Abstract
In this paper, the Takagi–Sugeno (T–S) fuzzy model is used to express dynamic systems, and an on-line identification algorithm for its parameters and structures is presented. A new multivariable fuzzy generalized predictive control approach is put forward based on the identified fuzzy model by means of Clark's principle of single-variable generalized predictive control, some of whose performances are analysed in detail. A simulation study of the multivariable load system of electric generating unit shows that the approach is superior to conventional load control systems.