Improved random forest method-based wind-solar power generation output short-term prediction method

Abstract

The invention belongs to the technical field of new energy power generation output prediction and discloses an improved random forest method-based wind-solar power generation output short-term prediction method. The method comprises the following steps of: obtaining an optimal initial parameter of a random forest method by adopting a genetic algorithm, and predicting short-term wind-solar power generation output by adopting the random forest method on the basis of wind-solar power generation and weather historical data, wherein the prediction process comprises an offline training process of the random forest method and online prediction of the improved random forest method. The offline training process comprises the steps of generating a training set, obtaining optimal characteristic parameters of a random forest by adopting a genetic algorithm, and carrying out model training by adopting a random forest method. According to the method, training data does not need to be additionally corrected, and the tolerance to abnormal data and system noise is high. According to the method, the genetic algorithm is adopted to optimally set the parameter values of the random forest decision tree. It is guaranteed that the random forest algorithm achieves the optimal classification precision and recognition rate.

Description

technical field [0001] The invention belongs to the technical field of prediction of new energy power generation output, and in particular relates to a short-term prediction method of wind power generation output based on an improved random forest method. Background technique [0002] At present, the closest existing technology: With the continuous adjustment of energy structure and the influence of ecological environment, climate change and other factors, the development and utilization of renewable energy has become the consensus of global energy development. Due to their fast growth rate and high technological maturity, wind power and photovoltaic power generation have become clean energy with the best commercialization prospects and increasing proportion of electricity supply. However, the random fluctuation of wind power and photovoltaic power generation will affect the safe and stable operation of the power system. Prediction of photoelectric power and photoelectric p...

AI Technical Summary

Problems solved by technology

However, the random fluctuation of wind power and photovoltaic power generation will affect the safe and stable operation of the power system. Prediction of photoelectric power and photoelectric power is a common technical means to solve this kind of problems.

[0003] At present, the wind power and photovoltaic power prediction technologies commonly used in the industry include the similar day method, neural network method, bee colony algorithm, crow algorithm, wavelet and neural network combination algorithm, etc., but these methods generally have such problems: first, they need to compare The preprocessing of data or training data will affect the accuracy of actual output data. Second, the learning ability of the prediction model is insufficient, and the algorithm is prone to fall into local optimal values, and the prediction accuracy is not high.

[0005] (1) The commonly used wind power and photovoltaic power forecasting technologies in the industry have factors that affect output, and the data accuracy is not enough

[0006] (2) The commonly used wind power and photovoltaic power prediction technology in the industry has insufficient learning ability of the prediction model, and the prediction accuracy is not high

[0007] (3) The industry lacks a universal and effective prediction method for both wind power and photovoltaics

Obviously, the problem of accurate prediction of new energy output needs to be solved urgently. However, the output characteristics of wind power generation and photovoltaic power generation are different, and there are many uncertain factors affecting their output prediction, which is often difficult to accurately model, and the effect of conventional prediction algorithms is difficult to meet actual needs.

Images