摘要
探索和尝试以轴心轨迹为依据 ,以全等度、相似度和扩散度三个特征参数为输入的神经网络捕获设备故障信号模型 ,可大大提高故障症兆提取的效率和质量 。
In order to greatly improve the efficiency and quaulity of extracting fault symp- tom,a neural network model forcapturing fault signals is developed by using three character- istic parameters defined in the paper as the input of the network,i.e.the equality,similarity and expansivity of the shaft centre orbit of a machine compared to the orbit of the shaft aquired under fault- free conditions.A practical application of the network to a monitoring system of a large turbin- compressor is made.
出处
《振动.测试与诊断》
EI
CSCD
1998年第4期252-255,共4页
Journal of Vibration,Measurement & Diagnosis
基金
河南省自然科学基金资助项目!(编号 :974 0 4 0 50 0 )
关键词
智能监测
神经网络
故障症兆
故障诊断
monitoring intelligently neural network fault diagnosis fault symptom