Sabp Z 018
Sabp Z 018
Sabp Z 018
1 Introduction..................................................... 2
2 Conflicts with Mandatory Standards............... 2
3 References..................................................... 2
4 Definitions....................................................... 3
5 Acronyms and abbreviations.......................... 3
6 Model Development....................................... 4
7 Thermodynamics............................................ 5
8 Components Slate Characterization.............. 8
9 Developing Petroleum Pseudo Components.. 9
10 Unit Operations………………………..…..…. 12
11 Logical Operations…………………….…….. 20
12 Flow Sheeting Topology.............................. 21
Appendices
Appendix B – Simulation Platforms:
Recommended Vendor List (RVL)....... 23
1 Introduction
Process Modeling and Simulation has evolved in the last twenty years from merely an
applied mathematical science into many corporates business decisions tools. It is
transcending the exclusive domain of the technologist to become an everyday
collaborative and decision support tool used by plant operators, supervisors, planners,
and managers. It is proving that using simulation on an enterprise-wide basis promotes
knowledge sharing and improves everyday production and business decisions.
Process modeling and simulation applied throughout the lifecycle of an asset has
traditionally been the domain of experts. The models that they develop embody and
codify process knowledge captured by an organization. As such, these models have
tremendous intrinsic value. Process models are developed at early phases of project
feasibility analysis and evolved through design, construction, and commissioning of the
plant life cycle. Process models increasing applications that go beyond this traditional
area of application to include evaluations for Process plant expansions, capital
investment options, cycle time reduction and safety.
Industrial processes can also be divided into continuous and discrete processes or
combinations of these. The discrete process simulation is related to products scheduling
and distributions. The continuous process which represents the day to day plant
operation is divided into steady state and dynamic state. The process simulation
covered in this work is limited to the simulation of continuous processes. Furthermore,
the simulation of continuous processes can be divided into steady state and dynamic
simulation. The field of process industry is broad covering such sectors as water
treatment, food industry, pharmaceutical industry, energy industry, metallurgical
industry, pulp and paper industry, oil refining industry and so forth.
The best practice is developed in alignment not to conflict with other Mandatory Saudi
Aramco Engineering Standards Requirements; the Mandatory Saudi Aramco
Engineering Standards Requirements shall govern.
3 References
In case of specific sections of this document are referenced within the body of the
document (SAEP-364). Material or equipment supplied to this best practice, shall comply
with the referenced sections of the latest edition of these specifications. Where specific
sections are not referenced, the system shall comply with the entire referenced document.
Page 2 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
4 Definitions
The term process modeling refers to the mathematical representation of the physical
system using physical-chemical (thermodynamics) relations of the process. The steady
state modeling refers to conservation of mass and energy (i.e., no. of system’s input
moles/energy is equal to the number of output moles/energy, no accumulations).
Simulation is a scenario of the mathematical models at specified conditions and
parameters.
This section contains definitions for acronyms, abbreviations, words, and terms as they
are used in this document:
APISOUR Wilson’s Activity Coefficient Model
ASME Compressor Design Based Association of Mechanical Engineer
BK10 Braun K-10 K-value Correlations
BP Boiling Point
BWRS Modified Benedict-Webb-Rubin Equation of State by Starling
DIPPR Design Institute for Physical Property Data
EOS Equation of State
FCCU Fluid Cat-Cracking Unit
GERG2008 Equation of State for Natural gas mixtures
GPSA Compressor Design Based Gas Processors Suppliers Association
GS Grayson-Streed
HCRU Hydrocracking Refinery Unit
HTC Heat Transfer Coefficient or U
Page 3 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
6 Model Development
The steady state model development can be categorized based on the objectives.
These include:
Design and ratings: first principle models are constructed for design and equipments
ratings. The Models workflow is based on the process thermodynamic equilibrium and
equipments geometrical induced effects. Initially, the model is developed to support the
process conceptual design. Once process equipments are selected the design and ratings
model is expanded to include vendor information such as vessel sizes, pump curves,
compressor curves, Heat Exchanger / Air cooler / Fired heater geometry, Column &
Tray details and catalyst reaction kinetics.
Page 4 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
Base case for dynamic model: steady state models built for further development into
the dynamic state do require additional data other than the operating conditions.
