A physical-statistical hybrid two-stage wind power prediction method

Abstract

The invention discloses a physical-statistical hybrid two-stage wind power prediction method. Wind power prediction obtained by a physical prediction method and a statistical prediction method is usedas input of a second stage, and through a BP neural network, a wind power sequence on a target day is predicted. The method comprises the following specific steps: a physical prediction neural network model is built and trained; with each day as the target day, wind speed general similarity coefficients of other days relative to the target day are calculated respectively; a statistical predictionneural network model is built and trained; with wind speed and wind direction data of 24 h per day as input, a wind power physical prediction sequence for each day in former 180 days is sorted; withwind power sequences corresponding to former five similar days of each day as input, a wind power statistical prediction sequence for each day in former 180 days is obtained; and a BP neural network prediction model in a second stage is built and trained until a training error is smaller than a given value, and a two-stage prediction model is obtained finally.

Description

technical field [0001] The invention relates to the technical field of wind power forecasting, in particular to a physical-statistical hybrid two-stage wind power forecasting method. Background technique [0002] With the increasingly prominent energy and environmental issues, wind power, as a new energy power generation technology with great development potential and relatively mature technology, has developed rapidly and extensively in recent years. As of 2016, the cumulative installed capacity of wind power in the world reached 486.75GW, of which my country accounted for 35%, and due to terrain and other factors, it showed more characteristics of large-scale and highly concentrated development. At the same time, because the inherent randomness of wind power brings difficulties to wind power forecasting, compared with load forecasting, the accuracy of wind power forecasting is lower. Running presents a challenge. Therefore, research on the prediction method of wind power...

AI Technical Summary

Problems solved by technology

[0004] Among them, the advantage of the physical prediction method is that the demand for historical data is small, but due to the complex physical mechanism of weather conditions affecting wind power power, the actual effect may not meet the accuracy requirements

The statistical prediction method does not care about the specific physical mechanism, but analyzes the statistical relationship between historical data (including wind power, meteorological data, etc.), so it is highly dependent on data

Images