To solve the problem of wind-induced vibration of heliostat mirror in the solar thermal power generation system, it is necessary to build a wind-induced vibration monitoring and control system and use a vibration absorber to track and suppress wind-induced vibration. In this paper, the wind-induced vibration signal of the heliostat is collected, displayed, analyzed and extracted based on the LabVIEW platform. The Python deep reinforcement learning algorithm is used to build the vibration state and control state data sets, and the optimal control decisions are obtained through training. The excitation current is precisely controlled to control the stiffness of the magnetorheological elastomer in the vibration absorber, and then the vibration frequency of the dynamic vibration absorber is adjusted and controlled in real time, so that the time varying vibration signal of the mirror can be quickly tracked and suppressed.
|