论文部分内容阅读
提出了主元和线性判别的集成分析算法以实施模拟故障数据的特征提取过程和方法.该集成分析方法首先对模拟故障数据进行主元分析,然后在主元变换空间实行线性判别分析,最后将所获得的最优判别特征模式应用于模式分类器进行故障诊断.仿真结果表明,所提出的方法能够充分利用线性方法的计算简便优势,增强单一主元分析或线性判别分析的特征提取性能,获取故障数据集的本质特征,简化模式分类器的结构,降低系统运行的计算成本.
An integrated analysis algorithm of principal component and linear discriminant is proposed to implement the feature extraction process and method of simulated fault data. The integrated analysis method first performs principal component analysis on the simulated fault data and then performs linear discriminant analysis in the principal component transform space. Finally, The obtained optimal discriminant feature pattern is applied to the pattern classifier for fault diagnosis.The simulation results show that the proposed method can make full use of the computational simple advantage of linear method and enhance the feature extraction performance of single principal component analysis or linear discriminant analysis, The essential features of fault data sets simplify the structure of pattern classifiers and reduce the computational cost of system operation.