摘要
对时间序列的长程相关性用于设备故障识别进行了可行性研究。提出将非趋势波动分析方法用于设备故障识别,对齿轮振动信号进行分类实验,结果表明非趋势波动分析能成功地对常见齿轮故障模式进行识别,与盒维数计算结果相比,非趋势波动分析具有更好的抗干扰性。
The feasibili .ty analysis of equipment fault recognition using long- range correlation of time series is discussed and the detrend fluctuation analysis is used in equipment fault recognition. A classification experiment to vibration signals of gear was done. The result shows that the general gear fault modes can be identified by detrend fluctuation analysis. The noise - resistance characters using detrend fluctuation analysis is better than using box dimension.
出处
《煤矿机械》
北大核心
2008年第7期202-204,共3页
Coal Mine Machinery
基金
国家自然科学基金资助(50705069)
关键词
非趋势波动分析
长程相关性
故障识别
detrend fluctuation analysis
long - range correlation
fault recognition
