SABP-Z-032

Best Practice SABP-Z-032

18 April 2012

Dynamic Process Modeling Best Practice Document Responsibility: Responsibility: Process and Control Control Systems Department Department

Saudi Aramco DeskTop Standards Table of Contents

1 Introduction.................................................... 2 2

Conflicts with Mandatory Standards............. Standards.... ........... 2

3 References.................................................... 2 4 Definitions...................................................... 3 5

Model Development................. Development...... ...................... ................... .......... .. 3

6

Model Testing.............. Testing.............................. .............................. ................ .. 17

7

Model Documentation............... Documentation..... ...................... ................... ....... 17 Appendices

Appendix A – Model Fidelity Definitions............. 19 Appendix B – Simulation Platforms: Recommended Vendor List (RVL)....... 20

Previous Issue: New

Next Planned Update: TBD Page 1 of 20

Primary contact: Yuk San Man on 966-3-880-8018 Copyright©Saudi Aramco 2012. All rights reserved.

Document Responsibility: Process and Control Systems Department Issue Date: 18 April 2012 Next Planned Update: TBD

1

SABP-Z-032 Process Modeling Best Practice

Introduction 1.1

Purpose

This Best Practice provides dynamic process modeling guidelines for upstream and downstream simulation models. models. However, this does not constitute constitute as a minimum requirement, but must be understood as “in add ition to the minimum” that may be required per prevailing Saudi Aramco A ramco Standards and Project Design Basis. 1.2

Scope

The aim of this document is to define a framework for developing and documenting dynamic simulation models for support or study purposes. This document is not intended to be b e used for Operators Training Simulators (OTS) as they are mainly developed by outside vendors. 1.3

Disclaimer

This SABP is not intended to detail de tail all aspects of the configuration and does not include adequate information to enable it to be used as an instruction manual. The software vendor manuals will also need to be referenced and utilized.

2

Conflicts with Mandatory Standards In the event of a conflict between this Best Practice and other Mandatory Saudi Saud i Aramco Engineering Requirements, the Mandatory Saudi Aramco Engineering Requirements shall govern.

3

References Specific sections of the following documents are referenced within the body of the document. Material or equipment supplied to this best practice, shall shall comply with the referenced sections of the latest edition of these these specifications. Where specific sections are not referenced, the system shall comply with the entire referenced document. 

Saudi Aramco Documents

Saudi Aramco Engineering Procedure SAEP-364

Process Simulation Simulation Model Model Development Development and Support Support

Page 2 of 20


4


Definitions This section contains definitions for acronyms, abbreviations, words, and terms as they are used in this document. 

5

Acronyms and Abbreviations csv

comma separated values

DCS

Distributed Control System

EOS

Equation of State

ESD

Emergency Shutdown

GS

Grayson-Streed

HTC

Heat Transfer Coefficient

OTS

Operator Training Simulator

P&ID

Piping and Instrumentation Diagram

PR

Peng and Robinson

PRSV

Peng-Robinson Stryjek-Vera

RVL

Recommended Vendor List

SABP

Saudi Aramco Best Practice

SRK

Soave, Redlich-Kwong

UA

Overall Heat Transfer Coefficient x Area

ZJ

Zudkevitch Joffee

Model Development 5.1

Model Definition Like every well-defined task, the objective(s) of the activity must be defined clearly from the outset. Whether the model will be used for study, process support or simply as a training tool, its functionality must be defined cle arly. Once its use is defined, the modeling scope can be adjusted to fulfill the objectives of the activity. Once the objective(s) has/have been defined, the fidelity of the model will need to be addressed next. For a detail engineering study, a high fidelity model which simulates equipment and piping volumes and pressure drops (for gas side on ly) accurately, may be required. However, for control system verification; such as Page 3 of 20



developing a model to test a particular modification to the control scheme, a medium fidelity may be sufficient. See Appendix A for Model Fidelity definitions. 5.2

Mark-up P&IDs Mark-up (Simulation) P&IDs for the entire scope should be produced. The simplest way to produce the simulation P&IDs is to mark up existing plant P&IDs. The main purpose of these simulation P&IDs is to illustrate modeling scope concisely on drawings everyone is familiar with. These simulation P&IDs are not intended to reproduce DCS displays and thus only limited color lines are proposed. It is suggested that the following convention is used: Table 1 – Simulation P&IDs Color Convention Modeled Drawing Items Process lines

Dark pink

Instrumentation (transmitters, 1 valves and controllers)

Yellow

Equipment

Outline in dark pink

Model boundary

Green circle with a green center point

Note:

5.3

Highlight In

1

: Valves include Control valves, ESD valves, Check valves and any modeled manual valves.

