摘要
为了解决风电功率预测易受各种因素影响产生异常数据导致预测准确度不高的问题,提出了一种基于遗传算法-反向传播(GA-BP)神经网络的风电功率预测方法。首先,通过数学模型中的四分位算法对异常数据进行识别,并通过加入带通滤波器剔除异常数据。然后,在风电功率预测的方法上设计新型GA-BP神经网络算法,通过自检验及循环检测的方式获得准确的风电功率预测结果。试验结果表明,该方法不仅有很强的异常数据识别能力,而且在进行风电功率预测时可以保持90%以上的准确率,具有良好的数据处理稳定性。该研究大幅提升了风电功率预测的工作效率,为风电功率预测技术的进一步发展提供了技术参考。
To solve the problem that wind power prediction is susceptible to the influence of various factors that produce abnormal data leading to low prediction accuracy,a wind power prediction method based on genetic algorithm-back propagation(GA-BP)neural network is proposed.Firstly,the abnormal datas are identified by the quartile algorithm in the mathematical model,and the abnormal datas are eliminated by adding a band-pass filter.Then,a novel GA-BP neural network algorithm is designed on the method of wind power prediction,and accurate wind power prediction results are obtained by self-testing and cyclic detection.The experimental results show that the method not only has a strong ability to recognize abnormal data,but also can maintain more than 90%accuracy when performing wind power prediction with good data processing stability.This study greatly improves the efficiency of wind power prediction and provides technical reference for the further development of wind power prediction technology.
作者
逯登龙
高鹏
范丽锋
郭彦飞
周维文
LU Denglong;GAO Peng;FAN Lifeng;GUO Yanfei;ZHOU Weiwen(Huabei Branch,CGN New Energy Holdings Co.,Ltd.,Shijiazhuang 050011,China;NARI Group Co.,Ltd.(State Grid Electric Power Research Institute Co.,Ltd.),Nanjing 211106,China)
出处
《自动化仪表》
CAS
2024年第3期97-102,共6页
Process Automation Instrumentation
关键词
风电功率预测
神经网络
异常数据识别
遗传算法
反向传播
循环检测
四分位算法
Wind power prediction
Neural network
Abnormal data identification
Genetic algorithm(GA)
Back propagation(BP)
Cycle detection
Quartile algorithm
