基于小波包分解和改进Elman神经网络的风电场风速和风电功率预测

摘要

Accurate prediction of wind speed and wind power is of great significance to the operation and maintenance of wind farms,the optimal scheduling of turbines and the safe and stable operation of power grids.A new method for wind speed and wind power forecasting based on the wavelet packet decomposition theory and an improved Elman neural network was put forward,and the concrete application steps of the method was given.Wavelet packet decomposition theory is firstly adopted to decompose wind speed data into several wavelet spaces,and according to the correlation,the optimal decomposition tree is persisted and random data are rejected.Then a new PSO training algorithm with disturbance is proposed to improve the training speed of neural networks and deal with the drawback of easily falling into local optimal solution of PSO.Finally,Elman neural networks with different structures are established and used to find the laws of wind speed in different frequency bands,wind speed and wind power prediction results are hence received.The forecasting results based on the wind speed data of a wind farm in south China show that the proposed method has higher forecasting accuracy and is able to reflect the laws of wind speed and wind power correctly.%准确预测风电场风速和风电功率对做好风电场运行维护、合理安排开停机计划以及确保电力系统的安全稳定运行具有重要意义.提出了基于小波包分解和改进Elman神经网络的新型风电场风速和风电功率预测方法并给出了具体应用步骤.首先利用小波包分解理论对经过初步处理的历史风速数据进行分解处理,根据相关性剔除随机数据,保留最优分解树;随后提出带扰动的PSO训练算法用以提高Elman神经网络的训练速度,并解决PSO算法易陷入局部最优解的问题;最后利用不同结构的Elman神经网络寻找最优分解树不同频段下的风速规律进而获得风速和风电功率预测结果.南方某风电场算例表明该方法具有更高的预测精度,能够正确反映风速和风电功率规律.