What techniques can you use to ensure reliable and valid plant operations simulation data?

There is a science of operations as first described by Factory Physics, as Deuard mentions. Operations science (OpScience.org) describes the basic relationships required to understand any processes’ behavior. Operations science should inform any process simulation. In my experience over 3 decades, I have seen simulations performed assuming that operations behavior has to be “discovered,” e.g., a simulation to determine the drivers of average time through the process(cycle time). This would be like building a bridge to determine how gravity works—we already know how gravity works.

A way to work through the model building and analysis of the fidelity of the model is the principles expounded up in Factory Physics by Wallace J. Hopp and Mark L. Spearman. The authors develop a systematic description of manufacturing systems.

It is likely that the modeling will result in a variety of interaction modules that are statistically constrained to interact in a coherently way to represent the factory. Each of the modules will be independently verified and statistically adjust to fit the pre and post modules. There are many software packages that can be used. Two to consider are Wolfram Mathematica (paid) and R. Additionally, a designed experiment could be conducted to analyze the behavior of the both the factory and the simulations to provide an estimate of the fidelity of the model.

You can use various techniques to validate and verify your simulation, such as comparing it with historical data, experimental data, or analytical solutions, conducting sensitivity analysis, testing extreme cases, or applying formal methods.

One of the simplest measures of the success of an optimization system is its ability to correctly identify the optimum process operations policy. In the case of a model-based real-time optimization system, few tools are available to the system designer for investigating the suitability of a candidate process mode for use in such systems. This paper presents point-wise model adequacy checking methods, both analytical and numerical, for determining the ability of the model-based optimization system to have an optimum coincident with that of the true process.

For nuclear power plans, we use full-scope simulators, which have to meet very strict regulations and standards for fidelity with the control room and plan behavior in normal conditions, anticipated transients and design basis accident conditions. See, for example, the requirements in the standard ANSI/ANS-3.5-2018 "Nuclear Power Plant Simulators For Use In Operator Training And Examination".