Performance Analysis of Model Predictive Control With Variable Weights in Optimization as Applied to PMSM

PMSMs (Permanent Magnet Synchronous Motors) have the major advantage of higher efficiency than induction machines. This is due to the use of permanent magnets that reduce the losses associated with the generation of magnetic fields and rotor windings. They also offer high power density and torque, which makes them suitable for applications requiring high performance and precision. Also, they are reliable and require less maintenance. PMSMs are being used increasingly in industries like automotive, aerospace, robotics, etc. Although PI control, Sliding Mode Control (SMC), and Intelligent Control are popular control algorithms for PMSMs. However, control vector optimization is lacking in these algorithms. Model Predictive Control (MPC) is the solution for this. MPC was primarily created for slower applications like process control and it can now be easily applied to electric drives due to the development of fast processors and quick-switching technology. The algorithm, which always uses an explicit model, predicts the system output for a limited number of future sampling intervals or future time horizons. Then it defines a cost function based on the estimated future errors. This cost function is minimized so as to minimize the error and a control input. The performance of an MPC-based drive depends on the weight matrices used in the optimization process. It has been found that MPC-based control not only takes care of errors and cost-effectiveness but also increases the robustness and stability of the drive. This paper presents a comparison of the performance of the MPC algorithm applied to PMSM. By selecting various optimizing weights in the cost function, the drive is tested for various operating conditions. Extensive simulations are carried out on MATLAB/SIMULINK platform to study the effects.