected retention was through residence-time distribution of the filler-loaded liquid with the adsorbing fibers. Experiments showed that the breakthrough curves of particles passing through a fiber bed matched the model well for the most part [Al-Jabari 94 and 94b]. Vengimalla et al. used a continuum approach, matching model equations to experiments in a fiber-packed column. Pressure and porosity as well as filler concentration profiles were measured in a fiber bed during filtration. The permeability and retention aspects were uncoupled in the model: retention was assumed not to affect local permeability [Vengimalla 961. Particle release was included in the model, but in a simplified form that did not consider hydrodynamically induced detachment. Steady models and survey data Webb has published a steady model, basically a species balance calculation, for inspecting the fluxes of the solids as well as chemicals that partly get adsorbed on fiber, fines, or filler, and partly remain in solution. The purpose of his model is to facilitate cost analysis of effluent treatment processes, and chemical or fiber gains from increasing retention. He states that the equilibration of the white-water system depends mainly on water consumption and single-pass retention (meaning the dynamics or time-constant of the system) [Webb 871. Britt and Unbehend published in 1983 a survey of 43 commercial machines, examining their retention levels. The data showed single-pass retention of total fines being about 29% for twin-wire machines, ranging between 18 and 43%. Corresponding data for single-wire machines had, on average, 39% pulp fines retention and 29% filler retention, suggesting that the filler retention on twin-wire machines is only about 20% (filler retention is poorer than the retention of pulp fines or total fines). The loss of fines (out of the system instead of into the product) was about 24% for unfilled twin-wire cases [Britt 831. The more recent survey of Korpi, concerning wood-containing grades in Finland, shows approximately similar retention levels [Korpi 921. From a figure included in this publication, filler retention can be roughly estimated to be of the order 30%. Hvdrodvnamic particle release in qeneral While papermaking related modeling has typically focused on particle capture and neglected particle release, the results have also been applicable only to handsheet forming. Typical “steady” handsheet forming will not show fines and filler depletion at the IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 52 Status Report wire side, instead, there will be enrichment due to more flow through the first formed layers of the web. To make the models compatible with the commercially observed depletion on the wire side, either on fourdriniers or on using loadable blades with twin-wire forming, we need to incorporate detachment or release of particles. The particles may be pulp fines, fillers, or aggregates composed of these. The fundamentals of hydrodynamically induced particle release have been studied by Hubbe, among others. After a discussion of the general phenomenology [Hubbe 841, he published results from a turbulent shear-flow experiment in a Couette-type device [Hubbe 851. He also compared the hydrodynamic effects with centrifugal effects, to conclude that the mechanism of detachment is particle rolling (not sliding or lifting). In accordance, tangential forces are the most efficient for releasing particles attached to some surface. Pendse used a rotating disk immersed in a beaker to impose hydrodynamic shear forces on particles adhering to the disk. He studied the effects of pH and salt concentration and how these would affect filler retention [Pendse 851. Sharma confirmed the results of Hubbe by experiments with laminar Poiseuille flow compared with centrifugation. In particular, he discussed the effects of surface roughness and elasticity on the critical shear stress required for hydrodynamic detachment-it appears that elasticity and roughness are very significant factors in somewhat idealized model systems [Sharma 921. Later on Das continued to refine the micromechanical theoretical aspects [Das 941. In general it is agreed that for a single particle, there is a critical torque that initiates the rolling motion and detachment. This critical torque correlates with the average wall shear stress due to fluid flow, and so a critical shear stress is a sufficient descriptor. Similarly, a critical flow rate in some fixed geometry could be given. If the flow is turbulent, the fluctuations in fluid-particle interaction will impose a random component, and with a constant average flow field, there will be a detachment probability over any time step-a particle can remain attached for a while and then be detached after an apparent time delay. On the other hand, the particle-wall interaction through a fluid layer is strongly dependent on the gap size, so after particle detachment its initial motion will be slow, possibly giving the impression of some time delay. Another component of randomness, also pertinent to laminar flow, is the distribution of attachment forces for a population of particles. The attachment forces depend strongly on local contact, with dependence on shape, roughness, and elasticity, even in a nearly ideal case (such as spherical particles on a plane surface). The critical shear stresses for the particle population have a corresponding distribution. Finally, the “collectors” in a fibrous bed are nearly cylindrical, not planar. Even with laminar flow, the impaction probability and detachment forces will have a distribution around the cylindrical collector, which is typically oriented transverse to the flow direction. A change in flow direction will change these distributions and may lead to immediate detachment of a large amount of particles, due to local elevation of shear stress. The particle collection process should have much slower dynamics and will not react as quickly to changes in the flow field. The references above provide some model equations that can be used in numerical simulations of particle detachment. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 53 Status Report Entrapment of particles Once particles are released to migrate with the flow, either they will be entrapped and fixed in position, or they will be purged out of the fibrous web with drainage water. Particles that are large relative to the pore constrictions can be entrapped individually even from a very dilute solution, while small particles are susceptible to multipatticle blocking provided that their concentration in the liquid is sufficiently high. Multiparticle blocking has been observed for particles significantly smaller than the pore constriction, down to about 1% in diameter. Concentration of the particles in the fluid is decisive for multiparticle blocking, also called bridging. Typically there exists a critical particle concentration in the suspension, above which a filter will tend to get plugged. Such plugging is observed by declining flow rate, which would translate to a drainage problem on a paper machine. Interaction of miqration with permeabilitv It is well known that pulp fines significantly affect the specific surface of a fiber bed and have a strong influence on drainage resistance-a reasonable assumption is that the surface area depends linearly on the amount of fines, and affects permeability correspondingly (Cozeny-Karman model). The work of Wildfong et al. has confirmed that fines content in stock and its retention will strongly affect the intrinsic permeability of the formed web. Also, it appears reasonable that the fillers attached on fiber or pulp fines will affect permeability only insignificantly. The only migrating component affecting permeability would then be the pulp fines. Due to their large surface area, these should be able to bond filler significantly, and the fines migration would also affect the filler distribution within the web. Since entrapment will affect the permeability, which in turn will affect the flow rate as well as the release and migration of particles, we have a “feedback loop” requiring an iterative solution eventually. The flow and migration equations should be coupled in a model of general applicability. Another aspect is the modification of void space by significant particle capture, leading to higher local flow rates in the pores even when the superficial flow rate is constant (e.g., in a constant-rate filtration experiment). The increasing flow rate should increase local particle detachment and limit capture. Connections with industrv practice Loadable blades are known to improve formation while retention deteriorates. Further, it has been noted with pilot experiments reported in literature that retention improvement due to increased roll dewatering may be set back by the drainage elements downstream in the wire section. These observations support the conclusion that strong flow (pulses) will deplete the migrating particles from the surface layers of the web. Bachand notes in his review that positive pulses by blades lead to poor retention, which may be improved by replacement foils having a sharp leading angle doctoring the wire [Bachand 831.This can be interpreted in two ways in terms of potential mechanisms. In a case where the formed web is incompressible, the positive pulse causes reversal of flow direction, releasing particles in areas where the shear stresses increased due to the change in flow. If, on the other hand, the positive pulse “loosens” the formed bottom IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 54 Status Report layers, the following drainage pulse would cause a sharp compression and an initially high flow rate, which will detach particles by strong shear forces. D. Conclusions Scientific Conclusions The fines and filler depletion observed on web surfaces created during forming is incompatible with results from models only considering particle collection. These models are sufficient to explain results from smooth slow drainage, with low fiber consistency, as in a regular handsheet mold. The apparent retention of fines and fillers is, theoretically, strongly dependent on fiber consistency during forming; realistic forming experiments should be performed with realistic stock consistency when retention is inspected. Even high drainage rate forming with steady vacuum and realistic stock consistency has shown almost uniform thickness direction profiles of fines and filler, in contrast with experience from machine formed samples. Apparently only pulsating forming causes realistic filler depletion on the water exit surface of the formed web. A theoretical basis exists for the phenomenological description of particle detachment caused by flow (hydrodynamically induced detachment), which could explain the phenomena due to flow pulses also. Such models have not been applied to papermaking but appear necessary for understanding the z-direction profiles of fines and filler created by various forming conditions. Experiments with turbulent agitation without mat formation abound, and they are descriptive of filler attachment to fibers prior to forming on the wire. During forming the porous flow is expected to be laminar, not turbulent, and particle collection will likely take place mainly in the case of pulp fines, while particle release can concern both these and adsorbed filler. The main causes of particle detachment include change in flow direction within the formed web, and fluid flow rates that exceed a critical detachment level. Detrimental effects on retention are associated with pulsating dewatering which may have two origins from the viewpoint of flow through a fibrous filter: the positive pulses on leading edges causing reversed flow and, possibly, loosening of the formed mat, and the initially high flow rates during mat compaction in the beginning of a suction pulse. A further possibility is that the formed fiber mat is redispersed by a positive pulse and then reformed. If the fiber bed were homogeneous and incompressible and the flow unidirectional (no positive pulses), particle depletion at the surface of the web appears very difficult to explain. In reality the fiber/wire interaction makes the web inhomogeneous in thickness direction, but this effect is difficult to model and a first modeling attempt should neglect it. Whether the effects of positive pulses are dominating needs experimental inspection, using the pulsating drainage tester. A homogeneous compressible fiber bed under pulsating drainage provides a reasonable explanation for filler depletion at the web surface, even without positive pulses. The surface layers get compressed and release water, leading to initially high flow rate at the surface. In upper layers there is initially not much relative motion of water and fiber, and the relative flow rates remain small because of the accumulated flow resistance of layers IPST Confidential information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 55 Status Report below. The surface layers get a different treatment from the inner layers, when compressibility is assumed. The only explanations for surface depletion that do not consider particle detachment are redispersion and inhomogeneity of fiber layers due to interaction with the wire. Modeling these aspects may be necessary if a simpler approach, using available fundamental and phenomenological knowledge, does not lead to reasonable results. Economics To facilitate actual dollar value estimates of improved retention, this section presents some data available from the literature reviewed. As the value of improved quality is typically difficult to quantify, the data below relates to cost reduction. Gavelin mentions several benefits of improved retention: decreased loss of solids lower headbox consistency improving formation lower load on saveall less renewed drainage with same fines and fillers recirculated less wire wear (due to quartz particles coming in with clay) He also states that high filler content of white water causes serious operating problems [Gavelin 751. Britt and Unbehend published data on the fines loss, meaning the difference of fines mass through the headbox and in the final sheet. This loss was of the order of 20% (somewhat lower for fourdrinier and higher for twin-wire) in unfilled cases, and increases with filler content [Britt 831. Further, fines recirculation gives them an opportunity to adsorb colorant materials, which can lead to brightness loss. Additionally the recirculation increases deposits and loss of solids from the saleable product. Webb’s calculations indicate that when the degree of adsorption is 70%, the value of lost additive equals that of additive in the product when water consumption is about 20 tons water per ton of product. The lost additive amount is about half of that retained in the product, if the water consumption is lowered to 10 tons water per ton product [Webb 871. The loss of chemical additives is then of the order of one third of purchase price. Webb also mentions the environmental cost of cleaning the mill effluent, and the effect of retention level on equilibration dynamics of the wet end. Deliverables The review presented is used to design experiments with the MBDT and to examine the results of these experiments. A simulation including detachment of particles will be constructed to examine the effects of drainage rate on filler profiles. It is expected that the simulation will show particle depletion on the wire side, and further there will exist a drainage rate above which this depletion leads to “frown-shaped” profiles. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 56 Status Report D. References Alince 91 Alince,B.; Petlicki,J.; van de Ven, T.G.M. Colloids and Surfaces 59(1991) 265-277 Alince 96 Alince, B. Tappi Journal 79, no. 3: 291-294 (March 1996). Al-Jabari 94 Al-Jabari, M.; van de Ven, T.G.M.; van Heiningen, A.R.P. Journal of Pulp and Paper Science 20, no. 9: J249-253 (September 1994). Al-Jabari 94b Al-Jabari, M.; van de Ven, T.G.M.; van Heiningen, A.R.P. Journal of Pulp and Paper Science 20, no. 10: J289-295 (October 1994). Bachand 83 Bachand, J.P. TAPPI Wet-End Operations Seminar (Appleton, WI) Notes: 263-268 (May 8-13, 1983). Britt 83 Britt, K.; Unbehend, J. ESPRI Research Reports, 79:1 SUNY College of Environmental Science and Forestry (Syracuse) (NY) 78 (October 15, 1983). Das 94 Das, S.K.; Schechter, R.K.; Sharma, M.M. J. Colloid and Interface Science 164, 63-77 (1994). Gavelin 75 Gavelin, G.; Odell, P.O.; Vyse, R.N. Svensk Papperstid. 78, no. 11: 392-399 (Aug. 25, 1975). Hubbe 84 Hubbe, M.A. Colloids and Surfaces 12(1984) 151-l 78. Hubbe 85 Hubbe, M.A Colloids Surfaces 16: 227-270 (1985). Kamiti 94 Kamiti, M.; van de Ven, T.G.M. Journal of Pulp and Paper Science 20, no. 7: J199-205 (July 1994). Korpi 92 Korpi, T. Wochenbl. Papietfabr. 120, no. 16: 638 (Aug. 31, 1992). Middleton 91 Middleton, S.R.; Scallan, A.M. CPPA Ann. Mtg. (Montreal) Preprints 77B: B35-46 (Jan. 31-Feb. 1, 1991). Pendse 85 Pendse, H.P. TAPPI Papermakers Conf. (Denver) Proc.: 259-264 (April 15-l 7, 1985). Ramarao 93 Ramarao, B.V. 1993 Engineering Conference (Book 2), 455-476 Sharma 92 Sharma, M.M.; Chamoun, H.; Sarma, D.S.H.S.R.; Schechter, R.S. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 57 Status Report J. Colloid and Interface Science Vol. 149, No. 1, March 1, 1992, 121434. Ven 84 van de Ven, T.G.M. J. Pulp Paper Sci. IO, no. 3: J57-63 (May 1984). Vengimalla 97 Vengimalla, R., Chase, G., Ramarao, B., Das, S. 1997 Engineering and Papermakers Conference, 1435-l 449. Webb 87 Webb, L. J. Paper Technol. Ind. 28, no. 3: 478-479, 481-483 (April 1987). Wei 96 Wei, H., Kumar, B., Ramarao, B.V., Tien,C. J. Pulp and Paper Sci., 22(1 l), November 1996, J446-J451. Wildfong 98 Wildfong, V.J., Shands, J.A., Genco, J.M., Bousfield, D.W. TAPPI 98 Proceedings Engineering Conference, 927-939. Wildfong 99 Wildfong, V.J., Shands, J.A., Genco, J.M., Bousfield, D.W. TAPPI 99 Proceedings Engineering Conference, 1219- 1229. Wildfong 99b Wildfong, V.J., Shands, J.A., Genco, J.M., Bousfield, D.W. TAPPI 99 Engineering/Process and Product Quality Conference & Trade Fair, 11739 1180. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F022 58 Status Report \ Some aspects of application potential of the MBDT A. Significance The Moving Belt Drainage Tester (MBDT) is being constructed at the IPST, and is expected to be available for initial testing before the end of current fiscal year. It provides exceptional capabilities in several respects, such as forming under conventional headbox consistency, applying controllable vacuum level and vacuum pulse rate during forming and dewatering, and providing timed mixing and agitation for chemicals addition prior to forming. In addition to final solids content after some dewatering sequence, the air flow through the wet web during high-vacuum dewatering (after the dry line) can be measured. The device has been demonstrated for laboratory-measurement based prediction of onmachine high-vacuum dewatering, and due to realistic (fourdrinier-type) sheet structure - except for the isotropic fiber orientation distribution - can be used to characterize the effects of stock or fiber properties on paper properties. An application in the project F022 is inspection of fundamental phenomena in fines migration during forming, but other aspects of value to the industry may be considered. The purpose of this document is to discuss earlier reported results that have a bearing on economic value estimates. B. The MBDT principles In the MBDT drainage takes place under pulsating suction. Instead of having a moving wire the tester has a stationary wire and stationary suction box between which there is a moving cogged belt. At intervals in the belt there is a row of punched holes which expose the wire to the vacuum, experienced as vacuum pulses by “an observer fixed onto the wire”. The belt speed is adjustable and the pulsation rate can be varied to a maximum of hundreds of pulses per second. It has been shown that a sheet formed under high pulsation rate possesses a similar structure, in terms of z-direction distribution of fines, to that of paper formed on a fourdrinier type paper machine. Vacuum level can be varied in a timed predetermined fashion during a single experiment. Three injectors are included in the design for the dosage of chemical additives enabling the use of multi component retention aid systems. A variable speed stirrer provides shear forces simulating those to which a furnish is subjected in the approach flow. Vacuum level in the suction box and air flow rate through the formed sheet are measured during an experiment. With individual suction pulses being very short, of the order of 1ms, the dewatering on suction boxes and couch roll can be simulated. An earlier version of the MBDT has been applied in Finland for several years, with excellent results, in the following fashion. With a fixed vacuum level (a controllable parameter) the dewatering is measured as a function of vacuum application time. A set of such curves is constructed, and used in a calculation model to predict the dewatering profile, based on given vacuum levels and effective vacuum application times on machine. In this manner the effects of changes in the vacuum profile, or changes in the headbox stock, on the solids off the couch can be predicted. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 59 Project F022 Status Report C. Economic aspects A consulting company (Papes Oy) made use of the MBDT vacuum dewatering characterizations, in their projects on “Optimization of the use of energy in water removal on a paper machine wire section” and on “Mill measurements of a vacuum system and development of the analysing method” [Maijala 961.The summary report on these projects indicates that using the current vacuum capacity, an average increase of 0.4%points in the solids content of the web after the wire section could be achieved in the analyzed paper machines. Such an increase reduces the water load in the press section by about 4%, with the reduction concentrating on the first nips. The report states that the dewatering capacity of the beginning of the press section often limits the machine speed. If no change in solids content is desired, on an average an energy saving potential of 15% was observed. The drive power requirements are of similar order of magnitude as energy used for vacuum pumping. In an energy optimization effort it is necessary to estimate the drive power requirement for the wire section, and how it depends on the vacuum profile. This is doable based on calibrating an estimation procedure with concurrent mill data. Economic aspects of retention have been discussed earlier on within this report. The MBDT has a significant bearing of retention studies, with timed mixing and agitation, and mat formation. While the 0.4% increase in solids may seem small, the wet strength of a paper web has a strong dependence on the solids content, and the inertially caused tensile loads depend on the moisture content. In combination these effects can have a significant effect on runnability, in particular wet end breaks. On a single machine the cost of breaks (energy, lost production,. . .), with 30-minute break occurring three times per day, is of the order of a million dollars. Even a slight improvement has a significant economic impact. The retention aspects are also connected with runnability and wet end breaks, through the stability of the white water recirculation. Selection of furnish composition, based on inspection of both paper properties and drainage/dewatering, the MBDT provides good potential. One published study compared the drainage properties of different newsprint furnishes, SC and LWC base paper grades, with the result that TMP possesses and advantage over PGW [Raisanen 961. Similar work can be envisioned with various fillers, retention, drainage or formation chemicals. No quantification of the value of such work is available at this time. D. References Maijala 96 Maijala, A.; Lahti, H., Report IO of the “Sustainable paper” program, KCL 1996 (in Finnish) Raisanen 96 Raisanen, K., Appita ‘96, pp. 655659. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s internal Use Only) Project F022 60 Status Report G. Figures The MBDT device at APPI has the same basic design as the device under construction at IPST. The belt loop is shown, with drive mechanism at the back wall and vacuum box inside the loop. IPST Confidential Information - Not for PGblic Disclosure (For IPSP Member Company’s Internal Use Only) 61 FLUID DYNAMICS OF SUSPENSIONS STATUS REPORT FOR PROJECT FO03 Cyrus K. Aidun (PI) E-Jiang Ding March 8 - 9,200O Institute of Paper Science and Technology 500 10th Street, N.W. Atlanta, Georgia 30318 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 62 63 Project FO03 Status Report DUES-FUNDED PROJECT SUMMARY Project Title: Project Code: Project Number: PAC: Fluid Dynamics of Suspensions FDS FO03 Papermaking Project Duration: 1999 - 2005 Project Staff: Faculty/Senior Staff: Staff: C. Aidun E. Ding Project Funding: $70,000 for 1999100 RESEARCH LINE/ROADMAP: Line 11. Improve the ratio of Product Performance to Cost Models, algorithms, and functional samples of fibrous structures . . ., break-through papermaking .. .processes. BENEFITS TO INDUSTRY: To increase the market share and profitability of the member companies by understanding the physics of fiber suspensions in turbulent flow and optimizing the paper forming and coating processes through break-through technologies in suspension transport to enhance quality and reduce cost. I. Investigate the effect of turbulent flow on individual and collection of fibers in the headbox and the forming section to find the most effective flow field for optimum forming process in various paper grades; II . Based on the information obtained through the first objective, develop an optimum forming process. PROJECT DELIVERABLES: 1. A direct method for investigation of individual the headbox and the forming section, results of fiber orientation and collection 2. Computational headbox, and interaction 3. Optimum shape and design of a fiber network forming for most effective fiber distribution in the sheet; of fibers in turbulent in the converging device section flow of of the and other flow parameters Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) - 64 SCHEDULE: Status Report The project schedule to tabulate the major tasks of the project are: Task Descriptions (example) 1. Literature Survey 2. Computational method 2.1 2-D 2.2 CPU optimization 2.3 3-D 3. Include turbulent flow modeling 4. Experiments with fiber in a converging nozzle with turbulent flow 5. Comparison between computational & experimental results 6. Final report 2000 Jan - Mar -1-e--1-w--1-11---X I-- 2000 Apr - Jun -I---------X 2000 Jul - Sep 2000 2001 Ott - Dee Jan - Mar 11----11- m----1-1111 X m-9 -9X -9-9-1111-w19 -9--1--m--- 111-11-1 X ----------1-1 -9-1-w-1-wX -------II X M-11-1 X The method being developed is a hybrid of the finite element and lattice-Boltzmann methods. In this method the motion of fiber in turbulent flow can be investigated in two steps: (1) first the fluid flow pattern is computed without any fiber recording the pressure and velocity at every node in the fluid; (2) the motion of the solid particle (i.e., the fiber) is obtained in the given flow. The first step of the study can be done by finite element method followed by an improved lattice-Boltzmann analysis of particle dynamics. The preliminary method for two-dimensional flows have been developed. Before running this code, two data files, FIINP and FIOUT, must be created by the finite element code, FIDAP. The first data file, FIINP contains locations of every element and their nodes, while FIOUT includes pressure and velocity at every node. The new code reads the information contained in these two files. The computational domain is confined in a small box containing the fiber under consideration. When the fiber moves in the fluid, the box moves simultaneously, keeping the fiber at its center. The pressure and velocity at the boundary are determined by the information contained in the two data files. When the computational domain is meshed by 4-node quadrilateral elements, the bilinear interpolation function is used to determine the pressure and velocity at any point on the boundary of the box. An important step in development of any computational method is the estimate of the computational time required for analysis of practical problems. Assume that the lattice size equals AX cwt , and the time step in lattice-Boltzmann method is At set, respectively. In this section, a subscript s is used for variables in CGS unit, while a subscript 1 is used for those in lattice-Boltzmann method. Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 65 Project FO03 Status Report The viscosity in CGS unit and lattice-Boltzmann method is related by Y =- vsAt 2 (1) (W And, the characteristic speed in the two unit systems is connected by u,At . Ul =Ax (2) In lattice-Boltzmann method, values for Y, and u, are confined in limited ranges. When Y z < l/ 6 the collision operator is over-relaxing, while Y z > l/ 6 corresponds to the underrelaxing collision operator. Y I = l/6 is the natural choice for the simulation, since the viscous stresses decay instantaneously. In computational analysis the range of Y is usually from l/30 to l/ 2. The characteristic speed, u,, should remain small, because, in the current hybrid method, the lattice-Boltzmann calculation is carried out only in a small box, which contains the fiber. The influence of the flow outside the box on the suspended fiber is transferred with the speed of sound, as pressure pulses travel through fluid. If the box moved too fast in the fluid, the flow on the boundary of the box would change rapidly. However, the flow in the box could not change immediately, and the fluid around the fiber would remain at its previous velocity and pressure. Then the results of the calculation would no longer be reliable. Hence the speed of the box must be less than 0.1 in normalized units. From relations 1 and 2, we find (3) and At (4) . Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) _.. 66 Project FO03 Status Report The total CPU time for lattice-Boltzmann simulation, T , is proportional to (At)-' and (Ax)-” where y1is 2 or 3 for two and three dimensional domains, respectively. Hence T oc @S (fs The following (5) ‘%)nc2 ly)“+l - examples illustrate the accuracy of the new hybrid method. EXAMPLE 1: A CIRCULAR CYLINDER IN A SHEAR FLOW To verify the accuracy of the new code, the results for a freely suspended circular cylinder in a shear flow at the particle Reynolds number Re = 0.335 have been obtained with three different computational approaches. The computational domain is shown in Figure I. The channel is 2Ocm long and 4cm wide. The fluid density and viscosity is p = 1.3g!cm3 and p = 0.040625g /cm see, respectively. Velocities of the two walls are -)-U = +O&m/sec , respectively. The radius and density of the circular cylinder is r = 0.161875cmand P = 1.3g / cm3 , respectively. Initially the cylinder is at rest, and its center is at (x, y) = (1.625cm,l.96875cm) . l S Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 67 Project FO03 Status Report Figure 1. A circular cylinder in a shear flow. (1) STANDARD LB , The first calculation is carried out by the standard lattice-Boltzmann method. The size of the channel is 320x64 lattice units. Results obtained by this calculation are presented with in Figures 2 and 3 showing the trajectory of the center of the cylinder, as well as the angular rotation rate of the cylinder, respectively. The rotation rate at steady state is 0.20lseC’. Since the standard lattice-Boltzmann method (ALD code) is accurate in simulating the motion of suspended circular cylinders in shear flow, the results of this calculation are reliable as well. Results obtained by the new hybrid method will be compared with these results, as well. Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 68 FO03 Status Report Standard LB Hybrid LB ........................“.. r Hybrid LB (smaller size) 1.95 1.94 1.93 0 25 X Figure 2: Trajectories of the circle in the channel. (2) HYBRID LB The purpose of this calculation is to test the validity of the new hybrid code. The pressure and velocity fields are obtained by FIDAP. The size of the box covering the circular cylinder is 2 cm x 2 cm, i.e., 32x32 lattice units. The radius of the circular cylinder is Y = 2.59 lattice units. Results obtained by this calculation, presented in Figures 2 and 3, show good agreement with the previous methods. The rotating rate at steady state is 0.207 set-’ , only 3% larger than the rate obtained by the standard LB. (3) HYBRID LB WITH SMALLER DOMAIN SIZE In order to test the influence of the domain size on the accuracy of the simulation results, the size of the domain is reduced to be 1 cm x 1 cm, i.e., 16x16 lattice units. Results presented in Figures 2 and 3 compares fair with previous calculations. The rotating rate at steady state is 0.20 se? , about 9% larger than the rate obtained by the standard LB. Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 69 Project FO03 Status Report 0.23 Hybrid LB(smaller domain) Hybrid LB Q) 0.21 5 .IF 5 H 02. Standard LB 0.19 0.18 20 25 Figure 3: Rotating rate of the circle in the channel EXAMPLE 2: DYNAMICS OF AN ELONGATED ELLIPSE IN A CONVERGING NOZZLE Flow of particles in a converging nozzle takes place in many applications including headboxes and coater heads. A simplified version of this flow is used here as an example for application of the hybrid lattice-Boltzmann method. The computational domain, shown in Figure 4, is a converging channel 40cm long, where the inlet and the outlet are 1Ocm and lcm wide, respectively. The fluid properties are that of water where viscosity is Y = 0.01cm2 /set , and density is p = Q/cm’. A parabolic velocity distribution for u, with a maximum velocity equal to lcm/sec is imposed, and the inlet velocity u, is zero. The major and minor axes of the elongated ellipse are b = 0.05~~ and c = 0.005cm , respectively. The ellipse is neutrally suspended with density, pc = lg/cm3 . Three initial positions at (x, y) = (O.lc~,lcm) , (x, y) = (O.lcm,2cm), and (x, y) = (O.lcm,4cm) are selected, where the initial angle from x-axis to the major axis of the is always x =JC/ 2. Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 70 FO03 Status Report * Figure 4. An ellipse in a converging nozzle. Positions and orientations of the ellipse at every 5 seconds are shown in this figure. Initial positions are at a=(O.lcm,lcm)(open ellipse), b=(O.lcm,2cm)(gray ellipse), and c=(O.lcm,4cm)(black ellipse), respectively. The finite element computations with FIDAP gives the flow pattern in the channel with streamlines as shown in Figure 5. A total of 4049 nodes are used to grid the computational domain. Following the finite element calculations, the lattice-Boltzmann computation of the particle dynamics with time step At is set to 1.5E-05 sec., k is 0.0015 cm, Y, = l/16, and z = 0.6875 is performed. The dimensions of the lattice Boltzmann computational domain are 0.2x0.2 cm , or 128x128 lattice units. The major and minor axes of the elongated ellipse are The particle at various initial positions generally b = 32 and c = 3.2 lattice units, respectively. accelerate to a maximum speed while drifting in the y direction to the centerline of the channel. The angle of attach x decreases until the major axis becomes parallel to the streamline. When Only the particle near the outlet, the particles fluctuate in orientation about the fluid streamline. at lowest initial position undergoes a full rotation by 180 degrees while turning with major axis parallel to the streamline. The same problem will be simulated with the regular lattice-Boltzmann method to evaluate accuracy and reliability of the hybrid technique. After this step, and after the computational Confidential Information - Not for Public Disclosure (For IPST Member Company’s internal Use Only) the Project FO03 71 Status Report demand is reduced with the modified hybrid method, problems more relevant to headbox and flows in a gap former will be investigated. Fi4gure5: Streamlines in the converging nozzle. Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 72 73 FUNDAMENTALS OF HEADBOX AND FORMING HYDRODYNAMICS STATUS REPORT FOR PROJECT FO05 Cyrus K. Aidun (PI) Paul McKay March 8 - 9,200O Institute of Paper Science and Technology 500 10th Street, N.W. Atlanta, Georgia 30318 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 75 Project FO05 Status Report DUES-FUNDED PROJECT SUMMARY Project Title: Project Project PAC: Code: Number: Project FUNDAMENTALS OF HEADBOX FORMING HYDRODYNAMICS FORM FO05 PAPER MAKING Staff Principal Investigator: Co-Investigators: Research Support Staff: PAC Subcommittee Cyrus K Aidun Paul McKay D. Anderson Chairman $219,000 FY 99-00 Budget: Time Allocation: Principal Research AND 25% 95% Investigator: Support Staff: Supporting Research: Ph.D. Students: Special Students: C. Park, H. Xu, M. Brown K. Ono Line #I 1 - Improve the ratio of product performance to cost for LINE/ROADMAP: pulp and paper products by 25% by developing break-through papermaking and coating processes which can produce the innovative webs with greater uniformity than that achieved with current processes RESEARCH PROJECT OBJECTIVES I. Investigate the fluid flow interaction with fiber network in a headbox and the forming section; improve designs for reduction of floe formation and improvement of fiber dispersion in headbox and the forming section; II. Develop methods for measurement of the velocity profile in CD and MD of the forming jet and influence on physical properties. Use this method as a diagnostics tool for process optimization to improve formation and reduce consumption of raw material. PROJECT BACKGROUND The two major areas in the project are to develop novel diagnostics characterization of the forming jet, as well as methods to control the One such system for direct measurement of the forming jet velocity been completed to be used with commercial machines by a process The other aspect of the project sections of the headbox. is to understand the behavior methods for forming jet hydrodynamics. and the turbulent /eve/ has engineer at the mill. of the fiber network IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) in various Project 76 FO05 SUMMARY Status Report OF RESULTS In this section, we summarize the results and their applications. IMMEDIATELY APPLICABLE RESULTS: 1. Development of the Surface Pattern Image Velocimetry (SPIV) method based on two-dimensional cross-correlation of the high-speed digital images of the forming jet has been completed and tested for forming jet velocity measurements. Various techniques to increase the accuracy of the measurements have been used to provide more accurate measurement of the forming jet velocity profile. This method is now capable of surface velocity measurements with one percent (1%) accuracy. 2. Automation of the SPIV method for the on-line measurement of the forming jet velocity has been completed. The mills can use this method for evaluation of the forming jet velocity profile. PROGRESS TOWARD GOALS: 1. The two-component laser-Doppler velocity (LDV) measurements of the streamwise velocity and azimuthal component of the mean and turbulent fluctuations through the step expansion tube has been completed as part of a student project (CSP). These measurements provide details of the flow characteristics in a headbox tube. With this information, and visualization of the floe breakup and dispersion mechanism, the effectiveness of the headbox design can be examined. 2. The floe dispersion mechanism at the step expansion section of the headbox tube has been examined. Methods to quantify the floe breakup and fiber dispersion as a function of flow rate and the dimension of the step change have been developed. The results show strong correlation between the mean axial velocity gradient through the transition from small diameter to larger diameter tube and the floe breakup and fiber dispersion. The fiber dispersion seems to be based on the turbulent eddy formation where the floe deformation, rupture and breakup are from the extensional flow and radial stresses. 3. An automated SPIV system is constructed with the necessary software to use for on-line optimization of process parameters in the forming section. 4. The SPIV method has been applied to a commercial system and results are used for optimization of the process parameters to improve the fiber orientation profile. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 77 DELIVERABLES Status Report FOR by June 00: 1. The velocity profile for the free-surface forming jet using the SPIV method; 2. An on-line mefhod for measurement of the CD velocity profile on the surface of the forming jet as a diagnostics tool for process optimization, DELIVERABLES FOR FY 00-01: 1. Results from visualization of floe formation and breakup at various sections of the headbox as a mean to understand and optimize process parameters for best formation; 2. An on-line method for measurement of the velocity profile along the thickness of the forming jet from the s/ice to the impingement zone as a diagnostics tool for process optimization. 3. Relation between the mean and averaged velocity profile and streaks in the forming jet with the physical properties such as fiber orientation and dimensional instability; 4. The relation between velocity profile and the design features of the headbox; 5. Practical methods to optimize process parameters in order to-minimize velocity profile nonuniformity, with the goal of improving uniformity of physical properties. SCHEDULE: The project schedule for each of the deliverables is provided Task Descriptions (example) 1. Floe dispersion/formation 2. On-line vel. system 2.1 Exp. setup 2.2 Experiments 2.3 data analysis 2.4 interim report 3. Mean vel. and surface patterns 4. Vel. profile 5. Minimize Vel. Nonuniformity 6. Annual Report IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) below. 78 Project FO05 Status Report RESULTS The results are presented in this section with additional detail. The focus is on characterization of the forming jet and the headbox hydrodynamics. The first section outlines a method based on two-dimensional cross-correlation of the digital images to obtain the surface velocity profile of the forming jet. This method provides more detailed information on the forming jet hydrodynamics. 10. Introduction The forming jet velocity profile has a significant impact on the physical properties of finished paper. It is believed that slight variations in the velocity profile can strongly affect fiber orientation, forming table streaks and other such properties. In order to achieve a desired jet-towire ratio, the forming jet velocity is typically controlled using feedback from a pressure transducer in the headbox with the jet velocity being calculated using Bernoulli’s streamline equation. However with high levels of turbulence in the headbox the total head at the point of the pressure measurement includes a non-negligible velocity head that varies with the flow rate. Thus an equation for the total head of a free jet at a nozzle must include a term accounting for the significant head losses resulting from the turbulent flow through the nozzle. These losses vary for different headbox designs, flow rates and slice openings and the head-loss equations must be determined empirically. In practice it is difficult to obtain jet velocity measurement with accuracy better than two percent. In order to study the impact of forming hydrodynamics on the physical properties of the sheet we need accurate measurements of the jet velocity profile. Laser Doppler Velocimitry (LDV) is a useful technique but it is limited in that it can only produce data for one physical point at a time and it cannot be used to analyze fiber suspension flows. Similarly there exist other methods that use point correlation to measure the surface velocity at a given point. In order to obtain the surface velocity profile for a section of the forming jet we have developed a Surface-Pattern Image Velocimeter (SPIV) using cross-correlation methods and wellestablished Particle Image Velocimetry (PIV) technique. The PIV technique is a particularly powerful method for making velocity field measurements. This method is, however, limited to transparent fluids with well defined seed particles. The fiber suspension in the headbox and the forming section is not transparent and therefore cannot be measured using PIV methods. SPIV takes advantage of the non-uniform flow patterns on the surface of the forming jet by recording the positions of distinct flow patterns using a high speed digital camera and application of crosscorrelation techniques between a sequential pair of images to determine the jet velocity profile. Various optimization methods are used in the calculations to maximize the signal to noise ratio. 20. The Theoretical Basis of SPIV The SPIV method is based on the comparison of a pair of sequential images and the tracking of recognizable non-uniform patterns between the two images of known temporal separation. This is illustrated in Figure 1 below. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 79 FO05 Status Report Find the pattern most closely Correlating to this in Image2 <..,,v..+.,.~ .A...,, K~.:~....QY(L(~,,.:.‘<~ ~(-~./,~.‘m~.‘~“‘.‘~:,~,,,~~ ..nii,;,n’,,c,,,“+, ‘+% r% Figure 1 - Illustration of the basis of SPIV 21. The Cross Correlation Method Eacj pattern in the image is represented by two images, the cross correlation between Equation I, below. The displacement during for the highest cross correlation value, O(jJ, an array of gray the two images, a time interval is for each point (i,i) scale numbers. To compare the 0((i), is computed according to determined by searching image 2 in image 1. D&j: Cross correlation value v(m, n) : value of gray scale of Image 1 (O-255) u(m,n) : value of gray scale of Image 2 (O-255) PVPU : Mean value of v(m, n), u(m, n) inside of the inspection box M,N : Size of inspection box Without further optimization of the method, the accuracy would depend However, the accuracy is greatly improved the pixels in the image. presented in the next section. 22. Estimation of the Displacement at the Sub-pixel on the physical size of by Sub-pixel analysis, Level This method has a limitation in resolution since the displacement pixels. In order to obtain more accurate velocity measurements the be improved or some additional filter must be added to the scheme method overcomes this limitation and realizes sub-pixel accuracy is always measured in whole resolution of the image must of the calculation. The SPIV in velocity measurement by IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 80 Project FO05 Status Report fitting the three points adjacent- to the peak of the correlation distribution function with a Gaussian profile, as illustrated in Fig. 2. This method is used to reduce the digitization error that occurs whenever a continuous field is projected onto a discreet domain. / Integer coordinate of peak J / Estimated peak location at subpixel level A I I I I X0 I I I i-l i i+l Figure 2 - Gaussian fit to the three points near the peak. The application of the Gaussian equations, given by f(x) = Cex xO =i+ 1nR(i-1,j) - 1nRQ+l,j> 2In Rcisl ,j> - 4In Rti3j>- 2In Rcj+l >j> YO 1nR(i,j-1) - In R(i,j+l) =j+ 21nR(ij 1)-4lnRcij, -2lnR. * , ,(w+l)(Eq. 2) (ii,j): integercoordinateof peak R(i, crosscorrelation value at (i, j) j): significantly improves the accuracy of the SPIV method. The range of the error is reduced by two thirds and the standard deviation of the error is reduced by approximately 80%. This increases the effective resolution of this technique by a factor of four to five over the resolution of the basic technique. The results of these calculations often contain questionable data points that do not appear correct in relation to their neighbors. These questionable data points are eliminated in order that the velocity profile should vary continuously in space. The value calculated at each point on the image is compared to the average value of it’s eight neighbors and is discarded if it deviates from this average by more that a certain threshold value. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 81 Project FO05 30. Status Report Automated SPIV method for On-line analysis We have implemented the SPIV method in an automated system, as illustrated in Fig. 3, suitable for use in a mill. This system consists of three components for image acquisition, image transfer and real time image processing and analysis. 31. Image Acquisition The images are acquired using a Kodak high-speed digital video camera connected to a Kodak Ektapro Model IOOOHRC Motion Analyzer. This system acquires and stores digital video images as a sequence of numbered eight-bit grayscale Bayer format images at a resolution of 512 by 384. These images may be acquired at a rate of up to 1000 frames per second and stored in the analyzer’s online memory. 32. Image Transfer Currently the images for analysis are transferred using a SCSI optical disk drive as a SCSI buffer; necessary since SCSI can not be used for direct machine to machine data transfer. A set of two numbered sequential images is downloaded through the analyzer’s built-in SCSI port and stored as TIFF files on a 1.3 GB optical disk. The SCSI drive is then reset to refuse interrupts from the analyzer and the files are read from disk by the image processing system. 33. Image Processing and Analysis The actual work of implementing the SPIV method has been programmed as a LabVIEW virtual instrument (VI) running on a Pentium Ill 500 MHz based Windows 98 notebook computer equipped with a PCMCIA based slim SCSI interface. The VI controls the analyzer’s functions for image acquisition, storage and export through the analyzer’s RS-232 compatible serial interface driven by the notebook’s COMI serial port. The images are acquired from the SCSI buffer through the PCMCIA slim SCSI card and read directly into the VI. Processing is handled through a series of sub-Vls and the final data is both saved as a text file for further postprocessing and displayed as a set of charts in near real time. Currently the system applies the SPIV method to a pair of images and calculates the surface velocity (in pixels per second) across the frame. It exports this data as a text file for postprocessing including the plotting of vector diagrams to represent the velocity field overlain on the original image and suitable for animation. This post-processing is done using a custom Visual BASIC PC program and customized image analysis software on an SGI graphics workstation to build the animation files. It also calculates the average velocity along lines of constant CD and MD and plots this data in the VI’s main window as plots of CD averaged velocity against MD and MD averaged velocity against CD. These plots may be printed or cutand-pasted into a report. They may also be reconstructed in a plotting program from the main text file for a more detailed analysis. Elapsed time from the initiation of image recording through the generation of the final plots and export of data is approximately four minutes with default settings. The VI allows the inspection box (that is the area of the image used for pattern recognition) to be adjusted from the default and for MD and CD offsets to be set to allow for imperfectly centered images. It also has a setting to allow for images with flow running either vertically or horizontally through the frame. Changing these settings, especially the inspection box size, from IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 82 Project FO05 Status Report the defaults can significantly alter the time required to process the images as well as affect the resolution of the velocity plots. 34. Future Plans Development of this system is a continuing effort. While all of the first objectives are met there are plans to greatly enhance the utility of this system. These can be divided into three areas; ease of use, instantaneous analysis and time dependent analysis. To improve the ease of use of the system we plan to look into eliminating the need to reset the SCSI bus through direct machine-to-machine transfer of the TIFF images. While this will not appreciably affect system performance it will eliminate one step requiring operator intervention and will remove one piece of equipment from the system. To the current system allowing instantaneous analysis (that is analysis of a set of two images) we plan to add features allowing the plotting of velocity along arbitrary lines and to allow the calculation of FFTs across lines of constant CD. We also plan to enable the system to deal with a time series of more than two adjacent images. It will repeat it’s set of calculations between each adjacent image in the set and display the velocity data as a function of time. This will be more suitable for a post-processing situation as it will appreciably increase the required processing time. IPST Surface Pattern Image Velocimitry PC Running LabVIEW System Highspeed Video Camera Kodak 1 OODHR C Analyzer RS232 Serial Image Data Link ‘1.3GB Optical Drive Figure 3. A schematic of the current on-line SPIV system. