基于RCC-GRU模型的超短期风电功率预测方法

Ultra-Short-Term Wind Power Prediction Method Based onRCC-GRU Model

准确高效的风电功率预测对于风电场和电网的稳定运行非常重要。提出了一种基于辐射分类坐标(RCC)和门控循环单元(GRU)的超短期风电功率预测方法。首先,分析了不同气象因素对风力发电的影响以及不同时间段的影响程度。其次,提出了一种辐射分类坐标方法对相似的时间段进行分类和选择,将所选相似时间段的数据集(包括发电量和多元气象数据)重建为训练数据集。然后,将GRU神经网络作为该模型的学习网络。实验结果表明,所提出的模型的预测准确率和确定系数分别为97.6%和98.99%,并结合3个误差指标和训练时间分析,RCC-GRU模型的准确性和效率均优于其它3个比较模型。

Accurate and efficient wind power prediction is very important for the stable operation of wind farms and power grids.This paper proposes an ultra-short-term wind power prediction method based on Radiation Classification Coordinates(RCC)and Gated Recurrent Unit(GRU).First,the influence of different meteorological factors on wind power generation and the degree of influence in different time periods arewere analyzed.Secondly,a radiation classification coordinate method iwas proposed to classify and select similar time periods,and reconstruct the data set(including power generation and multi-weather data)of the selected similar time period into a training data set.Then,the GRU neural network iwas used as the learning network of the model.Experimental results show that the prediction accuracy and determination coefficient of the proposed model are 97.6%and 98.99%,respectively.Combined with 3 error indicators and training time analysis,the accuracy and efficiency of the RCC-GRU model are better than the other 3 comparisons.model.