基于NMF的SVM故障诊断方法

SVM fault diagnosis method based on NMF

针对大维数系统故障诊断中存在特征提取困难和识别率低的问题,提出基于非负矩阵分解(NMF,Non-negative Matrix Factorization)的支持向量机(SVM,Support Vector Machine)诊断方法,避免了直接对故障特征的选择和提取,实现特征降维,提高故障模式分类的准确性和速度;对于NMF中的结果随机性问题,提出用前次分解所得系数矩阵求解样本降维特征矩阵的方法,保证多次NMF分解尺度一致.实验表明该方法能对故障特征有效降维,并具有较高的诊断效率和故障识别率.

For overcoming the difficulty of fault feature extraction and solving the low efficiency of fault feature classification in a large dimensions fault diagnosis system,an algorithm of support vector machine（SVM）based on non-negative matrix factorization（NMF）fault diagnosis was researched.It is to avoid the direct feature selection and extraction,to reduce the characteristic dimension,and improve the high-dimensional data feature mode classification speed and accuracy.In order to avoid NMF randomness,characteristics of fault samples dimensionality reduction by training samples coefficient matrix was calculated,so that the consistency of the scale of NMF decomposition times was ensured.The experiment shows that this algorithm can reduce the dimensions of fault feature.The method can enhance the running efficiency and the estimating accuracy.