Simulation Package Currently, P&CSD standardizes on AspenTech Hysys Dynamics as the simulation platform for dynamic simulations. However, this simulation platform may change from time to time due to technical and commercial reasons. All dynamic simulation models shall be developed using the applicable simulation platform from the approved department Recommended Vendor List (RVL), see Appendix B. A waiver, with justification, is required if an alternative software platform is to be use on technical grounds. The models shall be developed using the latest version/release of the approved simulation software package as agreed with P&CSD. For the purpose of this document, the simulation platform is assumed to be AspenTech Hysys.

Page 4 of 20


5.4


Thermodynamic Properties Selection For most single, two- and three-phase system, Peng Ro binson, PR, Equation of State (EOS) property method is recommended as it produces good stable results for these systems over wide range of temperatures and pressures. Another property method, SRK, Soave, Redlich-Kwong, will generally produce similar results as Peng Robinson. However, the applicable range for this method is significantly more limited than PR. In addition, this method is less reliable for non-idea systems. When SRK is selected, care should be taken to ensure the method continues to produce reliable results over the rang e of temperature and pressure to be considered. Other thermodynamic property methods are also available within the simulation package. If another property method is selected for the process reasons, cautions should be exercised to ensure the method is valid over the range considered. Hysys recommends the following correlations to be use on these systems: Table 2 – Recommend Property Methods for Systems Systems

Recommended Methods

TEG and Dehydration

Glycol or PR (Peng Robinson) - Hysys manual

Sour Water

Sour PR

Cryogenic Gas Processing

PR, PRSV (Peng-Robinson Stryjek-Vera)

Air Separation

PR, PRSV

Atm Crude Towers

PR, PR Options, GS (Gr ayson-Streed)

Vacuum Towers

PR, PR Options, GS (<10 mm Hg),

High H2 Systems

PR, ZJ (Zudkevitch Joffee) or GS (see T/P limits)

Reservoir Systems

PR, PR Options

Steam Systems Steam

Steam Package, CS or GS

Systems containing salts, electrolytes

Aspen Electrolyte NRTL

Amines based sour gas removal

DBR Amines property package

LNG

GERG 2008

In addition, Hysys also states over the ranges below, the following methods will rigorously solves for any single, two-phase and three-phase s ystem with high degree of efficiency and reliability:

Page 5 of 20



Table 3 – PR and SRK Validity Temperature and Pressure Ranges Method

Temp ( F)

Temp ( C)

Pressure (psia)

Pressure (kPa)

PR

> -456

> -271

< 15,000

< 100,000

SRK

> -225

> -143

< 5,000

< 35,000

For further detail on thermodynamic property methods, refer to the relevant section of the simulation software package selected for model d evelopment. For Amines systems, it is important to ensure that the model is tune d to one or two actual plant data points to confirm validity of the model. The column efficiencies for H2S and CO2 should be tuned in the absorber and regenerator. However, the property package selection ma y require revisiting, if unusually high or low efficiencies are required to match plant data. 5.5

Component-Slate Selection For multi-component systems, the streams shall be characterized with a range of components (component-slate) which will represent the process fluid closely. Usually, only one component-slate is used for a set of streams. It is not un-usual to have more than one component-slate in the model. However, streams using different component-slates are normally not allowed to be mixed with each other. If a design heat and material balance is available, the same component slate shall be used. However, if the slate is consisting of too many components (50+), as some steady state models often do, the speed of the dynamic model may be very slow. Depending on the situation, the slate may be distillate down to 30 to 40 components. For most dynamic simulation work, a slate of 30 to 40 is more than adequate. If the component-slate is reduced, it is advisable to build some test flashes in a test model to ensure the new slate still produces similar vapor and liquid split under the pressure and temperature ranges to be considered. For crude oil characterization, a mixture of pure and pseudo components are often used to characterize the D86 distillation curve. In most dynamic simulations, typical pure components used are ranging C1 to C6 with other pure components such as N2, O2, CO2, H2S, H2O and H2 also included. In addition, pseudo components starting from 120 F and go all the way up to 1650 F, in step of 10 to 20 F or higher may be used to characterize heavier constituent of the crude in the defined temperature range. The cut point range is depending on the level of accuracy required. After selecting the initial slate, it is advisable to 





