JIP Meeting Western Australia University

Flow Assurance
Workshop
1 June 2017
UWA Flow Assurance Workshop
Thursday 1 June 2017
8:30 AM. Coffee on arrival
8:45-9:00 AM. Overview of Flow Assurance Research Activities (Zach Aman)
Theme #1: Assessing Hydrate Risk
9:00-9:30 AM. Subcooling and Induction Time Measurements of Probabilistic Hydrate Formation (Vincent Lim)
9:30-10:00 AM. Scale Up Tests on Hydrate Formation and Deposition in High Pressure  
Visual Autoclave (Temi Kuteyi and Ting Huang)
10:00-10:20 AM. Improved Turbulence Modelling to Transfer Laboratory Results to the Field (Craig Booth)
10:20–10:40 AM. Morning coffee/tea break
10:40-11:10 AM. Rapid Simulation of Hydrocarbon Solid Formation and Deposition (Corey Baker)
Theme #2: Novel Management Methods
11:10-11:40 AM. Hydrate-in-Oil Slurry Rheology Measurements (Yahua Qin)
11:40-12:10 PM. Hydrate Cohesion in the Presence of Naturally Occurring Surfactants (Shane Morrissy)
12:10-12:30 PM. Guest presentation from Jill Stadjuhar (NERA) on collaborative research opportunities
12:30-1:30 PM. Lunch with Optional Laboratory Tour
Meeting location: Kensington Research Lab

Australian Resources Research Centre
26 Dick Perry Avenue
Technology Park
Kensington, WA 6151
The University of Western Australia

Overview of
Flow Assurance Research Activities
Zachary M. Aman
1 June 2017
Fluid Science & Resources Division

Academics Postdocs Collaborators
Eric Mike Zach Paul Brendan Einar Jianwei Mosayyeb Peter

May Johns Aman Stanwix Graham Fridjonsson Du Arjmandi Metaxas
PhD Students in Flow Assurance Research Areas

Gas Processing and LNG
Bruce Mauricio Yahua Shane Thomas Kwanghee Linda Separation and Adsorption
Norris Di Lorenzo Qin Morrissy Charlton Jeong Azizi
NMR and Fluid Properties
Hydrates, Flow Assurance

Zaki Craig Vincent Temi Angus Louis Shunsuke
Abidin Booth Lim Kuteyi Mckenzie Yu Sakurai
Laboratory Capabilities: microns to miles
Micro-DSC Raman Interfacial Tension

Oil Properties Dispersions Hydrate Plug Spectrometer & Wetting Angle
(GC, SARA) (NMR) Dissociation
3x Flowloop
Hydrate Sapphire
Cohesion Rheometer Autoclaves
1
2017+ Flow Assurance Research Themes
1) Assessing Hydrate Risk = Probability × Severity
+ + +
Hydrate Growth
Particle Bedding Slurry Viscosity Wall Deposition
Rate
• Across oil, gas, and water-continuous pipelines

• Horizontal, inclined, and complex pipe geometries
• Probabilistic flowline simulation with hydrates
2) Novel Management Methods
• Under-inhibition with MEG or methanol
• Novel hydrate inhibitors (e.g. AAs) for sensitive environments
• Identifying novel oil constituents that manage blockage risk
Assessing Hydrate Blockage Risk

Capturing probabilistic behaviour
• Hydrate nucleation and induction time (Vincent)
• Fluctuations in production rates and fluid conditions
Hydrate growth rate
• Modified kinetic equation used at present for oil systems
• Improvements required for transport-limited growth (Temi/Ting)
Impact on flow conditions
• Transferring knowledge from laboratory to field (Craig)
• Improving the accuracy of hydrate slurry viscosity (Yahua, Shane)
• Quantifying bed formation and behaviour (Masoumeh, Mauricio)
• Quantifying sloughing requirements (Mauricio, Tom)
Identifying and probing additional contributions (e.g. wax + hydrate)
Developing Assessment Tools

HyFAST 3 (Beta in 2017)
• Transient multiphase flow simulation with hydrate models
• Rapid screening capability for probabilistic assessment
OLGA Gas-Dominant Plugin (Beta Available Now)

• Hydrate growth rate, deposition, and pressure drop in horizontal flow
• Currently deployed with Woodside, Chevron
CryoFAST and ThermoFAST (Beta Available Now)

• Rapid simulation of wax buildup in LNG systems
• Stand-alone thermodynamic tool for hydrocarbon systems
2
June 2017 Presentation Schedule
8:45 AM Overview of Flow Assurance Research Activities
9:00 AM Subcooling and Induction Time Measurements of Probabilistic Hydrate Formation
9:30 AM Scale Up Tests on Hydrate Formation and Deposition in High Pressure Visual Autoclave
10:00 AM Improved Turbulence Modelling to Transfer Laboratory Results to the Field
10:20 AM Morning coffee/tea break
10:40 AM Rapid Simulation of Hydrocarbon Solid Formation and Deposition
11:10 AM Hydrate-in-Oil Slurry Rheology Measurements
11:40 AM Hydrate Cohesion in the Presence of Naturally Occurring Surfactants
12:10 PM Guest presentation from Jill Stadjuhar (NERA) on collaborative research opportunities
12:30 PM Lunch with Optional Laboratory Tour
THANK YOU
3
Subcooling and Induction Time Measurements
of Probabilistic Hydrate Formation
Vincent Lim, Dr. Peter Metaxas, Dr. Juwoon Park
1 June 2017
Personal Background
2015 UWA honours graduate in Bachelor of Chemical &
Process Engineering
Second year PhD student in FSR, ECM, UWA

• Probabilistic hydrate formation risk
• Interested in pursuing in flow
assurance field for future career
Supervisors
Associate Prof. Zach Aman
Prof. Eric May
Prof. Mike Johns 2
Conclusions
• With 700 rpm shear, mean subcooling of ethane hydrates at
superheating of 1 K is reduced by factor of 1.8
• Mean induction time of methane hydrates increased by a factor

of 2 when subcooling decreased from 7 K to 5 K
• Preliminary induction time data obtained is consistent with

Kaschiev’s proposed model for mononuclear mechanism
1
Motivation: 3.6 K Subcooling Serves as Industry
Heuristic to Simulate Hydrate Formation
• Benchmark against 5 field tests done by DeepStar on Werner Bolley line
• Minimum subcooling of 6.5 ˚F (≈ 3.6 K)
• Used in flow assurance & hydrate simulation software: OLGA, CSMHyK
Matthew et. al., Annals of the New York

