Testing of genetic-PI based controller for IPMSM drive

This paper presents a laboratory testing of genetic algorithm (GA) based self-tuned PI controller for the speed control of interior permanent magnet synchronous motor (IPMSM). A radial basis artificial neural network function is used for on-line tuning of the PI controller. GA has been used in this work in order to obtain the optimized values of the PI constants for precise speed control. An performance index has been developed using GA, whose minimum value ensures zero steady-state error, minimum speed deviation and minimum settling time of the IPMSM drive. The initial values of the radial basis function network (RBFN) are obtained through off-line learning. Training data for off-line learning are generated by simulating the IPMSM drive under various operating conditions and uncertainties. For on-line implementation, the PI constants are tuned by updating the parameters of the RBFN maintaining the genetic performance index at its minimum value. In real time implementation, the proposed controller has been realized using a digital signal processor (DSP) board DS1102 for a 1hp laboratory IPMSM. The control algorithm is written in C++, compiled and then down loaded to the DSP board. The agreement between the simulation and test results confirms the effectiveness of the proposed controller for the vector control of the IPMSM.