High-power rectification power supply real-time control method based on proxy model

The invention discloses a high-power rectification power supply real-time control method based on an agent model, and the method comprises the steps: 1), selecting a deep learning model according to a topological structure of a high-power rectification power supply; 2) digital PID parameters are determined according to power supply characteristics; 3) closing the switch K2 to control the rectification power supply through the traditional PID; 4) inputting and transmitting the key parameters rin, Kp, Ki, Kd and yout of the system as models to a deep learning model for online learning; (5) timing for K minutes, and repeating the steps (3)-(4) until the learning stop requirement is met; (6) when the set learning time is up, constructing a proxy model mapping relation, and completing the construction of the proxy model; according to the high-power rectification power supply control method, the problems that according to an existing rectification power supply control method, parameter calibration is difficult, the