February 2017

ESSENTIALS OF
DYNAMIC SIMULATION
An Essentials Guide, by the editors of Control

About the Control Essentials Series

The mission of the Control Essentials series is to provide process industry professionals with
an up-to-date, top-level understanding of a range of key process automation topics. Our intent
is to present essential engineering concepts in a practical, non-commercial fashion, together
with a review of the latest technology and marketplace drivers—all in a form factor well suited
for onscreen consumption. Check in at ControlGlobal.com/Essentials for other installments in
the series.
—The Control Editorial Team

This Control Essentials guide made possible

by MYNAH Technologies. See page 6 for more
information on the company’s full range
of simulation solutions.
A LIFE CYCLE VIEW

T
he process industries are by nature a complex, risky business. And that risk can
take many forms over a plant’s life cycle. First there’s schedule risk. From the Table I. Life Cycle Dynamic Simulation Tasks & Deliverables
first moment a new facility gets green-lighted, corporate stakeholders want the
plant up and running as quickly as possible so they can start to realize a return on the in- Capital Project Phase Operations Phase
vestment they’ve approved. Project delays can mean revenue lost forever, as in the case
Concept Safety
of a patented pharmaceutical. And in the energy industry, earlier “time to first oil” can
• Set goals for dynamic simulator use • HAZOP and process safety review
mean millions of dollars in cash flow. Then there are operational risks. Would an alter- • Develop vision for business impact • Interlock verification & training
nate control strategy make the process more efficient? Or would it give rise to instabili- • Risk analysis
ties or other unintended consequences? Are operators prepared to handle process upsets
and tasks that are not often performed, such as start-ups and shutdowns? Then again, Pre-FEED Training
• Develop project strategy • Safe operator experimentation
what happens when your most seasoned operators walk out the door, never to return?
• Develop functional requirements • Unit start-up, shutdown, trip recovery
Among all the software tools at the process manufacturer’s disposal, dynamic simu- • Develop execution plan • Infrequent process occurences
lation is uniquely suited to help address these various sources of risk from the earliest
stages of system design throughout a plant’s operating life. Dynamic simulation makes FEED (Front-End Engineering & Design) Knowledge Transfer
possible the early testing and troubleshooting of new control system code, when course • Review process design & control philosophy • Capture experienced operator lessons
• Identify operational & control issues • Accelerate new operator experience
corrections are easiest, fastest and least costly to make. Dynamic simulation makes pos-
• Efficient experiential training
sible the early hands-on training of process operators, which can be started even before
the physical process is commissioned. Once a unit is up and running, dynamic simu- Engineering Environmental
lation paired with an offline replica of the control system provides operators with ex- • Test control system configuration & graphics • Identify process risks
periential preparation for process upsets and procedures that are not often performed, • Advance evaluation of control system design • Determine control issues
• Analyze process event parameters
such as start-ups and shutdowns — an impossible or impractical task when only the live
production system is available. Dynamic simulation also is an effective tool for explor- Construction Regulatory
ing the implications of potentially more optimal “what-if” scenarios — without putting • Test control system for operational readiness • Operator procedure validation
actual production at risk. • Train operators, assess competency • OSHA, FDA preparation
And while the direct benefit of using a dynamic simulator is difficult to quantify, a • Training record development
recent survey by the Electric Power Research Institute (EPRI) indicated an average Start-up Optimization
yearly saving of about $4,500 per megawatt of generating capacity. These savings are • Test any final changes prior to implementation • Process improvement studies
attributed to reduced training costs, improvements in plant availability, fewer environ- • Review procedures, test operator competency • Loop tuning & optimization
mental excursions and reduced damage to equipment. A little bit of math indicates a
three-month payback for the typical power generation facility and begs the question: Early Operation
In what scenario would you not invest in a dynamic simulator? All dynamic simulation • Finalize operational procedures & documents
implementations, however, are not created equal. And on the pages that follow we’ll • Test process control improvements
explore the key considerations.
ALL MODELS ARE NOT CREATED EQUAL

