Application of an adaptive Takagi-Sugeno fuzzy identification approach for interaction analysis of MIMO nonlinear systems

Obtaining an accurate quantitative dynamic model for most industrial processes is a challenging task due to the interaction between their multiple input-output variables. In this paper, a systematic approach is presented to identify a nonlinear multi-input multi-output process using an online adaptive Takagi-Sugeno (TS) fuzzy method in order to enable its interaction analysis. For this purpose, a relative gain array (RGA) measure is used to evaluate the interaction characteristic. A new development is presented to calculate the interaction measure around the local operating point of MIMO processes based on the most recent fuzzy rule-based identified process model. The performance of the proposed approach is demonstrated on a distillation column benchmark case study in the LV-configuration.