Best practice for refinery flowsheets

Refinery-wide flowsheets can demonstrate the true representation of changes

in key operating variables and their associated impact on refinery operation

Jitendra Chellani KBC

Asma Al Sariyahi Orpic

D
ue to the integrated nature of process units balanced distillation models which use a sec-
within a refinery, change in key operating tion-by-section approach rather than simulating
variables has an impact on overall refin- each tray are calibrated to plant data.
ery operation and product blending as well. KBC and the company’s clients around the
Refinery-wide flowsheets demonstrate the true world have developed numerous refinery-wide
representation of this effect as all process units flowsheets. Petro-SIM based flowsheets are
are linked together. Non-linear kinetic or equilib- being used for identification and evaluation
rium models in the flowsheet represent the con- of margin improvement opportunities which
version units. Fractionation models capture the include optimisation of stream routings, blend-
efficiency of separation between different cuts ing strategies, molecular management, through-
or components. put maximisation, feedstock selection, and
KBC (A Yokogawa Company) has developed improvements in the unit operating conditions.
and used refinery-wide flowsheets for more than Refinery-wide flowsheets have also been used
30 years. Developed in the 1980s, Petrofine was for configuration studies for grassroots and
a FORTRAN based tool capable of refinery-wide revamp configurations.
flowsheeting. In 2004, KBC launched Petro-SIM
with additional features to simulate refinery Flowsheet development
process units. These standalone models are com- Standalone models for the process units are the
bined to create a complex-wide flowsheet which main building blocks for the refinery-wide flow-
includes all process units within the refinery. sheet. Detailed kinetic and equilibrium based
Petro-SIM is KBC’s process simulator used for Petro-SIM models are calibrated using test run
rigorous modelling of the entire refinery and data. Unit configuration, operating parameters
petrochemical complex, from crude to finished from historian and laboratory data are used to
products. Since each unit, including conver- calibrate standalone models for the process units.
sion units, is modelled meticulously, the over- The data are reconciled to close mass, sulphur,
all simulation suitably reflects the non-linearity nitrogen, carbon, and hydrogen balance. A cal-
of petroleum refining which enables sensitivity ibrated process model mimics the performance
analysis over a wide range of operating variables of the process unit. The models are valid over a
and feedstocks. wide range of operation as they are based on first
The conversion units are based on comprehen- principles and are non-linear in nature.
sive kinetic models that predict the unit yields To understand overall refinery operation, a
and product qualities. The kinetic models are cal- base month is selected. Base month operation
ibrated specifically to match the available plant provides an insight into marginal mechanisms in
data for a particular unit. This allows the simu- the refinery. The data used for standalone mod-
lation to be representative of the specific unit’s els are based on test run operating conditions
operation, independent of the  licensor. and these test runs may have been conducted at
Product separation is simulated using frac- a different time period. Due to this, it is essential
tionation technology that represents current to prepare a consistent basis for operating condi-
operation and heat balances. Heat-and-material tions of all process units.

