- By Gregory K. McMillan
- December 26, 2023
- InTech Magazine
- Feature
Summary
A bioreactor system design example shows a way forward for plant development and process optimization. This feature originally appeared in the December 2023 issue of InTech digital magazine.
We have exceptional opportunities offered by today’s automation system capabilities to meet the needs and challenges of sustainable and productive plant operation. Smart instrumentation performance and intelligence and the increasing capabilities of control algorithms, data analytics, diagnostics and simulation to deal with a wide spectrum of process situations offers incredible possibilities. This is particularly important considering increasingly prevalent control expertise retirements, tighter budgets and schedules, and equipment and professionals stressed to perform beyond original expectations.
It is increasingly difficult—in actual plants— to conduct the tests and experiments needed to develop, implement and continuously increase plant capacity and efficiency or to prolong equipment life. The digital twin offers a groundbreaking way forward by virtue of the virtual plant.
A virtual plant is a software model that encompasses all the latest capabilities in an adaptable real-time simulation that includes the dynamic responses of the process, the equipment and automation system. The digital twin sets up the understanding required to import the configuration and files that are used by the actual plant. The digital twin setup and use are normally within the functional realm of experienced process control engineers. Many can help improve and connect dynamic simulations. However, the models and advanced control tools related to virtual plants are more likely developed by specialists.
This article reviews the functional value of digital twins and uses a bioreactor system design example to show how digital twin modeling and simulation feeds virtual plant design and maximizes the synergy among scientists, operators, process control engineers and control systems to enable process optimization. It is an adaptation of material from two new books that can help process control engineers learn more about virtual plants, setting up and running dynamic simulations, and optimizing the operations of existing plants using digital twin technology. This knowledge is critical for addressing the many challenges in bioprocess and pH control systems and achieving the most reliable and proficient process performance.
Functional value
Figure 1 illustrates a digital twin’s functional value, highlighting the bidirectional flows of the control system and process/equipment knowledge of process control and analysis tools, including online key performance indicators (KPIs) and real-time accounting (RTA) metrics for greater analysis and justification of improvements. The two-way knowledge flow is the key to improving the process/equipment and the control system in addition to the dynamic model and data analytics. As the fidelity of the dynamic model increases, opportunities arise for these tools to get results from the digital twin that can be used in the actual plant. The dynamic model can be run faster than in real time with the tuning corrected by applying the speedup factors. New control functionality can be developed and included in the dynamic model for evaluation. If online metrics show significant improvements in control and process performance, the functionality prototyped can be added as new blocks or as improvements to existing blocks in the distributed control system (DCS).
The digital twin can accelerate the benefits gained by offering users the ability to use process and end point monitoring and control, continuous improvement, and knowledge management tools in an integrated manner.
The most familiar use of a digital twin is for testing and training. To check out batch sequences and train operators, it is important to be able to simulate batch phases repetitively and rapidly. The ability to stop, start, save, restore and replay scenarios and record operator actions is critical. For first pass testing and familiarization of sequences and graphics, an automated tieback simulation may be sufficient. To test and learn about the interaction and performance of control strategies and the process, the higher fidelity dynamics offered by process models is important. It opens the door to upgrading the process and control skills of technology, maintenance, and configuration engineers who support operations. A process simulation with high dynamic fidelity is also important for testing process and control system interaction and performance.
The dynamic models often used for training operators as part of an automation project have a much wider utility that is more important today than ever is a reality that is not well recognized. There is a great opportunity to use the digital twin to maximize the synergy between the operators, process control engineers and control systems. To start on this path, process control engineers must be given the time to learn and use a digital twin and set up online metrics for process capacity and efficiency. The digital twin offers flexible and fast exploring => discovering => prototyping => testing => justifying => deploying => testing => training => commissioning => maintaining => troubleshooting => auditing => continuous improvement showing the “before” and “after” benefits of solutions from online metrics. Figure 2 outlines the major steps in continuous improvement and maximizing innovation.
The capability of dynamic models to improve system performance has greatly increased, even though the use has focused mostly on training operators as an automation project nears completion. The digital twin should detail the tasks needed to address difficult situations based on the best operator practices and process knowledge and eliminate the need for special operator actions through state-based control. Advanced process control (APC) and model predictive control (MPC) can respond to disturbances and address constraints intelligently, continually and automatically with great repeatability.
