Version 9

October 1, 2009

by

Application Guide
Copyright 2009 ExperTune Inc.

This is the third of a set of four user guidance documents.

1. Project Planning Guide

Provides details on hardware, software, network, and communications requirements. Discusses the type of information/data needed for configuration. Suggests personnel resources that should be involved.

2. Installation Guide
Procedural manual for installing the software. Installing new versions and upgrades of PlantTriage. Instructions for configuring the server computer; networking, user IDs, OPC connectivity, and security. Establishing communications to OPC server. Verifying installed software is operating. Build first unit operation and point.

3. Application Guide
How to organize the system plants, units, points. Decide on key assessments, and economic priority selection based on site business drivers. Setting baselines and thresholds. Selection of sampling rate and assessment frequency.

4. Continuous Improvement Guide

Repository for your site specific configuration information, and project basis. Some thoughts on training resource types and skill sets. Integrating PlantTriage into the existing work processes. Suggestions for a configuration maintenance plan. Benchmark reporting.

2
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

Table of Contents
I. II. OVERVIEW ...................................................................................................4 ORGANIZING YOUR PLANTTRIAGE POINTS............................................5 A. Plant level................................................................................................................ 5 B. Unit level................................................................................................................. 7 C. Point level (Loop) ................................................................................................... 7 KEY ASSESSMENTS (LOOP HEALTH).................................................10 Business Objectives Drive Key Assessments ....................................................... 10 All Businesses....................................................................................................... 10 Quality................................................................................................................... 10 Throughput............................................................................................................ 11 Cost & Energy Reduction ..................................................................................... 11 Reliability & Maintenance.................................................................................... 12 BASELINE & THRESHOLDS..................................................................13 Steps for applying Baselines & Thresholds .......................................................... 13 Setting Initial Baselines & Thresholds ................................................................. 14 Setting Baselines & Thresholds Based on Historical Data................................... 14 Set Post Baselines & Thresholds....................................................................... 15 Fine-Tuning Baselines & Thresholds ................................................................... 15 USER CONSIDERATIONS .........................................................................16 Engineering Interface Access ............................................................................... 16 Browser Access..................................................................................................... 16 Browser Report Roles Management ..................................................................... 16

III. A. B. C. D. E. F. IV. A. B. C. D. E. V. A. B. C.

3
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

I.

Overview

This document provides guidance on organizing your points into the PlantTriage hierarchy. It explains in more detail some of the point and unit configuration parameters. The Key Assessments are explained and some suggestions are provided for different operational objectives. The Baselines & Threshold philosophy is discussed. All of these topics are sequenced as you would encounter them during your system configuration.

4
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

II. Organizing your PlantTriage Points

There are three levels of organization within PlantTriage; Plant, Unit, and Point(loop). Points are assigned to Units, and Units are assigned to Plants. Typically, you will follow the convention used in the control system for the plant. If there is no standard within the plant control system, then you can group points logically, according to the arrangement of unit operations at the plant. Some examples might be:
Plant: Utilities Unit: BoilerA Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: . Unit: BoilerB Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: . Unit: TurboGenA Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: . Plant: Feed Prep Unit: Preheat Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: . Unit: Splitter Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: . Plant: CityName Unit: BoilerA Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: . Unit: TurboGenA Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: . Unit: ProductLine1 Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: . Unit: Distillation Point: Flow Loop Point: Temperature Loop Point: Level Loop Point: .

Often the organization of the operator station control graphics can give clues as to the how to break points into logical groups.

A. Plant level
There can be several Plants on a single PlantTriage server. The main setting for the Plant level is the Sample Interval parameter. All the points assigned to units in that plant will be selecting from these sample intervals. There are 6 possible interval settings (A-F). The last F must be the slowest of the six. These are set through the engineering interface menu Tools, Plant Manager, Sample Intervals tab. The list below shows the point configuration parameter and the rate of sampling.

