An Improved Markov Chain Photovoltaic Output Prediction Method

An accurate and effective photovoltaic output prediction method has guiding significance for the planning and operation of photovoltaic power distribution networks. Aiming at the randomness of photovoltaics, an improved Markov chain photovoltaic output prediction method is proposed. The cloud layer state is divided based on the Cloud Cover Level index, and the cloud layer state transition matrix is established. The state duration sampling method is introduced, and an improved Markov chain model with time sequence is established to predict the cloud state change curve. The random sampling method is used to extract the irradiance sample value at each time from the cloud state time series curve, and the photovoltaic output time series curve is indirectly predicted according to the relationship between photovoltaic output and irradiance. Simulation analysis verifies the effectiveness of the proposed method.