Research And Application Of Fault Diagnosis For Mechanical Parts Of Numerical Control Machines

With the development of the social manufacture, fault diagnosis for the machinery has been playing a more and more important role. In mechanical field, the Numerical Control Machine is a kind of typical electric-mechanical integrated machine, which is consisted of computer, machine tool, motor, and so on. It is so highly integrated that even an accidental failure in one of its parts can possibly cause a breakdown, which usually means a great economic loss. Therefore, the importance of fault diagnosis for NC machines is highlighted.In order to address the practical problem of the cylinder body production line in Shanghai GM Corporation, this dissertation makes an intensive study in fault diagnosis methods the Mechanical part of NC Machine and designs a support database. The main points are listed as below.To begin with, based on the generation mechanism of NC machine's vibration signals, this dissertation makes clear the intrinsic relation between vibration characteristic, responsive signals and fault characteristic in combination with related analysis methods for signals. Different diagnoses sometimes need different analysis methods.Then, Local Wave method is introduced in consideration of the non-stationarity characteristic of the vibrating signals from NC machines. This method decomposes practical signals into several intrinsic mode functions with corresponding instantaneous frequency having physical meaning, in order to separate characteristic frequencies and identify fault information effectively.Finally, aiming at the question that the bearing model can't be searched effectively in the old database of fault diagnosis system, a new one is designed with PB as background support for the fault diagnosis analysis. .The new database's availability has been tested in pratical application.In practical diagnosis, Local wave method and diagnosis database are combined effectively. In this way, the fault of NC machines can be correctly identified. Therefore, the methods of NC machine diagnosis studied in this dissertation are proved to be effective.