Longitudinal photovoltaic power prediction method based on historical data mining

Abstract

The invention belongs to the field of research of solar energy utilization, and in particular relates to a longitudinal photovoltaic power prediction method based on historical data mining. The longitudinal photovoltaic power prediction method comprises the steps of: performing statistical analysis for historical photovoltaic power data in different seasons at first, so that six statistical indexes of photovoltaic power in corresponding seasons are obtained; clustering the statistical indexes by utilizing the Euclidean distance, so that similar day matrixes under different weather conditions in different seasons, and forming similar day typical curves and distribution intervals through a one-dimensional forward cloud generator; realizing longitudinal photovoltaic power prediction at the corresponding times in 24 hours in the future by utilizing the Markov Chain theory, and fusing the prediction value with the similar day typical curves and distribution intervals, so that a new prediction value is formed; and finally, performing weighted fusion of the new prediction value and the prediction value obtained through a continuous prediction method through a one-dimensional reverse cloud generator. Thereby, longitudinal photovoltaic power prediction based on historical data similar cloud fusion is realized; and thus, the photovoltaic power prediction precision is further increased.

Description

technical field [0001] The invention belongs to the field of solar energy utilization research, and in particular relates to a longitudinal prediction method of photovoltaic power based on historical data mining. Background technique [0002] Compared with water energy, wind energy, geothermal energy, biomass energy, etc., solar energy has become the focus of people's attention because of its outstanding and unique advantages. Abundant solar radiation energy is inexhaustible and inexhaustible, and photovoltaic power generation devices are noiseless, pollution-free, cheap, flexible in scale, and easy for human beings to freely and widely use. According to statistics, the energy of solar energy reaching the ground every second is as high as 800,000 kilowatts. If 0.1% of the solar energy on the earth's surface is converted into electrical energy, and the conversion rate is 5%, the annual power generation can reach 5.6×1012 kWh, which is equivalent to 40 times the world's total...

AI Technical Summary

Problems solved by technology

The above two methods have been applied in photovoltaic power prediction, but there are some limitations in the method. For example, for data information with strong regularity and periodicity, these two prediction methods can achieve high prediction accuracy, but photovoltaic power generation There are characteristics such as randomness and volatility. Using these two methods, the effect is often poor and cannot meet practical needs.

Images