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project FO05 83 IPST Confidential Information - Not for Publte Disclosure (For IPST Member Company’s Internal Use Only) Status Report 84 85 OVERCOMING THE FUNDAMENTAL WATER REMOVAL LIMITATIONS OF CONVENTIONAL WET PRESSING STATUS REPORT FOR PROJECT F039 Timothy F. Patterson (PI) lsaak Rudman Daniela Edelkind March 8 - 9,200O Institute of Paper Science and Technology 500 10th Street, N.W. Atlanta, Georgia 30318 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 87 F039 Status Report DUES-FUNDED Project PROJECT Title: SUMMARY Overcoming the Fundamental Water Removal Limitations of Conventional Wet Pressing F039 Papermaking Project Number: PAC: Project Staff Principal Investigator: Co-Investigators: Research Support Staff: Timothy Patterson None lsaak Rudman, Daniela PAC Subcommittee F. Cunnane D. Lacz T. Haller FY 99-00 Budget: Allocated as Matching $96,000 O/0 0 Funds: Time Allocation: Principal Investigator: Co-Investigators: Research Support Staff: Edelkind 20% O/0 I. Rudman 25% D.Edelkind50% Supporting Research: Special Students: None External (Where Matching Is Used): N/A RESEARCH LINElROADMAP: Line #7 - Increase paper machine productivity by 30% over 1997 levels via focus on breakthrough forming, dewatering, and drying concepts. PROJECT OBJECTIVE: Develop through theoretical, experimental and pilot scale studies a non-drying, dewatering technology that will produce sheet solids levels that equal the theoretical maximum for non-drying methods. PROJECT BACKGROUND: Project was initiated July I, 1999. MILESTONES: Identify and quantify opportunities for improving wet pressing water removal. (To be Completed by March 2000) Review published literature. Review unpublished IPST research from previous DFRC projects. Identify and prioritize alternative water removal methods and opportunities. 0 Evaluate potential methods at laboratory scale. (Schedule to be determined in consultation with sub-committee and PAC) IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 88 F039 Status a Implement method. method(s) at an appropriate scale to convincingly demonstrate Report the DELIVERABLES: Deliverables for March 2000 Overall Deliverable - Establish the path forward for the research project. Literature search directed at understanding the current limitations on wet pressing and possible methods for overcoming those limitations. Completed literature search will be submitted as an IPST Member Report. Review of previous IPST research, specifically; l Member company surveys completed for F021 - use information to establish pressing baseline. l Steambox (FO02) and Impulse Drying (FOOI) - for permeability data, pore size data, compressibility data. Use literature and previous IPST research to create a solids profile vs nip curve pressure profile and solids profile vs press position. - Identify and prioritize water removal methods and opportunities. STATUS OF GOALS FOR FY 99-00: . Review the literature and previous IPST research to establish the magnitude gain that can be obtained by overcoming the current fundamental limitations pressing (To be completed by March 2000). n of the on wet Prioritize, based on potential gain and technical feasibility, the possible methods for overcoming current limitations on wet pressing (To be completed by March 2000). m In consultation with the PAC sub-committee develop an experimental evaluating potential methods (To be completed by March 2000). . Goals for the remainder (March 2000). of the FY99-00 to be decided at Spring SCHEDULE:. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) plan for PAC meeting Project 89 F039 Status Report Task Descriptions 1. Literature Survey 2. Write up of Lit Survey 3. Review Prev. IPST Research 4. Develop Press Solids Profile 5. Prioritize Potential Methods 6. Develop Experimental Plan 7. Setup Experiment Equip. 8. Perform Experiments 9. Write yearly report SUMMARY OF RESULTS: A literature review directed at understanding the current limitations on wet pressing and possible methods for overcoming those limitations was initiated. Conventional pressing is not capable of producing sheet solids levels that approach the theoretical maximum for non-drying methods. Sheet and fiber compressibility limit the maximum sheet solids. - Attaining solids levels that approach the theoretical maximum for non-drying methods will require the use of alternative “driving forces” that are not the result of the application of mechanical pressure, e.g. conventional pressing. A review of the literature, previous IPST research, for sheet dewatering yielded several possibilities: Fiber modification. Ultrasound application during pressing. A modified form of displacement dewatering. Given the fundamental greatest potential. mechanisms operating, SUMMARY OF KEY CONCLUSIONS: . Conventional pressing is not capable of producing the theoretical maximum for non-drying methods. require the use of alternative “driving forces”. n and the potential displacement mechanisms dewatering has the sheet solids levels that approach Overcoming those limitations will A modified form of displacement dewatering appears to have the greatest potential, based on the fundamental mechanisms operating, for overcoming the current limitation on wet pressing water removal. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 90 F039 Status Report DISCUSSION: 1. INTRODUCTION The start date for this project was July 1999. Between the project start and the present time, the primary effort was to identify and prioritize potential methods for significantly enhancing the water removal of wet pressing. This effort had three parts: 1. A survey of the open Objective: Determine 1. Where in the 2. What are the 3. What can be 4. What can be literature. the following sheet is the water held during the pressing process? forces holding the water in the sheet during pressing? done to reduce the forces holding the water? done to increase the forces driving water from the sheet? 2. A review of an earlier paper machine Objective: Determine if 1. Sheet solids vs press section 3. survey (Project location F021) can be determined. A review of earlier steambox (FO02) and impulse drying (FOOI) Objective: Determine the following 1. Can representative average sheet pore diameters be determined from existing water permeability data? 2. Can average sheet pore diameter be used to determine the capillary forces in the sheet at various points in the pressing process? 3. Can sheet compression data, obtained from previous sheet displacement experiments, be used to estimate sheet solids vs nip position? 4. Can additional insight into the mechanisms controlling the press dewatering process be obtained? The primary results of the above work were: 1. Press dewatering is limited by sheet compression and the resultant decrease pore spaces. As the sheet compresses it becomes stiffer limiting further compression. The decreased pore size causes increased resistance to water flow. 2. Displacement dewatering, employed in a manner that takes advantage of initial sheet compression that occurs in the early part of the pressing process, has the greatest potential for yielding significantly increased water removal. 3. The paper machine survey data was not sufficient to obtain position in the machine vs sheet solids correlations. However, the data did show that many machines deliver low solids sheets to the dryer section. 4. Using theoretical considerations, water permeability data, and sheet displacement data it is possible to calculate the pneumatic pressure required for displacement dewatering, the optimum point in the nip to initiate displacement dewatering, sheet solids vs compression level, and sheet apparent density vs nip location (up to maximum comprassion). Compression data for the entire pressing process can be obtained and used to analysis the process. The work is described in the following sections. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 91 F039 Status Report 2. LITERATURE SURVEY At the beginning of the project, two different general means of enhancing press dewatering were considered. These were, change th.e sheet/fiber ability to hold water and increase the forces or use alternative forces to remove water from the sheet/fiber. Therefore, a first step was to understand the forces holding the water in the sheet and the forces acting during standard wet pressing. The following subject areas were investigated: wet pressing theory, wet pressing practice, web compression, fiber structure, fiber-water interaction, web-water interaction, WRV test, FSP test, rewet, displacement dewatering, and press felts. While a number of methods were found that could yield incremental increases in water removal, i.e., more uniform pressure application, heating, increased press load, optimized pressing pulse, longer pressing pulse, and improved felts, only three methods appeared to have any potential for significantly increasing water removal. These were, fiber modification to reduce fiber swelling and water content, application of ultrasound during pressing, and displacement dewatering. In standard wet pressing the amount of water removed from the sheet corresponds to the amount of water normally held in the inter-fiber spaces. This does not imply that only inter-fiber water is removed during wet pressing, earlier research has shown that intra-fiber water can be removed by wet pressing. However, if the fiber held less water the wet pressing process should result in greater solids contents at the end of the process. Thus, fiber modification (reduction in fiber swelling and water content) via the addition of chemicals to pulp prior to pressing was considered. There are a few disadvantages to this approach. Previous research indicated that the sheet strength tended to decrease as a result of treatments to reduce fiber swelling. There were also some contradictory results in the literature. In addition, it appeared that different treatments would be required for different furnish types. Several Russian researchers studied the use of ultrasound during and prior to the pressing process. These researchers filed a number of patents and claimed significant increases in water removal. The pressing application involved superimposing an ultrasound induced force on the standard pressure applied during wet pressing. A drawback to this research was that the fundamental mechanism causing the increased dewatering was not identified. The mechanism may have been a phenomena known as ultra sound induced capillary pressure. When a capillary is exposed to ultra sound the capillary pressure is increased. This could in theory, result in water being forced from inside the intra-fibers pores or from the inter-fiber pores. However, no mention of this phenomena was made by the Russian researchers. The subject deserves further investigation to determine the potential magnitudes of ultrasound induced capillary pressure. However, the lack of a clearly identified fundamental operating mechanism made this approach less promising. The final area was displacement dewatering. Several researchers have done work in this area. However, most of the work was directed at maintaining bulk and not at optimizing water removal. Based on some preliminary calculations there appeared to be the potential for potentially significant increases in water removal if displacement TPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F039 . 92 Status Report dewatering possibility derivation derivation permeability FOOI. were employed in the proper manner. There was also the additional of limiting rewet. The last section of this report presents a theoretical on how to employ displacement dewatering in an optimum manner. This was developed by lsaak Rudman, and utilizes the previous literature, water data from FOOI and FO02, and sheet displacement/compression data from A report covering the entire literature search is being written at this time. A section covering rewet is complete. The report will be submitted as an IPST Member Report. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F039 93 Status Report MACHINE SURVEY As part of project F021 a survey of a large number of linerboard, medium, and fine paper machines was undertaken. The objective of the survey was to obtain data on the operation and problems encountered in the dryer section. The survey included a number of questions about the operation of the press section. The data on the press section had not previously been studied. It was hoped that an analysis of the data would yield a press section position vs sheet solids relationship. This relationship would potentially help prioritize potential methods of enhancing water removal in the press section. A review of the data showed that there was not enough information to produce a position vs solids correlation. However, there were some interesting results. The level of solids at the press section exit was not high for any of the furnishes. In the case of the fine paper machines (see Table I), no machine produced 44% solids at the press section exit. Also, the sheet temperature tended to low, for most machines it was around 38 OC. The low level of solids and the temperature were probably the result of constraints on sheet quality and the limiting of press loading and sheet heating. In the case of linerboard (see Table 2) the exit than for the fine paper machines. However, the for two medium machines. Sheet temperatures than those for the fine paper machine, and are solids levels. greater than be rather low sheet resultant press solids tended to be slightly higher exit solids were all less than 45% except and press loading tended to be higher probably the main reason for the higher While no quantitative data was derived that could help with the current project, the results do demonstrate that there is significant room for improvement in water removal on most machines. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 94 F039 4 SUPPORTING THEORY FOR DISPLACMENT Status Report DEWATERING The work presented in the following subsections is an attempt to provide a motivation for the optimal use of displacement dewatering based on fundamental concepts and existing experimental data. The work is not intended to be exhaustive nor complete, but rather to demonstrate there is an optimum manner for implementing displacement dewatering. And, that its implementation can be guided by understanding the physical mechanisms that are at work and by utilizing data from carefully formulated experiments. The Mechanism of Water Removal - Limits Imposed bv Conventional Pressina. In conventional pressing, water removal is induced by compressing the sheet. Sheet compression results in a decrease in average pore size and increase in apparent density. These changes decrease sheet compressibility and as a result decrease the potential for water removal. Using peak pressures of up to 1000 psi, the maximum solids attainable in most press sections is 4550%. This solids level represents about the same amount of water as is found in the inter-fiber pores, i.e. inter-fiber water or free water, (Maloney et al. 1998). Inter-fiber water is contained in the pore spaces between the fibers, these pores generally have diameters of 1 micron or greater. Intra-fiber water (or swollen water) is the water contained in pores that exist in the fibers. These pores generally have diameters of less than 0.05 microns. The intra-fiber water consists of water that is bonded to the fiber through hydrogen bonding and water that is not bonded to the fiber and can be removed mechanically. The amount of intra-fiber water determines the Fiber Saturation Point (FSP) and is about 1.4-I 5 g/g (Lindstrom1986). If only free water is removed in press section the sheet solids should be 40-42%. Outgoing solids on most modern presses is in this range or slightly higher. Some of the latest machines may produce solids in the range of 50%, however, the quality constraints on many grades prevent the level of press loading required to attain those solids levels. Is only inter-fiber water removed in press section? Experiments indicate that intra-fiber water is also removed in the nip (Carlsson 1983, Laivins and Scallan 1993). Therefore, the low solids levels attained in conventional pressing imply that the water removal process is not a serial process - all the free water is removed and then the intra-fiber water is removed. As the sheet is compressed some intra-fiber water is pushed into the inter-fiber spaces and a portion of it may reach the felt. Some of the inter-fiber water also enters the felt, however some of the inter-fiber water may be absorbed by the fibers, thus becoming intra-fiber water. This process is beneficial for development of sheet strength but at the same time limits water removal by conventional pressing. A number of researchers have studied different approaches to increasing water removal in the press section. Peak pressure and nip dwell time can be increased to increase dewatering (Busker and Cronin 1982, Pikulik et al. 1996, Springer et al. 1991). The uniformity of pressure application can be enhanced with improved felts, thus increasing water removal for many grades (McDonald et al. 1999, Oliver and Wiseman 1978, Sze 1986, Szikla 1991, Vomhoff et al 1997). Also, the pressure pulse can be optimized (Schiel 1973). While the earlier work did show that improvements could be made, no IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 95 F039 Status Report significant breakthrough is expected by using conventional pressing at room temperatures. Some researchers have suggested fiber treatments to decrease fiber swelling and, therefore the amount of intra-fiber water in the sheet. These treatments may negatively affect the mechanical strength of paper (Swerin et al. 1990, Strom and Kunnas 1991). Limit of Water Removal bv Mechanical Means A portion of intra-fiber water (about 0.4 g/g) forms hydrogen bonds with the fibers and is contained in the fiber wall in pores smaller than 25 A (or 0.0025 microns) (Stone et al. 1966). This water is sometimes referred to as non-freezing water. The amount of hydrogen bonded water varies insignificantly for different pulps. The amount of this water is affected by neither beating, nor drying. It does not depend on the sheet treatment. This water cannot be removed mechanically, as its removal requires heating to break the hydrogen bonds. It constitutes the limit of water removal by mechanical means and represents a sheet solids content of l/(1 + moisture ratio)) =I/(1 +0.4) = 0.71. Thus, it appears that there is a significant potential to increase water removal from the present level of outgoing solids (40-45%) to magnitudes that are closer to the maximum limit on mechanical water removal, about 70% of solids. Alternative Water Removal Methods. Most alternative water removal methods rely on applying a pneumatic pressure differential to the sheet and on minimizing sheet compression. Earlier techniques associated with this concept involve displacement dewatering (Sprague 1986, Lindsay 1991), capillary dewatering (Chuang et al. 1997, Lee 1995) and blow-through dewatering (Kawka 1979). In the case of displacement dewatering the applied pressure differential acts against the capillary pressure resistance and drives the water out of the sheet. Capillary dewatering counteracts the sheet capillary pressure with a larger imposed capillary pressure. Blow through dewatering depends on water evaporation and water entrainment. All three methods minimize sheet compression in the nip to keep more pores open and increase the water removal rate. Pore size determines capillary pressure to be overcome by the driving pressure used to induce water removal. Pore size decreases with sheet compression. Thus, more efficient water removal can be expected if the sheet is not significantly compacted during application of driving pressure differential. In general, the pores of interest for all three methods are the inter-fiber pores. The intra-fiber pores are extremely small and result in correspondingly high capillary pressures, pressures that would be difficult to produce and maintain. Displacement Dewatering Concept. Displacement dewatering involves sheet prepressing and subsequent application of a driving air pressure. If displacement dewatering is to be efficient, it is necessary to determine the optimal combination of compressive and driving air pressure. Application of the proper compressive pressure ensures that the sheet is in saturated condition and serves to move some of the water in the sheet from intra-fiber pore spaces to inter-fiber pore spaces. Maintaining the sheet in a saturated condition can reduce the occurrence IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 96 F039 Status Report of fingering. Fingering results when there are small areas in the sheet that present less resistance to water flow than the remainder of the sheet. Given an equal pneumatic pressure applied to the sheet and the inter-fiber water, the water will follow these “paths of least resistance” quickly creating a path through the sheet that is not blocked by water. The driving air then flows through those paths creating fingering and localized blow through. The problem is the pneumatic pressure on the remainder of the sheet is relieved, stopping the water removal, and significant amounts of air can escape. If all the paths through the sheet were comprised of equal diameter pores, pores producing the same level of capillary pressure, there would be no fingering and blow through. Elimination of blow fingering can be also achieved by applying foam it the sheet surface (Skelton 1987). As the compressive pressure increases, the size of the pores available for water flow decreases, the capillary pressure resistance increases and the driving pneumatic pressure must increase. Inter-fiber water and intra-fiber water are held by capillary forces. This water can be removed mechanically when and applied pneumatic, P, exceeds the capillary resistance which is given by P = 2 y cos 8 / r, where y is surface tension of water (at room temperature y = 0.073 N/m); r is average pore radius; and 0 is the contact angle, usually varying within the range from 0 to 60 degrees (Hodgson and Berg 1988). Thus capillary pressure resistance depends on the wetting fluid, which determines surface tension and contact angle, and on average pore size. Effective use of displacement dewatering requires that approaches be found to determine necessary compressive and driving pressures in terms of their magnitudes and time distribution in the nip. Role of Water Permeability for Determining Pore Size. There is an experimentally observed and theoretically substantiated link between permeability and average pore size ( Bliesner 1964, Dullien 1986, Hoyland and Field 1976). This link makes it possible to obtain an estimate of average pore size and capillary pressure resistance using the water permeability test (Appendix 1). Thus, required driving pressure can then be calculated. An analysis of the available water permeability measurements conducted at IPST over the years was made. These measurements were routinely performed to estimate specific surface area of the water swollen fibers in the sheet, a measure thought to be correlated with critical temperature of impulse drying. No research specifically targeting effect of the applied pressure within the range encountered in the press nip (up to 500 1000 psi) was conducted using the water permeability test. However, tests were conducted at lesser applied pressures and the data is valuable to the current research. A summary of the available water permeability results is presented in Table 3. It indicates that at high applied pressures the pore size for the sheets, made from different furnishes, converges to approximately the same value regardless of their basis weights and permeabilities in uncompressed state. The coefficient of variation of average pore IPST Confidential Information -Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 97 F039 Status Report size decreases as the applied pressures increases from 10 to 1000 psi and then levels . off (Figure 1). The average pore size in highly-compressed sheets is about 36 A, which is close to the pore size in which hydrogen-bonded water is contained (about 25 A). This water cannot be removed mechanically and determines the limit of pressing. During the water permeability test, the overall sheet compression is measured by monitoring the separation of the press platens. Using this information, sheet compressibility results can be measured which enables a calculation sheet elastic modulus as a function of the sheet strain, or apparent density as a function of applied pressure. While the compression process used in the sheet is not identical to that encountered in the press nip. this information can be used to estimate sheet compression in the nip. Also the development of the solids as a function of compressive pressure can be described. Role of Thickness Measurements in the Nip A system developed to measurement the displacement of metal targets, placed on and within sheets, was used from April to June 1997 for an investigation of sheet compression and expansion during conventional pressing and impulse drying. A schematic is shown in Appendix 1. The system was used in conjunction with an MTS computer controlled hydraulic press. Various furnishes and impulses were tested, all tests used a felt to ensure the compression process was as realistic as possible. Some of the results obtained with this system were reported by Orloff et al. 1998, Table 4 shows the test conditions. Displacement measurement system, in general, worked well within certain limitations. The system used eddy current sensors to measure the displacement of the metal targets. Due to the errors inherent to this measurement method, reliable data could not be obtained from low thin sheets. The minimum sheet thickness for repeatable data was 100 - 150 microns. Thus, sheets or layers within a sheet that had basis weights less than 100 gsm had poor reproducibility. In addition, if the sheet layers were made from sheets of different basis weights cause the prepressing sample preparation process to produce a non-uniform sheet density profile. Figure 2 shows parameters that can be calculated using the displacement measurements. The case shown used a shoe press impulse. Sheet strain, function of dwell time, was calculated as follows: as a A&(t) = [Lo - L(t)] / Lo where Lo is initial (ingoing) thickness of the sheet; L(t) is the sheet thickness at a given time, t of the nip. The corresponding compression rate is dldt. Analysis of the pressure and compression rate curves makes it possible to single out four intervals within the nip which were qualitatively outlined in some publications associated with fundamentals of wet pressing (Carlsson 1983, Wahlstrom 1969, Wrist 1964). The first three intervals characterize the compressive phase of the nip, while the last characterizes the expansion phase of the nip. A more detailed description of the IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project pressure below. 