www.digitalrefining.com/article/1002322 PTQ Q3 2019 1

 The following guidelines are rather than using optimum
used to select the base month routings. Sensitivity analysis to
Round table discussion and
for the flowsheet: data collection validate optimum strategies is
• Crude blend for the month performed after completing the
represents the typical crude Unit data base case which represents ‘as-is’
blend used by the refinery Heat and mass
Overall yield data refinery operation.
balance
• Crude throughput should Historian
refinery-wide Understanding refinery stream
be close to the typical crude Lab analysis routing strategy can be a difficult
throughput of the refinery task if the refinery has multiple
• Most of the process units in Reconstruction of refinery trains or two to three process
the refinery should be operating and unit data
units for a purpose, for instance
at typical capacities and at nor- more than one diesel hydrotreat-
mal operating conditions Individual model calibration ing unit. In this case, routings
• Most of the process units are fixed based on the feedback
should operate continuously in from planning engineers and
Refinery-wide flowsheet
stable conditions base month data.
• Changes in the inventory of Petro-SIM’s blending unit
intermediate streams should not Unit throughput operation allows flexibility of
be significant. Production of different grades
for base month optimising refinery blends based
One of the major challenges on prices, product demand,
is with regard to inventory. Crude basket product specifications, and other
Inventory changes are not simu- Product demand constraints. Marginal mecha-
lated in the flowsheet as it rep- Operating targets nisms used by the refinery must
resents steady state operation of be reflected in the flowsheet
the refinery. Inventory changes Base case product blenders. The flowsheet
in feed and product are used to must hit key specifications for
estimate the net feed processed each blend, such as octane for
and net products produced in Figure 1 Work process for refinery- gasoline, flash for diesel, viscos-
the refinery. Inventory changes wide flowsheet development ity for fuel oil, and so on. The
for intermediate streams affect marginal streams used by the
throughputs of the process units and hence it is refinery should be reflected in the flowsheet as
essential that the base month operation has mini- well.
mum inventory changes for the intermediates. The refiner also needs to identify major
Mass balance for different process units may be changes which are planned in the near future
inconsistent for refineries which do not use data including the revamp of a process unit or a sig-
reconciliation tools, for instance, the through- nificant shift in the crude basket. The refin-
put of a delayed coking unit measured by the ery-wide flowsheet may be updated with these
feed meter may not be the same as the vacuum changes after which it represents the base case
residue production measured in the vacuum operation of the refinery.
distillation unit. Validating consistencies for a The work process to develop the refinery-wide
product which is routed to more than one pro- flowsheet is shown in Figure 1.
cess unit and blending is more challenging. Due
to these issues, data for the base month also Flowsheet complexity
requires  reconciliation. Before building a model, it is crucial to recognise
To build a refinery-wide flowsheet, all stan- its purpose. Refinery-wide flowsheets are used
dalone models are combined. Crude blend for to understand the interaction between different
the crude distillation units (CDU) and through- process units and the impact on product blend-
puts are updated so that the flowsheet feed ing. Including too many details in the model
represents base month operation. Routing strat- increases the time to build the refinery-wide
egies used by the refinery are replicated in the flowsheet and efforts required in analysing
flowsheet. During the flowsheet development the results.
phase, refinery strategies are used for routings Process engineers still need to build rigorous

