基于支持向量机和顺序前项选择算法的PNN风电机组液压变桨的故障诊断

针对液压变桨距系统的强耦合、非线性,以及液压变桨距故障发生原因复杂、故障单一造成的定位问题,该文提出基于支持向量机和顺序前项选择算法的概率神经网络诊断方法。首先,选取SCADA数据的特征值为输入,桨距角为输出,利用支持向量机进行模型的回归,得出桨距角输出的预测值;接着,将测量值与预测值带入顺序前项选择算法,挖掘和发现特征与故障之间的关系,评估各特征之间的重要性,并选出最好的一组特征集合;最后,建立变桨距概率诊断模型,将所选的数据送到故障诊断模型进行训练,再用所选数据进行测试,定位出变桨距系统的故障原因。实验分析表明:基于支持向量机和顺序前项选择算法的概率神经网络液压变桨距故障诊断方法可以有效地分辨出不同故障,并且诊断的精确度得到了提高。

Fault Diagnosis of PNN Hydraulic Variable Pitch Based on SVM and Sequential Selection Algorithm

In view of the strong coupling and nonlinearity of hydraulic variable pitch system,as well as the location problem caused by the complex and single fault of hydraulic variable pitch fault,this paper proposes a probabilistic neural network diagnosis method based on support vector machine and sequential selection algorithm.Firstly,the characteristic values of SCADA data were selected as input,and the pitch Angle as output.Then,the model regression was carried out using support vector machine,and the predicted value of pitch Angle output was obtained.Then,the measured value and predicted value are brought into the selection algorithm in the foregoing order,the relationship between features and faults is mined and discovered,the importance of each feature is evaluated,and the best set of features is selected.Finally,a variable pitch probabilistic neural network diagnosis model is established,and the selected data is sent to the fault diagnosis model for training,and then the selected data is tested to locate the fault cause of the variable pitch system.Experimental analysis shows that the probabilistic neural network fault diagnosis method based on support vector machine and sequential selection algorithm can effectively distinguish different faults,and the accuracy of diagnosis is improved.