In part I of this blog, the importance of having a good functional description was essential in developing a good simulation application, which then leads a good factory acceptance test. In this blog, the importance of simulation is discussed, along with simulation techniques.
Simulation of the software application is a critical step in any project in LSI’s view. The goal of simulation is to eliminate as much of the unknowns as possible before the most expensive phase of the project begins: startup and commissioning. As one of our customers puts it: “the goal is to eliminate the white knuckle ride of startup.” The other goal is to provide the right level of simulation to have a good factory acceptance test (FAT) or software acceptance test (SAT). The same LSI customer from above offers this perspective: “Simulation allows us to exercise all anomaly conditions of our process without real world damage.” Their application involves running gas compressors on a large pipeline. Emergency shutdowns can be catastrophic if not handled properly. Proper simulation allows for these conditions to be analyzed and taken care of before the control system is applied to the real world application.
A good simulation accomplishes things like:
- Confirms that interlocks are working properly
- Tests the sequences in the process or machine to make sure the steps are executing properly and that the proper transitions between steps are identified
- Confirms that any faceplates or pop-ups work correctly on the human-machine interface
- Any bugs in the code are detected and fixed
- Confirms the alarming strategy
Simulation techniques vary rather widely and the proper technique needs to be applied depending on project complexity and the project needs. As a user, you truly want maximum gain with minimum cost, so choosing the right technique is key. There are generally three types of simulation techniques applied and they are:
- Low Fidelity
- Medium Fidelity
- High Fidelity
In low fidelity simulations, simple “tiebacks” are used (i.e. when a valve is called to be opened, the open switch signal is triggered via a simple tie-back from the open output after a set time) and there is typically logic written in the control system to perform these “tiebacks.” Most of the time, these simulation routines in the PLC are removed after Factory Acceptance Testing. DCS systems often have a “simulation mode” built in. Many times it is up to the engineer doing the simulation to manually intervene to respond to process changes (i.e. if a pump that is filling a tank is called to run at a higher speed, then the tank level fills faster – there will have to be some interaction to make this happen). This technique is used for simple processes or machines where there is not a lot of loop tuning and loop interaction.
Medium fidelity simulation is much more process centric and takes things like mass balance, material balance, and heat balance into account. To continue with the example above, with the tank filling algorithm, a medium fidelity simulator would take into account that the pump is running faster and that the tank will consequently fill faster. This type of simulation is often used in process applications where flows, temperatures, and pressures need to be simulated, but the complex interactions that can take place between these variables does not need to be modeled. In medium fidelity simulation, it is not expected that loops will be tuned and loop tuning will need to be done at the time of commissioning.
High fidelity simulation is the most time consuming and cost intensive approach, but it can more than justify itself when doing a highly complex, integrated process. For instance, in simulating a power plant and its dynamic response to step changes in demand, it is critical to understand how the plant will react. The engineer and the operator would like to know what the impacts would be to steam flow, boiler drum levels, turbine speeds, etc. A high fidelity simulator literally simulates the entire plant and all of its interactions as it would happen in real time. The operator literally would see no difference in the simulated plant versus the real plant. Certain industries such as power and petrochemical require this level of simulation as it is critical to know how the processes will react together before the system goes live. It takes a lot of time and money to get the simulation completed in a high fidelity simulation. In industries like power and petrochemical, it is essential. In simple batch processes or machines, it is overkill.
The critical piece in all of this is to evaluate what makes sense for the application. LSI has used all of the techniques above. LSI writes a fair amount of simple tieback logic for simple batch processes and simple machine control. For medium fidelity simulation, LSI has used products like Mynah’s MiMiC (http://www.mynah.com/), Cape software (http://capesoftware.com/), and other products. For high fidelity simulations, LSI has used the above products, or has used more “home grown” systems from LSI’s partner company, Enero Solutions (http://enerosolutions.com/dynamic_simulation.php). (LSI and Enero partner together on steam optimization projects in power plants and cogeneration applications in plants like paper mills). During each project, it is imperative that LSI and the end customer evaluate the level of simulation required to streamline the startup and commissioning process.
More resources to understand simulation, how to apply simulation techniques, and what products are available can be found in the following locations: