论文部分内容阅读
以往开发的绝大多数故障诊断算法基于数据的平稳性假设,没有考虑机械某一运转周期内的时间相关细节特征。本文针对特定对象强调了非平稳模型用于信号分析的必要性,讨论了模型的时变,耐不变算法及相应的特征提取、工况判断过程。本文引入正交变换,一方面实现了数据的大规模压缩,另一方面完成了代表正常工况的母体模型的建立;其次,借助于模式识别理论的相似性判据得到对多个特征定量监测的方法。一关于柱塞泵振动监测的实例说明了方法的应用过程。
The vast majority of fault diagnosis algorithms developed in the past are based on the assumption of data stability and do not consider the time-related details of a machine’s operating cycle. This paper emphasizes the necessity of non-stationary model for signal analysis for specific objects, and discusses the time-varying, invariant algorithm of the model and the corresponding feature extraction, working condition judgment process. In this paper, the orthogonal transform is introduced to realize the large-scale data compression, on the other hand, the establishment of the parent model that represents the normal working conditions is completed. Secondly, by using the similarity criterion of pattern recognition theory, Methods. An example of piston pump vibration monitoring illustrates the application of the method.