Academy of Sciences, 2000, 912, 330-338
4
Recent Development of Apparatus to Measure

Hydrate Formation – HP-ALTA
Apparatus Brief Descriptions
High Pressure Automated Lag • Interfacial transmittance configuration
Time Apparatus (HP-ALTA) • Sudden decrease in transmitted light intensity
when solid forms
Advantages Limitations
• Can generate a • Only works in
lot of data points quiescent systems
• Large range of • Only works for
cooling rates transparent samples
Maeda et. al., Energy and Fuels, 2012, 26, 1820-1827 5
High-Pressure Stirring Automated Lag Time Apparatus
HP hydrate nucleation cells

Single-pass flow loop (CSIRO)
First
HP sapphire autoclaves
Generation
HP steel autoclave
Hydrate plug cells
Raman spectroscopy
Low-field NMR
High-pressure rheometer
HP differential scanning calorimeter
Interfacial tensiometry Second
Micromechanical force apparatus Generation
Hydrodynamic simulation
6
2
Hydrates Nucleation Studies
BASIC HPS ALTA

MEASUREMENTS OF SUBCOOLING
HPS-ALTA is more Efficient in Generating Hydrate

Formation Events
• Stirred system
• Direct measurements of PT data
Autoclave / Rocking Cells
• Able to generate large number of experimental runs
• Small volume (~3mL) HP-ALTA
Hydrate Formation Detected by Automated Analysis
Equilibrium
curve
2. Constant cooling Subcooling
1. Heating system
to Tmax of 40 °C
3. Sudden pressure Python automation

drop indicates
hydrate formation
3
Concept of Memory Effect - Lower Subcooling
Required for Hydrate Reformation
Importance of Memory Effect: Quiescent system
• Lower driving force required for hydrate
reformation (low subcooling / induction time)
• Troublesome reformation of serious blockages
after plugs have been decomposed in pipeline
Superheating 16 K
Pressure
Superheating 6 K
May et. al., Chemical Engineering Science, 2014, 107, 1-12

Temperature
Sloan and Koh, Clathrate Hydrates of Natural Gases, 2008 Superheating = Tregeneration - Tequilibrium 10
Memory Effect Easily Destroyed by 700 rpm Shear

Ethane Gas Hydrate Formation
Key Points:
• At 700 rpm, a superheating
of just 2.5 K produced
similar formation distribution
Tregen = 40 °C
Tregen = 13.5 °C Mean subcooling
as a superheating of 27.5 K
Mean subcooling Tregen = 15 °C = 8.1 K
= 4.4 K Mean subcooling
= 8.0 K • To unambiguously observe
the memory effect with C2 at
700 rpm, the superheating
had to be reduced to 1 K °C
11
Hydrate Formation in Quiescent System is Restricted

by Mass Transfer Limitations
Ethane Gas Hydrate Formation
rpm Data Pts. Mean (K) Std. Dev. (K) Max. (K) Min. (K)
700 129 4.42 2.20 10.09 1.30
0 105 11.4 1.20 13.82 7.87
(Both experiments should be subject to memory effect, 1 K superheating) 12
4
Hydrates Nucleation Studies
MULTI HPS ALTA

MEASUREMENT OF INDUCTION TIME
Schematic Diagram & Design of Multi-cell System

Magnetic Stirrer: Provide magnetic force National Instrument DAQ control system:
for stirring bar with required stirring speed • Communication ports between PWM and peltiers
(up to 1000 rpm) • Transfer signals from pressure transducer and RTDs
Chiller PWM (Pulse width modulation):

PSU (Power Supply Unit):
Control Cell Temperature with
Provide power to drive PWM
Glycol Chiller: Remove excessive large range of cooling rates (up to
and pressure transducer
heat generated by peltiers 6 K/min)
14
Measurements of Hydrates Induction Time
One complete
cycle Pressure drop indicates
hydrate formation
System heat up after

hydrate formation detected
Constant cooling trace
Induction Time
15
5
Exponential Distribution of Induction Time for
Methane Hydrate Formation
Methane Gas Hydrate Formation
7K subcooling
Key Points:
• Stochastic nature of hydrate
formation shown by large range of
experimental induction time
5 K subcooling • Lowering the subcooling from 7 K

to 5 K increase the mean induction
time by factor of 2
Subcooling # Data Median Std. Dev.

Gas Mean (s) Max. (s) Min. (s)
(K) Pts. (s) (s)
7 1 093 1 986 109 16 650 333 940 5
Methane
5 244 3 801 331 9 460 64 118 11 16
Large Data Points Required for Induction Time

Convergence
Methane Gas Hydrate Formation, 7 K subcooling
Repeatability test of Cumulative Mean

Rocking cells / Autoclave of HPS-ALTA
Repeatability test of HP-ALTA
17
Consistency with Mononuclear Mechanism at Low

Induction Times
Kaschiev’s exponential model:
ln(1-P) versus t data would be linear if

consistent with the exponential
distribution predicted for the
mononuclear nucleation mechanism.
18
6
Consistency with Mononuclear Mechanism at Low
Induction Times
1K increase in subcooling
increases JsV by a factor of
~2.2
19
Conclusions
• With 700 rpm shear, mean subcooling of ethane hydrates at
superheating of 1 K is reduced by factor of 1.8
• Mean induction time is increased by a factor of 2 when

subcooling decreased from 7 K to 5 K
• Preliminary induction time data obtained is consistent with

Kaschiev’s proposed model for mononuclear mechanism
20
Way Forward
• Develop predictive hydrate model for nucleation rate based

on Kaschiev’s “Nucleation”
• Rank the performance of inhibitors such as KHIs and AAs in

term of subcooling and induction time
• Predict hydrate formation in gas-oil-water system
21
7
ACKNOWLEDGEMENTS
Thor Frette
Martin Fernie
Gert Haandrikman
Daniel Crosby
George Broze
Questions
THANK YOU
8
Scale Up Tests on Hydrate Formation and
Deposition in High Pressure Visual Autoclave
Temi Kuteyi, Ting Huang
1 June 2017
Temi Background
• B.Sc. Chemical Engineering
- Curtin University of Technology
(2008 - 2011)
• Safety and Risk Engineer