98 F039 nip intervals is presented in Appendix Status Report 2. Some other results are described The MTS presses the sheet under dynamic conditions. The water permeability test apparatus presses the sheet under static, saturated sheet conditions. It is generally acknowledged that static pressing produces higher solids levels than dynamic pressing. Results from the two pieces of equipment can be used to compare sheet compression and solids during static and dynamic pressing. The development of the solids under static and dynamic pressing conditions is presented in Figure 3. It demonstrates the significant difference in outgoing solids attained by static and dynamic compression. An increase of platen temperature makes it possible to diminish this difference. Using the plot of sheet thickness as a function of time it is possible to determine the time and amount applied pressure which brings the sheet to a saturated state. Once the sheet is completely saturated, significant water removal can begin. This pressure can be considered as the minimum pressure required to avoid the fingering effect during displacement dewatering. Reaching this pressure can be thought of as a pre-pressing for displacement dewatering. The maximum of pre-pressing pressure is determined by the point at which the compression rate becomes small. After this point, water removal by sheet compression is not efficient and displacement dewatering by pneumatic pressure should be least susceptible to fingering and have the greatest potential for dewatering. The average pore size will dictate the magnitude of pneumatic pressure required. A point in favor of delaying the application of displacement dewatering until the this point is that as the sheet is compressed the pore size distribution decreases, thus further reducing the potential for fingering. Estimate of Water Removal in the Nip To determine the optimum relationship between pre-pressing and driving pressure in displacement dewatering, an estimate of water removal with respect to nip position should be made. Th.is would also allow a comparison of water removal in displacement dewatering and in conventional pressing. A theoretical derivation is presented in the following paragraphs. Within the limits of the assumptions made in the derivation, the result can serve as a guide for determining the point at which to initiate displacement dewatering. The velocity of the water flow during pressing is determined by the sheet compression rate. Since the area available for water flow in any cross section is the area not occupied by the fibers, the absolute velocity of water inside the sheet is given by the expression: U = (dL/dt) I (l-v c ), where v is the specific density of the sheet. The velocity volume of swollen of the water relative fibers and c = m, / (A L) is the apparent to the fiber network is IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 99 F039 Status Report U, = (dLldt) / (1 -v c ) - dL/dt = (dLldt) {v c / (I- v c )} . In the case of laminar flow, according to Poiseuille’s Law, viscous resistance compression-induced water flow through the fiber network can be expressed to the as P, = 8 L p U,-/ R* , where p is the dynamic pores for water flow. viscosity Since R can be expressed of water and R is the average hydraulic radius of the as R2=8K/& =8Kl(l-vc), where K is the permeability rewritten as for water flow, the equation for viscous resistance can be P v = L~U,E/K Using the relationship for Ur the expression for the viscous resistance becomes P, = L dL/dt p v c / K. It is important to note that although no pressure differential is applied to the sheet in the nip, the resistance to compression-induced water flow is the same as for Darcian flow. where L=Wlc, and W is the oven-dry differentiating yields, basis weight. Substituting the Darcian expression L and P, = (- p W* v) (dcldt)l (c* K). Water permeability K can be found as a function of apparent analysis of experimental data and the relationship density using regression K = k, c k2. The viscous resistance in the sheet is balanced by the portion of applied pressure equal to P - P,, where P is the total applied pressure and P, is the pressure compressing the fibers. Thus, P, can be stated as P” = P - P, IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) that is Project 100 F039 where P, is the compression . apparent density (Ingmanson Status Report pressure which is usually found et al. 1959, Wrist 1964) as a power function of P, = p, c p2. Finally, an equation for the calculation of apparent density as a function of the nip position at which water flow out of the sheet begins is obtained by combining the expressions for P, and P, (PI c p2 - p W* v) (dcldt)l (c* k, c k2) = P. Nilsson and Larsson (1968) concluded that measurements of flow resistance and compressibility do not give enough quantitative information to estimate water removal in the nip after the point of maximum compression. The relationship is suitable for estimates of water removal prior to that point. Additionally, time-dependent functions can be used if required (Wrist 1964). The solution of this equation can be easily found by coding it in an Excel Spreadsheet. As estimate indicate, the assumption about sheet saturation results in marginal error at high compressive pressures, because the volume occupied by compressed air is negligible. A sample calculation, for portion of compressive takes place, was performed for a sheet pressed parameters were used: phase of the nip where water removal at room temperature. The following k, = 0.01 x IO-“’ m* / (g/cm3); k 2 = -4.463 p, = 2.75 MPa / (g/cm”); p* = 3.175; Jo = 0.00995 (dynes/sq.cm) set; W= 0.01874 g/cm* v = 1.03 cm3/g. The result is plotted in Figure 3. This example calculation illustrates that the suggested approach produces a reasonable match of experimental and calculated results. Conclusions. An analysis of the amount water held in the inter- and intra-fiber of a wet sheet, indicates that there is an opportunity to increase water removal in the press from the current level of about 45% of solids to the levels approaching 70% of solids. The main factor which impedes attaining higher levels of solids during conventional pressing is the extent of sheet compression required. The compression is attained through high press loads. Compression of the sheet reduces the average inter-fiber pore size and restricts water flow. Thus, to achieve higher solids, water removal should not be accompanied by excessive compression of the wet web, which in turn requires an alternative driving force for removing the water. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F039 101 Status Report Displacement dewatering in the press section appears to be promising method for attaining higher water removal. Displacement dewatering involves sheet prepressing and subsequent application of a driving pneumatic pressure. Optimum displacement dewatering requires pre-pressing that saturates the sheet, removes some water from the fibers, but does not “excessively” compress the sheet. It then requires the application of the pneumatic driving force at the point in the nip when the compression rate is low, but not negative. The experimentally observed and theoretically substantiated link between permeability and average pore size makes it possible to use water permeability test for an estimate of average pore size and capillary pressure resistance. Thus, the required pneumatic driving pressure can be calculated. Additionally, an analysis of the available water permeability measurements shows that at high applied pressures the average pore size for sheets made from different furnishes converges to approximately the same value regardless of basis weights and permeabilities in uncompressed state. The amount of water contained in the pores at this convergent condition is approximately to the amount of water held by hydrogen bonds. This water cannot be removed mechanically and determines the limit of pressing. The convergence of permeabilities to approximately the same value at high applied pressures also supports the observation that the limit of water removal by pressing is independent of permeability in the uncompressed state. Dynamic measurements of the sheet thickness in the nip make it possible to determine the time and value of applied pressure which brings the sheet into a saturated state and the time and applied pressure that initiate significant water removal due to compression. This pressure can be considered as a minimum pre-pressing pressure for avoiding the fingering effect that can occur during displacement dewatering. The maximum prepressing pressure is determined at which the compression rate becomes small. After this instant of time, water removal by sheet compression is not efficient and displacement dewatering provides an alternative driving force for dewatering. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Table 1. Fine Paper Machines Operation Basis weight (g/m*) i i i; i 1 i Speed ; i ; : i ii Speed i i ; ; i ; i ’: ; i i1 i ; ;i Press 1 Press 1 Number Exit i Exit ; of i i (m/min) I Solids I Temp i Presses f Former Machine i i i i i i :i weight less than 50 gsm =ine Paper r I ------------c------------r--‘--“”””’------------~------------*----------.~-~~~~--~~~~-----. -Basis 47 45 : 2200 I 671 ; 39.1 i 95 f 2 .-~~~~~--~~~~~~~--. - ------------i------------~------------~-------------~------------~----------29 49 i 1401 I 427 i 38 90 i 2 .,,,,,,,,,,,,,,,,,.,,,,---------C------------,----------G-----------i ------------,----------49 1500 i 457 36 i : 2 .-----------------.-------------~------------i 32 f ------------c------------c-------------,----------f i Average 48 f 1700 ; 518 i 38 i 93 f 2 i Std. Dev. 2 t 436 I 133 j 2 f 4 1 0 t i i i: i ii i: i: ; ; i. i. : i I i Basis weight less than 70 gsm and greater than 50 gsm i :: Fine Paper .--____--_____--_-._____________C_______19 55 ! 2644----?------------r------------r--------------. : 806 ; 43 i 120 ‘------------‘.-------------i 3 : top wire ------------~-------------~------------~------------~-------------.-----------------.-------------c------------t I 1st Press ; i ; i i; ; Press 1 i Bottom Load i Top Felt f Felt i (l=yes) (Mpa) !. (l=yes) 1 f : i i: ------------i------------~-------------. :: yes i 0 ------------i------------‘--------------. 1.31 i -i Yes ------------~-------“““““““““” 1.03 : -: ------------i------------i--------------. Yes 1.17 j i 0.19 ; ;i i i i i: i : i ; ------------t------------“---------------.-------------2.76 i yes i 0 ------------+------------~-----------’ Basis weight less than 100 gsm and greater than 80 gsm :ine Paper -----------------..------------.r------------i------------~~-------------~------------~ “““““““““““““““‘c----------r------------.----1 83 : 2403 i 733 “““‘r”“““““c”“““““l”“” : 42.5 : 98 : 3 37 88 910 i 277 I 41 i -----------------.-------------:-----------i ~““‘4i’s”” “------------‘-----------r-----2-----?-ib;;;dri;;ier’ -------------t-----------33 90 i 1569 f 36 I loo 1 *go i 0 ,,,,,,,,,,,,,,,,,.,,,,,,,,,,,,-~------------,----------~-~~~~--------~~--~~~~~~-~~-,~~-~--------~--------------.““‘------“,””--------~-------------’ 20 93 ’ 2100 f 640 f 43 f 112 f 3 i fourdrinier 3.10 1 yes Average 90 ; 1526 i 465 i 40 f 109 f 2 *i 2.24 i Std. Dev. 2 ! 596 i 182 ! 4 i 8 f 1 i 0.75 i f----y-;;---i :i : i 0 Press Load (Mpa) I I I I : 1 Top Felt ! (l=yes) : ; ; f: 1 Bottom f Felt i (l=yes) i 4th Press , : I I I : 1 Top Felt 1 (l=yes) I :: i ii ; i :i f Bottom Press 1 Bottom f Felt Load I Felt 1 (l=yes) I (l=yes) (Mpa) : : i 1 :: ii :: :t i .i --------------,------------*------------. -------------~------------+-----------,------------~------------9-------------fI -i Yes -0 ; -I -BBswwBsss-----------------i-------------. wwmmmmwB----““““““~“““““” I------------i------------~---------2.41 yes i --() ; -f ---------------1------------G---------------. -------------~------------f---------------.------------‘----------‘-“---------------2.76 I yes i --: -i -0 ; -f ---------------i------------a--------------.-------------i------------------‘---------. ------------i------------~---------2.58 ; i: 0.00 ; i 0.00 ; i i i 0.24 1 i 0.00 1 0.00 1 i i ;. : i : i : i : :i I: ii :i II i : ------------~G------------. :: :-----i :: ---J-----------.. ’ w---s--s---- f ------------ic---------3.93 t yes i -2yi’4”” _‘-““’ f yes 1 -i ---------------.------------i--------------. -------------~------------*------------I ------------~------------*---------3.10 -f yes --1 -~~~~~~~~-~--~ 1““““““‘~““‘“““‘I~~~~~~~~---1‘---‘-‘-‘-“‘.“““““’ : -; --I -; --------------‘------------f-“--‘-’-”’ I------------:------------9----------: -: --I -i. --------------i------------~------------.------------i------------~----------i ---i --------------~-------------ic-------------.------------~-------------i-----------I : ;. : --------------~------------i--------------. --------------t------------G--------------. 0.15 ; yes i 0 3.10 : yes i 0 B-m- . ---------------i-----------<-s-s-s-M 2.41 ; yes i 0 --------------,------------~------------. 5.51 i yes f 0 : 3.68 1 i i 1.63 I : Press Load (Mpa) 3rd Press I I i: i 1 Top Felt i (l=yes) I -------------~-------------~------------, -i -------------~-------------i------------j ------------+---------; -ss--m---m---- 1------------i------------.------------: ---4.31 : i --l -““““““‘:““““““f”“““““’ yes ------------:------------“-------------i 4.14 yes i ---i ; : -------------,------------~------------.------------,------------i------------------------,------------$-------------,------------J------------~---------4.14 I yes I -4.47 I -i 0.00 1 0.00 1 i 1 i i i i 0*0° II 0*0° i :i : i f i: i I : 5.17 0.00 0.00 1 I i 0 ! i ; i yes ------------,------------*----------. -i 0 f 0.00 1 i i 0.00 i i --- --yes 0 Table 2. Linerboard and Medium Machine Operation I Basis Weight (g/m*) Machine I ; Speed 1 (mlmin) I ! 1st Press : : : i i i : i Press 1 i IPress Exit! Exit f Number of ! ! j Solids fTemp (F)! Presses :: :I i i 2nd Press 3rd Press I I Press Load (Mpa) Former I 1 1 Top Felt 1 (l=yes) I I i f Bottom I Felt f (l=yes) : I Press Load (Mpa) : 1 1 Top Felt ; (l=yes) I Bottom Felt (l=yes) Press Load (Mpa) i 1Pre Dryer i (l=yes) :: 1 1 I ! 1. Top Felt 1Bottom Felt! I (l=yes) i (l=yes) I i i Pre Dryer (l=yes) --~~~~~~~~. I~~---~~~~~~---~---. 17 t -----------~-------------~------------,-----------~----------------~---------------If-x .-. ! -71n.- .i A!?-- .f 175 .-- .i 3 i. Fmrdrinier 1 3 4!i I Shoe Press ,-----------. 3rd Press ,-----------. ‘2nd Press ,-----------. I 78 I 769 i 693 i 43 I 145 I 2 i Fourdrinier i 140 i 3 ; Fourdrinier 16 269 t 520 1 ii -------------------------------+-------------r------------J-----------~----------------~----------------. f 149 1 2 f Fourdrinier 24 298 I 76 f 42 --____---___---________________ +-------------r------------~----------~----------------~----------------. i 136 ‘i 2 Average 208 I 557 f 41 :I! : 1 Std. Dev. 62 1 215 1. 3 1 12, f i . i :I .i I8I : i ; Bleacehed Linerboard I I ------------~-----------,------------------~---------------~ i . ---‘-““““““‘T’-----------~--------~---~~ 205 : -------------c------------4------------427 i 37 i -i ----------------.----------------.-----------3 : Fourdrinier 15 --~~-~~~~~~----~-~-----.j@---f277 i 46 j 109 1 2 1 Fourdrinier 21 --------------------------------*--------------~------------~-----------~----------------~----------------: 247 1 352 i 42 1 109 i 3 Average i i : i i Std. Dev. I 1 ; i : I i .i t I : i ’ i Medium i I -----~~~~~~~~~-~~~~""1'3'i""T -*-- kmGk ----- f ------------i-----------r----------”------~----------------. 45 i 17 135 i 2i Fourdrinier -~~~~~----~~-~~-------MB---w I 11 It 137 .-. ,L-,-----------~---“--------!------------~----------------~---------------I. 763 --: A7 .I IRrl . -I 7 : Fmmhinier -----I WV --- - ---_--_-_-------_---___________ Std. Div. 1 14 I1 ---------------‘------------i-----------i-----------------,----------------.------------1-----------i------------------------t 98 f 2 i 8 f 0 ! :. :. : -_ 1 1 37.91 I ------------~------------~-----------1 1 f -1-~~~~~~~~~-~ i ------------~---------------;-----------------9.30 1 1 f 1 4.14 t 1 i 1 5.51 ; f 1 i -8.27 f-1---------------i----------------1 i ------------~:----------------------~------------i------------------------+------------r-----------c-----------I.45 1 ““““““f”‘-“‘““‘il”“-““’-1 1 1 --I -1---------------r,,,,,,,,,,,,,,--. -1 1.45 I -f 1 ~~~~-~~~~~~~ .L-m.---m.--..------------J,------------~----------------------I. I f. ! 4.9 I 10.1 1 18.8 1 !! 1 1 I ! 3.0 i 19.5 I i ; j i 1 12*0 f II 1 I I . .I I I I : i I i I ------------‘------------t-------------------------+--..--------1~~~~~~~--~--i ~~~~~~-~~~~1.90 1BBS--------f -----_____-_--__________ 1 1.90 1------‘-““f”““‘““‘,“““““” -j 1 1 7 2.41 -j-j 1 i -2.41 ; -i 1 ““““““‘“““‘““‘t”“““““’ -----------+-------““‘t”““““““----------. i 2.2 1 2.2 1 I f i : I I I i ! f I . : i i : I I f ! i : I I I : ; 1 1 : I I I I ------------,------------i-------------------------~ ------------i------------i------------““““““~ ------------i---------------~-----------------! 1 -I 6.89 I 1 ! 1 8.27 I 1 I 1 i -f ------------J------------i-414 --II :- 1 :. 1 14.1 i I. I I i:. I2nd Press1 I I Table 3. Average Pore Size Based on Permeability Measurements 1.2569 0.0773 6 7 8 I I I 1.0480 O.OlO6 1 0.6445 0.0047 I 05233 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 0.0018 n 0.4871 n I Project 105 F039 Status Report Conditions for Displacement Measurements Condition 1 Condition 2 Conventional Shoe Press. Pressing Peak pressure, MPa 4.82 5.39 Pressure impulse, 37.9 - 41.3 117.2 kPa s Dwell Time, ms 19-20 46 - 47 Ramp Duration, ms 0 0 Table 4. Experimental I Furnish Press Platen Temperatures, “C Ingoing Solids, % Layering, gsm Repulped Liner 30% occ 660 ml CSF 21 200 250 Repulped Liner 30% occ 660 ml CSF 21 100 200 300 35 51, 51, 51, 51 35 51, 51,51,51 IPST Confidential Information -Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Condition 3 Shoe Press WI Ramp 5.55 137.8-151.6 40 15 runs 4,5 26 runs I,2 43 runs 3 occ 620 ml CSF 275 250 300 325 350 40 15, 30, 30, 30, 100 Project 106 F039 Status Report a> .-> cn co a, k E 0 0 ch > s .-0 .-5 4-i t5 > 5 5.c) ;. I d z SUOJ31UJ d ‘JaJauwa. aJod aBwayf IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 107 F039 Figure 2. Displacement Status Report Results for MTS Pressing at Room Temperature Compression Rate = -dL/dt Strain = ( Lo- L) / Lo 800 - 0.95 -Pressure - 09. Comp Rate I I Strain A 0.85 08. Thickness 0.75 E E 550 07. f 500 -- 0.65 $ 450 --- 06. .-5 400 0.55 -- 05. 0.45 04. 0.35 03. 0.25 02. 0.15 01. _-- -_ 5 -200 -c 0 0.02 0 I II 0.04 0.06 Nip Time, set 0.05 0.08 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 1 -0.05 01. E E Figure 3. Compressive Pressure vs. Solids from Water Permeability (Static Data) and Sheet Displacement Measurements (Dynamic Data) -131 gsm / 1.8 mA2/g --r---435 gsm / 92 m*2/g - h= 79 gsm / 5.4 mA2/g * - -210 gsm / 9.7 m*2/g 187.4 gsm.! 21 degrees C 189.5 gsm /IO0 degrees C 06. Solids 07 l 08 . 09 . 1 109 Project F039 Status Report Figure 4. Experimental and Calculated Apparent Densities 800 08. Applied Pressure 700 600 06. 500 0 c) & .-s co 04. n5 05. 400 300 200 100 0 0 0.02 0.04 0.06 0.08 Nip Time, s IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 01. Project 110 F039 LITERATURE Status Report CITED Bliesner, W.C. A Study of the Porous Techniques. Tappi J., 47(7): 392-400 Structure of Fibrous (July, 1964). Burns, J.R., Conners, T.E., and Lindsay, J.D. Dynamic development during wet pressing. Tappi J., 73(4):107-l Sheets Using Permeability measurement 13( 1990) of density-gradient Burton, S.W., and Sprague, C.H. The instantaneous Measurement of Density Profile Development During Web Consolidation. Journal of Pulp and Paper Science:Vol.l3, No.5 September 1987. Busker, L.H. and Cronin, D.C. The Relative Importance Removal. 1982 Int. Water Removal Symp., 25-34. Carlsson, G. Some fundamental Stockholm, 1983. aspects of the wet pressing Chuang, S., Kaufman, K., Schlesser, R. Capillary U.S. pat. 5,701,682. Issued December 30, 1997. Dullien, F.A.L. Porous Media: of Wet Press Variables Fluid Transport Dewatering of paper. Method in Water Doctoral Thesis, and Apparatus. and Porous Structure. Hodgson, K.T. and Berg, J.C. Dynamic Wettability Properties of Single Wood Pulp Fibers and Their Relationship to Absorbency. Wood and Fiber Science, Vol.20 (I), 3-17, January, 1988. Hoyland, R.W, and Field, R. A review of transudation of water into paper - in five parts. Part 3. Some principles of flow & their application to paper. Paper Techn. and Industry, December 1976, 291-299. Ingielewicz, H; Kawka, DRYING OF POROUS 1979). W. INVESTIGATION ON INTENSIVE DEWATERING PAPERS. Paper (London) 191, no. 10: 618, 621-622 AND (May 21, Ingmanson, W.L., Andrews, B.D., and Johnson, R.C. Internal Pressure Distribution in Compressible Mats under Fluid Stress. TAPPI, Vol.42, #IO, 840-849, October 1959. Kawka, W., Szwarcsztajn, E. SOME RESULTS OF INVESTIGATIONS ON THE EQUIPMENT FOR INTENSIVE DEWATERING AND DRYING OF POROUS PAPERS. EUCEPA Conf. (London) 18, Paper No. 31: 17 p. (May 21-24, 1979). Laivins, G.V. and Scallan, A.M. Removal of Water from Pulps by Pressing - Part 1: Inter and lntra Wall Water. Tappi Eng.Conf, 1993, Orlando, Florida: 741 L 748. Lee, C. A. Method August 29, 1995. for Dewatering a Porous Wet Web. U.S. pat. 5,445,746. IPST Confidential Information -Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Issued Project 111 F039 Lindsay, J.D. Displacement 242, 1992. dewatering Status Report to maintain bulk. Paperi ja Puu, Vol.74, Lindstrom, T. The Concept and Measurement of Fiber Swelling. Pp.7597 edited by Bristow, J.A. and Kolseth, P. “Paper Structure and Properties”, #3, 232. in monograph 1986. Maloney, T.C., Todorovic, A., and Paulapuro, H. The effect of fiber swelling dewatering. Nordic Pulp and Paper Res. J., Vol.33, #4, 1998, 285291. on press McDonald, J.D., Pikulik, I.I., Ko, P.L., and Owston, T.H. Optimizing Market Pulp Felt Design for Water Removal. 85th Annual Meeting, PAPTAC, Jan.28,29, 1999, A243A250. Nilsson, P. and Larsson, K.O. Paper Web Performance of Canada, December 20, 1968: 68 - 73. Oliver, J.F. and Wiseman, Roughness. Transactions Orloff, D., Rislakki, Commercialization. in the Nip. Pulp and Paper Mag. N. Water Removal in Wet Pressing: The Effect of Felt of the Technical Section, TRI 04.TRI 09, December 1978. M., Rudman, I. impulse Drying of Board Grades: Status of IPST Technical Paper Series, #715, April, 1998, 8p. Pikulik, I.I., McDonald, Proc., 735-740. J.D., and Gilbert, D. Pressing of Market Schiel, C. Optimizing the Nip Geometry of Transversal-Flow Mag. of Canada, March 7, 1969, 73-78. Skelton, J. Foam-Assisted Dewatering - New Technology 28, no. 2: 431, 434-436 (March 1987). Sprague, C.H. New Concepts Paper Chemistry. in Wet Pressing. Pulp. 1996 Eng. Conf. Presses. Emerges. Final Report. March, Pulp and Paper Paper Technol. 1986. Institute Ind. of Springer, A., Nabors, L.A., and Bhatya, 0. The influence of fiber, sheet structural properties, and chemical additives on wet pressing. Tappi J., 221-228, April 1991. Stone, J. E.; Scallan, A.M.; Aberson, G.M.A.; The Wall Density Fibres; Pulp and Ppr. Mag. Can. T263-T268 (May 1966) of Native Cellulose Strom, G.; Kunnas, A. Effect of Cationic Polymers on the Water Retention Various Pulps. Nord. Pulp Pap. Res. J. 6, no. 1: 12-19 (April 1991). Value of Swerin, A.; Lindstrom, T.; Odberg, L.; Deswelling of Hardwood Kraft Pulp Fibers by Cationic Polymers: Effect on Wet-Pressing and Sheet Properties; Nord. Pulp Pap. Res. J. 5, no. 4: 188-196 (Dec. 1990). Sze, D. Measuring wet press felt pressure Tappi J., April 1986, 120-I 24. uniformity and its effects on sheet solids. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 112 F039 . Szikla, Z. Role of felt in wet pressing. Paperi ja Puu, 73 (1991):2, Status Report 160-168. Vomhoff, H., Norman, B. Model Experiments on Wet-Pressing-the Influence of FeltSurface Structure; Nordic Pulp & Paper Research Journal 1, no. 12: 54-60 (March 1997). Wahlstrom, PRESSING. P. B. OUR PRESENT UNDERSTANDING OF THE FUNDAMENTALS Pulp Paper Mag. Can. 70, no. 19: 76-96 [T349-691 (Oct. 3, 1969). Wrist, P.E. The Present State of Our Knowledge of the Fundamentals Pulp and Paper Mag. of Canada, T284-T296, July 1964. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) OF of Wet Pressing. Project 113 F039 APPENDIX 1 - Water Permeability, Average Status Pore Size, and Capillary Report Pressure In conventional wet pressing, the applied pressure drives water from the sheet by compression. Sheet compression results in a less compressible mat and a decrease in average pore size. Thus, pore diameter in the compressed state is a factor controlling dewatering. Average pore diameter can be evaluated based on an analogy between laminar flow in a channel and Darcian flow in the porous medium (Dullien 1986, Hoyland and Field, 1976). The expression for driving pressure in a circular channel is: AP = U p L / (2 D,*/64). The Darcy equation for driving pressure is: AP=U pL/K. Setting the two equations for AP equal to one another permeability to hydraulic diameter of the pores yields an expression which relates K = D,*/32. An expression, derived from by alternative means, that links permeability for a bundle of parallel capillaries is essentially the same K=E and pore size D,*/32, where E is porosity indicating volume of the mat available for water flow. This equation can be also obtained from the using the channel flow/Darcy analogy if one takes into account that flowing water is affected by the fraction of pressure differential equal to AP E. Capillary pressures, calculated as a function of pore diameter and contact angles of zero and sixty degrees are plotted in Figure Al .I. The intervals of pore sizes for interfiber water and intra-fiber water are also shown. Inter-fiber water is contained in the pores greater than 1 micron diameter. Intra-fiber water is contained in the pores having diameters below about 0.05 microns. The graph shows that in order to express free water, the driving pressure should be within the range 2-40 psi. Removal of the intrafiber water requires driving pressures in the range of 300 to 40000 psi. Removal of the water from the pores less than 25 A which would produce solids more than 70% requires pressure more than 8500 psi. Bliesner did some earlier work on determining average pore size. The method used for that work was to measure “breakthrough pressure”. This work produced pore size distribution curves and permeability measurements. From these results hydraulic diameter was calculated. A comparison of Bliesner’s results and the results obtained from previous IPST work is shown in Figure Al .2. The trends for the two sets of data are similar. