A plant simulator based on hybrid model of physical model and Just-In-Time model using statistical approach

摘要

This paper discusses the development of a plant simulator. In our previous work, we proposed a hybrid of physical and Mahalanobis distance-based Just-In-Time (MJIT) models to address the problem of difficulty in physical plant modeling for highly nonlinear plants. However, MJIT model has the problem that the estimation results will be discontinuous. This problem will lead to abrupt changes in manipulated variables in a control action. Therefore, we propose modified MJIT model referred to as mod-MJIT model to resolve this problem. This paper first describes a method for estimation by mod-MJIT model and then discusses how mod-MJIT model overcomes the problem of MJIT model. Furthermore, this paper provides a calculation method of the accuracy of estimation by mod-MJIT model.