Short-term interval prediction modeling of photovoltaic power plant output power based on combing Least Squares Method and weighted Markov chain

The paper examines a new method for photovoltaic power plant output power forecast modeling, in order to adapt to the impacts of new energy penetrated power system, realize the grid planning and operation of photovoltaic power generation system. It expands on the condition classify of photovoltaic power plant output power by taking full advantage of the matched curve of Least Squares Method, it builds the Markov chain model, and calculates the coefficients and weights of each order autocorrelation, it carries on the weighted Markov chain forecast modeling by means of combining the correlation analysis and Markov chain. The numerical results of Liaoning Jinzhou 3MW PV plant indicate that, the state interval division is reasonable, predictive analysis is accurate. The interval forecast of photovoltaic power plant output power supplies reliable basis for interval optimization modeling of uncertain variables after new energy into the grid. The method is easy to implement, it is more efficient use of time, and provides a broader outlook. Keywords: Least Squares Method; Weighted Markov chain; Interval forecast; Autocorrelation coefficient; Interval Optimization