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 114 F039 Status Report IPST accumulated a considerable data base of water permeability measurements during various projects over the past 4-5 years, a sample of these tests are tabulated in Table Al .I. The majority of these test were performed at compressive pressure of 0.02 to 2.0 MPa. From the available water permeability data three data sets were selected, the sets represented low, average, and high water permeability. Permeability as a function of compressive pressure was then determined using a power function regression K x10’* [m*] = a P b, where compressive pressure is taken in psi. Using the formula K = D,*/32, pore diameter as a function of compressive pressure for the sheets having low, average and high water permeability were calculated and plotted in Figure Al .I. Note that the plotted compressive pressures (70 MPa) are well above those used to obtain regression formulae which may result in some inaccuracies. The results in Table Al .I and Figure Al .I suggest that at high compressive pressures, permeabilities for any sheet converges to approximately the same number which is about (20-50) x IO-*’ m*. The average pore diameter at such permeabilities is about 25 40 A, which is the upper bound of the pores in the fiber wall which contain hydrogenbonded water. Using the permeability data, a regression function was determined to relate the values of apparent density to sheet solids. At high compressive pressures, apparent density of the sheet c = BWIL should approach the wood fiber density equaled to about 1.55 g/cc. As the experimental data were obtained within the range of 0.2 to 2.0 MPa, extrapolation to the pressures up to 7-70 MPa may result in inaccuracies. In some cases the values of apparent density exceeded their theoretical maximum. Plots of apparent density vs. compressive pressure can be translated into the functions of compressive pressure vs. solids in the nip which are frequently used in calculating water removal for wet pressing. An equation that links the apparent density and solids, s, can be obtained for a saturated sheet and has the form S =c/(c+p,-C&Y where pw = lg/cc and pf = 1.55 g/cc. apparent density and solids Which s = c/ (0.355 leads to the following c + I). IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) formula linking Project F039 115 Status Report Graphs reflecting effect of static compressive pressure on the average pore size and . outgoing solids for cases of different basis weights and specific surfaces (Data from Table 1) are plotted in Figure Al .2. Convergence of the average pore size is again confirmed. Even using static compression outgoing solids of about 70% can be obtained only at pressures in excess of 7 MPa. Thus, current machine solids levels of about 45% are quite understandable. The results plotted in Figure Al .I can be used to determine differential to displace water from the sheet when the sheet compressive load. The results can also be used to determine with a given permeability for displacement dewatering. the required pressure is also subjected to a the suitability of sheet For the purposes of illustration, let’s assume that in order to bring the sheet to a saturated state, the sheet to be subjected to displacement dewatering is precompressed by the compressive pressure of 0.7 MPa. The range of capillary resistance pressure at this compressive pressure is given in Table Al .2. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project I sure 116 F039 n I. I Parameter Low Permeability Oven-Dry Basis Weight, gsm Coefficient a in regression K*l O1*=a*(Pb) Coefficient b in regression K*l O1*=a*(Pb) Measured specific surface, m*/g Measured specific volume, g/cc Permeability*1 015, m* @I 0 psi Average pore diameter, A, @ 10 psi Permeability*1 015, m* @3000 psi Average pore diameter, A, @ 3000 psi 203.7 0.0052196 Average Permeabilitv 150.4 0.79691 High Permeabilitv 177.4 55.005 -1.1737 -1.9204 -2.5965 62.97 0.73 0.35 1060 0.000433 37.25 99. 1.03 9.572 5538 0.000167 23.16 2.156 1.405 1329.8 66799 0.000515 40.63 Low Permeability Average Permeability High Permeabilitv I 340 - 680 64 - 128 I I I ~ ~ ~ Table Al .2 Permeability I I Status Report I Capillary resistance pressure at compressive pressure of 100 psi I 800 - 1600 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 117 Project F039 Status Report % ;I - - 1 r 1 - I I I I - ..:i: ..: ,’ n- - --l-4-- $tL I I - - .Isd ‘amssq anlssadLuo3 :!sd ‘amssad . Aq~de=) . IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Onlvj Figure Al .2. Average Hydraulic Pore Radius vs. Permeability 8 7 __-------__ 0 6 5 0 0 _____ o Rh from Bliesner Exp Tappi 0 8 0 4 l 0 3 0 y’ - 0 2 ~__ 1 ___------. 0 1 2 Permeability x IO 8 (m*) 3 Rh = 2.83x (K)“1/2 Project 119 F039 APPENDIX 2. Displacement Measurement Status Report Results. The displacement measurement system used eddy current sensors and preforated copper target disks. These disks were either embedded in the sheet or place at the top and bottom surfaces of the sheet. This is the same approach as used by Burns et al. 1990, Burton and Sprague 1987. The displacement measurement system was used between April and June 1997 for an investigation of sheet compression and expansion in the nip during conventional pressing and impulse drying. A felt water receiver was used in all cases. A schematic of the experimental setup is shown in Figure A2.1. Typical results of displacement measurements temperatures of the platen for extended nip Figure A2.2 - A25 Analysis of the pressure possible to single out 4 intervals within the compressive phase of the nip, while the last Interval 1: From the entrance on the MTS press at (2nd set of experiments) and compression rate nip. First three intervals characterizes expansion room and elevated are plotted in curves makes it characterize phase of the nip. of the nip to the point at which a rapid decrease in compression rate occurs During this period of the nip, the compression rate in flow-controlled nip is usually very high, because sheet is easy compressible. Part of the applied pressure is balanced by fiber network, while the rest of applied pressure is balanced by air which is contained in the pores. No water or a negligble amount of water is removed from the sheet and, thus, the hydraulic component of total pressure is close to zero. At a known ingoing sheet weight and solids content, can be evaluated by using the formula the sheet thickness at zero porosity L=l /A (m(p, + m&J where m,and m, are mass of water and fiber in the sheet respectively; densities of water and fiber respectively. By comparing air is present the calculated and measured sheet thickness in the sheet at the end of this interval. p,,,,and pf are the it can be established that Interval 2: From the point at which a rapid decrease in compression rate occurs to the point at which slows to sliqhtlv areater than zero. At a point, when the pressure inside the sheet becomes higher than capillary pressure resistance, the water starts to flow out of the sheet. This point corresponds to a quasi-inflection point in compression rate curve. It is thought that during this period, water in the inter-fiber space is pushed from the sheet. Interval 3: from quasi-inflection point of compression rate curve to zero compression rate (or the minimum of sheet thickness in the nip). In this interval compression rate is very low which indicates that primarily intra-fiber water is removed during this period. As IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F039 120 Status Report sheet compressibility is low at high pressures, water removal in this interval is marginal. An increase in the duration of this interval will not produce a noticeable increase in water removal. Interval 4: from minimum of the sheet thickness to exit of the nip (pressure is about zero). In this interval the spring force of the compressed sheet exceeds applied pressure and sheet recovers. The compression rate becomes negative. Reducing the duration of this period should reduce rewet. Sheet recovery continues after the nip exit. If the sheet is not separated from the felt, post-nip rewet may occur. No thickness measurements of completely recovered sheet were conducted to determine total springback. IPST Confidential Information -Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 121 F039 Figure A2.1 Displacement Measurement Device IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Status Report Figure A2.2 Displacement Compression Results for MTS Pressing at Room Temperature Rate = -dL/dt, Strain = ( Lo- L) / Lo - 0.95 800 - 0.85 - Comp Rate - 0.75 600 500 - 0.65 400 - 0.55 300 - 0.45 200 - 0.35 100 - 0.25 0 - 0.15 - 0.05 - -0.05 0.04 Nip Time, set 0.06 0.08 0 . 1. E E uiu) i! -5 z 3 i! co $ b co 123 Project F039 Status Report s .-0 0 E 0 c) nl a a, “3 .F LL 0 0 0 0 0 0 0 0 00 o E g m o % oco E z r,as/LuuJ * 0 ‘0 1 qed aA!ssaJdwo=> r!sd ‘amssa-rd payddv IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 0 0 I 0 0 clI 0 Project F039 124 Status Report co - 0 d * - 0 d 0 0 00 0 0 b 0 0 (0 0 0 m 0 0 d- 0 0 m 0 0 N 0 0 \ 0 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Onlv1 0 0 I - 0 0 0 NI ki cn ai .-E I- .-Q z Project 125 F039 Status Report - 0 g 0 g 0 g YLa)ey 0 0 m 0 g 0 g anissaJduro=> :!sd . 0 E 0 0 ‘aJtWS3Jd 0 0 0 I payddv IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 0 0 NI 0 126 127 REMOVING IMPULSE LIMITATIONS ON DWELL TIME OF WET-PRESSES, DRYERS, AND DISPLACEMENT DEWATERING STATUS REPORT FOR PROJECT FO40 Paul Phelan (PI) lsaak Rudman Marcos Abazeri Edward Lindahl March 8 - 9,200O Institute of Paper Science and Technology 500 10th Street, N.W. Atlanta, Georgia 30318 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 128 129 Project FO40 Status Report DUES-FUNDED PROJECT SUMMARY Project Title: Project Code: Project Number: PAC: REMOVING LIMITATIONS ON DWELL TIME OF WET-PRESSES, IMPULSE DRYERS, AND DISPLACEMENT DEWATERING TECHNOLOGY LONG DWELL FO40 PAPERMAKING Project Staff Principal Investigator: Research Support Staff: P. Phelan I. Rudman, M. Abazeri, E. Lindahl PAC Subcommittee C. Kramer, F. Palumbo, N. Rudd FY 99-00 Budget: Allocated as Matching Funds: $143,052 None Time Allocation: Principal Investigator: Research Support Staff: 50% Technician (50%) Supporting Research: Special Students: External (Where Matching Is Used): None None RESEARCH LINE/ROADMAP: Line #7 - Increase paper machine productivity by 30% over 1997 levels via focus on breakthrough forming, de-watering, and drying concepts. PROJECT OBJECTIVE: The objective of this work is to improve productivity by increasing water removal while maintaining and/or improving bulk and other sheet properties. This will be achieved by increasing the effective dwell time of wet presses, impulse dryers, and displacement de-watering devices (FY 99-00 focus is on wet pressing). Technical approaches to implementing the technology may be evaluated. PROJECT BACKGROUND: This is a new project for FY 99-00. The P.I. was changed from D. Orloff to P. Phelan in November, 1999. MILESTONES: The following tasks were presented at the Fall ‘99 PAC meeting. 1. Survey the literature on shoe press technology with an emphasis on current and projected use in the manufacture of grades other than linerboard and corrugating medium. (Completed by first quarter, FY 99-00) 2. Determine how existing shoe press technology can be best utilized in the manufacture of bulk sensitive printing and writing grades. (Rejected by the PAC) IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 130 Project FO40 Status Report 3. Determine if there is justification for developing a “Super-Long Nip HighLoad” press. (Completed by fourth quarter, FY 99-00) Note: The PAC suggested that this task be the focus for fiscal year 99-00. 4. Identify and resolve the technical barriers to the development of a SLN-HL press. (Completed by fourth quarter, FY 00-01) 5. Design, build and evaluate a prototype with respect to technical and economic factors important to commercialization. (Completed by fourth quarter, FY 01-02) DELIVERABLES: The PAC agreed on experiments to justify the use of a SLN-HL press for a copy paper grade. Initial experiments will look at the impact of a third position SLNHL press on water removal and sheet property development. The experiments should be completed and reported to the PAC at the Spring ‘00 meeting with an economic analysis of the impact. The PAC will recommend continuation or termination of the project based on the results. STATUS OF GOALS FOR FY 99-00: The goal for the Spring, ‘00 PAC meeting is to complete initial experiments justifying a SLN-HL third press for copy paper at two refining levels and report the results. Experiments were completed in January, 2000. Proposed goals for completion by the end of the fiscal year are to conduct the same experiment with the SLN-HL press in the first press position and write a member report on all of the results obtained in 1999-00. Production of Formette sheets are to begin in February, 2000. SCHEDULE: Task Descriptions (example) I. Literature Survev 3a. Justification at Third Press 3b. Justification at First Press(proposed) 4. Identify & Resolve Technical Barriers 5. Build & Evaluate Prototype 6. Write yearly report (proposed) 1999 July - Sept ---m---m----X 1999 Ott - Dee 2000 Jan - Mar 2000 Apr-Jun IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) j 2-7 12-E j Project 131 FO40 Status Report SUMMARY OF RESULTS: Experiments were completed using a copy paper furnish at two levels of freeness and 40% ingoing solids. Five different cases of peak pressure and impulse, for the same basic shoe press profile, were tested for each freeness level. The following results were obtained. l The 400 ml CSF furnish was predominantly flow-controlled. l The 500 ml CSF furnish was predominantly pressure-controlled. l Density is a linear function l Tensile apparently l l SUMMARY l l l l is no apparent (outgoing with increasing density, increases Brightness and opacity scatter in the data. There of water removal are apparently trend for Sheffield unaffected, solids). but the change is slight. but there is a large amount of roughness. OF KEY CONCLUSIONS: As a third press a SLN-HL press can remove more water at higher impulses than a “standard” shoe press. There is a corresponding increase in density, but the increase does not appear to affect optical properties. Estimated years. payback periods for a runnability-limited Estimated years. payback periods for a dryer-limited machine machine range from 4 to 46 range from 0.5 to 6.2 There may be more potential for increased water removal, while maintaining bulk, if the SLN-HL press was in the first position, before the sheet is densified by roll presses. Additionally, a SLN-HL press, potentially, may be used to reduce the number of press nips, reducing rewet. DISCUSSION: Experimental Plan and Procedures Furnish: The furnishes made for this trial simulate copy paper and were made using the Formette Dynamic at 900 m/min. The pulp was a blend of 75% hardwood and 25% softwood. Each pulp was refined to two freeness levels, 400 and 500 ml CSF and then blended. Table 1 lists the additives used. Plan: The Formette sheets were cut into 5-inch diameter test samples for pressing. Table 2 shows the MTS conditions used. After pressing the samples were dried under constraint and tested as shown in Table 3. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project FO40 132 Status Report Figures 1 and 2 show examples of the type of generic shoe press pulses that were used. All of the pulses had the same relative shape, except that the depressuriation part of the pulse was the same for all cases. The exact shape was determined when the MTS is programmed. Cases 1, 2 and 3 had the same peak pressure, but the dwell time was varied to correspond to the different shoe lengths. Cases 1, 4 and 5 had the same impulse, but different dwell times and peak pressures. Results The outgoing solids levels, for each furnish, are shown in Figures 3 and 4. For the 400 ml CSF furnish, water removal is flow-controlled, as shown by the increase in solids with increasing impulse but with little change with increasing pressure. However, the 500 ml CSF furnish is predominantly pressure-controlled with a slight increase in solids with increasing impulse and a greater increase with increasing pressure. Figures 5 and 6 further indicate that the 500 ml CSF furnish is pressure-controlled by showing a stronger influence of pressure than impulse on density. On the other hand, the density for the 400 ml CSF furnish has little dependence on pressure, but increases with increasing impulse. Both furnishes have a linear relationship between density and outgoing solids as shown in Figure 7. Most strength properties increase with increasing density. However, the paper used in this experiment had low overall strength for reasons unknown at this time. Figure 8 shows that the geometric mean tensile index for both furnishes only has a slight increase with increasing density. Also plotted, for comparison, are the results of a commercially available copy paper tested at the same time. The MIT fold data were so low that the results are meaningless, most counts were less than 10 and the highest was 14. None of those data are in this report. In Figure 9 the Gurley porosity data for the 500 ml CSF furnish increases with increasing density and are relatively low indicating an open furnish. For the 400 ml CSF furnish the data are more scattered and have no correlation with density. Data for Sheffield roughness, IS0 brightness, and opacity do not show any significant trends or have large scatter. All of the data are summarized in Table 4. Conclusions Scientific Operating as a third press, a Super-Long Nip High-Load press must operate at high pressures and impulses (cases 2 or 3) to increase water removal. For a flow-controlled furnish this will result in a substantial increase in sheet density, while for a pressurecontrolled furnish the density increase is minimal. Data from this experiment are inconclusive about the detrimental effect of increased density on opacity, but there are indications that the change is insignificant. Therefore, the use of a Super-Long Nip HighLoad press would result in a net benefit based on increased productivity. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 133 FO40 Squeezing more water out of the web by using high increasing productivity. For a bulk sensitive furnish, without increasing density. Using a Super-Long Nip may accomplish this with a long dwell, low pressure Status Report pressures has economic benefits by it would be better to remove water High-Load press as a first press pulse (cases 4 and 5). Economic Using the IPST Economic Model developed for project ROCIT, two economic scenarios were investigated. Both assumed a papermachine producing 75.2 gsm copy paper. The base case assumed a machine speed of 3300 fpm and a production of 842 FMT/day. For each of these scenarios, it was assumed that the press section could be rebuilt at three levels of capital investment, $5M, $lOM, or $15M. The differential increase in gross profits was calculated for each improvement in capibilities. The estimated capital cost of rebuilding the press section was divided by the differential profit to obtain an estimated payback (in years). The first scenario was for a runnability-limited machine with 3.0% downtime due to sheet breaks. Higher ougoing solids from the press section would reduce the number of breaks resulting in higher overall productivity. Table 5 lists the estimated payback periods calculated. The second scenario was for a dryer-limited machine with press section outgoing solids of 42.0%. Using the rule-of-thumb that for a dryer limited machine, each percentage point of increased solids out of the press section results in a 4% increase in machine speed, payback periods were calculated for up to 4-percentage points of increased dryness. The results are in Table 6. Suggested Future Work The potential for removing water while maintaining bulk by using a Super-Long Nip High-Load press as a first press should be investigated by repeating this experiment with a lower ingoing solids level and no prepressing. Without the prepressing to densify the sheet, a low pressure SLN-HL press could gently squeeze out water while maintaining bulk. To maximize water removal, it is suggested that the experiment use double felted pressing with a grooved platen on the top and bottom. Another potential benefit of using a Super-Long Nip High-Load press would be to eliminating press nips. Each nip has the potential to introduce rewet as the web exits the nip. Eliminating nips has the potential of increasing the final outgoing solids from the press section, even with the same total impulse. Using a Super-Long Nip High-Load press with an optimized profile may reduce the press section to a single nip with resulting capital and space savings. A proposed experiment would compare two nips to a single long nip. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project 134 FO40 Status Report Figures 1000 900 800 1000 900 800 -: 700 g600 0.01 0.02 0.03 Time (s) 0.04 0.05 48 46 0 t 38!." 0.00 [y), ?tk I “‘: 0.04 0.08 ..‘: Impulse 0.12 ..‘: 400 ml CSFl 500 ml CSF 3 0.03 Time (s) 0.05 0.04 0.06 0.16 ‘..: 0.20 ..’ .E42-l 3%40-- : 0 : f 0.68-l 0.66-0.64-l 0.08 Impulse 0.12 4 Pressure (MPa*s) 0.70-- 0.62!...1..-1...:-.-:...:...~ 0.00 0.04 38:.~~.~..~~:..~.:...~~.~.~~.~..~~~.~ 3 0 1 2 I- 0.24 Figure 3. Outgoing solids for cases 1, 2, and 3. g.E E -g 5 :g 0.02 Figure 2. Pulses with same impulse. 44 i 40 0.01 t .g 42-F 9 1 41 51 j--Case 0 0.06 Figure 1. Pulses with same peak pressure. +j 0 m +-Case I-Case f\ -$$- 700 9600 0 g - --- 0.16 0.20 0.24 (MPa*s) Figure 5. Density verses impulse for all cases. 5 6 7 (MPa) Figure 4. Outgoing solids for cases 1, 4, and 5. g .g $ u i! :g u 0.70-L 3 0.62t,...~....1....:..-'~....~..'.:....! 0 1 2 0.68-l 0.66-l 0.64-l Pressure 3 4 5 6 (MPa) Figure 6. Density verses peak pressure for all cases. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 7 Project s E 9 9 s.-3 2 0” 135 FO40 Status Report 0.74 0.72 U z .-5 : .-G c s 0.70 0.68 0.66 0.64 0.621 1 . . 40 42 : . ; . . 44 46 Outgoing Solids (%) 1 . . 1 ! 48 50 Figure 7. Density verses outgoing solids for all cases. 45 r .- ‘;3; 40 z2 35 L ii530 2 z.- 25 e2!20 I; IOG z 8 ; .-z iii 5 n L,2 6 Conditioned Density (kg/m*3) Figure 8. Geometric mean tensile index verses density for all cases. ’ ‘. : ‘. : ‘. f.. . :. : “. 9-a-- H 7 -- H + @ 65- ; 4-. ii pzzF1; &I 3-- I '062. tI . 0.66 . 14 . . . II . . . . 0.70 Conditioned 0.74 0.78 I . . . 0.82 0.86 Density (kglm”3) Figure 9. Gurley porosity verses density for all cases. Tables Table 1. Sheet additives. Amount Additive 15% of dry weight PCC 1 Optical Briqhtener 1 2 Ibs./ton 1 3 Ibs./ton 1 AKD March I 12 Ibs./ton 2 Ibs./ton Retention Aid IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 136 Status Value Condition . Platen Beloit J 1 10, 20, and 30 inches 1 Nip Lenqth w 3000 ft/min (914 m/min) I Machine Speed Basis weight 67 gsm OD Single-felted wet pressing simulated Pulses Standard’ Press load (Case 1) 4500 pli (788 kN/m) Freeness 400 and 500 ml CSF 40% Ingoing solids Preheat none Fine Paper felt, preconditioned Felts Ingoing Felt Moisture 20% 30 Repeats per Case Table 3. Testing. Non-destructive Tests Gurley Porosity (T460) Opacity (T519) IS0 Brightness (T534) Sheffield Roughness (T538) Table 4. Average I CasF! -v-v Freeness(ml) Dwell Time (ms) PeakPressure(MPa) lmpulse(MPa*s) Basis Weight (gsm) ingoing Solids (%) Outgoing Solids (%) Soft Platen Caliper(pm) Cond. Density(kglm"3) Cond. Bulk(mA3/kg) Gurley Por. (s/lOOml) Top Sheffield (ml/min) Bot. Sheffield (mI/min) MDTensile lndex(N*m/g) CDTensileindex(N*m/g) GM Tensile lndex(N*m/g) IS0 Brightness . I Ooacitv Table 5. Estimated Breaks (4OO 30. 25. 20. 15. 10. 400 17 6.17 0.062 66.74 39.02 44.37 105.7 0.671 1.493 8.09 1778 2190 34.83 15.69 23.38 87.72 I 88.86 payback shoe press Destructive Tests MD and CD Tensile, 4-inch span (T494) MD and CD MIT Fold (T511) OD weight (basis weight, bulk) values for physical Ii Report properties. 12131415111213l4 400 400 35 52 6.19 6.21 0.126 0.197 66.21 66.69 39.29 39.47 46.73 48.55 100.9 97.8 400 33 3.18 0.063 67.05 39.45 43.95 105.5 0.696 1.436 7.87 1763 2151 36.48 16.38 24.44 87.92 I 88.23 01674 1.483 6.73 1780 2077 32.05 15.70 22.43 88.27 89.07 I 0.723 1.385 7.53 1692 2078 34.93 14.83 22.76 88.47 88.69 I I I 400 49 2.06 0.063 66.85 39.67 43.34 106.8 0.667 1.504 5.13 1519 1732 31.14 13.68 20.64 88.59 89.36 I 500 17 6.20 0.063 68.41 40.13 47.67 108.3 0.669 1.495 4.88 1598 1959 29.37 12.64 19.27 88.66 88.96 I 500 35 6.19 0.127 68.02 40.49 49.21 106.5 0.675 1.483 5.00 1648 1958 29.18 13.21 19.63 88.51 88.84 500 52 6.20 0.196 65.90 40.69 49.54 102.3 0.683 1.465 4.65 1647 1928 30.00 13.65 20.24 88.10 1 88.15 periods (years) for a runnability-limited Capital Costs 5,000 k$ 10,000 k$ 154. 77. 52. 39. 309 15’4 IO.3. 77. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 500 33 3.17 0.063 67.35 40.34 45.66 110.3 0.649 1.543 3.78 1750 1904 28.71 12.76 19.14 88.63 1 88.54 151Com.I 500 49 2.05 0.063 68.34 40.39 44.31 114.0 0.637 1.570 3.64 1786 1912 30.58 12.37 19.45 88.53 I 88.70 machine. 15,000 463 23’2 15’5 11’6. k$ # 72.71 94.1 0.828 1.208 8.16 1492 1347 64.43 26.88 41.62 84.15 I 89.11 I Project 137 F040 Table 6. Estimated 1 Solids Increase (% Points) 0 1 I 2 I I 3 payback 1 I I I periods (years) for a dryer-limited machine. Capital Costs 5,000 k$ lb,000 k$ 21. 10. 07. 05. I I I 41. 21. 14. IO. ~~ I ~~ I IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 15,000 I 62. 31. 21. 