Design of Control System of Once-Through Steam Generator Based on Proximal Policy Optimization Algorithm

Abstract

Because of the characteristics of the small water volume of OTSG, it is hard to control the outlet steam pressure when the load is changed or disturbed. This study is devoted to the control of the once-through steam generator (OTSG). A double-layer controller based on the PPO algorithm is proposed to control the outlet steam pressure of OTSG. The bottom layer is the PID controller; it directly regulates the OTSG feed water valve and then controls the steam pressure. The top layer of the controller is the agent based on the PPO algorithm, which is responsible for optimizing the parameters of the PID in real time to obtain better control performance. The agent chooses PID parameters as actions to the environment, and then, the reward value is obtained through the reward function of the environment which enables online learning of the agent. Compared with the PID controller, the simulation experiment result shows that the method not only has a good control performance but also has a good anti-interference ability.