The point of using simulation in operator training is simple. It reduces operator error. "If they are well trained, they make fewer mistakes," said Torgrim Schiefloe, area manager for ABB AS in Norway. Schiefloe presented "Building Operator Confidence and Competence Through Training Simulation" this week at ABB Automation and Power World in Orlando.
Better operator competency is a major contributor toward meeting productivity challenges, Schiefloe explained. "Training simulators prepare operators to run their processes efficiently and in an optimized manner."
Research, too, shows that simulators help make better operators. The ASM Consortium reports that about 40% of abnormal operations are caused by human error. In another study it was shown that 28% of the 170 largest property-damage losses over a 30-year period were caused by operational error or process upsets. Even in the nuclear power industry, the issues are significant. In a World Association of Nuclear Operators (WANO) study of 1,540 operating event reports, 891 were related to human error. "Human error is still one of the major causes of accidents in nuclear power plants," Schiefloe said.
These errors can be reduced by continuous use of simulators for operator training. As early as 1989, numerous studies have reported that a good process simulator is a powerful tool throughout the complete lifecycle of a plant—from pre-engineering and commissioning planning to operator training and troubleshooting. This contributes to operational readiness. Startup time can be reduced and maximum efficiency at startup can be more readily achieved, Shiefloe reported.
The better qualified a plant's operators are, the more profitable they are, too. The American Petroleum Institute (API) did a study, Shiefloe said, in which they noted that the competency of operators follows a Gaussian distribution, with the median skill level being an Operator II. The report determined that the payback from moving an Operator II to the skill level of an Operator III was on the close order of $350,000 per year for each operator.
"High-fidelity models are much better in representing ‘transitions,'" Schiefloe said, in describing the types of training operators need. "High-fidelity models will be better at simulating what is going on during startup, plant upsets and shutdown of a plant." So if your main concern is to train operators in handling process upsets and plant startup, then a high-fidelity process model will give better training. High-fidelity models also are more accurate. "If you plan to train operators in process optimization then hi-fi process models will give a more realistic response."
Simulators need maintenance, too, Schiefloe noted, just like control systems. When the plant deviates from the model, when the hardware and software evolve, when training requirements change, and when the operator training use cases expand, the simulator must be updated and maintained to correspond with reality. Schiefloe said that a simulator with no maintenance loses its relevance to actual plant operations in less than 24 months.
And when the simulator isn't really how the plant works, the operator gets exposed to "negative training," Schiefloe said. Repetitive experiential training is critical for operators to respond correctly and successfully in process upset conditions. Studies in aerospace, medicine and the military, among other areas, have consistently shown that repetitive experiential training works. The bad news, Schiefloe reported, is that training an operator on a simulator that doesn't mirror exactly the conditions and operation of the plant he or she is to be running makes them do the simulator action, even on the real control system—which can be dangerous and even deadly.
"If developed correctly and used effectively," Schiefloe said, "a simulator will provide realism to your training program. Defining a goal and plan will ensure you specify the right solution and provide a benchmark for measuring success."
Using the simulator from a lifecycle perspective, as ABB's customer Kiruna, Sweden-based LKAB (an iron ore mining company) is doing, makes it possible to have the dynamic process model developed in parallel with the process design; the control system can be tested and further developed on the simulator; the plant can be virtually commissioned on the simulator, and process optimization and modification studies can be done on the simulator so that they have considerably greater probability of effectively working when they are introduced to the actual control system.