Study on wind-induced fatigue of heliostat based on artificial neural network

Heliostat is a reflection device in tower solar power station. It is generally arranged in a flat and open area, with an independent column structure. Since the angle needs to be adjusted continuously during use, and the mirror panel suffers a large wind pressure, its support structure is prone to wind-induced fatigue. In this paper, a multivariate joint distribution model was first established based on the law of movement of the heliostat with the sun and the distribution of wind speed and direction. Through a combination of wind tunnel test and finite element analysis, the wind-induced fatigue of typical working conditions was analyzed, and then artificial neural network was used to predict the fatigue life of unknown working conditions. The neural network was improved and tested to make it more consistent with the actual situation, so that it can replace a large amount of work of finite element analysis. Finally, the distribution law of wind-induced fatigue life of heliostats under different factors was summarized, and statistics were performed on the working conditions with high probability fatigue, so as to provide a basis for practical engineering application.