Page 6 of 20



carry out some test flash to confirm the acceptability of the slate. Adjust the cut point range to achieve the level of accuracy required. To define each pseudo component, the following (minimum) base properties must be supplied: - Boiling point - Molecular weight - Ideal liquid density Once these data is entered, the simulation shall estimate other properties using these parameters. It is worth noting that, through previous crude ch aracterization work; many Saudi Aramco crudes have been characterized using a given component slate. The user may use this component slate as a starting point. Moreover, component slate selection for refinery systems can be simplified through use of Refsys options in HYSYS (steady State). Predefined crude oil slate can be imported from csv (comma separated values) file such as arablt.csv. HYSYS also provides predefined component slates such as petroleumcomp1.cml which can be useful for modeling Crude units, Hydrocracker, FCC (Fluidized Catalytic Cracking) downstream separation section and in fact any heavy end of the refinery. For the light end of the refinery such as Platformer and Isomerization, it is recommended to use catrefisom.cml as the component slate. 5.6

Naming Convention To allow easy cross reference between the p lant/equipment/instrument items and model, each unit operation of the model shall be named after the plant item’s tagname. For example: if an exchanger is tagged as D75-E-111 on the plant, then the model shall be named as D75-E-111 instead of some random simulation tagname as such E-100. If the model confines to a smaller plant area then the plant number such as D75 may be dropped. Furthermore, the dashes in the tagname may also be omitted. For unit operations involving the same piece of equipment or instrument, consider adding a suffix to differentiate them; i.e., E111M1. For streams, use PDF stream number where possible. For other non-PFD intermediate streams, use the closest associated PFD with suffixes appended to them or name stream after the upstream item. E.g., 123-1 or E111LiquidOut.

Page 7 of 20



Whichever convention is applied, it is important that one consistent convention is used. Do not mix different conventions together. Note:

Naming of streams and unit operations can be partly automated by specifying user defined pre-fixes through:

Tools | Preferences |Naming sheet (under simulation tab), see ex ample below:

Figure 1 – Stream and Unit Operation Pre-Fix Definitions

5.7

Flowsheeting Where possible, each plant area model shall be building using a sub-flowsheet (sub-model). Flow paths in each model should be built from left to right; similar to how P&IDs are drawn-up. For larger models with multiple trains, each train shall be built within its own

Page 8 of 20



sub-flowsheet with streams entering and leaving each trains being distributed in the main flowsheet. 5.8

Other Model Set-ups 5.8.1

Units of Measurement It is advisable to set-up simulation to use the same units of measurement as the plant. This way makes comparison with plant data, either through plant readings or PI system, that much easier. With some exceptions, Saudi Aramco, generally use British units and therefore it makes sense to set up the simulation to use British units.

5.8.2

Simulation Step Size Simulation step size – simulation time wait before another loop of calculation is performed – is important to model stability. Generally speaking, smaller is the time step, the more calculations the simulation will do per second and slower the simulation will be. The benefit for this is greater model stability. Conversely, larger is time step, faster the simulation will run, but this may be at the expense of model stability and accuracy. Typically for fast compressor transient runs; such as surge studies, the step size is around 0.01 seconds or lower. For general plant support work, a step size of .1 second may be sufficient.

5.8.3

Compressors Curves When available, the actual as tested performance curves should be entered into the compressor model. If the compressor is a variable speed machine, then more than one speed curves should be inputted. If as-tested curves are not available, use predicted curves. These curves should be sufficiently close for most simulation purpose. Finally, if no performance curves are available; ge nerate a new set of curves by specifying design head (pressure rise), flow, and efficiency in the model. Do not model a compressor with fixed head rise (or pressure) as the compressor head will not change with flow if specified this way.

