Research And Application On Fault Diagnosis For Headstock Of Numerical Control Machine

In recent years, numerical control machine has got important status and predomination in manufacturing of China. Numerical control machine has grown as key and deciding factor in many plants. Without timely fault diagnosis and service, serious economic loss can be caused. According to the Stat, about 60 percent of machinery fault in numerical control machine is fault of headstock. Therefore, headstock is the key segment in fault finding field, and the fault diagnosis on the headstock is becoming more and more important.In this paper, the research is mainly carried out on fault diagnosis on the headstock of numerical control machine, which is extracted from the project of Shanghai GM Ltd. and Institution of Vibration Engineering of Dalian University of Technology. It can be divided into several segments listed below:First, the mass antilogy of fault diagnosis to headstock of numerical control machine is expressed in the paper. Likewise, collect the data from local equipments by data collecting instruments, utilize the database of fault diagnosis and software of PDM2000 system, then the total fault diagnosis process on headstock of numerical control machine is realized.Then the time domain analysis methods and frequency domain analysis methods which are frequently used for fault diagnosis have been discussed. Principles of the methods, advantage as well as disadvantage also been introduced. On the basis of the methods above, the method of time-frequency analysis: local wave method is introduced. Compared with ordinary frequency analysis's limitations, the fundament of local wave method is disserted, including instantaneous frequency, local wave components, local wave decomposition, the time-frequency demonstration of local wave and local wave boundary spectrum based on local wave. At last, the feasibility of the local wave method applied on the headstock of numerical control machine's fault diagnosis has been discussed after the emulation experiment.The basis of methods above is integrated data of equipment parameter and character, so a self-contained and complete database of fault diagnosis is mostly necessary. In this paper, the research flat for the database, key techniques such as data pipeline and data windows, and numerical control resource is worked out, while the construction of the database is particularly expressed.Finally, the two methods of traditional and local wave are applied in a headstock of numerical control machine in some workshops in Shanghai GM Ltd. respectively, and the diagnosis examples are presented in the paper. The examples not only prove that the local wave method can diagnose the headstock of numerical control machine effectively and cut the shortage of traditional method, but also testify the practicability and validity of fault diagnosis database.