Research on Solar Radiation Estimation Based on Singular Spectrum Analysis-Deep Belief Network

Because of its clean, safe, convenient and efficient characteristics, the photovoltaic power generation act as an important part to effectively improving the climate change, energy security as well as environmental pollution. To effectively use solar energy resources and increase the photovoltaic power generation, accurate solar radiation estimation is an indispensable basis and premise. A new estimation algorithm based on singular spectral analysis-deep belief network (SSA-DBN) is proposed in this paper. Firstly, considering meteorological factors, the air pollution data is mainly studied. A deep belief network model is established based on the input quantity selected by the maximum correlation minimum redundancy algorithm (mRMR). Afterwards, the singular spectrum analysis of historical data sequences, including singular value decomposition and reconstruction, is applied to the radiation estimation model improvement, and a daily solar radiation estimation model based on singular spectrum analysis-deep belief network (SSA-DBN) is established. It could improve the accuracy of radiation estimation and provide more effective information for photovoltaic grid-connection and power system operation and scheduling.