- Lloyd’s Register
(2012 – 2015)
• M.Sc. Oil and Gas Engineering

- University of Western Australia
(2015)
• PhD. Oil and Gas UWA (2016 Feb)

Prof. Zach Aman
Prof. Eric May
Prof. Michael Johns
Conclusions
Improved mass transfer limited model can predict

hydrate growth rate in gas-water systems
• Gas transport based on the simple film theory
Reynolds Number (calculated from impeller properties) is

not a comparable parameter in autoclaves
Plugging-type behavior above 30 vol% hydrate

• Bed formation and wall deposition enhanced hydrate growth
1
High Pressure Visual Autoclaves

HP steel autoclave
Hydrate plug cells
Raman spectroscopy
Low-field NMR
HP differential scanning
calorimeter
Interfacial tensiometry
Micromechanical force apparatus
4
Motivation: Deposition of Hydrates Leads to Plug

Formation
ϕtransition
Joshi et al., Chem. Eng. Sci, 2013
Hydrate particles agglomerate and form moving bed

• Detected by increase in motor torque
• Critical stage (ϕtransition) observed before plug formation
• Increase hydrate growth rate
5
High Pressure Visual Autoclave

1” Autoclave High Pressure Cell
• 1” Sapphire cell
• 77 mL Volume
• Vane-blade mixing geometry
• 100 – 1500 RPM
• Constant pressure or volume
2” Autoclave High Pressure Cell

• 2” Sapphire cell
• 500 mL Volume
• Vane-blade mixing geometry
• 100 – 1500 RPM
• Constant pressure or volume
6
2
Experimental Methodology
Basic Procedure Sudden
pressure drop
1. Fill the cell with liquids (water/oil)
2. Purge 3 times with selected gas Model Region
3. Pressurize with stirring

4. Cool below hydrate equilibrium
5. Record temperature, pressure and
motor torque
Hydrate Formation
Experiment starts
at initial conditions
Heat system up to Constant cooling

initial condition rate to 1°C / h
Hydrate formation
point detected
7
Ting Huang
HYDRATE GROWTH MODEL
My Background
B.Sc. Oil and Gas Storage and Transportation

-Southwest Petroleum University
(2009-2013)
PhD. Oil and Gas Storage and Transportation

-Southwest Petroleum University
(2013 Sep)
Visiting Student Flow Assurance

-UWA
-Under the grant of China Scholarship Council
(2016 Oct)
Prof. Zach Aman
9
3
Hydrate Growth Rate Predicting Models
§ Kinetic model (used in industry OLGA) § Mass transfer limited model (OLGA)
dm
dm
dt
= kB Ag-w ΔT VS dt
= kmass Ag-w CCH
water
4
(
- CCH
water-hyd
4
)
Different
kB Kinetic constants resistance kmass Mass transfer coefficient
water Gas concentration in water

CCH EoS describes the G-W equilibrium
Different 4
DT Subcooling
driving force
water-hyd
CCH EoS describes the water phase
4 vdW-P model describes hydrate phase
Ag-w Gas water interfacial area Ag-w Gas water interfacial area
10
Comparison to Autoclave Experiments

Mass transfer limited model Kinetic model
6000
Relative deviation %
5000 140%
4000
3000
2000
1000
0
200 400 600 800
Stirring speed/rpm
§ Mass transfer limited model has an average § More than 140% deviations still exist
relative deviation of 509% between experimental data and mass
§ Kinetic model has an average relative transfer limited model
deviation of 3400%
§ Mass transfer limited model is more accurate § How to improve the model?
11
Solving the Mass Transfer Limited Model
water viscosity hydrate slurry viscosity

••Determine kmass 4.0
Step 1 = f (T , r L , rg , µL , vL ) 3.5
3.0
viscosity/cp
••Determine Ag-w 2.5
Step 2 = f ( dB , H L ) 2.0
1.5
1.0
••Calculate driving force 0% 5% 10% 15% 20%
Step 3 hydrate volume fraction
(Joshi, Sanjeev V. 2012)
§ Use the hydrate slurry viscosity instead of water

viscosity to determine mass transfer coefficient
12
4
§ Measure the height of the liquid phase to get HL

••Determine Ag-w
§ Measure the size of gas bubbles in the liquid phase
Step 2 = f ( dB , H L )
§ Compare to 4 correlations to calculate dB
H L = -0.000514 ( rpm ) + 0.997

R 2 = 0.991
13
§ Use different equations of state to calculate

••Calculate driving force methane concentrations in water under G-W
Step 3 and H-W equilibria
G-L CPA H-L CPA H-L VPT G-L CPA H-L CPA H-L VPT
0.0025 0.0029
0.0025
Gas Solubility
Gas Solubility
0.0021
0.0021
0.0017
0.0017
0.0013
0.0013
0.0009 0.0009
274 276 278 280 40 50 60 70 80 90
Temperature/K Pressure/bar
60 bar 274.15 K
14
Model Predictions Comparable Autoclave Experiments
Current mass transfer limited model (OLGA)

900
Relative deviation %
Improved mass transfer limited model
600
300
0
200 400 600 800
Stirring speed/rpm
§ The average relative deviation drops from § Need to be tested in gas-water pipelines
509% to 62%
§ High water cut experimental data from flow
§ Hydrate growth rate is well predicted by loops could be utilized to verify the improved
the improved mass transfer limited model model (Sanjeev Joshi’s data)
15
5
Temi Kuteyi
SCALE UP TESTS ON HYDRATE FORMATION AND

DEPOSITION IN HIGH PRESSURE VISUAL AUTOCLAVE
Comparison between Two Autoclaves – 800 RPM

Two repeat trials at 800 RPM
• Regional boundaries (I – III) identified from relative torque data
• Region II onset corresponds to ϕtransition
Region I Region II Region III
Joshi et al., Chem. Eng. Sci, 2013
Stirring speed not an appropriate parameter to account for transition point in

two different geometries 17
Observation of ϕ transition From Three Independent

Sources
Pressure Consumption No hydrates 7.6 vol%
• Sharp decrease in pressure

– slope change
Motor Torque
• Slight increase in torque
first observed
Visual Observation 18 vol% 45 vol%
• Homogenous and
heterogeneous distribution
of hydrate particles
• Deposition on wall
18
6
Definition of Reynolds Number in Stirred Tank
Stirred Tanks (Nauman, 2008)
ρ N I DI
Re =
µ
ρ = Fluid Density
NI = Rotational Speed
DI = Impeller Diameter
DI
μ = Viscosity
19
Comparison between Two Autoclaves – Reynolds

