定日镜积尘因素分析及反射率变化预测模型OACSTPCD

In order to reduce the loss of power generation efficiency caused by natural dust accumulation,the solar thermal power station needs to clean the heliostat regularly.How to select the appropriate cleaning node is the key to improve economic benefits.In view to this,a prediction model of heliostat reflectivity change based on meteorological forecast data is proposed,which can be used for the solar thermal power station to flexibly arrange manual cleaning according to the reflectivity change trend.The factors affecting the dust accumulation of heliostats are analyzed.An artificial dust accumulation chamber is developed to carry out relevant qualitative experiments,and on this basis,a reflectivity prediction model based on neural network is constructed.Through outdoor dust accumulation experiments,the effect of the model is preliminarily evaluated.The experimental results show that the model can effectively predict the change trend of reflectivity.The average relative error with the measured data is 1.28% ,and the root mean square error is 1.34% ,which is of reference significance for the formulation of cleaning plan of solar thermal power stations.