Artificial Neural Network to Improve Speed Control of Permanent Magnet Synchronous Motor

This paper presents the design of a Neural Network based approach to improve efficiency in a vector control of Permanent Magnet synchronous Motors (PMSM). The conventional Proportional-Integral (PI) controller is largely used in industry because of the robustness this regulator procures. But in some case, when the dynamics of the system vary over time or with operating conditions, the performance of the controller will be spoiled. The Artificial Neural Networks (ANN) used as a speed controller seems to be a promising solution in this purpose. In this study we apply a multivariable (MIMO) neural network regulator in place of PI controllers of the vector control scheme of the PMSM. Analysis and simulation results are presented to demonstrate the validity of the proposed controller to ensure robustness against load and parameters variations and to achieve the required performances.