16. k$ I I 138 139 EXTENDING HIGH INTENSITY WATER REMOVAL PRINCIPLES INTO THE DRYER SECTIONS STATUS REPORT FOR PROJECT F041 Fred Ahrens (PI) Paul Phelan lsaak Rudman Edward Lindahl March 8 - 9,200O Institute of Paper Science and Technology 500 10th Street, N.W. Atlanta, Georgia 30318 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 140 141 Project F041 DUES-FUNDED PROJECT Status Report SUMMARY Project Title: Extending High Intensity Water Removal Principles into the Dryer Section Project Number: F041 PAPERMAKING PAC: Project PAC Staff Principal Investigator: Co-Investigators: Research Support Staff: Subcommittee FY 99-00 Budget: Allocated Time as Matching Funds: Allocation: Principal Investigator: Co-Investigators: Research Support Staff: Supporting Research: Students: External (Where Matching F. Ahrens P. Phelan, Lindahl I. Rudman Babinsky, Kaufman, $86,000 OY0 15% 25% 20% Is Used): N. Alaimo None (M.S.) RESEARCH LINE/ROADMAP: Line #7 - Increase paper machine over ‘97 levels via focus on breakthrough forming, dewatering, concepts [faster drying] PROJECT OBJECTIVE: Demonstrate/verify a high intensity the basis for a feasible, high productivity, capital/space/energy Provide the data and understanding needed for development. PROJECT BACKGROUND: Watson productivity by 30% and drying drying concept that provides effective dryer system. New DFRC project in FY 99-00 (started July 1999) MILESTONES: a Identification of promising conditions and configurations intensity drying experiments [by Mar. 20001 for initial laboratory high Completion of high intensity drying experiments to demonstrate the drying rate potential and guide concept development [First phase by Mar. 20001 0 Development fundamentals of promising concept(s) [ based on systematic application of and analysis of experimental results] [review - Mar. 20001 0 Development of plan to refine the concept(s) [Q4, FY 99-001 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 142 Project F041 0 Identification l Completion of equipment of preliminary DELIVERABLES: 0 High intensity Definition of preliminary assessment [Q4, FY 99-001 [Q4, FY 99-001 [Mar. 20001 simulation and technical analysis technical and economic analysis of most promising Data and understanding issues/requirements technical/economic dryer concept(s) Results of laboratory [Q4, FY 99-001 Results and clothing Status Report concept needed and applications for scale-up of promising [Q4, FY 99-001 [FY 00-011 [FY 00-011 Model of PAPRlCAN’s high intensity (impingement) drying concept collaboration with PAPRICAN] {proposed new task} STATUS OF GOALS concept(s) [via proposed FOR FY 99-00: Identify promising Fall ‘991 conditions and configurations for drying experiments: [completed Conduct high intensity drying experiments to demonstrate the drying rate potential and guide further concept development: [ First phase completed] Develop analysis promising concept(s) based on systematic application of fundamentals of experimental results: [Progress review at March PAC meeting] Develop plan to refine the concept(s): Submit pre-proposal to Agenda [Spring 20001 2020 Capital Effectiveness Program: [Nov. 19991 Model of PAPRICAN’s high intensity (impingement) dryer concept [via proposed collaboration with PAPRICAN]: [proposed new task - under development] IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) and 143 Project F041 Status Report SCHEDULE: I Task Defineinitial experiments . Perform initial experiments Develop promising concept Plan for concept refinement I Identify equipment/clothing issues Preliminary technical/economic evaluation Data for scale-up Model of PAPRICAN concept I SUMMARY I FY 99-00 Ql Q2 X Q3 X X X X X X _ Q4 I FY 00-01 Ql Q2 Q3 X 1 I I I Ixl~lxlxIxI OF RESULTS: Initial MTS experiments demonstrated very large (heat input based) drying rates (I 000’s of kg/hr/sq m) for simulated copy paper and linerboard, over an applied pressure range of 10 to 1000 psi (applied in one or two 20 ms pulses), with surface temperatures of 149 to 233 OC, for initial solids levels of 50% and 70%. In response to an inquiry from PAPRICAN, the PAPRICAN high intensity (impingement) been developed. Focus is on development computer model of the concept. SUMMARY a preliminary proposal for interaction with drying effort (for printing grades) has and pilot machine verification of a OF KEY CONCLUSIONS: According to the initial laboratory simulation results, it should be possible to accomplish (via multiple cycles of intense heat input and vapor removal on a high temperature cylinder) drying equivalent to that normally associated with several/numerous dryer cans and their open draws. l A wide range of “shoe press” loadings and surface temperatures appear to be useful, offering the opportunity to identify a technically and economically feasible design. The vapor removal development. , X X X Preliminary/simplified concept definition: Heat input via multiple long, moderate pressure nips, with interspersed vapor removal areas, on high temperature cylinders. The sheet would be restrained by a suitable fabric or felt. 0 Q4 and sheet restraint strategies still need attention IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) and 144 Project F041 DISCUSSION and ADDITIONAL Status Report DETAILS INTRODUCTION The dryer sections prevalent for most grades of paper and board are based on use of contact heat transfer from numerous steam heated dryer cans and movement of large volumes of heated air for vapor removal. Dryer sections have undergone evolutionary changes in recent years [e.g., recent developments include the single-tier configuration and the incorporation of impingement units at selected stations in the dryer section], but are still limited to rather low overall drying rates, resulting in very long, expensive systems and only modest-to-good energy efficiency. In view of the industry need for increased productivity, improved capital effectiveness, and reduced energy use, it is timely to seek to apply the principles and techniques associated with high-intensity heat transfer and water removal (including, but not limited to, adaptations of technology developed for impulse drying) to the challenging task of making beneficial step increases in drying rates for paper and board. In this project we intend to bring together available technical tools and new ideas for implementing high-intensity heat transfer and water removal principles, in order to overcome the heat and mass transfer impediments associated with current approachesto drying. The goal of the work is to develop a feasible high-intensity drying concept having the potential to make the paper machine dryer section at least an order of magnitude smaller, and significantly more energy efficient, than a conventional dryer section of the same capacity. Preliminary analysis and literature review in Fall 1999 indicates that it should be possible to achieve drying rates well in excess of those offered by the best currently available technologies (impingement drying and Condebelt), and points toward a particular implementation concept. The benefits from successful commercial implementation of expected project results are as follows. For retrofit of existing machines, it will be possible to install significant extra capacity in a smaller space and at lower capital cost than for comparable conventional capacity. For new machines, greatly reduced space and capital cost, and increased energy efficiency, will be achieved; the very low drying rates and enormous dryer sections of today will be avoided (e.g., using 2 to 4 larger, high drying rate cylinders rather than 40 to 100 standard dryer cans). TECHNICAL BASIS FOR SUCCESS Conventional paper drying involves low-pressure (cl psi) thermal contact between the sheet and numerous steam-heatedcylinders for heat input, with low intensity convective mass transfer (drying via evaporative cooling) in the draws between these cylinders (as well as on the lower cylinders, in single-tier dryer sections). Although paper drying is always a combined heat and mass transfer process, in the conventional dryer section there is only a relatively minor amount of evaporation while the sheetis covered by a fabric on the dryer cylinder. Therefore, conventional paper drying can be viewed as a cyclic process [cycles of heat transfer to the sheet (heat-up) followed by mass transfer from the sheet (accompanied by cooling of the sheet)]. The use of a hot air impingement unit over the fabric-covered sheet (on the dryer cylinder) would introduce a greater component of simultaneous heat and mass transfer to the drying process, via increasing the extent of vapor removal while the sheet is on the dryer cylinder. Additionally, the unit would increase the total rate of heat transfer to the sheet.Consideration of these principles (cyclic vs. simultaneous heat and mass transfer) should be useful in helping to establish and discuss a preliminary concept for a high intensity dryer. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F041 145 Status Report Mention of a key concept relevant to development of a high intensity dryer is believed important, at this point. That is, there are two distinct modes by which mass transfer from a paper web can take place. The dominant one in nearly all current drying processesis diffusion of vapor within the web, coupled with convective diffusion from the web surface to the surroundings. The driving force for this process is a vapor partial pressure gradient. This mode occurs when the web temperature is below the ambient boiling point (which is the case in the conventional dryer section). The web temperature is, of course, influenced by heat transfer rate. The resistanceto mass transfer in the web and in the adjacent boundary layer is a major factor limiting drying rate. The other mode for mass transfer can be called bulk vapor flow. Here, the driving force is a total pressure gradient. It occurs if the web temperature exceeds (even slightly) the ambient boiling-point. In this case, the only resistance to mass transfer is inside the web, and it is inversely related to the vapor permeability. This resistanceis typically negligible compared to the-diffusional resistance-mentioned-above.Therefore, it tends to be only heat transfer factors that limit the drying rate. However, very intense heat transfer to the web is needed to enter into this regime (which can, therefore, be defined as the high-intensity drying regime). An early description of the transition from conventional drying to high intensity drying can be found in Ref. 1. The Condebelt process is the only known commercial process in which the bulk flow mode of vapor removal occurs (in it, the vapor leaving the sheet condenseson a cooler adjacent surface). A consequenceof the bulk flow mode of vapor transport is that the energy transport within the web is enhanced by intense ‘heat pipe effects’ (evaporation-flow-condensation). This implies that temperature differences in wet portions of the web tend to be small (provided the permeability does not become too low). As a point of technical interest, one other statement can be made concerning the drying process. That is, pure impingement drying (i.e., simultaneous convective heat and mass transfer) can, at best, provide wet web temperatures equal to the ‘wet bulb temperature’, which can only approach the ambient boiling point. Thus, although relatively high drying rates may be possible with this technique, it is not an example of high-intensity drying, as defined here. This statement is not intended to imply that impingement heat and/or mass transfer should be disregarded as a potentially useful tool in achieving a successful high intensity dryer concept. It might be noted that example calculations of mass transfer rates for both modes, and other information on transport phenomena in the web, were included in the Fall 1999 Project F041 presentation to the Papermaking PAC. This material will not be repeatedhere, but it is believed to support the preliminary concept definition given below. Although, the high intensity mode has excellent potential for increasing drying rate, its limitations should be noted. If web vapor permeability does become small (perhaps due to the web moisture content being high and/or due to high levels of compression), the internal vapor pressure can become large (relative to the ambient pressure).In this case, if the web has insufficient restraint (relative to its cohesive strength), delamination can occur. A less serious phenomenon (liftoff, a loss of thermal contact between the sheet and the hot surface) is also believed possible under some conditions. A further limitation to high intensity drying is that a dry layer tends to develop in the sheet (adjacent to the hot surface); its thickness increaseswith contact time, tending to gradually reduce heat transfer rate (2). A successful dryer concept would need to manage this effect. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 146 Project F041 PRELIMINARY Status Report CONCEPT DEFINITION Since mass transfer impediments can be removed by operating in the high-intensity drying mode, concepts compatible with that regime have been given primary attention. With heat transfer-related factors tending to control drying rate in the high-intensity drying regime, they represent a major focus of the preliminary concept definition effort. The two most important factors for high heat transfer rate to the web are large temperature difference (driving force) and low thermal resistancebetween the hot surface and the web. These factors point toward use of high-surface-temperature dryer (using an as yet unspecified heat source) and higher than conventional mechanical pressure (to increase the contact heat transfer coefficient). Loading considerations will, of course, place a limit on the combinations of pressure level and contact area at pressurethat are feasible. A further, direct restriction on mechanical pressurelevel could come from either its impact on sheet bulk or its related impact (via web vapor permeability) on the possibility of blistering/delamination. The phenomenon mentioned earlier concerning a dry layer, of continuously increasing thickness, developing during the high intensity drying process suggeststwo other considerations. It may be desirable to interrupt the regions of intense heat transfer periodically to allow some z-directional redistribution of moisture to occur. It may also be desirable to alternate heat input from one surface of the sheetto the other at reasonable intervals. The first of these points is compatible with the use of shoe press-like devices at intervals around the cylinder. In this context, the term ‘shoe press-like’ refers to the general configuration, but is certainly not intended to imply that the pressurelevels and clothing employed should be similar to those associatedwith shoe pressesfor wet pressing or impulse drying. In fact, it is believpedthat the best range of pressure for high intensity drying may be we1.1below that for wet pressing. The purposes are quite different. Ii the drying application, we are not seeking to squeezewater from the sheet,but to improve heat transfer. It can be expected that vapor removal via the bulk flow mechanism (i.e., flow from the sheet into the felt) would occur while the sheet is being intensely heated in the shoe press nip zone. After the sheet leaves the nip, it may evaporatively cool to temperaturesbelow the ambient boiling point. In this between nip region, an appropriate sheet-restraining fabric may be needed to provide restraint against shrinkage effects, as well as to promote acontinuation of heat transfer to the sheet (possibly at a reduced rate due to the reduced applied pressure).Also in the between nip region, impingement could potentially be employed to increase the local mass transfer (drying) rate. INITIAL HIGH INTENSITY DRYING EXPERIMENTS As suggestedby the above description of high intensity drying principles, and the preliminary concept definition, adaptation of shoe pressing techniques, coupled with higher than conventional dryer surface temperatures, should offer an opportunity to achieve large heat transfer .anddrying rate increases.The experiments described here were intended to provide an initial, relatively broad, overview of the effects on water removal of “shoe press” variables (pressure level, number of pressure pulses [ 1 vs. 2]), surface temperature, initial sheet solids level and grade. The particular objective of these initial (heated MTS press) experiments is to demonstrate and quantify the drying rate (heat transfer aspects) potential, and to guide the development of high intensity drying concepts that integrate multiple “shoe pressing” nips (to intensify heat transfer) and higher than conventional surface temperatures into the drying process. A limitation of this initial work is that short dwell times between pressure pulses were not achievable on the MTS [2 set was shortest]. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 147 Project F041 Experimental Status Report Plan Test Conditions: Samples: These were 4” diameter circular discs cut from handsheets. 1. Linerboard-like sheets: l 205 gsm l 100% unbleached softwood kraft l 500-550 CSF l 50% initial solids (via pressing to at least 45%, followed by air drying, if necessary) Initial sheet solids: 2 levels: 50% and 70% Surface Temperature: 2 levels: 300 F and 450 F [ 149 C and 232 C] Dwell times: l single nip: 20 ms l multi-nip: 2 hits at 20 ms each, with 1 level of dwell (minimum possible) in between [about 2 set] Applied Pressure Pulse: 3 levels: 10, 100, 1000 psi Repeats: 10, for each test condition 2. Copy paper-like sheets: l 70 gsm l 75% BHWKl25% BSWK l 450 CSF l 50% initial solids (via pressing to at least 45%, followed by air drying) Initial sheet solids: Mainly 50%; limited 70% solids runs at one pressure Surface Temperature: 2 levels: 300 F and 450 F [ 149 C and 232 C] Dwell times: l single nip: 1 level: 20 ms l multi-nip: 2 hits at 20 ms each, with 1 level of dwell (minimum possible) in between [about 2 set] Applied Pressure Pulse: 3 levels: 10, 100, 1000 psi Repeats: 10, for each test condition l l Procedures: Each sample was supported on a “dry” felt, which was resting on the lower platen of the MTS press. The dry felt pressesthe sheet against the hot platen, serves as a water receiver, and permits some vapor venting to occur. In a limited number of runs, fine thermocouples were placed between the sheet and felt, to provide some insight on the heat transfer process. Experimental Results and Discussion The above experimental program has recently been completed. The primary results and observations are presented here. Analysis and interpretation of these results is still in progress, but some preliminary discussion is included. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 148 Project F041 StatusReport For each experimental condition, the water removal data were used to calculate a drying rate based on residence time in the shoe press nip period(s) (i.e., either 20 ms or 40 ms). The drying rate defined in this manner representsan upper bound on the actualdrying ratethat occurred in the experiment, or on that achievable in practice. It representsthe (heat transfer based) drying rate for an ideal situation in which vapor removal (evaporation) is instantaneous. In effect, this rate definition also corresponds to the assumption that the sample is in good thermal contact with the hot surface only during the pressure pulse(s) (perhaps not always the case in practice). The drying rates, as defined above, are presentedin Figures 1, 3 and 4 [some limited additional data, for the bleached, 70 gsm samples, at 10 psi peak pressure and 70% initial solids, appear in Table 11. For the bleached, copy paper-like sheets,the total water removal was sufficiently large as to make it interesting to present the final sheets solids data, as well (see Fig. 2). Over the entire range of the experimental program, the nip residence time based drying rate ranged from 1100 to about 9700 (kg/h/sq m)’ . For reference, note that typical ‘TAPPI drying rates’ are in the 15 to 30 kg/h/sq m range, impingement units provide rates of about 60 to 110 kg/h/sq m, and Condebelt rates can likely be as high as 150 kg/h/sq m. The actual drying rate based on total residence time in the dryer, comprising both periods in the shoe press nip(s) and time between/after nips [e.g., available for vapor removal from the web], will be less than that based on nip residence time, alone. The relationship is: DR actual = DRnrt * NRTI TRT where: DR actual = actual - drying - rate D&l rt =nip- residence - time - based - drying - rate TRT=NRT+ANRT . TRT= total residencetime NRT= nip residence time ANRT= after-nip residence time (= between nip time, for a multi-nip dryer) The rates presented in Figs. 1,3 and 4 are all in the range of about 10 to 100 times greater than the actual rates typical for either impingement drying or Condebelt drying. Therefore, if the time needed for vapor removal (ANRT) can be kept to less than about 10 to 100 times the nip residence time, respectively, the actual drying rates for the high intensity dryer concept envisioned here should exceed those for the fastest currently available technologies (impingement drying and Condebelt). Intuitively, it would seem likely that ANRT/NRT ratios considerably shorter 10 should be adequatefor vapor removal. This is one topic meriting investigation in the near future. The amount of heat transfer occurring (and contributing to the actual drying rate) via post-nip contact also needsto be investigated. ’ Note: dividing drying rate in (kg/h/sq m) by 4.88 converts it to (lb/h/sq fi). IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F041 149 Status Report Inspection of the drying results in the cited figures reveals the following: l l l l l The rate increases with surface temperature and applied pressure.These results are compatible with the elementary heat transfer concepts of thermal driving force and of pressure-dependentcontact resistance,respectively. The rate decreasesfor two nips, compared to one nip. This is compatible with the existence of a dry layer near the hot surface impeding heat transfer in the second nip. The rate is less for a 70% ingoing solids level than for 50% ingoing solids. This is compatible with the expectation that drier paper is a better insulator than wetter paper. The drying rate for the 70 gsm web is greater than that for the 205 gsm web. This may be due to a greater part of the heat input being needed for sensible heating, in the case of the heavier web. For the bleached, lighter weight sheets,the high intensity drying experiments led to very high final solids levels (Fig. 2). Another aspect of the experimental program involved tracking the sticking and blistering tendencies of the paper samples. A summary of these observations is provided in Tables 1 and 2. The first point to be made about these results is that blistering/delamination was found to occur at only the most intense condition (232 C, 1000 psi), and only for the 205 gsm samples at 50 % initial solids. Thus, there seems to be plenty of opportunity to apply high intensity drying ideas without undue concern about delamination. Sticking was more frequently observed, but the practical significance of these observations is not yet clear. The tendency for sticking was greater for the linerboard than for the copy paper-like samples. The tendency was also greater for 50% initial solids cases than for 70% initial solids. In any of the double nip cases where only between nip sticking was observed, there is probably no cause for concern relative to potential applications, since such sticking would likely promote improved heat transfer in the between nip section. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 150 Proiect F041 Status Report Fig. 1. Drying Rate (based on Nip Time ) for Bleached, 70 gsm, 50 % Initial Solids Samples, 20 ms/nip E, 8000 I-* 149 C, 1 nip I+232 C, 1 nip I+ 149 C, 2 nips 1-E 232 C, 2 nips I 6000 5000 3 4000 .Ip 6 3000 2000 1000 0 IO 100 1000 Nominal Peak Pressure, psi 1 Fig. 2. Final Solids for Bleached, 70 gsm Samples 50% Nominal Initial Solids, 20 ms/nip --n-+- 232 C, 1 nip 149 C, 2 nips 60 50 IO 100 1000 Nominal Peak Pressure, psi IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 151 Project F041 Status Report Fig. 3. Drying Rate (based on Nip Time) for Linerboard, 50 % Initial Solids, 20 mship 205 gsm, 8000 E 7000 5 6000 : 5000 PC : 4000 -Cl- 2 3000 l5 2000 232 C, 1 nip 0 10 100 1000 Nominal Peak Pressure, psi 1 Fig. 4. Drying Rate (based on Nip Time) for Linerboard, 205 gsm, 70 % Initial Solids, 20 mship 5 $ 2000 0 IO 100 1000 Nominal Peak Pressure, psi IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 152 Project F041 Status Report Table 1. Bleached 70 gsm Samples: Degree of Sticking Degree of Stickinq varied edge varied 50% 50 50 Peak Pressure IO psi 100 1000 1 1 50 50 50 232 232 232 IO 100 1000 6257 8439 9673 severe very slight slight 2 2 2 50 50 50 149 149 149 IO 100 1000 3663 3605 4760 moderate between hits between hits 2 2 2 50 50 50 232 232 232 IO 100 1000 5077 5818 6890 none between hits none 1 70 149 IO 2961 varied 1 70 232 IO 3740 none 2 70 149 IO 2183 none 2 70 232 10 2775 none 1 Nominal Initial Solids Drying Rate ka/h/sq m 4290 6610 6769 Surface Temperature 149 c 149 149 nips 1 1 1 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 153 Project F041 Table 2. Linerboard: nips Nominal Initial Solids 50% 50 50 Surface Temperature 149 c 149 149 Status Report Degree of Sticking Peak Pressure IO psi 100 1000 Drying Rate ka/h/sq m 2069 3198 4082 Degree of Sticking severe moderate mod/sev 1 1 50 50 50 232 232 232 10 100 1000 3321 5385 7238 none slight/mod hi severe* 2 2 2 50 50 50 149 149 149 10 100 1000 1800 2218 4506 slight varied severe 2 2 2 50 50 50 232 232 232 IO 100 1000 3076 3687 6258 slight severe severe* 70 70 70 149 149 149 IO 100 1000 1412 2381 2628 none varied varied 70 70 70 232 232 232 IO 100 1000 2442 4421 3267 none very slight varied 2 2 2 70 70 70 149 149 149 10 100 1000 1102 1531 2382 none none varied 2 2 2 70 70 70 232 232 232 IO 100 1000 2320 2578 2938 none none between hits IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) *1 of 10 blistered *most shts blistered 154 Project FO41 CONCEPT DEVELOPMENT Status Report ISSUES AND PLANS The work to date has likely generatedmore questions than answers. Hopefully, it has also given some initial support for the value and direction of this project. The following list, which is neither complete nor prioritized, is intended to give an idea of the types of issues that need to be addressedas the project goes on. Work in the coming period to refine the concept presentedhere will be guided by these issues and questions. Number of shoe press nips per cylinder vs. length of nips Number of cylinders vs. size of cylinders Ratio of total nip length to total contact length Temperature-pressuretradeoffs (drying rate and paper properties aspects) Vapor removal requirements Clothing issues and desirable characteristics (interaction with sheet surface properties, sheet restraint strategies, operating temperatures, vapor removal, etc.) Equipment issues and limitations (including technical and economic impacts of mechanical pressureand surface temperature levels, etc.) Heat sources Paper properties impacts Fillers and additives impacts Economic issues, tradeoffs and feasibility PROPOSED NEW PROJECT TASK: Development/Verification of a Model of the PAPRICAN High Intensity (Impingement) Dryer Concept It is quite well known that PAPRICAN is active in the development of a ‘high intensity dryer’ concept (although they are not necessarily using the same definition as given in this report) involving impingement heat transfer to the sheeton a large diameter cylinder. The work is focused on drying of printing grades. Recently, they initiated communication with IPST regarding potential collaboration on one aspect of this work (development and verification of a computer model of the concept, which would be useful for scale-up, etc.). Since such collaboration could provide a way to gain understanding of the merits and potential of impingement techniques in the dryer section, it would seem compatible with the other aspectsof this DFRC project. A proposal for moving forward in this endeavor has been prepared (see APPENDIX A). Feedback from the Papermaking PAC on this proposed work would be appreciated. REFERENCES 1. Ahrens, F., Kartsounes, G. and D. Ruff, “A Laboratory Study of Hot-Surface Drying at High Temperature and Mechanical Loading”, PULP & PAPER CANADA 85(3):T63-67 (March 1984). 