W
hile every process simulation is based on underlying mathe-
matical models, it’s important to remember that all models are
not created with the same goals in mind. The models used for
process design have their place, but typically are ill-suited to the dynamic
requirements of control system testing and operator training. We’d hardly
be surprised to hear that the workhorse CAD models for Boeing’s latest air-
plane don’t also drive the airplane’s flight simulator. So why should our in-
dustry be any different?
In contrast with most process design tools, the process models used for
operator training and control system testing have to be dynamic and they
have to execute in real-time. They must converge on a result every second
and provide realistic results over a wide range of process conditions and
transients. This isn’t to say that dynamic models are less rigorous than their
steady-state counterparts; it’s just that the absolute accuracy of process vari-
ables takes a back seat to accurately depicting the dynamic behavior of the
process over time.
As such, dynamic models used for operator training and control system
testing still rely on chemical engineering “first principles.” That is, the ther-
modynamics, reaction kinetics and heat and mass transfer equations that
govern process behavior. Within a dynamic simulation software package, Industry standards such as the ISA/ANSI 77.02 specification on func-
basic and advanced function blocks model the dynamic behavior of valves, tional requirements for fossil fuel power plant simulators correctly apply a
transmitters and entire unit operations — configurable with process-spe- different accuracy standard to dynamic performance (20%) than to steady-
cific information such as known reaction kinetics. It’s also important to note state (2%). Further, the specification states that for a dynamic simulator the
that same level of complexity need not be used to model all aspects of plant “observable change in parameters shall correspond in direction to those ex-
operation: a tank farm requires a simpler model than a series of interacting pected from a best estimate for the simulated transient and shall not violate
distillation towers. Using the appropriate model for each unit operation ulti- the physical laws of nature” and that “the simulator shall not fail to cause
mately will provide the best balance of simulator performance, cost and fu- an alarm or automatic action if the reference plant would have caused an
ture flexibility. Once operating experience has been gained, dynamic mod- alarm or automatic action, and, conversely, the simulator shall not cause an
els can be further tuned using actual data. Models based solely on statistical alarm or automatic action if the reference plant would not have caused an
correlations of process data, however, are unlikely to behave well through- alarm or automatic action.” This is, after all, the true measure of a dynamic
out the full range of possible conditions, especially at the edges of the oper- simulator’s utility: Does it correctly model the process behavior such that
ating envelope such as during an upset, startup or shutdown, making them operators and process engineers can better understand how the process will
inappropriate for use in testing and training. react to specific actions and operator interventions?
IMPLEMENTATION MATTERS

T
he right choice of process model is only the first step in the successful
application of a dynamic simulator. Just how the models are Training
implemented and the manner in which the production control system
Operations/engineering
is recreated offline will make or break how well the combined system works HMI
initially and how easily it can be maintained over the process life cycle.
HMI HMI HMI Dynamic
Perhaps the most important architectural consideration is the logical separa- simulator
tion of the dynamic simulation package from the offline replica of the produc-
tion control system. Resist the temptation to wedge a dynamic simulator into High level
the control system environment, or to emulate the control system within the Virtual controller process models
dynamic simulator. Both functions will suffer — initially and into the future Training
manager
—from this sort of compromise. Process control systems are designed for con- Control modules
trol: they don’t include the tools needed to readily build and manage a dynamic IO & equipment
simulation, and attempting to do so will likely cause the performance of the off- models
line control system to diverge from its production counterpart. A dynamic sim-
Control system
ulator, in turn, will likely fall short in trying to emulate the subtleties of a given IO
manufacturer’s process control system — initially and over the course of time. Real-time IO Simulated IO
Maintaining these domains as separate applications allows the offline con-
trol system to faithfully duplicate the operation of the production control sys-
tem complete with identical operator graphics, alarms and control strategies. Figure 1. For best performance and maintainability, a dynamic process
Most all control system manufacturers today offer a soft controller or simulator simulator should be implemented separately from the offline control system
that emulates the operator stations, engineering station, process controllers and with which it works.
other higher level system functions. These OEM products provide an environ-
ment where the control system can run in a manner identical to the actual
plant, and updates to the production control system can be readily synchro-
nized across the two platforms.
In turn, modeling of the transmitters, final control elements, and the process
itself in a dedicated dynamic simulation environment allows the user to easily
develop and maintain process models to the level of complexity or fidelity re-
quired by the task at hand. Further, software functionality for training logistics
should reside in yet a third integrated domain (Figure 1) that allows supervisory
control of both the control system and the process simulator.
DYNAMIC SIMULATION’S ROI