5
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

PV SP CO MODE Online Status P I D Filter Range Limits

The chosen rate A-F The chosen rate A-F The chosen rate A-F The chosen rate A-F The chosen rate A-F The slowest rate F The slowest rate F The slowest rate F The slowest rate F The slowest rate F

Choosing a sample interval is a trade-off. For the best model quality and analysis, you want the fastest sample time possible. However, to minimize the load on your DCS system, you may want to reduce sample times. The following table provides some recommendations for setting point sample times. Loop Type Flow Control Loops Liquid Pressure Control Gas Pressure Control Temperature Control Level Control Analyzers Recommended Scan Rate 1-10 seconds 1-10 seconds 5-30 seconds 5-30 seconds 5-30 seconds 10-100 seconds

Loop scan rate selection is very important to developing accurate, meaningful assessments of control performance. For accurate determination of controller dynamics, the scan rate must be fast enough that it does not obscure the relevant process dynamics. Ideally, this means that the sample rate should be 2-4 times faster than the process dead time. Assessment Process Model Parameters Robustness, Harris Index, ExperTune Index Oscillation Detection Statistical information: Variability, Variance, Opportunity Gap Noise Band, IAE, Valve Reversals, SP Crossings Tuning Effects of a Slow Scan Rate Deadtime and time constant will appear longer than expected. These assessments depend on the process model, they will be inaccurate. Opportunities to improve will be obscured. Higher frequency cycles will not be detected. Increased risk of aliasing. Loss of high frequency data contribution.

Loss of high frequency data might mask a problem

When using integrated tuning capability, user must manually adjust sample time with each loop before gathering bump responses. 6

PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

Sampling faster than the recommended rates will result in only marginal accuracy increases for performance assessments.

B. Unit level
Units are intended to contain points pertaining to a common function; for example the measurements and control loops on a distillation column. Units maybe configured as batch or continuous type units. The assessment interval is set at the unit level. This is the frequency at which PlantTriage will perform assessments on the points assigned to that particular unit. In the case of a batch type unit, the assessments are done over the entire batch processing period. The time setting for the batch units just specifies a minimum run time. In some cases, individual loops with slow dynamics (long oscillation periods) can be removed from their logical unit grouping and placed in a special unit with a long assessment interval. When deciding on the assessment period a trade-off is being made; the number assessments per day and/or early alerting to problems (short assessment period) vs. failure to detect some longer oscillations (long assessment period). Units have statuses, either online or offline. When a unit is offline assessments are not performed for any of the points assigned to the unit. If it is common for a unit to be out of service for periods of time a point maybe used to indicate the service condition (on/off line). It maybe as simple as a temperature > 100 for 90% of the assessment period. Or, more involved logic can be used in the Equation Builder to create a point that reflects the status of the unit. Units also have there own performance assessments: Throughput, Quality, MaterialCost, EnergyCost, Reliability. These values are not calculated by PlantTriage they are read from other sources via OPC. An example might be a calculation of EnergyCost for the unit. This could be a sum of the steam flows in mmBTU/Hr for the unit multiplied by the $/mmBTU. This calculation could be performed in the Equation Builder and linked to the EnergyCost unit assessment.

C. Point level (Loop)

Some plants start out with a small PlantTriage system, and increase the point count over time. When there are more process points than PlantTriage points some decisions must be made as to which points will get supervised with PlantTriage. ExperTune recommends configuring the all the loops in a processing area, and adding processing areas as the PlantTriage system is expanded. In the absence of specific process unit concerns it is best to start with the front of the processing chain and work towards the end. If the site includes a utilities area (boilers, generators, hot oil systems) this area should be included early in the PlantTriage growth. Since you do not know in advance where you will find the root cause of problems, it is important that you blanket each unit operation. This will enable you to drill down into the root cause more quickly. It can be frustrating when tracking down a load 7
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

disturbance and running out of possibilities because upstream points have not been included. Even if the site does not have a utilities area, configure points for: ambient air temperature, plant air header pressure, steam supply header pressure, and fuel gas supply pressure. Often these utility systems are the source of upsets and disturbances. There are two options when configuring points; individual point building and bulk building. Individual point building starts out with selecting an existing unit or creating a new unit. The process is handled by a wizard, and it is quite easy to follow with context sensitive help available. The bulk building process is very fast for systems that have maintained a consistent tagname convention. This process is handled with a comma separated file (.csv). ExperTune has provided an Excel spreadsheet (Loop Configuration Data.xls) with all the point configuration fields, and some notes for guidance. This file is located in the \xtune\Import folder. The first tab is for collecting all the data, the second tab is linked to the proper cells in the first and this tab can be saved as a .csv file. The PlantTriage import program can process the .csv file and create the points in PlantTriage. Please make a working copy of this file to avoid corrupting the original. Setting the economic significance of each control point will help PlantTriage to prioritize information. PlantTriage allows you to rank each point as Low, Average, or High economic importance. This section provides some guidance on how to identify which points belong in each category. As you consider a point keep in mind the points connection to business objectives. Economic Significance High % of Points < 10% Which Points to Include Only the most important points. Typically key product attributes, and points involved directly in energy and material consumption. Most of the points in your plant. Points that are relatively unimportant, such as waste streams, surge tank level controls, and ancillary systems

