【作者】 张鸽； 高金玉； 孙海蓉；

【Author】 ZHANG Ge;GAO Jin-yu;SUN Hai-rong;Hebei Technology Innovation Center of Simulation & Optimized Control for Power Generation,North China Electric Power University;

【机构】 河北省发电过程仿真与优化控制技术创新中心(华北电力大学)；

【摘要】 从风电机组状态评价的角度出发,提出一种齿轮箱油温组合预测模型。首先利用互信息技术对参数之间的相关性进行分析,选取最优的神经网络输入向量;其次,分别建立了基于反向传播神经网络和非线性自回归神经网络的单一预测模型;最后,选取两种单一模型的最优组合权值,建立了齿轮箱油温的组合预测模型。根据实际采集的数据进行仿真验证,为防止数据过拟合现象,随机选取机组1和机组2的数据对模型分别进行验证。结果表明,所提出的组合预测模型具有较高的预测精度,同时也为风电机组的状态评价奠定了基础。更多还原

【Abstract】 From the point of view of wind turbine condition evaluation,a combined algorithm for predicting oil temperature of gearbox is proposed.Firstly,the correlation between parameters is analyzed by mutual information technology,and the optimal input vector of neural network is selected.Secondly,the prediction models based on dynamic neural network and back propagation neural network are established respectively.The weights of two single models are selected,and the combination model of the two models is established.Finally,the model is validated according to the actual data collected.In order to prevent data over-fitting,the data of unit 1 and unit 2 are randomly selected to validate the model.The results show that the proposed combined forecasting model has high forecasting accuracy and lays a foundation for the state evaluation of wind turbines.更多还原