Compare this with what operators can do in terms of constant attention, deep knowledge and timely predictive corrections considering dead time, multivariable situations and uncertainty in human behavior. Some operators may do well, but not all operators. Then, of course, an operator can have a bad day. Automation enables continuous improvement and recognition of abnormal conditions by a much more consistent operation. A better understanding by the operator of control system functionality and process performance from online metrics greatly reduces disruptions by an operator unnecessarily taking a control system out of its highest mode and/or making changes in flows. Furthermore, procedure automation can eliminate manual operations during startup when the risk is the greatest compared to steady-state operation.
While we have singled out operators and process control engineers, the need for knowledge to attain the best performance extends to maintenance technicians, process engineers, mechanical engineers and information technology (IT) specialists. Think of what can be realized if we are all on the same page understanding the process, operational opportunities and the value of the best measurements, valves, controllers and software.
Possible digital twin opportunities to increase plant knowledge and performance include:
- Cause-and-effect relationships
- Interactions and resonance
- Valve and sensor response
- Process safety stewardship
- Control system and safety instrumented system (SIS) knowledge
- Validation and regulatory support
- Code checkout
- Process and equipment knowledge
- Process equipment degradation
- Startups, transitions, shutdowns and batch operation
- Optimum operating points.
Before the configuration even starts in the front end of a project, the process models can be used to evaluate control strategies and advanced control tools. In the past, this was done with offline dynamic simulations. Having ready access to an industrial tool set for basic and advanced control and simulations that are adapted to benchtop or pilot-plant runs offers rapid prototyping opportunities. This can lead to control definitions that have better detail and potential performance.
Benchtop or pilot-plant systems with a mini version of the industrial DCS are now available that greatly facilitate developing and scaling up the control system. Benchtop systems and pilot plants that have all the functionality of the main manufacturing systems are not yet prevalent because the development groups of these types of companies traditionally do not have the expertise (and, more importantly, the interest) to configure, maintain and engineer these systems. The digital twin enables synergy between scientists and control engineers to make the incredible capability of DCS part of the process R&D.
Bioreactor system design example
Digital twin use is particularly beneficial in pH and bioreactor system design. A pH system offers many orders of magnitude greater hydrogen ion concentration control precision and rangeability than any other concentration measurement. However, this is accompanied by extraordinary process gain nonlinearities and instrumentation response requirements. A digital twin can greatly increase system performance and decrease system cost as detailed in “Advanced pH Measurement and Control,” fourth edition, ISA 2023.
Bioreactors used to produce biologics for most modern-day new pharmaceuticals require incredibly tight pH and temperature control. There are exceptional digital twin opportunities to increase pH and temperature system performance but also develop innovative glucose and glutamine control systems to improve batch cycle time and yield by advanced control including batch profile and endpoint control. The digital twin offers the ability to improve batches worth more than $10 million by better process development and control without testing or experimentation within the actual plant as detailed in “New Directions in Bioprocess Modeling and Control,” second edition, ISA, 2020.
In addition, a dynamic model can be nonintrusively adapted to improve virtual plant fidelity by matching the virtual and actual plant manipulated flows. This can be done by an MPC developed offline whose controlled variables are the virtual plant’s manipulated flows, targets are the actual plant’s manipulated flows and manipulated variables are the corresponding virtual plant’s model parameters.
The adaptation is done without affecting the actual plant because the plant’s manipulated variables are being read by, but not changed by, the digital twin. It is critical that the digital twin has the same setpoints and tuning settings as the actual plant and that the digital twin is started with controller outputs initialized to match the actual plant.
Figure 3 shows a bioreactor model adaptation and consequential optimization by an MPC using inferential measurements and KPIs. The optimized setpoints from MPC with inferential measurements of growth and production rate are done in an advisory mode that does not affect the actual plant. Not shown in Figure 3 is that an MPC is run in automatic mode in another digital twin that is a duplicate of the adapted digital twin to generate and study the optimized setpoints. The setpoints are only eventually used to automatically optimize the plant if they prove more accurate, beneficial and reliable per KPIs than an MPC with inferential measurements computed from online and at-line analyzers.
Final thoughts
The digital twin and virtual plant provide a revolutionary opportunity to conduct the tests and experiments needed to develop, implement and continuously improve plant capacity and efficiency or to prolong equipment life. Process control engineers need to learn more about setting up and running the dynamic simulations and how they can be used to develop new plants or optimize the operations of existing ones.
This feature originally appeared in the December 2023 issue of InTech digital magazine.
About The Author
Gregory K. McMillan, the author of more than 30 books and 400 articles, is a retired Senior Fellow from Solutia and retired principal software engineer from Emerson. He won the InTech magazine “Most Influential Innovators” award in 2003 and the International Society of Automation “Life Achievement” award in 2011.
Did you enjoy this great article?
Check out our free e-newsletters to read more great articles..
Subscribe