Re = 6150 Re = 4980
• 1000 RPM in One inch cell • 800 RPM in One inch cell
• 255 RPM in Two inch cell • 205 RPM in Two inch cell
Reynolds number is not a representative parameter to account for

turbulence in hydrate formation 20
Conclusions
Improved mass transfer limited model can predict

hydrate growth rate in gas-water systems
• Gas transport based on the simple film theory
Reynolds Number (calculated from impeller properties) is

not a comparable parameter in autoclaves
Plugging-type behavior above 30 vol% hydrate

• Bed formation and wall deposition enhanced hydrate growth
21
7
Way Forward
Plug the improved mass transfer limited model into

OLGA and apply to high water cut pipeline systems
Instead of calculating Reynold’s Number from impeller

properties (diameter and mixing speed), calculate
Reynolds from fluid properties (CFD simulation).
22
Questions
THANK YOU
8
Improved Turbulence Modelling to
Transfer Laboratory Results to the Field
Craig Booth
1 June 2017
CFD: Computational Fluid Dynamics or Colourful Fluid

Diagrams?
2
El Khoury, G.K. et al, 2013, Flow Turbulence Combust, 91:475-495
Conclusions
Simulations show no correlation between turbulence and

Reynolds number for autoclave systems
Autoclave Reynolds correlation updated

• Turbulent Kinetic Energy (TKE) accuracy increased 54%
• Turbulent Dissipation Rate (TDR) accuracy increased 150%
Results suggest new approach to estimating Reynolds number for

dispersing systems
• Initial effort improved overall TKE/TDR accuracy by 550%
1
Craig Booth
BEng and BSc at UWA
• Mechanical Engineering & Physics
• Graduated November 2015
PhD at UWA
• Simulation of Deepwater Blowouts
• A/Prof. Leggoe & A/Prof. Z Aman
Work presented at
• AFMC 2016
• GoMOSES 2017
http://www.news.com.au/technology/environment/bps-gulf-of-
mexico-deepwater-horizon-oil-spill-left-a-bathtub-ring-on-the-sea-
floor-scientist-david-valentine-says/news-
story/00d095f5581328af7d67609a198b6347 4
UWA Capabilities
HP steel autoclave
Hydrate plug cells
Raman spectroscopy
Low-field NMR
Use CFD to Explore Turbulence in Dispersing Systems

We can use CFD to inspect the resolved fields
Use turbulent properties to categories and compare flow fields across

apparatuses
Turbulent Kinetic Energy (TKE, k) is the energy of the fluctuations of the

mean flow
Turbulent Dissipation Rate (TDR, ε) is the rate that the turbulence is

converted into heat
2
Choice of Turbulence Model is Complicated
• DNS
Direct Numerical Simulation
Very Expensive, out of reach for practical flows
No modelling
• LES
Accuracy, cost
Large Eddy Simulation

Expensive, Difficult to get correct
Practical Reynolds Numbers possible on Supercomputing Equipment
• SAS
Scale Adaptive Simulations (Augmented RANS)
Experimental
Possibility to be a good balance between RANS and LES.
• RANS
Reynolds Averaged Navier Stokes
Models all turbulence scales
Inexpensive, Steady-State is possible, Turbulence Models are highly tested
7
Simulation Geometry Simplified as First Approximation

Autoclave Pipe
12.7mm
254 mm
8
CFD Parameters Varied to Ensure Convergence
Maximum TDR for different Maximum TDR for Mesh M2

Meshes at varying timesteps
200 200
Maximum TDR (m2/s3)
Maximum TDR (m2/s3)
180
180
160
160
140
140
120
100 120
80 100
0 0.02 0.04 0.06 0 0.02 0.04 0.06
Time (s) Time (s)
M1 (Coarse) M2 (Medium) M3 (Fine) 750µs
T1 = 0.00075s 375µs 187.5µs
T3 = 0.0001875s 93.75µs
T4 = 0.00009375s
T2 = 0.000375s
9
3
Results Enable Visualisation of the Flow Field
10
You Can’t Directly Match Re to TDR or TKE

Reynolds Number (Autoclave): Reynolds Number (Pipe):
∗
()*+, (2*/
!"#$%& = !"/0/1 =
- -
1 1000
Turbulent Kinetic Energy (m2/s2)
Turbulent Dissipation Rate (m2/s3)
Autoclave Pipe
Autoclave Pipe
100
0.1
10 2.304
1.57
0.01
1
0.001
0.1
500 5000 50000 500000
500 5000 50000 500000
Reynolds Number Reynolds Number
11
Mean Flow Fields Appear Dissimilar

Pipe Autoclave
Velocity
3.11 Velocity Magnitude
1.75
0.00
;/=
Turbulent Kinetic Energy

0.00812 Turbulent Kinetic
Energy
0.00406
0.00
;,/= ,
12
4
Maximum TDR Identified in Autoclave and Pipeline
Pipe Autoclave
• In the Autoclave maximum TDR occurs across a very small patch

~40 to 50 microns long by 2-3 microns wide
• 1-2 microns off the tip of the vane
• In the Pipe Maximum TDR occurs in the wall region and is not
periodic
13
LES Resolves Fluctuations in Pipe Flow
14
LES Resolves Fluctuations in Pipe Flow
15
5
Autoclave Reynolds Number Updated
• Dtank = Circumference of blade tip (instead of diameter)
• Recovers tip velocity as U
• Improve correlation on TKE by 54.2% and 150% for TDR
1
Turbulent Kinetic Energy
Autoclave Pipe
0.1
(m2/s2)
0.719
0.01
0.001
1000 10000 100000
Reynolds Number
16
Future Work: Using Resolved Data to Modify

Reynolds Number
• Breaking occurs due to the stresses caused by velocity
fluctuations → TKE
• Amount of time exposed to high turbulence ~ 1/U
• Use mean velocity and hydraulic diameter for U and D.
• New estimate of ‘velocity’ based on turbulent contributions:
BC,
2A =
2
• Proposed modification to Reynolds number for scaling:
2A * (
!"ADE0F01E =
-
17
Future Work: Log of TDR vs Modified Reynolds

