摘要
传统的风力预测模型中有线性预测模型和非线性预测模型,在具有强非线性记录数据的局部风力预测中,绝大部分的非线性预测模型比线性预测模型预测准确度更高.对于短时间内风力大小是非线性、随机性和难以准确预测的特点,该文运用一种基于长短期记忆(long short-term memory neural network,LSTM)循环神经网络的短期局部风速预报技术预测风力.首先建立以LSTM神经网络为基础的短期局部风速预测模型,然后采用TensorFlow深度学习平台进行模型参数调试,在此基础上,结合华东某局部风电场的历史数据作为输入,对模型进行训练和测试.研究结果表明,LSTM循环神经网络预测风速与实际的风速吻合较好,预测效果较好,并且深层神经网络具有强大的拟合能力,在数据预测方面有很强的应用性.
There are linear prediction models and nonlinear prediction models in traditional wind prediction models.In local wind prediction with strong nonlinear recorded data,the vast majority of nonlinear prediction models have higher prediction accuracy than linear prediction models.In this paper,we apply a short-term local wind speed prediction technique based on long short-term memory(LSTM)recurrent neural network for the characteristics that wind speed is nonlinear,random and difficult to predict accurately in a short period of time.Firstly,a short-term local wind speed prediction model based on LSTM neural network is established,and then the TensorFlow deep learning platform is used to debug the model parameters.On this basis,the model is trained and tested by combining the historical data of a local wind power plant in East China as input.The research results show that the LSTM recurrent neural network predicts wind speed in good agreement with the actual wind speed,and the prediction effect is good,and the deep neural network has powerful fitting ability,which has strong applicability in data prediction.
作者
熊龙祥
涂佳黄
廖惠惠
XIONG Longxiang;TU Jiahuang;LIAO Huihui(School of Civil Engineering,Xiangtan University,Xiangtan 411105,China)
出处
《湘潭大学学报（自然科学版）》
CAS
2023年第4期118-128,共11页
Journal of Xiangtan University(Natural Science Edition)
基金
湖南省自然科学基金(2021JJ50027)
湖南省教育厅科学研究项目(21A0103)。
关键词
短期局部风速预测
循环神经网络
长短期记忆网络
深度学习
short-term local wind speed prediction
recurrent neural network
long short-term memory network
deep learning