Data required are the realistic pressure profile across the flowsheet and driven by
equipments rating base. Additionally, reasonable volume hold up in each of the
equipment is very vital especially if the dynamic model is required for process control
and compressor surge analysis.
The settings of pressure and temperature profiles across the flow sheet and also for
distillation column initial solution set up are very important to achieve accuracy and
reaching a converged solution. The pressure profile set up is either assumed based on
accurate assessment or set up the model hydraulic calculation across the flowsheet to
enable enough flexibility in the model to cover a wide range of operating scenario.
7 Thermodynamics
Thermo dynamical correlations and physical properties are the basic driving force that
determined the process flowsheet structure and meeting the required separation and/or
reactor’s conversions and throughput. The selection of thermodynamic correlation
depends on the stream mixture of components, operating conditions and type of process.
Most hydrocarbon tend to behave ideally with the exception of mixture contains
Hydrogen.
Thermodynamic data include vapor-liquid equilibrium ratios (Kj values), enthalpies, and
entropies. These properties are necessary for the unit operation calculations. Entropies
are used for isentropic calculations such as compression and expansion.
The Kj values of any given system of components are a function of the temperature,
pressure, and composition of the system. The calculations of Kj using thermodynamic
physical properties include:
Equation of State (EOS) (e.g., Peng Robinson, SRK, BWRS, Lee-Kesler Plocker,
etc…)
Liquid Activity Coefficients (e.g., NRTL, UNIQUAC, Wilson, etc…)
Page 5 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
Physical and transport properties for the feed and product streams may be calculated
using the individual component contributions. Those properties include; total flow,
molecular weight, and stream density. Transport properties such as the viscosities,
thermal conductivities, and liquid surface tension are computed using the above physical
correlation. ASPEN PLUS provides several set of options or property methods, that is,
combinations of methods for calculating the thermo physical and transport properties.
The property methods can be modified by replacing one or more of the methods for use in
a simulation. It is also possible to define an entirely new property method. Furthermore,
each process unit in the flowsheet can be simulated using a different property method. In
this case, the enthalpy and pressure of a connecting stream are preserved, but the other
properties (temperature, density, etc.) differ for the stream in the models for the process
units. HYSYS can facilitate to have several sets of physical properties. The feature of
stream cutter tool would allow similar feature.
Selecting the specific physical property method would depend on the process as well as
the mixture of components. Commercial software the like of Aspen Tech’s HYSYS
and ASPEN PLUS and Invensys’s ProII provide a comprehensive guide on the matter.
Following are some of specific applications with what is appropriate:
a. Light hydrocarbon System (i.e., Gas Plant and NGL); Peng Robinson (PR), Soave
Redlich Kwong (SRK). Sour option of PR and SRK are used if the system
contains hydrogen sulfite H2S.
b. Heavy Hydrocarbon System (i.e., Refining Processes); Grayson Streed, BK10, PR
and SRK both should be utilizing Lee Kesler enthalpies.
c. Hydrogen Rich System (i.e., Hydrogen Plant and Hydro-processing; PR and SRK
d. Sour Water System (sour water with NH3, H2S, CO2); Sour PR, Sour SRK, and
APISOUR
e. Gas Compression System; PR, SRK with Lee Kesler enthalpies, and BWRS for
conditions of pressure higher than 600 psig.
f. Light Hydrocarbons Dissolved in Water; Kabadi-Danner SRK.
g. Hydrocarbon Dehydration System; Glycol package
h. Hydrocarbon Sweetening System (removal of H2S and CO2); Kent Eisenberg
Amine model.
i. Steam Systems; ASME and NBS packages
j. Cryogenic and Air Separation Systems; RefPROP, GERG2008
Page 6 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
Page 7 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
All commercial simulators have large pure components libraries containing data for the
various classes of chemical components. The Common source for pure components is
Design Institute for Physical Property Data (DIPPR) libraries. For non-library
components like pseudo components correlating parameters such as normal boiling
point (BP), specific gravity (SG), and molecular weight (MW) are used to estimate the
thermo physical properties. The sources of those correlating parameters are obtained
either from design process flow diagram or equipments data sheets. Other sources
include data from other simulation software such as PVT and PVTSim in the upstream
applications. The PVT input converts black oil data to pseudo components. Both of
these sources are a result of sample testing at some point.
Page 8 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
For the flow-sheeting unit operation processes, it is necessary to represent all process
streams with discreet components for which thermo physical properties may be defined.
