A Simulation Method of Solar Irradiance Data Based on Feature Clustering and Markov Transition Probability Matrix

Solar irradiance is one of the significant influential factors of solar photovoltaic power generation and it is necessary to model and simulate abundant solar irradiance data. In this paper, we propose a simulation approach of solar irradiance data based on feature clustering and Markov transition probability matrix. We introduce the features of solar irradiance data, k-means algorithm and Markov transition probability matrix of solar irradiance conditions, which make up simulation algorithm of solar irradiance. According to this method, a simulation example of National Renewable Energy Laboratory (NREL) one-minute data is presented and the paper gives analysis and evaluation of the results. Finally, there are the conclusion and some possible extensions.