On Accuracy in Modeling and Simulation for System Control

Abstract

This paper addresses the accuracy requirements in the modeling and simulation of thermal systems. For the appropriate parameter identification of these systems, machine time constants and system time constants are discussed. The former is primarily useful for sensor selection whereas the later determines the constants on modeling accuracy for system performance predictions. Also the modeling and performance prediction via simulation of several components, systems and perturbations are considered in order to identify accuracy requirements. Several examples are reviewed to show the type of accuracy available in current models and their effect on simulation performance predictions. It is noted that the accuracy is quite different for parameters, inputs and measurements. More importantly, the level of accuracy is also dictated by the level of validation that is possible with real systems. Although this issue is not addressed in this paper, results are used to generate the level of modeling accuracy necessary (or needed) as a function of end-use.