Research on Load prediction of solar photovoltaic power system based on Extreme Learning Machine
Solar photovoltaic power generation is affected by a variety of factors,including solar irradiance and panel tempera-ture,when the outdoor environment changes,the photovoltaic array can not continue to work,resulting in the reduction of the energy conversion efficiency of photovoltaic power generation,the load and energy injected into the grid will be reduced.Only when the sur-rounding environment changes are analyzed accurately,can the solar energy resources be utilized rationally and the output load of solar photovoltaic power generation be fully mobilized.Therefore,the load prediction method of solar photovoltaic power system based on extreme learning machine is studied.The photovoltaic power system is affected by solar radiation.In order to predict the system load,the peak load model is established by beta function to obtain the radiation intensity of photovoltaic cells and determine the distribution probability density of solar radiation.The probability density is used to construct the load prediction function set,and the dependence of the prediction function is defined on the basis of seeking the optimal function.Based on the expected value of extreme learning ma-chine associated prediction function,the prediction weight was determined by the least square solution to realize the load prediction of solar photovoltaic power system,and the method design was completed.The experimental results show that the new method can effec-tively predict the output value of parallel photovoltaic power system,and the average relative error can be controlled below 1％,which has application value.
photovoltaic power systemextreme learning machineload predictionsolar energy
万方数据
维普