Average Low

60-70% 20-30%

Do not spend a lot of time to determine economic significance. Chances are your process engineers or operators can tell you exactly which points should be ranked High, without too much thought. The Low loops can also be found quickly. Everything else should default to Average. In fact for the initial pass you might set all loops to average (easily done with bulk loading). Then make changes individually, as users make the information available on the priorities. Note: The points of Low importance should still be configured in PlantTriage. There is likely a large amount of interaction between these loops and the more important loops. If you do not track them, it will be much more difficult to identify root causes.

8
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

A couple of tips: 1. A point may only be assigned to one unit, if it is desired to use the point in another unit a duplicate point must be created with a slightly different name. 2. If you wish to control the way the points show up in drop down lists, you should prefix the name with an alphabetic sequence (ex. A1_xxxx, A2_yyyy..).

9
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

III. Key Assessments (Loop Health)

PlantTriage calculates over 70 assessments for every point in the system. You can designate some of these assessments to be Key Assessments. These assessments receive special treatment within PlantTriage. For example, Key Assessments are used to calculate the composite Loop Health measure, and drive the rankings in Biggest Payback Loops. By identifying a few Key Assessments, you can help to focus the efforts of plant personnel.

A. Business Objectives Drive Key Assessments

Use your business objectives to drive the selection of key assessments. Everyone is likely to agree that cost, quality, reliability, and production rate are important. However, most businesses go through cycles when they focus on one or two of these strategies at a time. For example, if you are in a sold-out position, throughput becomes more important than energy cost. More production = more profit. But if sales are lagging, you will likely be focused on quality or cost. You should select Key Assessments based on the current business environment. The sections below can provide some guidance to select key assessments based on some of the most common business conditions.

B. All Businesses
Time in Normal Time in Normal Mode is a very good high-level indicator of loop performance. This is because many operators will put a loop in manual if they have a problem with any part of its performance. For example, an operator will likely put a loop in manual if any one of the following conditions occurs: The sensor has failed completely. The sensor is periodically spiking. Too much valve movement The tuning is too aggressive, creating oscillation. The tuning is too sluggish. Process response too slow. I dont trust the controls

As a high-level indicator, Time in Normal is highly recommended for most every plant.

C. Quality
Quality may be an important business objective if you are experiencing waste, rejects, recycles, customer complaints, product downgrades, or similar quality problems. Recommended Key Assessments for quality: 10
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

Variance, Normalized Normalized variance is a good measure of the process variability. Because it is normalized to the instrument scale, many people find it to be somewhat more meaningful than the raw variability, or other statistical measures such as PV standard deviation. Making improvements to variance will help to stabilize the process and reduce process excursions that result in quality incidents. Monitoring points with a sudden increase in normalized variance will provide insight into the source of process upsets. Oscillation Significance One major cause of process variation is oscillation. PlantTriage has many assessments to quantify oscillations in your plant. Oscillation Significance is a good choice for a key assessment, because it includes basic oscillation detection, combined with the % of variance attributable to oscillation, and the overall variance. Reduce the oscillation significance to see improvements in Quality, Energy, and Cost.

D. Throughput
If throughput is important, you will want to uncover all situations that could help to improve production rates. A few Key Assessments for this are included below. Output at Limit Output at limit calculates the percent of the time that the controller output is at a limit. If a valve is fully open, then higher production rates are not possible. As your business grows, and production increases, different parts of the process start to become the bottleneck. Service Factor Service factor combines three factors to ensure that each control loop is being used to its full potential. These three factors are: Time in Normal Mode Output at Limit PV at Limit This is high-level measure of the way that individual constraints are affecting the process.

