Feature Extraction In Complex Mechanical System Fault Information Based On Topological Dynamical System

With the development of modern industry, the rapid development of science andtechnology, and the improvement of the automation, the structure of mechanicalsystem become more and more complicated. So the requirements for complexmechanical system fault diagnosis are increasingly high. While the feature extractionof fault information is an important part of the system fault diagnosis. Domestic andforeign scholars study a variety of feature extraction methods to solve some problemsof the complex mechanical system fault diagnosis. And there are many problemsworthy of discussion and study. In this paper, an extraction problem of faultinformation for the complex mechanical system has been researched, which is basedon the topology and topological dynamical system. The detail content is as follows:First, on the basis of the theory of the topology and topological dynamical system,the topological space of the state characteristic information for a complex mechanicalsystem is built, as well as the topological dynamical system model is constructed, andthe description for the implementation of the system feature is realized by the method,which has laid the foundation for the system failure information feature extraction.Second, the characteristics of the system are analyzed by researching the trackproperties of the topological dynamical system. Therefore, this method can realize toextract the system feature. In order to describe the track information of the systemthrough the symbol sequence, the track of the system should be symbolized. Thevibration state of the system is recognized by analyzing the complexity of the symbolsequence and calculating the symbolic entropy, which is the topological entropy,namely the system fault feature, of the topological dynamical system. The extractionmethod of the fault information characteristics for the complex mechanical system isbased on the topological dynamical system. It is applied to the extraction issues of thefault information characteristics for the rotary machinery system, which can be provedto be feasible by analog simulation.Finally, a large of measured signal information can be obtained through the fault diagnosis experiment on rotation experiment platform. The realization of the effectiverecognition for the different system operation state, which is received by extractingand analyzing the measured signal feature, can verify the effectiveness of theextraction method of the fault information characteristics, which is based on thetopological dynamical system.