基于BP神经网络和多因素权重法的风电机组载荷预测和分析

Load prediction and analysis of wind turbine based on BP neural network and multi-factor weight method

基于BP神经网络,建立了风电机组关键部位载荷的快速准确预测方法。以风电机组关键参数风速、空气密度、湍流强度、入流角、风切变、偏航误差角等为自变量,以机组关键部位载荷作为输出变量,建立用于快速预测机组关键部位载荷的BP神经网络模型;然后基于多因素权重法对风电机组不同参数的影响权重进行分析,获得影响风电机组载荷的关键变量。结果显示:基于叶素-动量理论模型计算得到不同风况下风电机组关键部位载荷,然后设置合理的神经网络结构以及合适的神经网络参数,可以实现对不同风况下风电机组关键部位载荷的预测;对于叶根和塔底的平均载荷和极限载荷4个不同的变量,风速、空气密度、湍流强度、入流角、风切变、偏航误差角等参数影响的权重各不相同。

Based on BP neural network,a quick accurate prediction method for load of key parts of wind turbine is proposed.By taking key parameters of the wind power units such as wind speed,air density,turbulence intensity,inflow angle,wind shear and yaw error angle as independent input variables,andthe loads on key locations of the unit as output variables,a BP neural network for rapid load prediction is established.Moreover,based on multi-factor weight analysis,the influence weights of different wind power unit parameters are quantitatively analyzed to obtain the key variables affecting the load of wind turbine at specific site.The results show that,the loads of key parts of the wind turbine under different wind conditions,which are calculated based on BEM theoretical model,can be predicted by by setting reasonable neural network structure and appropriate neural network parameters.For different variables,including theblade root mean load,blade root ultimate load,tower bottom mean load and tower bottom ultimate load,the influence weights of parameters such as wind speed,air density,turbulence intensity,inflow angle,wind shear and yaw error angle are different.