摘要
针对复杂机械故障的模式分类问题,提出一种基于非线性判别的多故障分类方法。与线性判别分析相比,基于核的判别分析更适于处理线性不可分的分类问题。分析了基于核的判别分析方法与核函数主元分析方法之间的联系与差异,指出了两者不同的应用背景,核函数主元分析适于检测机械设备异常状态的出现,而基于核的判别分析则适于在积累历史故障征兆基础上对多种机械故障进行分类识别。将上述方法应用于风机工作状态的分类识别与齿轮故障模式分类,结果表明该方法对于多种复杂的故障模式分类具有出色表现。
To deal with pattern classification of complicated mechanical faults, an approach to multi-faults classification based on non-linear discriminant analysis is presented. Compared with linear discriminant analysis (LDA), kernel-based discriminant analysis (KDA) is more suitable for classifying the linear non-separable problem. By analyzing the connection and difference of KPCA (Kernel Principal Component Analysis) with KDA, the different application background is pointed out. KPCA is good at detecting machine abnormality while KDA performs well in multi-faults classification based on the database of historical faults symptoms. When the above method is applied to air compressor condition classification and gear fault classification, an excellent performance in complicated multi-faults classification is presented.
出处
《振动工程学报》
EI
CSCD
北大核心
2005年第2期133-138,共6页
Journal of Vibration Engineering
基金
国家自然科学基金资助项目(50475095)
广东省自然科学资金资助项目(04020082)
振动
冲击与噪声国家重点实验室开放基金资助项目(VSN-2004-03)
关键词
故障诊断
分类方法
非线性判别
核函数
风机
Classification (of information)
Compressors
Failure (mechanical)
Gears
Nonlinear systems
Operations research
