Qualitative modeling and simulation of dynamic systems

Abstract

Qualitative simulation is a promising technique for design and analysis, particularly in model-based reasoning systems. The purpose of qualitative simulation is to explain process observations by reasoning from physical descriptions to behavioral descriptions. Qualitative simulation has the ability to yield partial conclusions from incomplete knowledge of the process. In this work, the qualitative simulation algorithm, QSIM, is used to model qualitatively several systems from chemical engineering. The QSIM algorithm successfully generated qualitative descriptions for the open-loop responses for all of the systems studied including linear, nonlinear and multivariable processes. Some models required the use of redundant contraints to eliminate otherwise ambiguous parameters. The closed-loop behavior of a mixing tank has also been successfully analyzed with qualitative versions of feedback control. The correct dynamic behavior and the qualitative features of proportional control, such as offset, are captured by the QSIM algorithm.