Since most of the actual components are unknown for petroleum streams, petroleum
pseudo components are developed to represent the unknown components in these
streams.
Three (two in some) correlating parameters are needed to define the pseudo component
and to estimate the thermo physical properties for petroleum pseudo-components:
Normal Boiling Point (NBP), Specific Gravity (SG), and Molecular Weight (MW).
Correlations are available to enable the prediction of any of these three parameters from
the other two.
Initially, the first step is to develop the pseudo components cut-point ranges on the
stream TBP distillation curve. Usually, the TBP distillation curve is a predicted
one, as the TBP distillation is rarely run due to time required and the associated
high cost. Alternatively, another type of laboratory distillation is used and a
correlation is used to convert the test data to the corresponding TBP distillation.
The NBP for each pseudo is defined as the weight average temperature for its
cut-point range of the actual or predicted TBP distillation.
Page 9 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
The following cut point ranges are reasonable for most refining petroleum
stream and are:
The refinery reactor simulation components can be a real or pseudo. Usually, all
the light components, such as Hydrogen, Hydrogen Sulfide, Ammonia and light
hydrocarbons ranging from C1 to C4, are all modeled as real components.
Hydrocarbons with higher carbon numbers are lumped and simulated as pseudo
components because of the extremely large number of isomers and the difficulty
of individual identification. One of the common ways to generate these pseudo
components is to make selective cuts from the distillation curve, these cuts are
not based on the boiling intervals specified by the user, but either on the
correlated boiling ranges of hydrocarbons that have the same number of carbon
atoms or on the approximate boiling ranges of the product streams that are
directly separated from the reactor effluent “reaction pseudo components”.
Another common way to define the pseudo components is to select a set of
model components from the real compounds and use them to represent the C5+
portion of the hydrocarbon mixture “model components”. Though these model
components are real, most of them are applied as pseudo components in the
Page 10 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
Both methods (reaction pseudo components and model components) have taken
account of the different reaction tendencies of the paraffinic, isoparaffinic,
olefinic, naphthenic and aromatic structures by either further breaking up each
pseudo component into these subgroups or systematically choosing various
model compounds with these chemical structures. The distribution of these
subgroups or model components is dependent on the feed quality, and the goal is
to have a suitable set of pseudo components that can be used to construct the
reaction network through which the reaction behavior of the feed mixture is
described.
Lumping and de-lumping of components involve; the feed stream data and
reactor operation data are transferred from the flowsheet simulator to the
refinery reactor simulator and, conversely, after the reactor unit is solved, the
product stream data and the reactor results are transferred back to the flowsheet
simulator. The product stream data are then mapped into the product stream of
the flowsheet.
Mapping Feed Stream Data; Reactor feeds stream data transferred to the
reactor simulator include temperature, pressure, flowrates of real components,
and flowrates of model components or assay data and the required refinery
inspection properties. Based on these data the refinery reactor simulator can
characterize the feed stream and create the reaction pseudo components as
necessary.
Page 11 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
10 Unit Operations
The normal situation in distillation design is that the feed conditions are given:
flowrate, composition, temperature and pressure. The desired compositions of
the product streams are also typically known. Separation is dictated by the feed
thermodynamics and subsequently affecting the number of required equilibrium
stages. Column’s liquid and vapor traffic will be calculated by the simulation to
provide the required separation. Column’s hydraulics, length, diameter will be
calculated. The column pressure profile is assumed and the feed thermal status
will determine feed tray location.
Page 12 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
For a given feed composition and assumed column pressure profile, two
parameters must be specified for a two product stream column. Some examples
are the distillate product purities or distillate flow rate, bottom product
impurities or bottom flow flowrate, or condenser and re-boiler duties. Reflux
ratio is a function of distillate flow rates. Note that there are multiple
combinations of two specifications that can be provided; however, some
combinations are more adaptable to changes in the simulation conditions, and to
real plant operation. Note that more complex multiple draw fractionators
included in Refinery processing plants have pump-around and side withdrawal
streams and need more specifications provided to solve.
While this section focuses on existing columns, the same principles may be
applied to column design. Certain steps are always necessary in the setup,
solution, and interpretation of results for column models. Simulation for column
design are devised to establish for sizing columns and its associated equipments
based on the feed characterizations, flow rates and required separation.