E. Cost & Energy Reduction

To reduce costs and save energy, focus on the Key Assessments noted below. Realtime Savings Advisor Realtime Savings Advisor is one of the most direct ways to deliver savings. It is a critically important assessment for cost & energy savings. The six sigma properties and Opportunity Gap will need to be configured. Oscillation Significance When processes are oscillating, they are inefficient. There are many tools to diagnose oscillation issues with PlantTriage. Oscillation significance is a good choice as 11
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

a Key Assessment to identify issues with oscillation. Reduce the oscillation significance to see improvements in Quality, Energy, and Cost.

F. Reliability & Maintenance

Time in Normal Process equipment, instrumentation, and valves tend to fail in a wide variety of ways. As explained above, this is a high-level indication of which loops are experiencing equipment problems. Valve Travel When valve travel is excessive, the valve will fail sooner. Monitoring valve travel is a good way to predict which valves will have more problems. Noise Band Monitoring the noise band for a measurement can identify trouble before equipment fails. A drop in the noise band might indicate transmitter plugging.

12
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

IV. Baseline & Thresholds

PlantTriage will need to know what is considered normal and what is unacceptable for the results of the assessment calculations. The settings that define normal and unacceptable are what PlantTriage refers to as baselines and thresholds. Baselines and thresholds (B&Ts) apply to all assessments not just the key assessments. The baseline is the normal or typical value of the assessment. The threshold is the value of the assessment that would be considered unacceptable. There can be high and/or low thresholds. B&Ts fall into two categories; fixed and statistically determined. Fixed are values that are commonly agreed upon without having to analyze measurement data. For example a baseline for %Time in Normal Mode might be 100. The other category, statistically determined, requires some review of the measurement data to decide on an appropriate value. For example a baseline for %Noise band would require a review of the typical or best conditions value of %Noise. ExperTune provides several templates for automated setting of the B&Ts. These templates are broken down by process measurement type (Flow, Pressure, Temperature, etc). When the system is first configured and started up there are no B&Ts set. In order to start identifying, and prioritizing performance issues the B&Ts must be set. Expertune also provides sets of templates called initial templates. These templates have entries for the baseline and threshold values of the fixed type assessments. By applying the initial type templates to the points, users are able to quickly see some of the performance issues needing attention. Then, after a short time (3-6 weeks), the standard templates can be applied. These templates make use of the data collected by PlantTriage to determine the proper settings for assessment B&Ts. These would be the statistically determined settings. There are also a couple of Post templates. These are special templates that are applied after all the B&Ts have been set. They only change certain assessment B&Ts for special situations. There are two Cascade loop related Post templates. These templates apply to loops involved in a cascade arrangement. There is a template for secondary loops and one for primary loops. Depending on the number of loops involved in cascade arrangements it might be effective to create groups for these type loops. (see: Display Group Manager) Select by: [Group automatically contains all loops where: Cascade = Primary] and another display group for [Group automatically contains all loops where: Cascade = Secondary].

A. Steps for applying Baselines & Thresholds

ExperTune suggests the following approach to preparing and maintaining B&Ts: 1. Configure points in PlantTriage. 13
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

2. 3. 4.

5. 6. 7. 8.

Review initial templates for each process measurement type make any changes (Template Editor). Apply initial templates. Observe results for 3-6 weeks. (depending on site operating conditions, need to obtain unit operation data from normal, or better time periods. This data will be used as the basis for the B&Ts setting of statistically dependent assessments) Review templates for each process measurement type make changes if needed (Template Editor). Apply templates. Observe results, and refine settings on an as needed basis. Annual review of B&Ts (at minimum the B&Ts for the key assessments), consider tightening limits as part of a performance improvement program.

B. Setting Initial Baselines & Thresholds

PlantTriage comes with a set of BL & TH initial templates. These templates are based on the experience of ExperTune and our field service partners. We recommend that you start by using the initial templates for each type of control loop (Flow, Level, etc.). They will set BLs & THs, based on the type of control loop. Set All Initial Baselines & Thresholds From the Engineering interface: 1. Choose the button for Set Baselines & Thresholds. 2. Choose a loop. 3. Click OK on the Select a Performance Assessment for Trending window. 4. At the bottom center of the screen, click on the Set Initial button.