Shows an Almost Linear Trend
1000
100
10
0.1
Modified Reynolds Number
Autoclave Pipe 18
6
Future Work: Modified Reynolds Number Improves
TDR Correlation by 550%
1000
Autoclave Pipe
100
10
0.1
100 1000
Modified Reynolds Number
19
Conclusions
Simulations show no correlation between turbulence and

Reynolds number for autoclave systems
Autoclave Reynolds correlation updated

• Turbulent Kinetic Energy (TKE) accuracy increased 54%
• Turbulent Dissipation Rate (TDR) accuracy increased 150%
Results suggest new approach to estimating Reynolds number for

dispersing systems
• Initial effort improved overall TKE/TDR accuracy by 550%
20
Questions
THANK YOU
7
Rapid Simulation of Hydrocarbon
Solid Formation and Deposition
Corey Baker
1 June 2017
Corey Baker
BEng Chemical & Process at UWA
• Graduated in 2013
PhD (Final Year) at UWA

• Avoiding Hydrocarbon Solid Formation
• Software Development
Internship with ExxonMobil in 2016

• Altona Refinery in Melbourne
• Business Development Group
International Conferences Attended

• Porto (ECTP), San Antonio (GPA)
Thesis Supervisors
• E. May, Z. Aman, M. Johns, T. Hughes
UWA Capabilities | Hydrocarbon Solid Modelling

HP steel autoclave
Hydrate plug cells
Raman spectroscopy
Low-field NMR
1
UWA Capabilities | Hydrocarbon Solid Measurements
HP steel autoclave
Hydrate plug cells
Raman spectroscopy
Low-field NMR
Conclusions
New tools for hydrocarbon solid formation and deposition

ThermoFAST Thermodynamic Calculation Tool
CryoFAST Hydrodynamic Simulation Tool
Inbuilt models can be applied to any hydrocarbon solid

• Including waxes, aromatics and lighter alkanes
Models are tuned regularly to new hydrocarbon solid data
Lack of experimental hydrate data at extreme conditions

• Low temperature, low water content
Risk of Hydrocarbon Solids in Oil and Gas

Hydrocarbon Solids Comps. Tmelt
Paraffin Waxes C20+ > 309 K (36°C)
Benzene 278 K (5°C)
Aromatics Toluene 178 K (-95°C)
p-Xylene 286 K (13°C)
Light Alkanes (C1 to C20) < 309 K (36°C)
Crude Oil
Natural Gas
Feed changes
Slug Flow Solid formation
↑ Heavy Components Cold Spots
↑ Heavy Components + deposition
↓ Temperature
↑ Pressure Drop
2
ThermoFASTThermodynamic Calculation Tool
Calculating Hydrocarbon Solids
Predicting the Formation of Hydrocarbon Solids
Find the fugacity of a pure solid

-→;
∆12,$ 45,$ ∆9(,$ 45,$ 4 ∆<$-→; . − .5
ln #$% = ln '()*+,$
-
. − −1 + − 1 + ln −
345,$ 4 3 4 45,$ 34
Calculate for phase equilibrium
#$% = #$-
Cubic EOS to calculate the liquid composition
#$%
#$% = =$ '$- . =$ =
'$- .
Optimise the BIP (kij) inside the '$- to tune to binaries
| User Interface
UWA (2016) Mix 1 Mix 2
n-Hexane 0.6727 0.9037
Hexadecane 0.3273 0.0963
277.0 K 261.6 K
Tmelt
3.9°C -11.5°C
Measured by UWA
3
| Tuned to Solid Formation Temperatures
n-Hexane 0.6727 0.9037
277.0 K 261.6 K
Tmelt
3.9°C -11.5°C
| Input Multiple Mixtures

n-Hexane 0.6727 0.9037
277.0 K 261.6 K
Tmelt
3.9°C -11.5°C
| Perform Rapid Flash Calculations

n-Hexane 0.6727 0.9037
277.0 K 261.6 K
Tmelt
3.9°C -11.5°C
4
| Locate Solid Formation Temperatures
n-Hexane 0.6727 0.9037
277.0 K 261.6 K
Tmelt
3.9°C -11.5°C
| Plot Solid Formation Graphs

n-Hexane 0.6727 0.9037
277.0 K 261.6 K
Tmelt
3.9°C -11.5°C
| Plot Solid Formation Graphs

n-Hexane 0.6727 0.9037
277.0 K 261.6 K
Tmelt
3.9°C -11.5°C
High Risk of
Deposition
Safe Zone
5
Industry Case Study | LNG
RasGas LNG Train 4 Case Study:

Adapted from Ismail & Al Thani, LNG-18
Blockage occurred here

100 kg of water was removed
Plant restart in Aug-2014
4 months later…
Water Content into the Cryogenic Heat Exchanger
How did water begin accumulating?

No operational changes
Restart here
Figure adapted from Ismail & Al Thani, LNG-18, Perth 2016
6
Heavy Carryover into the Cryogenic Heat Exchanger
Blockage occurred here
Operator’s lost reflux for

50 seconds upon restart!
C5+ Content into the Cryogenic Heat Exchanger
Could a tool predict this +

provide remedial action?
Figure adapted from Ismail & Al Thani, LNG-18, Perth 2016
Requirements for a Solid Risk Assessment Tool
Will solids Where and when Operational Remediation

freeze out? will solid’s deposit Symptoms? Methods
7
Case Study Simulation Results

Will Solids Freeze Out?

RasGas Case Study Feed Composition to MCHE
Normal
Upset Condition
Operating
(C5+ Carry Over)
Condition
CO2 0.000040 0.000040
CH4 0.909329 0.907510
C2H6 0.057492 0.057377
C3H8 0.021953 0.021910
iC4H10 0.004239 0.004230
nC4H10 0.006847 0.006833
iC5H12 0.000038 0.000796
nC5H12 0.000026 0.000544
C6H14 0.000021 0.000441
C7H16 0.000010 0.000241
0.000004 0.000078
Benzene
[3.7 ppm] [78 ppm]
Case Study Simulation Results

8
Simulating the Cryogenic Heat Exchanger
Expansion Valve (Further Cooling)
Mixed Refrigerant Liquefied Gas (LNG)
Temperature Profile ~122K (-151ºC)
~(110K to 200)K MODELS OUTPUTS
Reduced Pipe Increased
insulation Equation of State Vapour, Liquid +
Diameter
+ Solid Equation Solid Fractions
velocity Internal Tfluid
DP
Liquid Only Mass Flow Rates

Material Balance + Mass Deposition
Vapour + Momentum
Liquid Region Pressure Drop +
INLET: Balance
Phase Velocities
(99 to 100) (Steady-State)
mol% vapour
Distillation Column Exit Temperatures,

Process Upset: Energy Balance
Overhead Product Limitations due to
Heavy carry-over! (Heat Transfer)
>85mol% CH4 (210 to 240) K Deposition
| Simulating the RasGas Case Study
Determines Where in the Heat Exchanger Solids Form
Benzene deposited in
the upper cold bundle!
9
Instantly Determines the Operational Symptoms
0.05% Change
- Not Measureable
+8 K Change
With Temperature Control

+3% Refrigeration Duty
Benzene Should’ve been Removed Under Normal

Operational Conditions
~18g of Benzene Deposited per tube

Deposit should’ve been removed within 37 hours
But water kept building up…
Hydrates or Ice Form During Normal Operation?!