The following guidelines should be considered:
a) Commercial simulators require the user to supply flow estimates for all
column products except one, either the top, or the bottom liquid product.
The reflux quantity must also be estimated, the first (or second tray if a
partial condenser) and last tray temperatures are also required for most
simulators.
b) It is always good practice to allow the distillate rate to be a degree of
freedom for the calculations, with its rate calculated to meet a performance
specification. For condensers with small vapor product flows, it is best to
specify the vapor flow and allow the condenser drum temperature to be
calculated. This produces the most stable model, particularly if it is
planned to explore alternate operations that may alter the composition of
the overhead vapor to the condenser. For condensers with significant
vapor product flows, the condenser temperature may be used for the
performance specification and the vapor product flow calculated to satisfy
the equilibrium balance with the liquid product flow.
c) Setting up a custom configuration for the re-boiler(s) system within
distillation template that is not represented in the generic model template
may require to follow the following steps:
c.1) Set up the generic model and get a converged solution. Copy the
converged solution to be used as initial estimate for the new set up.
c.2) Delete re-boiler model configuration and re specify the pressure
value for the last try.
Page 13 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
c.3) Add a pump with pressure differential increase equal or more to the
new detailed heat exchanger pressure drop. Add a separator after
the heat exchanger.
c.4) Check the required specification to solve the heat exchanger before
running the column. The heat exchanger degree of freedom will be
added to the column one.
d) The user is required to supply values for the first and last trays pressure in
the column model. A linear pressure profile is normally assumed for the
intermediate trays. Note that for columns with condenser, a pressure
should always be supplied for the second tray, since there is a substantial
pressure gradient between the column top tray and the condenser.
For crude and vacuum columns, it is especially important that the pressure
of the flash zone of the column be accurate and an additional pressure
should be supplied for this purpose.
e) Identification of column separation zone is crucial to the solution. The zone
is defined as a column section where split between two products occurs.
Multi draw column such as the crude and vacuum distillation columns have
multiple separation zones. Simple column with only overhead and bottom
products have only a single separation zone. Note that both rectification of
heavy components and stripping of light components may occur in a
separation zone.
f) A proper selection of the separation key components (light and heavy)
should be accurately defined especially within distributed components.
The fractionation is based on a desired split between the key components.
The light key is a light distributed component for which a small, but definite
quantity is present in the bottoms product. The heavy key is a heavy
distributed component for which a small, but definite quantity is present in
the overhead product. Note that trace components do not meet these criteria.
The key component concept is most useful for column sections in which the
distributed components are defined as real components.
g) Accurate representation for the distributed components in a separation zone
is required to achieve solution. Distributed components are those
components that exist in both products leaving the separation zone.
These are the components that fractionate or split. The other components
are non-distributed components, with the heavy non-distributed components
exiting in the bottom product and the light non-distributed components
exiting in the overhead product. Non-distributed components have almost
no influence on the separation, and they may be combined or averaged as
desired with little effect on the calculated results for the column.
Page 14 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
Page 15 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
Page 16 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
There are two model approaches to calculate the heat exchanger LMTD
depending on the passing fluid state of phase condition:
c) Weighted. The heating curves are broken into intervals, which then
exchange energy individually. An LMTD and UA are calculated for each
interval in the heat curve and summed to calculate the overall exchanger
UA. The Weighted method is available only for Counter-Current
exchangers. This approach is used where there is a phase change in the
system as in condensing steam or vaporizing liquid.
d) Endpoint. A single LMTD and UA are calculated from the inlet and outlet
conditions. For simple problems where there is no phase change and CP is
relatively constant, this option may be sufficient. This approach is used
where there is no change in the system phases passing through the
exchanger.
For Air-Cooled Heat Exchanger the modeling may be simplified and the
LMTD calculation is assuming having fixed medium temperature (ambient).
All Commercial simulators provide several types of heat exchangers (i.e., shell
and tube, double tube, etc…). Modeling of heat exchanger in a flow sheet
would highly depend on the process, fluid, and the weighted importance
between rigorous and simplified ones.
For Fired Heater, design data are used to establish the heat duty split between
the convection and radiant sections. The fuel gas composition and operating
conditions of flow rate, temperature, and pressure must be specified. The flue
gas thermal conditions must also be known. The process side of inlet and outlet
conditions of composition, flow rates, temperature and pressure must be
provided. The Heater duty efficiency, heat flux, and heat transfer coefficients
are estimated in the convection and radiant sections.
