摘要
针对磨损监测过程中获得的大量参数之间存在冗余及关联影响自动识别这一问题,首先运用粗糙集理论和主元分析2种不同的数据约简方法对监测数据进行约简,然后采用支持向量机建立滑动轴承磨粒信息和磨损表面信息之间的映射关系识别器。应用示例表明建立的模型对识别滑动轴承的磨损表面信息和磨粒信息映射关系具有较好的效果。
Among the information on wear debris and worn surfaces during wear condition monitoring, many parameters are redundant and correlative which influences the implement of automatic recognition. In order to solve this issue, rough sets and principal components analysis （PCA） was firstly applied to reduce the amount of attributes of the information of wear debris and worn surfaces. Support vector machine （SVM） was then adopted to seek the mapping relationship recognizer between wear particles and worn surface information. The application example demonstrates that the developed recognizer is feasible to obtain the mapping relationship between worn surface features and wear debris information in sliding bearings.
出处
《武汉理工大学学报》
CAS
CSCD
北大核心
2009年第12期123-126,共4页
Journal of Wuhan University of Technology
基金
教育部博士点新教师项目(20070497029)
关键词
滑动轴承
磨损
磨粒
粗糙集
主元分析
支持向量机
sliding bearings
wear
wear debris
rough sets
principal components analysis
support vector machine