2 PTQ Q3 2019 www.digitalrefining.com/article/1002322

standalone models for various analyses, includ- sheet. Workflows can be configured by the user
ing heat integration, column hydraulics, identify- and key constraints in the refinery process units,
ing constraints for throughput maximisation, and blends and product logistics can be added in
so on. It is recommended to simplify the stan- the flowsheet. With Workflows available in the
dalone model as much as possible before using flowsheet, quality checks of the results can be
them in the refinery-wide flowsheet. The refin- performed quickly. Hand-over of the flowsheet
ery-wide flowsheet usually includes a kinetic from one process engineer to another can be
and equilibrium based reactor model followed done fast. Workflows ensure that the results are
by ‘section-by-section’ fractionation columns. always consistent and constraints have not been
Simulating heat exchangers, pumps, compres- overlooked.
sors, valves, and other items is not recommended For most cases, including extra equipment
for refinery-wide flowsheets. increases the time required for building mod-
Scenarios or problems which can be eval- els and troubleshooting but has minimal impact
uated using the refinery-wide flowsheet are on the time for flowsheet convergence. The time
different from those which require detailed stan- required for the convergence is a function of log-
dalone models of the process units. The follow- ical ops, recycle and adjust ops. Equipment or
ing should be evaluated using the refinery-wide process units within a recycle loop increase the
flowsheet: time necessary for convergence. Few of the recy-
• Optimum feed selection of the process units cles include process units and product blending
• Incentive of debottlenecking a process unit as well, such as light cycle oil used as cutter stock
• Trade-off between unit throughput and and processed in a diesel hydrotreater, heavy
conversion naphtha used as low flash blending component,
• Optimum product distribution from a process and feed for a naphtha hydrotreater (NHT).
unit The hydrogen network is one of the recycle
• Molecular management of naphtha compo- loops which may consume a significant amount
nents between different process units and blends of time for flowsheet convergence. Kinetic based
• Estimation of the optimum unit operating models for hydroprocessing units in the flow-
target sheet estimate the hydrogen demand. Kinetic
The following should be evaluated using based models in the flowsheet estimate the
detailed standalone models of the process units: hydrogen available from the producers. Using
• Identification of constraints to achieve opti- the supply-demand gap, the required through-
mum conversion or operating target put for the hydrogen generation unit is esti-
• Required changes in the process unit for mated in the flowsheet. If the refinery uses low
achieving targeted throughput vs conversion purity hydrogen in the make-up for hydropro-
trade-off cessing units, the convergence of the low purity
• Investment required for unit debottlenecking recycle also increases the time for flowsheet
• Detailed analysis of the equipment in process convergence.
units which includes unit hydraulics, heat inte- Nesting is one of the techniques available in
gration, and so on. Petro-SIM which can reduce flowsheet conver-
• Troubleshooting and root cause analysis. gence time. Using nesting, users can set the pref-
Models in the refinery-wide flowsheet should erence for the process unit convergence. One of
still be able to monitor key operating constraints. the best options is to give preference to atmos-
Petro-SIM reactor models allow monitoring of pheric and vacuum residue processing units
major operating variables in the process units followed by middle distillate  processing units
such as main air blower limitation in catalytic and then naphtha processing units. Vacuum res-
cracker or gas/oil ratio in hydroprocessing units. idue processing units produce naphtha due to
If the constraints are specific to any equipment cracking which affects naphtha processing units
such as fired heater tube metal temperatures, whereas vice-versa is occasionally possible.
this equipment may be simulated in the refin-
ery-wide flowsheet. Flowsheet utilisation
Petro-SIM Workflows provide a technique for KBC and the company’s clients have used Petro-
monitoring key constraints in a complex flow- SIM refinery and complex-wide flowsheets for

