Study On The Methods Of Two Channels Signal-fusion Analysis To Diagnose Bolt Looseness

Bolt connections with the properties of strong bearing, easy maintenance and replacement is the most important mechanical joints, it is the most extensive form. It is one of key components of the mechanical system which can affect the system working properly. Because of the disassembly of screw bolt, bolt connection structure looseness sometimes occurs. Bolt connections looseness in mechanical system can affect system work normally, shorten service life and even cause casualties.Bolt connection looseness related many factors, such as the number of bolts, the position of bolt, the diameter of bolt and so on. To establish a general mathematical model to reflect the real bolt junction characteristics is impractical. Bolt looseness fault in vibration environment is studied in this article by the means of signal analysis without dynamic model. Use some common signal processing methods in time-domain and frequency-domain to study the vibration characters of bolt connection structure, in order to find the most suitable method to diagnose and identify looseness.According to the characteristics of bolt connection structure, some methods to identify and diagnose looseness which integrate two legs vibration signal of the bolt connection structure was put forward in this paper. One method identifies bolt connection looseness by displacement fusion curve of two signals under different conditions. One method identifies bolt connection looseness by a model of binaural hearing. Another recognition method is based on principal component analysis. These three methods can easily diagnose looseness, especially the method which is based on principal component analysis. The method which is based on principal component analysis can realizes intelligent recognition.