D
espite a paucity of on-record reports of cost savings due to the use of dynamic simu- Similar benefits of dynamic simulation are evident in reports from other plants using
lators, anecdotal figures in an ARC Advisory Group report from 2006 continue to various automation and simulation systems:
impress: up to $500,000 per day; $1 million in savings per production run due to • A leading pharmaceutical company undertaking a control system modernization proj-
higher on-spec product; $100,000 per hour in operating cost savings due to more proficient ect reported that dynamic simulation for testing and training “served to minimize sur-
operators; and $1 million in risk mitigation per incident through better identification and prises during commissioning and qualification.”
addressing “dormant errors” in automation systems. • A major operation extracting oil sand for processing into refinery feedstock reported
For example, Ron Cisco, DCS training specialist at Salt River Project’s coal-fired Coro- training simulations led to greater safety, environmental compliance and efficiency,
nado Generating Station, reported at the 2013 ARC Advisory Group Industry Forum that faster startups and reductions in downtime, energy costs, equipment damage, produc-
the company was able to bundle an operator training simulator into a larger control system tion loss and more.
migration project because it represented a relatively small additional cost. Even so, the simu- • A nd a recent survey of 250 simulation users in Norway by World Oil Online attributed
lator was justified based on the demographic crunch of retiring operators that has only inten- a 31% increase in operator effectiveness to the use of dynamic simulators for operator
sified in recent months. Indeed, in the two years since the project was completed half of the training.
plant’s control room operators had retired and 100% turnover is expected within five years. In the end, an investment in dynamic simulation is much like an insurance policy. It rep-
What surprised Cisco and the rest of company management even more was the 144 con- resents a known upfront cost that can pay for itself many times over when a hidden automa-
trol logic issues that were discovered and resolved with the aid of the dynamic simulator tion system error is exposed or when an operator in unfamiliar territory finds he knows what
before operation began and “before we even started operator training,” Cisco said. These to do because of the training and preparation he’s received on the plant’s dynamic simulator.
were issues that were not seen during the factory acceptance test with simple, non-dynamic
I/O tiebacks. It was only when the true process dynamics were simulated that these issues
appeared. Management consensus was that the dynamic simulator “already paid for itself”
before the new control systems even came on line.
And at Barrick Gold’s $3.8 billion Pueblo Viejo mine in the Dominican Republic, dy-
namic simulators on 10 major processes are helping operators to deal with ore variability,
including hardness, mineral content and other characteristics that can help or hurt recovery,
according to Paul Yaroshak, process systems engineer. “Our processes need robust logic to
effectively respond to these changes, and simulations help us to better understand process
interactions, remove bottlenecks, and provide insight for managing variability.”
“Likewise, our training simulations include using actual screens and graphics with realis-
tic process dynamics and responses, which force operators to react in real-time to different
alerts, alarms and process interlocks, and prepares them to handle many situations and fail-
ures,” says Yaroshak. “The old way would have been to wait until the plant was running to do
this kind of training, so the simulations buy us a lot of time. We’re also using our simulations
to develop testing of configuration and function logic within our DCS, so we can evaluate
and fine tune interactions between different elements of the control system, and make sure
they’re working together as designed and intended.”
MADE POSSIBLE BY

This Essentials guide on Dynamic Simulation was made possible by MYNAH

Technologies, a leading provider of a life-cycle dynamic simulator that is used
for a variety of purposes from automation system acceptance testing to opera-
tor training, and industrial Ethernet solutions. MYNAH’s Mimic simulation soft-
ware is used in more than 1,300 sites in 68 countries worldwide ranging from
hydrocarbon production and refining to chemical, pharmaceutical and biotech
industries.

Learn more about MYNAH Technologies’ dynamic simulation offering.