www.digitalrefining.com/article/1002322 PTQ Q3 2019 3

 ery-wide flowsheet along with
Base case Opportunity cases the price set are used to assess
Cut point optmisation Opportunity case, the economics of each opportu-
cents/barrel nity. Using this methodology,
Stream routings
Opportunity case techno-economic analysis is
Molecular management
performed for each opportu-
Opportunity case nity. KBC’s clients also use this
Blending approach to identify and evalu-
Opportunity case
ate opportunities for improve-
Unit feed selection
Opportunity case ments. One of the user case
studies is discussed later in this
article.
KBC has also used refin-
ery-wide flowsheets for con-
figuration and reconfiguration
Figure 2 Opportunity evaluation using refinery-wide flowsheet studies. Flowsheets can provide
an in-depth analysis of changes
numerous objectives. The refinery-wide flowsheet in the product slate with upgradation of bottom
is one of the main tools used to evaluate opportu- of the barrel and its impact on existing units.
nities under the company’s margin enhancement Non-linear equations used for kinetic and equi-
programme (Profit Improvement Program, PIP). librium models of the existing units are essential
KBC uses the methodology shown in Figure 2 to for reconfiguration studies in which considera-
evaluate identified opportunities. ble changes are expected in feed for the existing
The refinery-wide flowsheet (the base case) is process units. The flowsheet linked with the Excel
used to assess the impact of a change in operat- interface along with capex and opex estimation
ing conditions, routings, blend management, and tools provide inputs for financial analysis.
so on. As it is difficult to analyse results when Users have applied the refinery-wide flowsheet
all independent variables are optimised at the to evaluate different feedstocks available for the
same time, only one opportunity is assessed at a refinery and petrochemicals complex. Petro-SIM
time using the base case. This approach is used allows seamless propagation of detailed naph-
to identify improvements in operating strategies tha components available in the crude assays.
rather than achieving optimum numbers from a Evaluation of feed for an aromatics complex
simulation. The flowsheet provides a quantitative requires rigorous simulation of the reformer,
analysis of an opportunity, but it is essential that trans-alkylation, isomerisation, and component
the focus is on identifying improved strategies based fractionation which includes extractive dis-
rather than targeting specific numbers. The abso- tillation. A case study for feedstock evaluation
lute value of the KPIs may change in an operating at an aromatics complex is discussed later in this
refinery due to various changes such as different article.
crude basket, varying product demand, and cat- Refinery engineers usually perform quality
alyst change-over. The improved strategy iden- give-away (QGA) analysis on a monthly basis.
tified using the flowsheet is usually valid for all QGA analysis focuses on the properties which are
these changes but the values of KPIs are different. constrained in the blends, such as RON and RVP
Petro-SIM is capable of generating an auto- for gasoline, or flash point and recovery for die-
mated Excel interface. The Excel interface is sel. One of the main inputs for QGA analysis is a
linked with Petro-SIM flowsheet and provides loss of margin per unit of give-away in key con-
mass and volume balance for the overall flow- straint, for instance gasoline RON of 92 compared
sheet and for each process unit as well. Properties with a spec of 91. The loss of margin can be easily
of streams within the flowsheet and all operating estimated for some of the constraints, for instance
variables of the reactors and fractionation are also gasoline RON if naphtha is used as the low
available in the Excel interface. The Excel inter- octane blending component. Estimation of the
face can be tailor-made based on user preferences. loss of margin for some of the properties requires
Mass or volume balance estimated by the refin- rigorous simulation, say diesel recovery which

4 PTQ Q3 2019 www.digitalrefining.com/article/1002322

 Product
CDU/VDU Conversion
Stream blending
units
Feeds routings Products

Suhar
complex

Muscat
refinery

Figure 3 Orpic complex-wide flowsheet

can be corrected by increasing the cut point of distribution from the unit. If few items of equip-
heavy gasoil (HGO) in the CDU which affects ment are underperforming, say damaged trays
feed for the diesel hydrotreater and fluid catalytic in a column, a model may need to be recalibrated
cracker (FCC). KBC has used the refinery-wide for these changes.
flowsheet to estimate the impact of QGA reduc- KBC uses Petro-SIM’s collaboration functional-
tion in numerous PIPs. ity to update kinetics and fractionation models in
the flowsheet using updated standalone models.
Flowsheet maintenance Collaboration can be used offline or through a
Maintenance of the flowsheet is as important as common server which connects standalone mod-
developing it. Results from the flowsheet may els with the flowsheet.
deviate from actual operation if changes are not
made in the flowsheet on a regular basis. Refinery Case study
routings, marginal mechanisms, blending strate- Orpic is one of Oman’s largest and most rapidly
gies, and so on do not change on a frequent basis. growing businesses. Orpic’s refineries in Suhar
Changes in the infrastructure of process units, and Muscat, as well as the aromatics and poly-
catalyst replacement, and under-performance of propylene production plants in Suhar, produce
major equipment affect the overall performance fuel, plastics, and other petroleum products.
of the refinery. Due to this, the main changes are In a collaborative effort, KBC and Orpic devel-
usually made in the standalone models which oped a complex-wide flowsheet which includes
need to be updated in the flowsheet as well. the Suhar and Muscat (MAF) refineries. The com-
Recalibration of the standalone models is plex-wide flowsheet also includes an aromatics
required for replacement of the catalyst. Catalyst plant. The snapshot shown in Figure 3 provides
deactivation in fixed bed reactors usually does an overview of the complex-wide flowsheet.
not require recalibration. If users can provide the The envelope of the flowsheet starts at the feeds
age of the catalyst, then Petro-SIM can predict the for the complex and the finished products pro-
status of the catalyst and its impact on the yield duced by Orpic are the final products from the