Other types of heat exchangers (i.e., Plate, Kettle, etc…) can be modeled in the
same way provided that the degree of freedom is zero and if more a
specifications of constraints must be specified.
Note that detail heat exchanger design and rating software are integrated or
linked to general flowsheet simulators. These are normally used for detailed
analysis of the heat exchanger performance where the effects of exchanger
physical design are significant and high accuracy is required. Heat transfer
specialized software from Heat Transfer Research Institute (HTRI) is one of the
industry available to address all types of Heat transfer equipment simulations
Page 17 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
10.3 Compressors
Page 18 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
tolerance. SRK and BWRS are favored by compressor vendors while LK and
PR are favored by software vendors. SRK and or sour SRK with LK enthalpies
is recommended for compressors system in operation.
10.4 Pumps
Pump moves liquid that is not compressible, and are designed for a Net Positive
Suction Head (NPSH) based on fluid properties, suction and discharge
conditions. Simulating pumps requires the specifications of performance curve
and efficiency. Pumps are normally not required in steady state simulations
since stream flows are not dependent on realistic pressure profiles through the
flowsheet. However, they can be added to estimate energy use or to provide
realistic stream pressures.
Several test data are required to tune the model which will estimate the
individual rate of reactions, catalysis deactivation rates, throughputs, and yield
qualities. The estimated reaction parameters may be used to predict operating
scenarios such as maximizing throughput or yields and specifications of specific
streams.
Page 19 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
11 Logical Operations
In process synthesis when product streams are fed back to a previous point in the
simulation or the reaction operations involve reversible reactions or competing
reactions, the split fractions of the species leaving the separators are complex
functions of the operating conditions such as the temperatures, and pressures,
then iterative calculations are necessary. In these cases, the simulation flow-
sheets usually contain information recycle loops, that is, recycles for which too
few stream variables are known to permit the equations for each unit to be
solved independently. For these processes, a solution technique is needed to
solve the equations for all of the units in recycle loop. One solution technique is
to tear one stream in the recycle loop, that is, to guess the variables of that
stream. Then, a non-sequential iterative solution is employed for solution.
Flowsheet recycle can be of a single or multiple loops. The guidelines are:
a) Use Nested type for flow sheet with a single recycle or not connected
multiple recycles.
b) Use Simultaneous type for flow sheet with multiple inter-connected
recycles.
It is the absolute error resulted from the recycle operation solved inlet and
assumed outlet streams. The recycle tolerance is set for each given stream
properties (i.e., temperature, pressure, flow rates, compositions, etc…).
The actual tolerance of the Recycle operation is calculated as the product of the
absolute (default) tolerance for the given property (Simulation software internal
fixed value) and the relative tolerance (specified by the user). Relative tolerance
specified values for each specific property should be specified depending on its
importance. For instance the specified actual tolerance for individual
composition should be within 10-3 to 10-5. However, components with high
compositional value should have very stringiest tolerance of 10-5; on the other
hand components with low compositional value a tolerance of 10-3 should be
reasonable not to cause the recycle solution to fail.
Page 20 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
Page 21 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
The Equation Oriented (EO) approach solves the entire flowsheet (mass and
heat balance equations) simultaneously so there are no requirements for
recycles; however, it is more complicated and difficult to set up. The EO
approach does require initial system parameter estimates. This is normally
provided by first running the SM approach (without the need to converge) and
then changing the simulator mode to EO. The EO approach is most suitable for
online and offline optimization and systems with many or “back-to-front
specifications that requires iterative type of solutions. Other examples include
calculating pressure profiles through a unit or piping system. Distillation
columns and refinery / petrochemical reactors use this approach but the details
are normally hidden from the user’s view. ASPEN PLUS EO, HYSYS EO, and
ROMeo use this approach.
Revision Summary
30 September 2012 New Saudi Aramco Best Practice.
16 November 2014 Editorial revision to transfer responsibility from Process Control to Process Optimization
Solutions Stds. Committee.
Page 22 of 23
Document Responsibility: Process Optimization Solutions Standards Committee SABP-Z-018
Issue Date: 16 November 2014
Next Planned Update: TBD Steady State Process Modeling
For the full list of P&CSD approved simulation platforms, please refer to Engineering
Procedure SAEP-364, Exhibit I.
Page 23 of 23