Normal Hydrate
0.12 ppm Operating (Type I) Pressure = 3.2 MPa
Condition CSMGem
H2O 0.000012 85.1913
CO2 0.004000 0.0008 The pressure close to
CH4 90.932900 13.4526 the benzene deposit
C2H6 5.749200 1.3553
C3H8 2.195300 - Solid Formation Temperatures
iC4H10 0.423900 - [Hydrate I] CSMGem 181.3 K [-91.8°C]
nC4H10 0.684700 -
[Hydrate II] Infochem CPA 206.8 K [-66.3°C]
iC5H12 0.002800 -
nC5H12 0.002600 -
[Ice] ThermoFAST 206.9 K [-66.2°C]
Peng-Robinson EOS
C6H14 0.002100 -
C7H16 0.002100 - [Hydrate I] ThermoFAST 191.7 K [-81.5°C]
Benzene 0.000400 -
Helmholtz Eq. (TREND*)
* Span, R.; Eckermann, T.; Herrig, S.; Hielscher, S.; Jäger, A.; Thol, M. (2016): TREND. Thermodynamic
TOTAL 100 % 100 % Reference and Engineering Data 3.0. Lehrstuhl für Thermodynamik, Ruhr-Universität Bochum.
10
Reducing the Likelihood of Blockage

Remedial Actions that the Operator’s Could’ve Taken

Maximum Window of Opportunity ≈ 37 hours
- Water crystallisation then becomes self-sustaining
(a) Increase the temperature by 2K ≈ 7 hours

(b) Reduce the feed’s C5+ Content by half ≈ 5 hours
(c) Both (a) and (b) ≈ 3 hours
Conclusions
New tools for hydrocarbon solid formation and deposition

CryoFAST
Hydrodynamic Simulation Tool
Inbuilt models can be applied to any hydrocarbon solid

• Including waxes, aromatics and lighter alkanes
Models are tuned regularly to new hydrocarbon solid data
Lack of experimental hydrate data at extreme conditions

• Low temperature, low water content
11
Way Forward | What’s Next?

CryoFAST
Hydrodynamic Simulation Tool
Get ThermoFAST Now: http://fsr.ecm.uwa.edu.au/capabilities/software/thermofast/
Get CryoFAST Now: http://fsr.ecm.uwa.edu.au/capabilities/software/cryofast/
Inclusion of waxes for subsea pipeline simulation
• Ability to specify pseudo-components
Install hydrate models into CryoFAST
• Simulate hydrate/ice build-up in the Cryogenic Heat Exchanger
ACKNOWLEDGEMENTS
Stanley Huang
Dan McCartney
Paul Pickering
Darren Rowland
Jordan Oakley
Questions
THANK YOU
12
Hydrate-in-Oil Slurry Rheology
Measurements
Yahua Qin
1 June 2017
My Background
• B.Sc. Oil and Gas Storage and Transportation
– China University of Petroleum
(2009-2013)
• M.Sc. Oil and Gas Storage and Transportation

– China University of Petroleum
(2013-2014)
• PhD. Oil and Gas UWA (2014 Aug)

Prof. Eric May
Prof. Michael Johns
A/Prof. Zach Aman
2
Conclusions
Current rheology model in OLGA differs from experiment data
from two crude oil systems of different viscosity by more than
70%
AA can be used to study underlying suspension behaviour

• AA slows down the hydrate growth by approximately 50%-90%
• AA reduces hydrate slurry viscosity by approximately 50%-70%
A new model is developed by replacing the underlying model in

current hydrate rheology model and found to greatly increase
the predictive accuracy
1
High-Pressure Rheometer

HP steel autoclave
Hydrate plug cells
Raman spectroscopy
Low-field NMR
HP differential scanning
calorimeter
Hydrodynamic simulation 4
Motivation: Current Hydrate Viscosity Model in OLGA

is Not Accurate Enough
!" = !$%& ⁄!'() +,% = .// ∗ (!"234 −!"6'&2) )⁄!"234
1100psi 1100psi
Light oil: 60 cP 20 cP Heavy oil: 598 cP CH4
108 cP
CH4
Average relative deviation: 73% Average relative deviation: 171%

Camargo and Palermo 2002 5
Current Hydrate Viscosity Model is Based on

Suspension Model and Aggregation Model
a) hard-spheres suspension model b) aggregation model
b
models must be tested by deconvoluting contributions

Mills (1985)
The effective volume fraction of
.−8 the aggregates:
!" =
8 :
(. − 8 ) 82?? = @(8$%& , BC , &4 , Ḋ , ?, !/ )
693
Φ particle volume fraction;
Genovese 2012
Camargo and Palermo 2002
Φ<=> the maximum volume fraction, 4/7; 6
2
AAs Used to Test Underlying Suspension Model
Turner 2006
Amphiphilic molecules of AA (surfactants):
Genovese 2012
Zerpa et al. 2011
Simpler and more homogeneous hydrate slurries

7
METHODOLOGY
Emulsion Preparation
1. Water Cut Confirmation : Prepare Oil and 1.2
De-ionized Water by volume ratio: 0.30,

0.25, 0.20, 0.15, 0.10, 0.05 1
0.8
2. Mixing of Oil and Water Phase: Shear the
Norm. probab.
crude oil at 17500 rpm and add DI water

drop by drop within 5 minutes 0.6
0.4
3. Ripening Process: Leave the emulsion at
room temperature for at least 2 hours
0.2
4. Dropsize Measurement: Put the emulsion 0

sample into 1T NMR to measure the 0.001 0.100 10.000
dropsize distribution Droplet radius (µm)
9
3
High-Pressure Procedure
Gas charging
1. Emulsion Saturation: !
• 20 ºC,1500 psi, shear rate 217 1/s
2. Hydrate formation: F
• 1 ºC, shear rate 43.5 1/s
• another 12 hours after hydrate first detected
3. Steady-state flow curve: !

