Modeling, Identification and Validation of the Modified Magnetic Levitation Model

Abstract

This paper presents the nonlinear system identification methodology on the Modified magnetic levitation model (MLM). The proposed modification of the original CE 152 MLM replaces the outdated MF624 laboratory card with the Board51. This modification enables the integration of the MLM into the Distributed Control Systems (DCS) architecture and moves the control loop closer to the system. The integration into the DCS architecture requires a detailed description of the system, which along with well-known physics laws is used by the analytical identification method to derive a mathematical model. The method of experimental iden- tification is used to estimate all four unknown parame- ter values of the derived mathematical model. Parame- ters are estimated in the structure of output prediction error using the nonlinear least squares method. The resulting gray-box model is used to design a stabilizing optimal state control algorithm (LQI). This control al- gorithm is used for indirect validation of the identified model by comparing the outputs of the real and simula- tion models in a closed-loop setup. The validated model can be used as the basis for the creation of the digital twin.