5.8.4

Pumps Similar to compressor, the pump model should be furnished with the

Page 9 of 20



actual pump curve. If such curves are not available, use the simulation to generate a curve with a generic curve passing through the design point. Likewise, do not model a pump with fixed discharge head. 5.8.5

Tray Efficiencies in Column Unless a column is modeled using theoretical tra ys, tray efficiencies should be used. Generally, it is quite difficult to estimate efficiency on each tray. For this reason, it is recommended to use a single tray efficiency for a section within the column. It should be noted that, if the tray efficiency is set too low, the user will notice a big difference in vapor and liquid temperatures leaving the tray (assuming no external factors presence such as pump around, feed streams or side-draws). For amine systems, if DBR amines package is used, HYSYS will calculate tray efficiencies for H2S and CO2. It is recommended that these efficiencies are validated with plant data for at least two points. And if required, override them so that plant data can be matched. Beware of very low or high efficiencies in order to achieve plant results.

5.8.6

Exchanger Heat Transfer Coefficients (HTC) and Pressure Drops Entering all the necessary data into a Hysys exchanger model may not necessary produce the same results as given on the vendor datasheet. This is partly because the equations used in Hysys are different to the ones used by the vendor.

To set-up important heat exchangers, it is recommend ed that a test model is set-up for HTC tuning. In the test models, the hot and cold streams are set-up as per datasheet as far as flows and inlet temperatures are concerned. The UA (HTC * surface Area) is adjusted to achieve the required outlet temperatures. Depending on situations, in some case, adjusting the duty may be a better option. It should be noted that when adjusting UA together, the sizing data should be set to overall (a Hysys requirement). Set sizing data to either shell or tube will not allow adjustment on UA. For exchangers with extra surface area margin bui lt-in, increase the UA by the corresponding over-surface margin afterward, unless the surface area given on datasheet include such margin. The correlations within the simulation should be sufficient to correct HTCs with flows.

Page 10 of 20



For exchanger pressure drops; use the same test ex changer model to calculate flow resistances of the exchanger. It should be noted that, where calculated pressure drops are given, u se these to calculate exchanger resistances. If only use, allowable pressure drops are given, then use these to estimate flow resistances, but be mindful that the allowable data are the maximum expected pressure drops so flow resistances calculated will be slightly higher than the actual. 5.8.7

Fired Heaters Fired Heaters can be modeled using the fired heater unit operation. This unit operation will do basic calculations on the process side as well as the flue gas side. This is useful to get quick estimates of the air and fuel requirement. However, it will not calculate the flame temperature correctly nor will it be able to model the thermal inertia of fired heaters.

5.8.8

Heat Losses Unless heat losses represent for a major portion of overall heat balance, heat losses are generally not modeled (or ignored). For long sub-surfaces pipelines, heat losses should be accounted on a sectional by sectional basis. Heat losses from vessels or piping can be modeled simply (based on fixed UA and temperature difference between vessel and ambient temperature) or more detailed method (by specifying appropriate HTCs).

5.8.9

Vessels -

Nozzle Position The current version of Hysys, V7.3, like all its p revious versions, models feed (inlet) and Product (outlet) nozzles “horizontally” irrespective to the orientation of the vessel, see figure below extracted from Hysys Unit Operation Manual. It is, therefore, not possible to drop feed “vertically” into vessel via the vessel’s highest point or drain the vessel “vertically” from its lowest exit point.

Page 11 of 20



Figure 2 – Hysys Nozzle Set-up

This means, even with outlet nozzle is set at 0%, the lowest possible position; vapor will leave the Product outlet nozzle if the liquid level in vessel falls below of the position of Product nozzle + Product nozzle diameter. For example: if the liquid Product nozzle is set at 0% and the diameter of the nozzle is 4 inch, vapor will start to leave the vessel with the liquid when the liquid level falls to lower than 4 inch; i.e., before the vessel is completely drained of liquid. -

Recycle Efficiencies Use recycle efficiencies in vessels with caution. Adjusting the efficiencies may help simulation to converge during transient, but it may lead to vapor and liquid becoming not in equilibrium with each other, thus, resulting in temperature differences between the vapor and liquid phase.

5.8.10 Columns -

If in a column where the stripping and rectifying section have different diameter from each other, each section h as to be modeled as a separate section.

-

For multi-pass trays, the total weir length will equal to length of each weir multiply by number of passes. For example, on a 2-passes tray, n = 2, with a weir length of 10ft, w l = 10, then the total weir length to be entered on the tray will be = n x wl = 2 x 10 = 20ft. Entering an incorrect weir length will lead to incorrect weir height being calculated on the tray.

