Simplified Finite Control Set Model Predictive Control Strategy Based on Support Vector Machine

In recent years, Finite Control Set Model Predictive Control (FCS-MPC) has been widely used in power converters due to its advantages of fast dynamic response, no regulator and multi-constraint control. However, due to traversing all possible voltage vectors, the computational burden of traditional FCS-MPC is relatively heavy. In order to reduce the computational complexity of FCS-MPC strategy in rolling optimization, a simplified control method based on support vector machine ( SVM ) is proposed in this paper. The SVM model is trained by collecting the input and output data of the controller during the steady-state operation of the system, and the mapping relationship between the system state and the optimal vector is established. In the control process, SVM is used to quickly predict the optimal vector and construct a candidate vector set, thereby reducing the search space and performing optimization. The simulation model of three-phase two-level LCL grid-connected inverter is built in MATLAB / SIMULINK. The simulation results show that the consistency rate of this method with full vector search is 98 % in steady state, which significantly reduces the computational burden and improves the control efficiency.