C. Setting Baselines & Thresholds Based on Historical Data

After you have sufficient historical data (a minimum of one full week), use PlantTriage templates to update baselines & thresholds. From the Engineering interface: 1. Choose the button for Set Baselines & Thresholds. 2. Choose a loop. 3. Click OK on the Select a Performance Assessment for Trending window. 4. Use the trend window to dial in a time span with good overall process operation. 5. On the Template executor window lower left select a pairing of point groups from the drop down selections (Example: EntirePlant / Flow). 14
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

6. In the lower right of the same window (Template) select the appropriate template to go with the point grouping (Example: Flow). 7. Click Calculate button, then the Apply template for entire group button. 8. Repeat step 5-7, going through each template type and corresponding grouping. 9. In the Browser interface select the Problem Solver report Admin-Key Baselines to scan for any points that may have been missed.

D. Set Post Baselines & Thresholds

From the Engineering interface: 1. Choose the button for Set Baselines & Thresholds. 2. Choose a loop. 3. Click OK on the Select a Performance Assessment for Trending window. 4. Use the trend window to dial in a time span with good overall process operation. 5. On the Template executor window lower left select a pairing of point groups from the drop down selections (Example: EntirePlant / Casc-Primary). Note: you must have already created the display group. 6. In the lower right of the same window (Template) select the appropriate template to go with the point grouping (Example: Post-CascPri). 7. Click Calculate button, then the Apply template for entire group button. 8. Repeat step 5-7, for the cascade secondaries.

E. Fine-Tuning Baselines & Thresholds

With some time and experience, you will start making individual adjustments to baselines and thresholds, increasing the accuracy of diagnostic information. For specific information on how to do this, consider additional on-site training, or contact ExperTune for more information.

15
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

V. User Considerations
The PlantTriage performance supervision system has three types of users; Administrator, Analyst, and Browser. Before an individual can access either the engineering interface or the browser interface they must be registered. The one exception to this would be if no users are registered or all users are registered as administrators. The user management functions are performed from the administrator window (Menu: Tools, Administrator, Users tab). An individual must always log-in to the browser interface regardless of registered level. The engineering interface does not require log-in unless there are non-administrator level users registered.

A. Engineering Interface Access

An administrator has access to all facilities in the engineering interface and the browser interface. The analyst user has access to all facilities in the Browser interface, but a limited set of facilities in the Engineering interface. Analysts can view data, trends, invoke the Tuner/Analyzer and set loop engineering notes. The only configuration changes an analyst can make are related to loop tuning and analysis: Process Variable and Controller Output spans, process direction (up/down) and Controller Output limits. The browser users have access to all facilities in the Browser interface. They have no access at all to the Engineering interface.

B. Browser Access
When a new user (Analyst, or Browser) is added to the system, the Administrator should assign that user to a reporting role, and establish which Plants and Units will be visible to the user. This is done in the browser interface using the left hand menu: Support/Manage user properties. The Report Role governs which reports and dashboards the user can access. Administrators can create new Report Roles, see the next section for more information. A users scope can also be limited by editing the Visible Unit Operations tab. This can help reduce the volume of information that may appear on reports. Plants and unit operations that are outside of the user's visible unit operations do not appear in reports, dashboards, treemaps, etc. A unit control engineer in one area of the facility may have no interest in loops showing up in reports that have nothing to do with their area or responsibilities.

C. Browser Report Roles Management

Administrators use the Manage Report Roles (Left Menu: Support/Manage Report Roles) screen to define a number of Report Roles and the reports allocated to each Report Role. Several standard Report Roles are provided: View All Reports, Control 16
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.

Engineering, Maintenance, Operations, Plant Management, and System Administration. You are free to change the standard report roles or create your own. As users create reports and dashboards and make them public an administrator might be asked to add them to certain roles. This only applies to public reports and dashboards, a user may always view the private reports they have created themselves. Note that if a user created public report has a user created drilldown dashboard they must both be added to the report role. Administrators can also set which dashboards appear in the specific drill down situations shown at the bottom of the manage report roles page. An administrator may customize the three gauges that appear in the upper left of the Plant Performance Dashboard (Home Page). The assessment shown, the scale, and the drilldown report or dashboard can all be modified for different user roles.

17
PlantTriage and the PlantTriage logo are registered trademarks of ExperTune, Inc.