• 1 ºC, shear rate 4-217 1/s and ramp down
• 5 ºC, shear rate 4-217 1/s and ramp down
Ḋ
10
RESULTS
AA Slows Down Hydrate Growth and Reduces

Hydrate Slurry Viscosity in Two Oil Systems
Light oil Heavy oil
2
3
No AA 2 3
1 With AA
12
4
Infinite Shear Viscosity Shows Current Underlying Rheology
Model is not Appropriate to Hydrate Suspensions
G = 0.187
Genovese 2012
Mills 1985
infinite shear viscosity Chong et al. 1971
Krieger et al.1959
Dabak et al.1986
Mooney 1951
13
Model for Rough Particle Suspensions is Appropriate

to Hydrate Suspensions
differential effective medium theories
(DEMT):
8
!" = [. − ]P[!]
. − 8N
.−( )∗8
ΦQ = 0.52 8N
S =5
[S]: the fitting factor (intrinsic viscosity)
which is determined by the particle shape
ΦQ : filling fraction, the critical

concentration at which the suspension
loss its fluidity, 8N ≤ 8693
I. Santamaría-Holek, Carlos I. Mendoza 2010
Genovese 2012 14
Accuracy of the New Model is Higher than Current One for

Hydrate-in-Oil Slurry Without AA Over Whole Shear Rate Range
1100psi 1100psi
Light oil: 60 cP 20 cP Heavy oil: 598 cP 108 cP
CH4 CH4
G = 0.220 G = 0.151
Average Relative Deviation (over whole shear rate range):

133% 3% 145% 25%
5
Accuracy of the New Model is Higher than Current One for
Hydrate-in-Oil Slurry without AA over Whole Shear Rate Range
1100psi 1100psi
Light oil: 60 cP CH4
20 cP Heavy oil: 598 cP CH4
108 cP
33% 48%
New Model Predicts More Accurate Values than Current One

During Restarting Process for Hydrate-in-Oil Slurry Without AA
1100psi 1100psi
Light oil: 60 cP CH4
20 cP Heavy oil: 598 cP CH4
108 cP
125% 59% 276% 28%

17
Camargo and Palermo 2002
Conclusions
Current rheology model in OLGA differs from experiment data
from two crude oil systems of different viscosity by more than
70%
AA can be used to study underlying suspension behaviour

• AA slows down the hydrate growth by approximately 50%-90%
• AA reduces hydrate slurry viscosity by approximately 50%-70%
A new model is developed by replacing the underlying model in

current hydrate rheology model and found to greatly increase
the predictive accuracy
18
6
Way Forward
• Rheo-NMR flowloop test (Mosayyeb Arjmandi)
NMR test section
• Optimize current rheology model in terms of maximum aggregate size

19
ACKNOWLEDGEMENTS
Paul Pickering
Nino Fogliani
Hosi Sabavala
Questions
THANK YOU
7
Hydrate Cohesion in the Presence of
Naturally Occurring Surfactants
Shane Morrissy, Angus McKenzie, Danyal Rasheed
1 June 2017
About Me
Graduate of UWA
• BE (Chemical & Process Engineering) &
BSc (Chemistry)
• 3rd Year PhD Candidate
Research Areas
• Chemical adsorption to hydrate
interfaces
• Low dosage hydrate inhibitors
My research is focused on the

development of novel quantitative
methods for testing the interfacial
behaviour of oilfield chemistries
2
Conclusions
Studying oil fractions presents reproducibility issues
• Difficult to control for the effect of biodegradation vs fractionation
Surface-active species can reduce hydrate cohesive force 99%
Surface active species may introduce safe time before shut-in

hydrates anneal
• Lower baseline cohesive force
• Increase time before sintering behaviour dominates
Unmodified surface-active oils can exhibit comparable

performance to industry anti-agglomerants
3
1
Motivation: Hydrates Sinter During Shut-in
Sintering will be driven by hydrate film growth along
water-oil interfaces
Oil
Bridge will convert
/sinter to hydrate
Hydrate
Particle
Oil-Water IFT
(γOW ) Liquid Bridge
Hydrate
Particle
Hydrate-Oil IFT
(γHO)
Aman et al. PCCP. 2011. 4
Motivation: What Makes Chemical Magic Oils Work?

Two types of magic oils
1. Mechanical (density, viscosity)
2. Chemical (hydrate active surfactants)
Chemical magic oils are biodegraded

• High concentration of carboxylic acids No Magic Oil
(surface active)
• Acid composition may play an
important role
• Minimal consideration given to other
active components in oil
Borgund et al. E&F. 2007
Aman and Koh, Chem. Soc. Rev. 2016
Aman et al., E&F 2017 3 wt% Magic Oil 5
Hydrate Cohesion in the Presence of Naturally Occurring Surfactants
METHODOLOGY
2
Cohesive Force and Film Growth Using
Micromechanical Force Apparatus
HP steel autoclave
Hydrate plug cells
Raman spectroscopy
Low-field NMR
MMF Quantifies Cohesive Force And Film Growth

Cooling
Carbon Fibre Jacket
Cantilever
Cantilever Hydrate
Sleeve Particle
Hydrate
Particle
Micromanipulator Cantilever Hydrate

sleeve Epoxy Particle
Capillary Carbon-
Holder fibre
8
Methodology: CyC5 Hydrate Preparation
(i) Place water droplets onto

the ends of cantilevers
(ii) Freeze in liquid nitrogen to
form solid ice particles
(iii)Transfer to CyC5 bath
which is below the ice
point
(iv)Raise temperature of
CyC5 bath above ice point
3
Methodology: Pull-Off Trial
Hooke’s Law
F = k ⋅ Δx
k Δx
F – Hydrate cohesive
force
k – Spring constant
Δx – hydrate break-off
displacement
500 µm
250µm&
10
Methodology: Measuring Film Growth Rate
Linear Film Growth Rate
Δx
RH =
Δt
200 µm
250µm&
11
RESULTS
4
Last Time: Naturally Occurring Surface-Active
Species Inhibit Film Growth
3
Hydrate Film Growth Rate [µm/s]
Hydrate Film Growth