Page 12 of 20



5.8.11 Turbine Currently, Hysys does not have a unit operation for turbines. If a turbine model is required, consider modeling the turbine u sing a spreadsheet unit operation within Hysys. In the spreadsheet, power output should be calculated using fuel or steam flow. The speed of the turbine will be calculated using a detail power balance between the driver (turbine) and its connected loads (compressors, coupling and gearbox, etc.). The total inertia of the connected equipment (referencing to a given speed – usually highest speed shift) shall be used to determine the rate of speed change. Furthermore, mechanical losses of the connected equipment, usually expressed a function of compressor speed ^2, shall also need to be included in the overall power balance. 5.8.12 Elevations It should be noted that all equipment and unit operation have a default elevation of 0 (if static head calculation is enabled). Often these default elevations causing some confusion with pressures and flows; i.e., a forward (upstream to downstream) flow through a valve is observed when inlet pressure appears to be lower than outlet pressure. However, when the relative elevations are taken into account, the upstream pressure is not lower than downstream. 5.8.13 Other Simulation Options (Simulation | Integrator) -

Set unit for time in seconds, minutes or hours depending on situation

-

Check “Enable static head contributions” option if static head is required.

-

Uncheck “Truncate large volume integration errors”. Leaving it checked will cause mass imbalance during transient. Be-sure to uncheck this if model is used for detail studies.

-

Check “Close component material and energy balances” for closed loop systems.

Page 13 of 20



Figure 3 – Example of Integrator Option Settings

5.9

Data Reduction 5.9.1

Piping Volume and Pressure Drop Calculations (Gas Side Only) For some studies, accurate modeling of the gas side piping volume and line losses is important; such as compressor surge studies. When required, and data available, a detail isometric analysis of the gas flow route is performed to calculate piping volume, which is determined by multiplying all the straight line length by the internal diameter of piping. Line losses are calculated by converting all pressure drops due to fittings (elbows, tees, branches, check valves, valves, reducers, enlargers, entrances and exits) into fitting equivalent lengths to which total straight length are added to produce an overall equivalent length for the section. Once the overall equivalent length is determined, this length can be used directly in a piping model or convert the equivalent length in a valve Cv so that the user can incorporate line losses as an extra valve in the system. Page 14 of 20



If such isometric drawings are not available, the piping volumes may be estimated from plot plan and apply a factor for routing. Typically, the routing factor may be between 5% and 50%. 5.9.2

Vessel Volume Calculations – Detail Studies If the latest Hysys is being used, the software should have the option to calculate the head ends volume by specifying the geometry of the head ends (Flat cylinder, Sphere, Ellipsoidal or Hemispherical), diameter of the vessel, height and head (ends) height. If the total actual vessel volume, as given in the datasheet, is sufficiently different to that the value calculated by the model, then the vessel straight length or diameter may be altered to achieve the requirement volume.

5.9.3

Valve Data Where available, the actual valve Cv and characteristic should be used. For ESD or isolation valves, valve data, other than stroke time, is of lesser important than for control valves. If control valve datasheet is unavailable, use the simulation to size the control valve. Typically, Cv of a control valve is calculated based on design flow through the valve, valve’s design pressure drop using the design fluid density. Once Cv is calculated the final selected valve will be 30 to 40% bigger. See default valve data if data is unavailable.

5.9.4

Default Data In the absence of real engineering data, the following default may be used. As with all assumptions, these default data should be recorded in the model documentation Table 4 – Suggested Assumptions

Item Valves

Parameters Characteristics

Assumptions For ball valves using equal percentage (=%) characteristic. For butterfly valves, opening/closing characteristic is similar to that of =%. Linear characteristic may be used.

Page 15 of 20


Item

Parameters


Assumptions Note: Do not use quick opening characteristic, unless stated otherwise. Only blowdown valves tend to have this characteristic.

Valve stroke time

Use 1 second per inch of valve body size, with a minimum time of 5 seconds. E.g., a 10” valve will close approximately in 10 seconds. However, for a 4” valve, use 5 seconds as the stroke time

Exchangers

Tube and shell volume

Do not change these volumes as the model does not fully use these volumes to determine dynamic of the system. It is advisable to model these volumes explicitly using piping or vessels.

Compressors / Pumps

Polytropic Efficiencies

Use 75% if no data is available

Pipe Segments

Pipe Friction Model

Use Full range Churchill Option

Typical Control loop tuning (starting values).

