Induction motor speed-regulating control system based on nonlinear kernel ridge regression

Abstract

To deal with the inverse model identification problem in motor variable frequency speed-regulating system, a new control method based on nonlinear kernel ridge regression is proposed. On the basis of reversible analysis of the original system, the-α order reversible model approximated by kernel ridge regression is connected with original system, and the pseudo-linear combined system is constructed. Experimental results show that the combined method can realize the linearization of the induction motor control system successfully, and high performance of speed control can be ensured for systems with inverse model identification errors and changeable load. © 2011 Springer-Verlag.