Baseline 2.2 ± 0.2 [µm/s]
2
Free Resins
Binding Resins
1
Residual Asphaltenes
0
1 10 100 1000
Morrissy et al. E&F. 2016
Additive Mass Fraction [ppm] 13
Different Fractionations Can Lead to Different Active

Concentrations
This work only compares species from the same fractionation
process
3
Hydrate Film Growth Rate
Hydrate Film Growth

Baseline 2.2 ± 0.2 [µm/s]
2
Free Resins 1st
[µm/s]
Fractionation
1 Free Resins 2nd

Fractionation
0
1 10 100 1000
Surface-Active Species Reduce Hydrate Cohesive

Force by Over an Order of Magnitude
6
Radius Normalised Cohesive
5
Hydrate Pull-Off Force Baseline
Force (mN/m)
4.3 ± 0.4 mN/m

4
3 Free Resins
1 Residual Binding
Asphaltenes Resins
0
1 10 100 1000
5
Free Resins Supress Sintering Between Hydrates
20.0
No Surfactant
15.0
Force (mN/m)
10.0 70ppm Free

250ppm Free Resins
Resins Hydrate Pull-Off Force
Baseline 4.3 ± 0.4 mN/m
5.0
0.0
0.01 0.1 1 10 100 1000
Aman et al., PCCP 2011 Contact Time [min] 16
Surface-Active Species can Outperform Industry

Anti-Agglomerants
6
Current Gen Industry
5 Anti-Agglomerant
Hydrate Pull-Off Force
Force (mN/m)

4
3 Crude Oil
(Unmodified)
2
1
Isolated Free
Resins
0
1 10 100 1000 10000
Unmodified Surface-Active Oil Performs Comparably

to an Industry AA without the Emulsification Tendency
Unmodified oil presents lower emulsification risk than AA
6.00 60
Oil-Water IFT Baseline
Interfacial Tension (mN/m)
55 ± 5 mN/m
5.00 50
Hydrate Pull-Off Force
Force (mN/m)
4.00 40
3.00 30
2.00 20
1.00 10
0.00 0
0.0001 0.001 0.01 0.1 1 10
Addtive Mass Fraction [wt%] 18
6
Conclusions
Studying oil fractions presents reproducibility issues
• Difficult to control for the effect of biodegradation vs fractionation
Surface-active species can reduce hydrate cohesive force 99%
Surface active species may introduce safe time before shut-in

hydrates anneal
• Lower baseline cohesive force
• Increase time before sintering behaviour dominates
Unmodified surface-active oils can exhibit comparable

performance to industry anti-agglomerants
19
FUTURE WORK
Hydrate Adsorption Column

Past Work
• Work at SINTEF and University
of Bergen used freon hydrates Peristaltic
to identify hydrate-active Pump
petroleum fractions.
Cyc5
Hydrates
Future Work
• Adsorb surface-active Circulation
components onto hydrates Fluid: Cyc5
+ 1,10 wt% Oil
• Extract hydrates and analyse
using GC-MS
Erstad et al., E&F 2009 21 21
7
Questions
THANK YOU

JIP Meeting Western Australia University

Uploaded by

Copyright:

Available Formats

JIP Meeting Western Australia University

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

JIP Meeting Western Australia University

Uploaded by

Copyright:

Available Formats

Flow Assurance

8:30 AM. Coffee on arrival

8:45-9:00 AM. Overview of Flow Assurance Research Activities (Zach Aman)

Theme #1: Assessing Hydrate Risk

10:20–10:40 AM. Morning coffee/tea break

Theme #2: Novel Management Methods

11:10-11:40 AM. Hydrate-in-Oil Slurry Rheology Measurements (Yahua Qin)

12:30-1:30 PM. Lunch with Optional Laboratory Tour

Meeting location: Kensington Research Lab

The University of Western Australia

UWA Flow Assurance Workshop

Fluid Science & Resources Division

Eric Mike Zach Paul Brendan Einar Jianwei Mosayyeb Peter

PhD Students in Flow Assurance Research Areas

NMR and Fluid Properties

Hydrates, Flow Assurance

Laboratory Capabilities: microns to miles

Micro-DSC Raman Interfacial Tension

• Across oil, gas, and water-continuous pipelines

Assessing Hydrate Blockage Risk

Developing Assessment Tools

OLGA Gas-Dominant Plugin (Beta Available Now)

CryoFAST and ThermoFAST (Beta Available Now)

9:00 AM Subcooling and Induction Time Measurements of Probabilistic Hydrate Formation

10:00 AM Improved Turbulence Modelling to Transfer Laboratory Results to the Field

10:20 AM Morning coffee/tea break

10:40 AM Rapid Simulation of Hydrocarbon Solid Formation and Deposition

11:10 AM Hydrate-in-Oil Slurry Rheology Measurements

11:40 AM Hydrate Cohesion in the Presence of Naturally Occurring Surfactants

12:30 PM Lunch with Optional Laboratory Tour

Vincent Lim, Dr. Peter Metaxas, Dr. Juwoon Park

UWA Flow Assurance Workshop

Second year PhD student in FSR, ECM, UWA

• Mean induction time of methane hydrates increased by a factor

• Preliminary induction time data obtained is consistent with

Matthew et. al., Annals of the New York

Recent Development of Apparatus to Measure

Maeda et. al., Energy and Fuels, 2012, 26, 1820-1827 5

High-Pressure Stirring Automated Lag Time Apparatus

HP hydrate nucleation cells

BASIC HPS ALTA

HPS-ALTA is more Efficient in Generating Hydrate

Hydrate Formation Detected by Automated Analysis

3. Sudden pressure Python automation

May et. al., Chemical Engineering Science, 2014, 107, 1-12

Memory Effect Easily Destroyed by 700 rpm Shear

Hydrate Formation in Quiescent System is Restricted

MULTI HPS ALTA

Schematic Diagram & Design of Multi-cell System

Chiller PWM (Pulse width modulation):

Measurements of Hydrates Induction Time

System heat up after

Constant cooling trace

5 K subcooling • Lowering the subcooling from 7 K

Subcooling # Data Median Std. Dev.

Large Data Points Required for Induction Time

Repeatability test of Cumulative Mean

Repeatability test of HP-ALTA

Consistency with Mononuclear Mechanism at Low

ln(1-P) versus t data would be linear if