Flow (fast acting)

Proportional Gain = 0.1 Integral time = 0.1 minutes

Pressure (slow)

Proportional Gain = 1 Integral time = 5 minutes

Level (measurement in %)

Proportional Gain = 0.5 Integral time = 10 minutes

Temperature

Proportional Gain = 0.1 Integral time = 10 minutes

For cascade loops, the master controller should be slower than slave controller)

5.10

Model Building and Commissioning As with all mathematical models, outputs are calculated using inputs. For this reason, it is recommended that the model should be built section by section. Furthermore, if an entire flowsheet is built without first initializing (set-up / define) individual feeds or unit operations along the way, the final model will unlikely to run, as well as making debugging more difficult. By building the model section by section in the direction of flow, the devel oper will have more control over the model. Ideally, all feed streams to a vessel or equipment should be defined and initialized first before running the model to commission the item. Once the item and its outlet stream(s) are initialized, it can be connected to the next plant item. Build up the flowsheet this way the user can be sure that everything is initialized correctly before adjusting the model to match the required data.

Page 16 of 20



If a stream is recycled from the back of the model, consider either ignoring the stream for now or use a temporary dummy stream with the expected composition and conditions to initialize upstream equipment. Once the relevant downstream recycle flow is modeled and commissioned, this stream can then be tie-back with upstream equipment or replacing the temporary dumm y stream. After the model is built, key process parameters; such as flows, pressures, temperatures and controller outputs, should be plotted on graphs to establish of the state/stability of the model. Controllers tuning will have a big influence on the operation of the model. 5.11

Plant Data Matching Depending on the final requirement, the model will need to be tuned to match a set of agreed data. Typically, the model will need to first match design data; i.e. design case PFDs, before further fine-tuned to match plant data. Matching design match would confirm the model (and its data) is a close representation of the process which is being modeled. Typically, the model should match design values to ±2% on all major flow streams. Once the design case condition is matched to the required tolerances, the mod el is adjusted to match plant data. The aim of this exercise is to tune the model to plant data over a stable period of operation. Usually, the model aims to match hourly averages. It should be noted that this task should not be considered as any data reconciliation exercise.

6

Model Testing Prior to using the model, some preliminary testing shall be carried ou t and results from the tests documented. As a minimum, the model shall be subjected to following tests:

7

-

Flow (feed) turndowns (between 40% and 50%)

-

Main control set-point step changes (30 to 60%)

Model Documentation As a minimum, the following documentation shall be provided: -

Scope and objective(s) of the model

-

A set of mark-up simulation P&IDs

-

Assumptions and Simplifications used Page 17 of 20



-

A summary of instrument and equipment data used including any default data. Where additional engineering calculations are performed, the method, equations and intermediate calculation steps shall also be given

-

A data register summarizing all documents/drawings used together with their document/drawing number and revision

-

Result comparisons between the model and the baseline data

-

Documented test results

18 April 2012

Revision Summary New Saudi Aramco Best Practice.

Page 18 of 20



APPENDICES Appendix A – Model Fidelity Definitions Table A1 – Model Fidelity Fidelity High

Definition

1. Model developed based on first principle of chemical engineering 2. Mass and heat balances are performed rigorously 3. Piping (for gas side only) and vessel volumes will be correctly accounted for

4. Piping pressure drops will be accounted for in detail. Typically, piping volumes are calculated from isometric analyses

5. Equipment will be modeled in more detail with actual performance data and dimensions

6. Streams will be modeled using multi-components (except in single component system)

7. Liquid static head will be accounted for Medium

1. Model developed based on first principle of chemical engineering 2. Mass and heat balances are performed rigorously 3. Piping (for gas side only) and vessel volumes will be approximately accounted for

4. Piping pressure drops will be approximately. Typically, piping volumes are calculated from plot plan

5. Streams will be modeled using multi-components (except in single component system)

6. Liquid static head will be accounted for Low

1. Black box (simple mathematical) model approach, may not base on any chemical engineering first principle

2. Model will be directionally correct. Process values may not be strictly accurate

3. Often only one component or reduced components slate may be used to model streams.

Page 19 of 20



Appendix B – Simulation Platforms: Recommended Vendor List (RVL) For the full list of P&CSD approved simulation platforms, please refer to Engineering Procedure SAEP-364, Exhibit I.

Page 20 of 20

SABP-Z-032

Recommend Documents