Simulation needs to overcome perceived complexity and trust issues
Industry experts offered views about the benefits simulation technologies might offer in the process and manufacturing sectors.
Control Engineering Europe asked industry experts for their views about the benefits simulation technologies might offer in the process and manufacturing sectors, and about progress on adoption of the technology.
Q: Is simulation technology being used widely in process or manufacturing sector operations today? If so, why and in what type of application?
Dr. Mathias Oppelt, head of simulation center for process automation at Siemens: Yes, simulation technology can be considered state-of-the-art along the lifecycle of process and manufacturing plants. It starts with support of the process first with process simulation tools. Next the engineering of the plant needs to be supported by simulation. For example, for validating the construction or to support the engineering and validation of the automation system by virtual commissioning. Further, operators of the plant can be trained in a realistic simulator environment, consisting of a high-fidelity model of the plant and process and the original automation logic and graphics. Finally, during plant operations simulation can be applied. What they all have in common is that they are based on a rigors mathematical model of the plant process, which is then applied to numerical solvers which undertake optimization calculations.
The overall reason to use simulation or model-based technology along the life cycle is to support the decision-making processes and thus to lower risks and increase efficiency.
Ronnie Baines, director, process simulation & digital twin Europe, Emerson: Simulation has been used in the process industries for many years. However, historically it was viewed as an expensive technology, difficult to maintain and so was only applicable to a subset of end-users.
Recently, we have seen the technology become much more accessible, applicable to all the process industries as well as easier to maintain. This has resulted in it being applied to a number of different applications. Emerson’s Digital Twin lifecycle concept outlines how this technology can be applied. It describes how an initial steady state design model of process (which almost all processes will have in some form) can be utilized and built upon to develop a plant wide dynamic simulation of the process. This allows it to be used for applications such as design review, engineering studies, procedure development, control system validation, virtual commissioning, competency assessment and most importantly as an operational decision support tool.
Prith Banerjee, chief technology officer at ANSYS : We are seeing simulation being used widely in the process and manufacturing sector to model various processes in order to optimize design parameters in both continuous process automation, such as pulp and paper or oil and gas, and discrete manufacturing automation like automotive and aerospace. Simulation is also used to model the control systems such as DCSs, PLCs and drives, driving such process plants and discrete automation.
Q: How is simulation technology being integrated into legacy process and control systems? Is any addition technology/hardware needed to ensure its successful integration?
Oppelt: The first is to use the original hardware controllers and interface them to a simulator, representing the connected I/O, devices and processes. This setup is called hardware-in-the-loop and it requires a hardware device to emulate the bus participants the controller is expecting to communicate with. This setup is dependent firstly on the bus protocol and then on the controller. The second option is to use emulated controllers which are able to run the original controller logic. This setup is called software-in-the-loop and it does not require additional hardware. Emulators can typically can run on a normal PC, which would also be needed to execute the simulation model for the plant behavior. The simulation is the same as in the hardware-in-the-loop configuration but it talks directly with the virtual controllers.
These virtual controllers are typically proprietary as they can only emulate a specific controller type. This second configuration typically comes with some additional benefits, as an emulated controller can also support functions like snapshot, to save and load states at any time, and to run in a virtual time, supporting fast- and slow-modes too.
Sometimes you can also find a third configuration, where the control system is completely reimplemented within a simulation system. This configuration is used for more generic simulators, where the exact plant and control behavior does not need to be represented.
Baines: Simulation provides two unique solutions for legacy systems. Firstly, we can develop a simulation of a process and integrate it into a copy of the legacy control system. This provides a ‘replica’ of the legacy system which can be used for training and removed bottlenecks from the process. Emerson’s Mimic simulation framework allows native connection into a large number of third-party control systems.
Banerjee: For legacy systems, users are building digital twins of their plants by taking pictures of the shop floor and building up the CAD models. This allows them to build simulation models to match the experimental results using artificial intelligence and machine learning technologies. Once these digital twins are built, they are being used to improve the Operating Equipment Efficiencies (OEE) metrics of various plants.
Q: What are the barriers to simulation technology implementation in industrial applications, and how do you see these barriers being overcome?
Oppelt: There are a few barriers to the application of simulation technology. The first barrier is the effort needed to create simulation models. This can be very time consuming and requires highly skilled domain experts. So, everything that reduces the effort for model generation is likely to support wider use of simulation. For example, models for virtual commissioning can now be generated to a high degree by reusing information from the control system or the plant engineering tool such as P&IDs. Other valuable tools are ready-to-use library components which allows the model generation exercise to focus solely on the setup of the overall system, rather than on the modeling of individual components.
Further benefits can be generated simulation tools can be used, more intuitively, by non-modeling experts. If, for example, models are equipped with simplified GUIs, which support dedicated simulations by non-experts, it would open the space of simulation-based decision-making to a whole new group of people within an organization.
The underlaying simulation technology – such as numerical solvers – are already quite mature for most engineering problems, nevertheless if these can gain more accuracy with less computing power consumption and high robustness it is likely to lead to more simulation projects.
Finally, the most critical factor as always is trust. Decisions made by simulation need to be trusted – by users and by management and certification agencies such as the TÜV or FDA. There are already good examples where early collaboration with certification agencies delivers high returns. Today it is a lack of confidence that seems to be one of the biggest hurdles to adoption of simulation technology, even as the technology in most fields is already very advanced it is the human factor that now needs to be addressed.
Baines: I believe the biggest barrier is awareness of how simulation technology has evolved in recent years and the perception of how difficult it is to develop. Traditionally, it has been deemed as expensive, limited in its uses and only supported by few specialist resources. This has changed with the advent of better simulation technology, improved hardware performance and tools which have been developed to create, utilize and maintain simulation solutions to ensure the return on investment is maximized at each stage of a facilities lifecycle.
Re-use of simulation models – from design through operations support – making the alignment with a production system an automated activity, are some of the ways to remove these barriers. Doing this helps to highlight the tangible values simulation will bring to both projects and operations, independent of the process.
Banerjee: The barriers to the adoption of simulation technologies in industrial applications is the legacy process plants where the original CAD models are not available. Another common barrier is plants that have not been connected using IoT technology to the IoT platforms. However, once they are connected it is possible to build accurate simulation-based digital twin models.