2. Ahrens, F. and Astrom, A., “High-Intensity Drying of Paper”, DRYING TECHNOLOGY 4(2): 245-270 (1986). IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 155 Project FO41 Proposed Interaction Status Report APPENDIX A with PAPRICAN in the Area of High Intensity Drying F. Ahrens, Water Removal Unit, IPST The purposes of this document are: to present an opportunity for a limited (but significant) interaction between IPST and PAPRICAN in an area of mutual activity (high intensity drying processes)and to outline an approach to implementing this interaction that provides benefits to both of our organizations (and their memberships). An overall Proposed Path Forward is also provided. Background There are several background points that contribute to this opportunity and/or to its value: PAPRICAN is working on a “high intensity dryer” concept involving inserting larger than conventional diameter steam cylinders (e.g., one or two), fitted with hot air impingement hoods, at selected points in the dryer section (for printing grades). The sheet is believed held to the dryer surface by adhesion rather than by fabric tension. In essence,the configuration is that of a Yankee dryer. A major purpose is to achieve a large increase in drying rate. They will begin evaluating the concept on their new medium/high-speed pilot machine this year. l l Several supplier companies (Beloit, ABB, Valmet) are promoting various ways of introducing impingement units into the dryer section, as well. IPST has initiated a DFRC project (F041) in July 1999 on “Extending High Intensity Water Removal Principles into the Paper Machine Dryer Section.” The scope of this project is currently broader than that of the PAPRICAN project (i.e., it is not limited to use of impingement techniques). In spite of that, it would be desirable to gain accessto the PAPRICAN work and, more generally, to develop a better understanding of the potential and the limitations of impingement techniques in the dryer section. l At the recent (5 Jan. 2000) ‘IPST Forum,’ one of the highly rated areasfor IPST improvement in the new decade was “improving interactions with our sister organizations” (of which PAPRICAN is an important one). The author has extensive experience in computer modeling of the Yankee dryer (for towel and tissue machine applications). The author’s current model is based on fundamental heat and mass transfer equations, together with various empirical correlations developed from published (non-proprietary) information. However, the model would likely require some modification to make it better suited to simulation of the drying of printing grades. In particular, the variation of sheet temperature and moisture in the z-direction can be (and was) neglected for tissue drying, but these variations should be included in modeling the behavior of heavier grades. Also. some of the empirical correlations used in the model may need to be adjusted. l Most importantly, Dr. Ivan Pikulik (responsible for the PAPRICAN project) has approachedthe author with the question as to whether the author would be able to work with PAPRICAN in applying/adapting the Yankee dryer model to IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 156 Project F041 Status Report the evaluation of their high intensity dryer concept and its further development and application. Part of the effort would be to use data from trials with the pilot machine high intensity dryers to verify and tune the computer model. A copy of Dr. Pikulik’s initial e-mail inquiry is in the Attachment. According to that email, it appearsthat the PAPRICAN management is inclined to be favorable to the proposed interaction. A subsequente-mail, received in early January of this year, reaffirms the PAPRICAN interest in pursuing this interaction. Apparently, they do not currently have accessto a suitable model, and cannot justify developing one of their own ‘from scratch.’ Dr. Pikulik has expressed interest in having long term direct accessto the version of the model resulting from the proposed collaboration. Proposed Approach In view of the background presented, it seems that there is a sound basis for mutually beneficial cooperative work in development of a computer simulation of the PAPRICAN high intensity hryer concept. The proposed approach would include the following steps: l l l l l Initial meeting at PAPRICAN to review the current Yankee dryer model, to inspect the pilot machine dryers, to establish requirements for the proposed high intensity dryer model, and to develop a more detailed project scope and timing. Create an initial high intensity impingement dryer computer model that is incorporates useful portions of the current Yankee dryer model, but is focused on the PAPRICAN application. Some of the new features believed needed were mentioned in the Background section. It is proposed that this work would be considered part of the DFRC F041 project. It is expected that this step would take about one man-month of effort. Conduct pilot machine trials at PAPRICAN to provide data sufficient for verification and tuning of the model. It is expected that this experimental work would be performed and paid for by PAPRICAN. However, the author would participate in planning the trials. Use the experimental data to verify and tune the model. This would likely be a joint effort, with the model tuning being done by the author. Make the model user-friendly, and train the PAPRICAN team in its use. Expected Benefits For IPST and its members, the potential benefits of this work include: l A deliverable (validated computer model) would be produced that could be made available to interested member companies. The model would be structured in a general way, to make it useful for scale-up and investigation of design variations. As a result, the model could later be applied/adaptedto the evaluation of the impingement ideas being promoted by the various equipment suppliers. l Access to the PAPRICAN pilot machine dryer system and associatedperformance data. The extent of this would have to be negotiated. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F041 157 Status Report Establishment of working relationships that could lead to additional cooperation in the water removal area. For example, if work on Project F041at IPST produces promising concepts beyond impingement, they could potentially be evaluated within the PAPRICAN pilot machine dryer system in the future. l For PAPRICAN and its members: l Availability of a validated computer model that adequately simulates their high intensity drying concept, without having to fully pay for its development. l Long term accessto the model so that they can apply it to the design and economic assessmentof future (commercial) installations. The form of this accesswould have to be negotiated. l Hopefully, establishment of working relationships with IPST that could lead to additional beneficial cooperation in the water removal area. Proposed Path Forward The following steps appear to be needed prior to beginning the proposed work: l l l l l l IPST management reviews this proposal and approves proceeding to next step (completed 21 Jan. 2000) Papermaking PAC F041 Subcommittee reviews this proposal (Jan. 2000) Initial meeting with PAPRICAN to develop more detailed scopeandtiming (Feb. 2000) Papermaking PAC reviews scope of proposed interaction (Mar. 2000) Veda Christmas (Legal) and Marsha Gill (Contracts) develop appropriate agreements with PAPRICAN (Mar. 2000) PAPRICAN and IPST reach final agreement and proceed with the work (April 2000). Acknowledgement: Thanks to David White for encouragement and for helpful comments on a draft of this proposal. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 158 Project F041 Attachment: Status Report e-mail from I. Pikulik From: [email protected] To: [email protected] cc: [email protected], [email protected], jrogers @paprican.ca Date: Tue, 2 Nov 1999 17:07:45 -0500 Subject: Cooperation on modeling high-intensity drying Dear Fred, During the Engineering Conference I told you that we would be very interested in using your Yankee simulation model to evaluate the drying rate during High-Intensity drying (HID) of printing papers. I have checked with Paprican management, and there is no objection to us establishing a co-operation on a project of our mutual interest. At this time I am not sure what form would could take such a co-operation. Presently, our pilot paper machine is in the start-up phase. By the end of the year we intend to start up the dryer section, consisting of two HID units. One of machines first tasks will be further to develop the HID technique and to demonstrate these units at commercial speeds. Our dryers are highly instrumented - much more than any commercial installation. This will allow us to examine the effects of various parameterson drying rates, paper quality and energy efficiency. This might be an excellent opportunity for you to verify your model. We might gain a tool to estimate the effect of an HID installed in a commercial machine, before the decision about such an installation is made. Can we find on this basis a room for co-operation? Perhapsit would be useful to meet and to discuss these possibilities in more details. Perhaps the PAPTEC Paper Week held in Montreal at the end of Janu,ary, beginning of February 2000 would be a good opportunity for such a meeting. Friday morning of that week we might demonstrate the machine for various visitors and it might be an opportunity to seeit in action. Ivan Pikulik IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 159 DRYING PRODUCTIVITY STATUS REPORT FOR PROJECT F021 Fred Ahrens (PI) Tim Patterson (PI) Hiroki Nanko Yulin Deng Shana Mueller Marcos Abazeri March 8 - 9,200O Institute of Paper Science and Technology 500 10th Street, N.W. Atlanta, Georgia IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 160 Project F021 Status Report 161 DUES-FUNDED PROJECT SUMMARY Project Title: DRYING PRODUCTIVITY Project PAC: Number: F021 PAPERMAKING Project PAC Staff Principal Investigators: Co-Investigators: Research Support Staff: F. Ahrens/T. Patterson H. Nanko, Y. Deng S. Mueller, M. Abazeri Subcommittee Reese, Time as Matching Funds: 70% Allocation: Principal Investigators: Co-Investigators: Research Support Staff: Supporting Research: Students: External (Where RESEARCH Beck $141,000 FY 99-00 Budget: Allocated Worry, Matching 25% 10% 50% Is Used): None Project 4253 (DOE-$400,000) LINUROADMAP: Line 7 - Increase paper machine productivity by 30% over ‘97 levels via focus on breakthrough forming, dewatering, and drying concepts [faster drying and improved runnability/quality] PROJECT OBJECTIVE: Understand and reduce the impediments (e.g., picking/sticking, surface deposits, cockle, sheet sealing) to the use of higher surface temperatures in the first dryer section. Provide fundamental knowledge and tools needed to design new technologies that will allow ultra high speed web transfer. PROJECT BACKGROUND: Project F021 was connected to DOE Project 4253 in Oct. 1998. 70% of F021 funding represents cost share to DOE project. MILESTONES: 0 0 Establish baselines on current operations: 0 Drying strategies and problems [Ql, FY 99-001 0 Roll/dryer can contamination/topology [Q3, FY 99-001 Develop experimental equipment: Contamination Test Stand (CTS) [Q3, FY 99-001 0 Web Adhesion and Drying Simulator (WADS) [Q3, FY 99-001 l IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 162 0 Status Report Develop empirical data and models: 0 contamination 0 picking/adhesion 0 cockle [Ql, FY 00-011 Develop and verify web transfer model for high speed machine [Q3, FY 00-011 Develop/evaluate: 0 surface conditioning l improved operation technologies drying strategies [Q4, FY 00-011 DELIVERABLES: Summary of baseline Experimental information equipment: [Q4, FY 99-001 CTS and WADS [Q3, FY 99-001 Understanding and data on factors influencing deposits/contamination [Ql, FY 00-011 Dryer section contamination Surface conditioning Web transfer STATUS operating problems strategies relative to picking/sticking, [Q3, FY 00-011 technologies cockle, cockle, and surface and surface [Q4, FY 00-011 model for high speed operation OF GOALS picking/sticking, [Q3, FY 00-011 FOR FY 99-00: Summarize and analyze the Questionnaire results to guide project work and to characterize the opportunity for improved productivity by grade [further analysis put on hold at Fall ‘99 PAC mtg.] Collect meaningful surface contamination [complete for 3 grades] and topological data from mills Design, construct and debug a Web Adhesion and Drying Simulator for use in investigating effect of sheet and drying conditions on sticking and picking [complete] Design and construct a Contamination Test Stand for use in providing contaminated surface coupons for use in the WADS [complete; working out contamination procedures] Use WADS to quantify sticking/picking [trial plan in place; ready to begin] for clean and contaminated Investigate the extent to which drying-related variables influence collaboration with the Project FO20 team] [first trial complete] IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) model surfaces cockle [in 163 Project F021 Status Report FY 99-00 Task Summarize/analyze Questionnaire results Collect dryer contamination/topology info. Develop experimental equipmt: WADS, CTS Web adhesion/picking experiments Cockle experiments Develop web transfer model Develop/evaluate improved drying strategies Develop/evaluate surface conditioning technologies SUMMARY Ql Q2 Q3 X X X X X X X X X X FY 00-01 Q4 Ql 7. Q2 Q3 Q4 X X X X , X X X X X X X X X X X OF RESULTS: Questionnaire results indicated that picking/sticking, surface deposits, and cockle are believed to be the biggest problems influenced by drying strategy in the first dryer section, and confirmed that some degree of temperature graduation is used on most machines, regardless of grade. Mill visits [conducted under Project 42531 have provided data on the nature and extent of surface deposits on dryer cans, for three grades. Contamination Test Stand [designed and constructed under Project 42531 is now available to provide model contaminated surface coupons for use in the WADS. The checkout and development of experimentation Adhesion and Drying Simulator is nearly complete. sticking can begin (trial plan in place). procedures for the Web Investigation of picking and Data on effect of drying variables (surface temperature, applied pressure, fabric design, proportion of drying under restraint, and heat flux uniformity) on cockle severity have been obtained. SUMMARY l 0 OF KEY CONCLUSIONS: The Questionnaire results support the value of Project F021 The Web Adhesion and Drying Simulator checkout process revealed that accurate measurement of the low level adhesion forces occurring under typical drying conditions will require integrated analysis of tension data and video recordings of the peel event IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F021 164 Status Report The cockle results to date are consistent with (but do not yet prove) the statement that high surface temperature in the first dryer section can aggravate cockle problems l Non-uniform heat transfer (e.g., from surface deposits) was shown to increase cockle severity. Surface deposits problems appear to be prevalent in the first dryer section IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 165 Project F021 DISCUSSION Topology Status Report and ADDITIONAL of Dryer DETAILS Roll Surfaces Significance It is anticipated that the roughnessof underlying cast iron surface has a role in determining the topology of contaminated surfaces and the amount of contaminant that deposits. These relationships have as yet to be confirmed. Results Work reDorted at the fall’99 PAC In previous work, reported at the fall 1999 Paper making PAC, we have made roughness measurements on cleaned dryer cylinders from two mills making fine paper. For this purpose, we utilized an epoxy material to obtain replicas of the surfaces. We utilized the services of CyberMetrics to characterize the topology of these surfaces. They used a noncontact method to measure the one-dimensional roughness of the replicas. Analysis of the results suggestedthat the surfaces on paper machine B were generally rougher than was observed on paper machine C. Work during this period During this period we have extended our topological measurementsto include the cleaned dryer cylinder surfaces from a paper machine making three-ply linerboard (designated PHL) and a paper machine that produces medium (designated PHM). In addition, we have also characterized the topology of curved cast iron coupons that are being used in our laboratory experiments. These included two cases,coupons with a smooth finish (designated CI (smooth)) and coupons that exhibited machining scratches (designated CI (scratched)). Topology measurements were obtained from epoxy replicas for the linerboard and medium machines while direct measurements were made on the cast iron coupons. In an effort to’obtain more accurate data, we obtained both one and two-dimensional roughness measurements. Figure 1 shows one-dimensional measurement of some key roughness parameters for the earlier reported fine paper machines as well as the data from the linerboard and medium paper machines and the laboratory coupons. We observe that the topologies of B and PHL are similar and the topologies of C and PHM are similar. We also observe that the smooth cast iron coupons show similar topological characteristics to the linerboard machine and the scratchedcoupons are similar to the medium machine. Figure 2 shows the two dimensional measurements for the cast iron coupons as well as for the linerboard and medium machines. We observe similarities between the coupons and the commercial surfaces. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 166 Project FO2 1 Status Report Conclusions We haveshownthat the topologyof our curvedcastiron couponsis within the rangeof topologiesobservedon commercialpapermachines.Hence,we may utilize thesecoupons in future laboratoryexperiments. Figures: .l d I Rq . Ra Topology I RqfRa of Dryer I RP One-Dimensional Roll Surfaces I . Rku VW I Rz . Measures of Roughness Figure 1. One Dimensional Measurements of Roughness . . - 100 ‘3 Ia . CI(scratched) PHL PHM 10 1 .l s; s; si it [Ssk] Sl;u S; Two-Dimensional Figure 2. Two Dimensional St& Sh Shl & Spk S;k Measures of Roughness Measurements of Roughness IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Sbi Sci S;i 167 Project F021 Dryer Roll Deposit Status Report Composition Significance It is anticipated that chemical deposition on dryer rolls can play a significant part in reducing paper machine productivity and paper quality. Deposition of chemical contaminants onto dryer roll can lead to sticking and picking which adversely impacts paper surface properties and paper machine runnability. To reduce these effects, papermakers routinely reduce dryer roll temperatures, which negatively impact machine productivity. Contamination is often found to be non-uniformly deposited on dryer cylinders. These non-uniform deposits can lead to non-uniform heat transfer, which may lead to paper defects such as cockle. Ultimately we wish to be able to determine the adverse economic impact of dryer deposits and to develop technologies to reduce these chemical depositions and their impact. Results Work reported at the fall’99 PAC Samples of deposits were taken from the surface of dryer cylinders from two commercial fine paper machines. Our purpose was to estimate the range of contaminant loading and identifying the chemical composition of the deposits. We found that contaminant loading on paper machine B ranged from 3.8 g/m2 to 8.7 g/m2, and that on paper machine C was fro m 0.01 g/m2 to 1.4 g/m . We observed that the contamination was uniformly deposited on each cylinder of paper machine B and was deposited in MD streaks on cylinders of paper machine C. The primary purpose in characterizing the chemistry of the deposits was to develop a model contaminant mixture for future laboratory experiments. The work at this early stage was primarily qualitative. A battery of tests including wet chemical analysis, inorganic elemental analysis and microscopy were conducted. On both machines we observed large amounts of fiber and fines, latex, starch, kaolin clay, titanium dioxide and calcium carbonate. In some caseswe also found smaller amounts of glue, styrene, talc, and rust. An attempt was made to also characterize the surface energy of the collected deposits. For this purpose, samples of deposits from each cylinder were compacted and formed into flat surfaced pellets. The contact angle (water) of deposits from paper machine B ranged from 75’ to 78O,while that from paper machine C ranged from 98’ to 115’. For comparison we measured the contact angle of clean cast iron as 77’. Work during this period . The focus of work during this period was on characterizing the deposits that were sampled from the surfaces of dryer cylinders of a linerboard machine and a medium machine. General Background The linerboard machine used a double felted first press and a double felted Beloit extended nip second press followed by a smoothing press. The Black Clawson dryer section was configured such that the bottom dryer cylinders (even numbered) were without dryer fabrics while the top dryer cylinders (odd numbered) had dryer fabrics. Doctors were located at the end of each dryer section. Hence, cylinders 5, 17, 3 1,43, 55,67 were doctored. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 168 Project F02 1 Status Report The linerboard that is produced is three-ply with 50% OCC and 50% Virgin Kraft in the middle ply. A small amount of recycle is in the side of the liner that is in contact with the top dryer cylinders, while the bottom cylinders contact the side of the liner that is 100% Virgin Kraft. The mill reports the use of the following chemicals in the manufacture of the linerboard; Alum, Sulfuric Acid, a defoamer (Callaway 8514 supplied by Calloway Chemical Company, Columbus Georgia), and felt cleaners (feltECCel23 1,455, and 465 supplied bY ECC International, and TWI C505 supplied by The Way, Metairie, Louisiana), The mill reports picking problems only when the steam pressureis increased in the first dryer section. Cockle problems have not been reported. Typically they do not clean the dryer cylinders. When there is a break they will use caustic on the dryer fabrics. Also, once a month they spray Fast Foam (a commercial caustic) on all parts of the machine and then rinse with water. The dryers on the medium machine are configured into four sections. The first section includes cylinders 1 through 9. The top cylinders (odd numbered) in the first section were all doctored. A single dryer fabric wraps each of the cylinders in the first section. In later sections, separatetop and bottom dryer fabrics are utilized. Dryer cylinders 10, 12, 14, 23, 37, and 51 are doctored. Chemicals utilized in the manufacture of medium include a defoamer (Callaway 8514), a wetting agent (Callaway 5507), and felt cleaners (feltECCe123 1,455, and 465 and TWI C505). Dryer Cylinder Surface Temperatures Dryer temperature surveys were conducted by the mill on the linerboard machine in Otto ber of 1999 while it was producing 42# liner and on the medium machine in August of 1999 while it was producing 26# medium. Surface deposits were collected by IPST during shutdowns in November of 1999. Average infrared surface temperatures are shown in Figures 1 and 2. Dryer Cylinder Deposit Loading During machine shutdowns, we scrapped a one square foot area of some of the cylinder surfacesto determine deposition loading, deposit chemistry and underlying cylinder topology. The deposit loading results are shown in Figures 3 and 4. We took samples in the first dryer section on the linerboard machine and found a range of loading from 1.2 g/m2 to 6.5 g/m2. We took our samples later on in the drying process on the medium machine as the area around cylinder 16 showed particularly heavy contamination. Unfortunately, the region of greatestcontamination on cylinder 16 was not close enough to the catwalk for us to safely obtain a scrapping of a fixed area. As a result we do not have a measurement of loading on cylinder 16. We were able to obtain a sample for chemical analysis and estimate the thickness of the deposit to be of the order of a sixteenth of an inch. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 169 Project F02 1 Status Report Dryer Cylinder Deposit Chemistry Samples of the deposits from the liner and medium machine were analyzed by IPST research services. The weight percent of various organic and inorganic components are shown in Figures 5 and 6. Figure 5, for the linerboard machine, shows that Polyvinyl Acetate was a main component of the deposit on cylinder 1 while it was absent on cylinders 2 and 4. This is consistent with the fact that cylinder 1 was in contact with a side.of the sheet that contained OCC while cylinders 2 and 4 were in contact with a side that contained 100% Virgin Kraft. In contrast, we observe a high concentration of calcium carbonate and silicates on cylinders 2 and 4 while these are in much lower concentration on cylinder 1. Figure 6, for the medium machine, shows a high concentration of Styrene-Butadiene in the sample from cylinder 16. This component is absent from cylinders 22 and 26. Similar to the linerboard results, we observe high concentrations of silicates in the samples from cylinders 22 and 26. The styrene-butadiene could be coming from pressure sensitive adhesivesbrought in by waste materials, 20-3O%OCC, used in making up the medium furnish. Conclusions The loading and chemical composition data should give us sufficient information to establish model contaminant mixtures for each of the three grades. Once we have defined these models we can use them to develop an understanding of the deposition process and allow us an opportunity to prepare contaminated coupons for adhesion and picking experiments on the laboratory web adhesion and drying simulator (WADS). IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Status Report 170 Project F021 Fig u res: Dryer Roll Deposit Composition Paper Machine Note: Deposition samples were taken from cylinders 0 5 10 15 20 25 30 Cylinder Figure 1. Average Surface 35 40 45 50 55 60 65 70 Temperature for Dryer Cylinder Machine Making Linerboard on the Commercial Note: Deposition samples were taken from cylinders 11’1’1’1”‘1’1’1’1”‘1’1’1’11”111’1””11”’11’1’111”111a’1111” 0 5 10 15 20 25 30 35 40 Cylinder Figure 2. Average 1,2, and 4 75 80 Order Paper Paper Machine 50 PHL 45 50 55 60 PHM 16,22, and 26 “1”11”11’1 65 70 75 Order Surface Temperature for Dryer Cylinder Machine Making Medium on the Commercial IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Paper 80 171 Project F02 1 Status Report Paper Machine PHL 8 7 Cylinder Number Figure 3. Deposit Loading for Dryer Cylinders on the Commercial Linerboard Paper Machine Making 10 9 8 7 6 5 4 3 2 1 0 26 22 . Figure 4. Deposit Cylinder Number Loading for Dryer Cylinders on the Commercial Medium Paper Machine IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Making 172 Project F02 1 100 - 75 n Tacky Polymer - Mostly Polyvinyl Acetate q Wood Fiber q Misc. Including Starch, Felt Fibers, Hard Plastics q Solvent Extractable Hydrocarbons 0 I I I Status Report q q q I Polystyrene Calcium Carbonate Silicates Including Clay Rust Paper Machine PHL I 50’ I I 25 I 0* 2 Cylinder Number Figure 5. Composition 100 of Deposits from Dryer Cylinders Making Linerboard on the Commercial Paper Machine - Paper Machine PHM Wood Extractives (Resin & Fatty Acids) 75 - I I d 10 Rust I 503 25 - 16 22 26 Cylinder Number Figure 6. Composition of Deposits from Dryer Cylinders Making Medium on the Commercial IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Paper Machine I 173 Project F02 1 Status Report Web Adhesion and Drying Simulator (WADS) Project F02 1 was initiated in order to study methods of increasing paper machine productivity by using high temperatures in the first dryer section. Currently, the use of elevated temperatures is limited by various interactions between the paper and drying roll surfaces which cause picking/sticking of the web. Furthermore, contamination of the dryer roll surfaces may have an effect on the degree of picking/sticking. A review of the relevant literature has revealed that no quantitative analysis of this problem has been undertaken. A Web Adhesion and Drying Simulator (WADS) was thus designed and constructed (via Project 4253 funding) to better understand the mechanisms behind the picking/sticking phenomena so that solutions can be developed to increase product quality and paper mill productivity. The WADS unit provides a direct measure of the peel force under simulated dryer conditions. Through the use of the Mardon equation shown below, the peel force can be correlated to the Work of Adhesion [ 11. Tf- W11 - I-co@ T’ = + mv,’ Tension (g/cm) W” =Work of Adhesion (g/cm) m = Mass per unit area (g/cm2) VI= Velocity (cm/s) 4 = Peel Angle The Work of Adhesion is a more fundamental parameter which depends on a number of factors including surface topology, surface materials, surface energy, contaminants, fiber characteristics, application pressure, surface and sheet temperature, and moisture content. Picking/sticking occurs when the adhesion between the web and the dryer roll surface is of the same order of magnitude as the cohesion of the web. By characterizing the work of adhesion, ways of preventing and reducing picking/sticking can be determined. WADS Unit The WADS system, shown below in Figure 1, consists of a belt driven flywheel to which a removable cast-iron “coupon” is attached. The coupon can first be installed in the Contamination Test Stand (CTS) to acquire a layer of surface contamination, and then can be transferred to the WADS system for peel testing. Therefore, the WADS can be used to test the effect of surface conditions, particularly contamination, on the peel force and work of adhesion. Various surface materials and surface treatments can be investigated as well. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 174 Project F02 1 Status Report Figure 1: WADS Unit Web Adhesion and Drying Simulator SAMPLE TRAY PEEL AKE-UP TENSION ROLL END OF CYCLE PRO XIMITY SENSOR START PROXIMITY OF CYCLE SENSOR--/ Unlike Mardon’s sheet stripper apparatus [2], the WADS has the ability to control and vary the surface temperature which is necessary in order to simulate realistic dryer roll conditions. A range of peel angles, peel velocities, web application pressure (controlled by belt tension), and dwell times are available. The sample inlet (web) temperature, which plays a role in the degree of sticking, can also be adjusted. During an experiment, the coupon makes one revolution from its start position at a set peel speed. The dryer belt acts to laminate the web (pulled in from the sample tray) onto the coupon as it makes its way around the wheel. The speed of the wheel and length of this dryer belt, both user specified parameters, determine the dwell time. In addition, web application pressure is adjusted with the belt tension. After passing through the applicator section, the paper is pulled from the coupon at the specified peel angle and peel point with a length of tape that is tracked over the tension sensor. The sensor records the peel force (tension) required to pull the paper from the coupon. Upon completion of the peel event, a proximity sensor triggers the brake. The data acquisition program collects the peel speed, tension, belt application pressure, and coupon surface temperature throughout the course of an experiment. Initial experiments on the WADS unit showed a significant level of noise with the tension sensor. Viscoelastic vibration damper material was used to mechanically isolate the frame which supports the sensor. In addition, a low pass filter was installed into the system to electronically reduce noise. Static calibration of the sensor demonstrates that readings are accurate within +/- 0.1 g. Dynamic tests show a consistent lg frictional drag of the tape over the sensor apparatus which increases the tension output. This factor will be incorporated into the data analysis to determine the real peel force. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 175 Project F02 1 Status Report WADS Checkout Runs Paper samples of a bleached pulp blend (75% BHWIU25%BSWK) were prepared in the Formette for the development and testing of the WADS system. Linerboard samples were also used for initial development becauseof its superior durability and sticking ability. For preliminary runs, equipment settings are shown below in Table 1. Table 1: Initial WADS Setup Setting Parameter 1 Peel Angles 60°, 45O,and 15’ Peel Speeds 50 fVmin & 150 ft/min (0.25 & 0.50 m/s) Coupon Temperature 250 OF(121OC) Sample Temperature approx. 150’ F (66 “C) Application Pressure 0.47 psi (3240 Pa) Dwell Time 0.271 s - 0.407 s (depending on speed) I A plot of tension vs. time for a typical experimental run using the bleached sample is shown in Fig 2. There is a start-up and shutdown period of transience in each run before and after the actual peel event which is indicated by the arrows. Figure 2: Typical Experimental Run (Tension vs. Time) Bleached Wet Paper: Raw Data of Tension vs Time 8 7 6 0 1 2 3 Time (s) 4 5 6 l/6/00 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 176 Project F02 1 Status Report Figure 3: Extracted Peel Event (Tension vs. Time) Bleached Wet Paper: Tension vs Time 2 2.2 2.4 2.6 Time 2.8 3 3.2 (s) l/6/00 For the data analysis, the actual peel event, shown in Fig. 3, is extracted from the raw data file (which records a user specified number of points and sampling rate). For the preliminary runs, extraction of the peel event was difficult becausethere was no time stamp in the data showing the onset and end of the peel. A method of extraction was found while performing the simulated experiments (as explained in the next section) and was used on previously gathered data. The data must be corrected for various factors including zero value, paper/tape weight and friction. During an experiment, the tape is being retracted and the paper sample is being added as the paper sample is pulled off the coupon. These two factors contribute to the tension sensor output and must be accounted for when determining the actual peel tension. Given the constant speed of the paper sample around the wheel, a linear function describes the change in weight of the paper (and tape) sample. The resulting weight is subtracted from the tension value to give the peel tension. The contribution of friction mentioned earlier must also be accounted for in the final data analysis. Initial results revealed problems with producing adequate sticking to the clean coupon surface. A study was done to correlate solids content with optimal sticking for the bleached Formette sheets. Results showed similar sticking behavior for a range of inlet solids content, 30.60%. Tension values appearedto be quite low, in the 0.5-3.0 g range, making it very difficult to assessequipment performance. Setup conditions were modified to maximize sticking by moving to low angles and intermediate temperatures. However, only minor improvement was observed. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 177 Project F02 1 Status Report Simulated Runs In order to verify equipment capability and understand the peel, simulated experiments were performed using dry paper (copy and blotter paper). By using double-sided tape on the coupon, an artificial sticky surface was created. Video documentation was made of each run in order to correlate the data with the actual peel event. A number of important observations resulted from these simulated tests. First of all, at higher peel angles, frame-by-frame video analysis showed that the actual peel angle varied greatly from set point. For example, for a peel angle of 60”, the actual angle ranged from 18’-24’. This is primarily due to the fact that the paper does not peel from the coupon surface at a sharp angle. Rather, there is some curvature due to lack of adhesion and inherent stiffness of the paper creating this discrepancy. Since the peel angle is required in the Mardon Equation to solve for work of adhesion, an accurate angle measurement is necessary. Similar results were observed by Mardon who found that at set points of 90°, the actual angle was about 45’ [3]. He used a Fastex camera to account for the actual angle in the calculations. For the WADS system, the implication is that videotaping of each run is necessaryto record accurate peel angles. The Mardon equation must then be applied locally for each discrete point. In addition, the actual peel event can be extracted by using the drop in the speed signal as shown in Fig.4. The motor cut off is triggered by a proximity sensor at which time the speed signal begins to drop. The distance between the sensor and the peel point allows the start time of the peel to be determined. The length of the peel is simply based on the linear speed and length of paper sample. This method of analysis proved to be highly accurate in locating the actual peel event, with the start time varying only +/- 0.02s. Recently, another proximity sensor has been added to the system in order to provide a measure of the exact time at which the peel event takes place. Figure 4 : Speed Signal Drop Bleached Wet Paper: Speed & Tension vs Time 6 Time(s) l/6/00 IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 178 Project F02 1 Status Report Tension and peel angle data collected using this method were used in the Mardon equation to calculate the work of adhesion. A plot of the work of adhesion is shown in Figure 5. Units of work are shown in g/cm which is numerically approximately the same as the SI units of J/m2 (0.98 conversion factor). Figure 5 : Work of Adhesion during the Peel Event Work of Adhesion vs Time : Dry Paper on Sticky Surface 5 4.5 3.5 0.5 0 I 6 .lO I 6.20 6.30 6.40 6.50 Time 6.60 6.70 6.80 6.90 (s) Figure 6 is a video image of an actual peel event for one of the dry paper runs. Tension values of IO-40g , significantly higher than with the wet paper on the clean coupon, were found with the tape runs. Figure 6 : Video Image of Peel Event IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 179 Project F02 1 Picking/Sticking Status Report Experiments The first part of the WADS testing involves clean coupon measurements of tension and a characterization of the work of adhesion at a range of speeds,coupon and sample temperatures, and dwell times. Initial experiments were used to verify and define the operating range of the WADS equipment. Currently, the dwell time has a range from 0.14s to 1.20s. The speed can likewise be varied from 25ft/min to approximately 200 Nmin. While the WADS does not provide peel speedssimilar to actual paper machines, it does nevertheless provide dwell times in the range of the paper-to-dryer contact times at real machine speed. Consequently, the picking/sticking data can be expected to be highly meaningful. In addition to determining work of adhesion, another goal of this work is to quantitatively describe the picking/sticking phenomena. Previous work by Meinecke [4] showed that the degree of picking depends on surface temperature, peel speed (dwell time) and inlet solids. However, the amount of picking was not quantified nor was it related to the work of adhesion. The WADS, on the other hand, should be capable of providing data for a quantitative map relating the regions of greatest picking/sticking. A detailed procedure is being developed in order to collect and measure the amount of picking. A qualitative visual rating will also be used. The Clean Coupon experiments will employ the following conditions: I ~~ Parameter Coupon Surface Temperature Dwell Time Initial Sheet Temperature Initial Sheet Solids Peel Speed Peel Angle Paper Type Setting 194’ F, 248OF, 302’ F 1(90°C, 120° c, 150° C) 0.17s 0.5 s 100’ F (38 “C), 150’ F (66°C) 40%, 50%, 60% 50 ft/min, 150 ft/min (0.25 & 0.50 m/s) fixed at 30 degrees Copy Paper and Liner Board I Once these experiments are complete, contaminated coupons from the CTS can be installed on the WADS to test the effect of surface condition on work of adhesion as well as on picking/sticking. The same conditions as those listed above will be analyzed. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) I I Project F02 1 180 Status Report References 1. Mardon, J., “The Release of Wet Paper Webs from Various ‘Papermaking Surfaces’,“APPITA, Vol. 15, No. 1, pp. 14-34. (July 1961) 2. Ibid 3. Mardon, J., “Theoretical and experimental investigations into the peeling of paper webs from Solid Surfaces,” Paperi ja Pm, No. 11, pp.797.815 (1976) 4. Meinecke, A., Chau Huu, T., and Loser, H., “Neue Erkenntnisse iiber die Papiertocknung mit Trocken-zylindern,” Das Papier, 42(10A), pp.159.165 (1988) IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 181 Project F02 1 Cockle Status Report Investigation Background One of the several reasonsgiven for the use of temperature graduation in the dryer section (i.e., use of low temperatures in the first dryer section) is that heating/drying the sheet too quickly can increase the likelihood/severity of cockle. Another problem associatedwith high initial dryer temperatures is increased surface deposits. Since deposits can easily be non-uniform, they are expected to contribute to cockle due to the associatednon-uniformity of heat flux to the sheet. Qualitative studies of the influence of initial moisture and heating non-uniformities on the cockle developed in paper as a result of drying, were performed by Brecht, et al (1). A discussion of some of the results is given in the paper by Gallay (2). D. Coffin’s team has developed a technique to quantify the degreeof cockle in a paper sample (shadow Moire technique), and applied it in initial experimental studies of the factors believed to influence cockle [Project FO20]. Some of the factors investigated were the degree of restraint during drying, formation uniformity, dryer surface temperature (over a rather narrow range), pressing level and basis weight. The capability of an existing dryer simulator has been upgraded to aid in conducting a further investigation of the extent to which drying-related variables influence cockle. The dryer simulator now has two heated surface options, a uniform cast iron platen and a cast iron platen having a pattern of epoxy-filled depressionsto simulate non-uniform surface deposits. The simulator offers control over surface temperature, dwell time, and restraint pressure. Furthermore, the sheet can be lightly pressed against the hot surface with samples of real dryer fabrics (which apply somewhat non-uniform pressure to the sheet). Initial Cockle Experiments In keeping with the input from the Papermaking PAC, some new cockle experiments relevant to Project F021 have been planned and executed, in collaboration with D. Coffin’s team. The basic strategy developed for the trial is to generatesamples that have been dried to various extents (in terms of moisture ratio) under selected conditions in the dryer simulator (i.e., with some restraint), and to let the drying be completed via unrestrained air drying. It was hypothesized that the amount of cockle in the final sheetswould be dependent on the proportion of drying occuring in the simulator, as well as on the drying conditions. The trial plan developed is as follows. Trial Obiectives .To quantify the effects of surface temperature and applied pressureon cockle .To investigate the effect of heat flux non-uniformity on cockle Experimental Plan There are 4 parts to this trial: (1) generation of drying curves (i.e., moisture ratio vs. time) to provide input to Parts 3 and 4; (2) control runs; (3) drying runs with uniform heat flux; (4) drying runs with non-uniform heat flux. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F02 1 182 Status Report Test Conditions: Samples: These will be cut @“x5” squares)from some of the existing Formette sheetsthat were initially made for use in WADS adhesion experiments: l 70 gsm l 75% BHWK/25% BSWK; with appropriate additives to simulate a copy paper furnish (seeTable 1) l 450 CSF l Tensile Ratio: about 2.0 l Initial Solids Level: 45% (via pressing) Table 1. Sheet additives. 1Additive 1 Amount PCC 15% of dry weight Optical Brightener 2 lbs./ton 3 Ibs ./ton I 12 lbs./ton Starch Retention Aid 2 lbs ./ton I Part (1): Drying Curves: l Dwell time in dryer simulator: 10 levels, TBD l Surface temperature: 2 levels, 110 C and 170 C l Mechanical pressure: 0.2 psi and 1.0 psi [covering the range corresponding to the pressurecreated by fabric tension on a dryer can] Platens: uniform surface and non-uniform surface l l No. of runs per condition: 1 Part (2): Controls: 0 entirely air dried without restraint: 10 samples 0 completely dried with high restraint in the dryer simulator: l 5 samples using uniform platen [for one combination of surface temperature, mechanical pressure and fabric] l 5 samples using non-uniform platen [for one combination of surface temperature, mechanical pressure and fabric] Part (3): Drying runs with uniform heat flux (uniform surface): Final moisture ratio (for dryer simulator runs): 3 levels, about 0.7,0.3,0.1 g/g l l Surface temperature: 2 levels, 110 C and 170 C {possibly one extra level at the final moisture condition giving worst cockle} l Mechanical pressure: 0.2 psi and 1.0 psi l Dryer fabric: 2 fabric designs l Completion of drying after removal from dryer simulator: unrestrained air drying in a 50% RH/22 C lab No. of runs per condition: 5 l IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 183 Status Report Part (4): Drying runs with non-uniform heat flux (non-uniform surface): Final moisture ratio (for dryer simulator runs): 3 levels, about 0.7,0.3,0.1 g/g Surface temperature: 2 levels, 110 C and 170 C Mechanical pressure: 0.2 psi and 1.0 psi Dryer fabric: one design Completion of drying after removal from dryer simulator: unrestrained air drying in a 50% RH/22 C lab No. of runs per condition: 5 Test Runs: Part (1): Trial and error (and consultation with F. Ahrens) should be used to select about 10 different dwell times that define the moisture ratio vs. time curves over the range from about 1.2 g/g down to about 0.05 g/g. Parts (3) and (4): The drying curves will be used to determine the dwell times needed to achieve the target moisture levels for the dryer simulator runs. The sheet and fabric MD’s need to be aligned. After initial drying in the dryer simulator, to their established moisture targets, the completion of the drying of samples will be done via air drying without restraint, in a room at 50%RH/22 C. Data Collection/Handling: Record all test conditions; give samples unique, easily understood code designations; record initial and final (after dryer simulator) weights. After the air drying (unrestrained) step is complete (24 hrs. at 50% RH), submit samples to Kennisha Collins for cockle analysis. After cockle analysis, oven dry samples and calculate initial and final (after dryer simulator) solids levels. For Part (1), plot moisture ratio vs. time curves. For the Part (2) control sheetsdried on the simulator, after cockle measurement at 50% RH, expose sheetsto high humidity (95%) and again measure cockle. Shadow Moir6 Measurements: From visual inspection pick the most cockled and flattest sheets.Prepare Moire system for measurements (Clean dust from glass and paperholder, use 50 lines per inch glass, make sure holder is flat, make sure light source i S in working condition) Set the Moire system up to get the maximum x-y coverage and still get reliable z measurements. It would be good to make measurements on a square area of 3 to 4 inches. Also make sure your window of interest has at least 200 pixels in each direction. Once the camera is focused on the area to be measured do not change the settings until all sheets are measured. First make sure that you get good measurements for the least and most cockled sheets. Adjust as necessary. If the sheetsare very flat, consider taking a second measurement using the 100 lines per inch plate. Measure all sheetswith the MD direction facing the same way. Make sure that the distance between the glass and the paper is the same for each measurement (i.e., move the holder to the same starting position each time, without severely cockled sheetstouching the glass) By looking at a graph of the height data make sure the image is reasonable(high frequency wrinkles or abrupt changes indicate that poor fringe patterns were obtained). Analyze the raw data so that the height is with respect to the best-fit plane (averageheight is zero). If the samples have curl, the low frequency responsewill have to be filtered using FFT. Calculate the standard deviation of the height, the average of the absolute value of the height, and the average gradient squared. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) 184 Project F021 Status Report Results of Cockle Experiments The drying runs have been completed for nearly all of the trial conditions. The cockle analysis is in progress, but results for a portion of the trial conditions are given in Fig. 1. 1 The “cockle standarddeviation” is a measureof the sheet-averagedeviation of the cockled co&l .eh sheet surface topology from that of a flat sheet (i.e., a larger number implies greater cockle). The “leaving moisture ratio” refers to the moisture in the sample upon removal from the dryer simulator, prior to final air-drying. The data point at 1.22 g/g is a contra control1 case,in which the samples (initially at 45% solids were entirely air-dried without restraint. restraint. Fig. 1. Cockle Standard Deviation Vs. Moisture Leaving Restrained Drying Phase [initial MR Ratio =I-221 - -+- d 6 - Uniform Heating,1 10 C, 1 psi Uniform Heating, 170 C, 1 psi - -& - - Nonuniform, 170 C, 1 psi I ~~---~-.~~,-~-~.~..~,.. O-' 0 0.5 Leaving W-N, 1 moisture ratio, ww The results in Fig. 1 show several interesting features. Let us first consider the end points at high and low moisture. The cockle associated with the 1.22 g/g control is, by definition, not dependent on drying conditions. Instead, it probably reflects the effects of sheet formation and basis weight non-uniformity. The end point at lowest moisture has a very low cockle (i.e., corresponding to a flat sheet). This reflects the fact that the entire drying occurred with the sheet under restraint. The most interesting behavior is associatedwith the intermediate points. The cockle is seen to be far greater in the intermediate region than at the ends, in spite of the fact that sheet was restrained for a portion of the drying process! The general trends exhibited by the peak values seem to be physically reasonable. That is, the uniform heating at higher temperature results in more cockle than uniform heating at lower temperature. This result is consistent with (but does not yet prove) the statement that high surface temperature in the first dryer section can aggravate cockle problems. The peak for non-uniform heating (e.g., representing effects of non-uniform surface deposits) is seento further increase cockle. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F021 185 Status Report The potential for shrinkage during (unrestrained) drying is a function of the moisture change. Thus, one would expect cockle severity to be associatedwith moisture differences (and levels) in the sheet at the beginning of the unrestrained drying period. Examples of the free shrinkage behavior of paper (vs, moisture content) are contained in Refs. 3-6. Over the range of sheet moistures occuring in the dryer section (e.g., 1.5 g/g down to about 0.05 g/g) the shrinkage function is quite non-linear, with most shrinkage occuring between 0.7 and 0.05 g/g. The shape of the cockle vs. moisture ratio curves in Fig. 1 can, therefore, probably be explained as the competition between two phenomena: 0 if the sample stays in the dryer longer (to lower moisture ratio) there is, on the average, less shrinkage potential in the subsequentunrestrained drying phase, decreasing the tendency for cockle. 0 if the sample stays in the dryer longer, the local moisture deviations may be greater at the end of the restrained period, at least over part of the final moisture range, increasing the tendency for cockle. One other observation from the experiments seems notable. The sheetstended to exhibit a small-scale “cockle” pattern matching the fabric pattern. This may be due to a fabric scale non-uniformity in the heat and mass transfer processescreating fabric scale moisture differences in the sheet. Potential Next Steps The results to date appear to provide some new insights into the effects of drying conditions. There would seem to be merit in continuing these collaborative cockle investigations. One direction to be considered is that of trying to more closely simulate the drying process occuring in the dryer section. A sequenceof short heat input periods (in the dryer simulator), with brief periods of unrestrained or unidirectionally restrained evaporation/shrinkage in between, to simulate the open draws, could be performed. Unfortunately, currently available laboratory drying devices are not sophisticated enough to give a truly realistic simulation of the time-varying drying and restraint conditions experienced by a sheet as it progressesthrough a typical dryer section. Development of a more realistic, versatile laboratory simulator for the drying process could be undertaken, if there is sufficient interest. In any case, from the point of view of Project F021, the overall goal of future work in this area should be directed toward: l Pinpointing the drying conditions and sources of non-uniform shrinkage in the conventional process that create the greatestdegreeof sheet surface non-uniformity l Proposing and demonstrating strategies for reducing the cockle problem, that also provide an increase in drying rate. REFERENCES 1. Brecht, W., Muller, F., and H. Weiss, “Uber das ‘Blasigwerden’ von Papieren”, Das Papier 9 (7/8): 133-142 (1955) IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only) Project F021 186 Status Report 2. Gallay, W., “Stability of Dimensions and Form of Paper”, Tappi J. 56 (12): 90-95 (1973) 3. Wahlstrom, T. and C. Fellers, “Biaxial Straining of Paper During Drying, Relations Between Stresses, Strains and Properties”, Proceedings 1999 TAPPI Engineering Conference, pp. 705-720 4. Wahlstrom, T., Adolfsson, K., &tlund, S. and C. Fellers, “Numerical Modelling of the Shrinkage Profile in a Dryer Section, a First Approach”, Proceedings 1999 TAPPI International Paper Physics Conference, pp. 5 17-531 5. Waller, M. and Singhal, A., “Development of Paper Properties During Restrained Drying of Handsheets”, Proceedings 1999 TAPPI Engineering Conference, pp. 72 l-732 6. Wedel, G., “Drying Restraint in a Single-Tier Dryer Section”, Proceedings 1989 TAPPI Annual Meeting, pp. 23-29. ACKNOWLEDGEMENTS Thanks to John Chabot, Warren Davis and Kerin Strange for their efforts in design, construction and debugging of the WADS, and to Georgeta Maghiari for assistancein the WADS checkout experiments. Thanks to Paul Phelan for suggesting the use of a patterned platen to provide non-uniform heat flux for a portion of the cockle study. Thanks to David Orloff for guidance and participation in many aspectsof the reported work, and for preparation of the sections on dryer surface topology and deposits. IPST Confidential Information - Not for Public Disclosure (For IPST Member Company’s Internal Use Only)