www.digitalrefining.com/article/1002322 PTQ Q3 2019 5

 flowsheet as well. Crude to the CDUs is defined Conclusion
using the crude assay, and few of the import Refinery-wide flowsheets can demonstrate the
streams are defined using the required properties true representation of changes in key operating
and detailed composition. variables and their associated impact on overall
Stream routings in the flowsheet represent the refinery operation and product blending. They
routing strategies used by Orpic. All conversion are being used for identification and evaluation
units in the refinery and aromatics complex are of opportunities for margin improvement which
simulated using Petro-SIM reactors which are include optimisation of stream routings, blending
based on non-linear kinetics and equilibrium. The strategies, molecular management, throughput
process units in the flowsheet include fractiona- maximisation, feedstock selection, and improve-
tion models as well. ments in unit operating conditions.
Petro-SIM blenders are used to optimise Detailed kinetic and equilibrium based models
product blending based on prices of different for process units are the main building blocks for
products, market demand, and specifications. a refinery-wide flowsheet. These are connected
Petro-SIM workflows configured in the flowsheet and calibrated using test run data, and operating
ensure that the major constraints of the refinery parameters from historian and laboratory data.
and process units are not violated. Development of the flowsheets is critical as
The flowsheet provides a detailed under- including too many details increases the build
standing of the interaction between the different time as well as the efforts required in analysing
refineries and the petrochemicals complex. As the results.
all process units are interlinked, the impact of Adding the price set data of the complex allows
change in the operating condition of any process for techno-economic optimisation of opportu-
unit affects all process units in the downstream, nities, whether this is part of a large improve-
recycles, and product blending as well. ment programme or just part of troubleshooting
The Orpic flow sheet has been utilised for many solutions.
studies. One of them was to utilise the excess To ensure top performance, workflows and reg-
amount of condensate stream from the MAF com- ular check-ups are required. Changes in the infra-
plex to the aromatics (AP) complex to maximise structure of process units, catalyst replacement,
the Suhar isomerisation unit (ISOM). under-performance of major equipment, and so
In normal operation, OLNG condensate is on affect the overall performance of the refinery
imported feedstock used as the main feed in MAF and therefore need to be reflected in the refin-
NHT to maximise PENEX feed since it comprises ery-wide flowsheet as well.
approximately 65% of light naphtha components.
In some cases, and mainly when there is a limita- Jitendra Chellani is Senior Consultant with KBC (A Yokogawa
tion in MAF NHT, the agreed amount of this con- Company) in Abu Dhabi. He has over 14 years of experience
densate is not fully processed in the unit so the in refining, process simulation and consulting and holds a
excess amount of it is shipped to Suhar and then master’s degree in chemical engineering from Institute of
sold out as light naphtha stream or blended in the Chemical Technology, Mumbai, India.
Suhar gasoline pool. Asma Al Sariyahi is is senior Process Engineer with Orpic in
It was found that processing of this stream to Sohare. She has over eight years of experience in refining and
process simulation and holds a bachelor’s degree in petroleum
the AP NHTs will increase light naphtha feed to
and natural gas engineering from Sultan Qaboos University,
the isomerisation unit, so it will be recovered as
Muscat, Oman.
gasoline with RON of 91. The current throughput
of ISOM unit ranges from 80% to 100% and it can
increase to 110%. LINKS
On the other hand, in some cases the availabil-
ity of AP imported feed in the market contrib- More articles from: KBC (A Yokogawa Company)
utes to AP NHT utilisation which affects ISOM More articles from the following categories:
feed. Processing OLNG condensate in the AP Energy Efficiency and Energy Management
NHT in this case could recover reduction of the Instrumentation, Automation and Process Control
light naphtha stream and sustain isomerisation Process Modelling and Simulation
throughput in the higher side.

6 PTQ Q3 2019 www.digitalrefining.com/article/1002322