The Fault Prediction Technology Research Of CNC Machine Based On Built-in Test

In this paper, built-in test(BIT) was used to the fault detection and prediction in terms of the fault handling situation of milling spindle and feeding system in CNC machine by changing self-diagnostic processing mode. First of all, in order to obtain key components failure mode,we adopted failure mode and effect analysis(FMEA) to the elaborate analysis of all kinds of fault happened to CNC machine, and classified the fault according to FMEA hardware method. Then we determined the severity, frequency, detection of each fault by the way of expert questionnaire.In the end, the risk priority number was divided according to the theory of fuzzy comprehensive evaluation. The secondary fuzzy evaluation model was established and mechanical failures of spindle system and feed system were analyzed quantitatively. The validity of fuzzy comprehensive evaluation and FMEA was verified for comprehensive analysis of CNC machine tool mechanical failures. Secondly,sensors used to CNC machine tool fault detection were selected and optimized by genetic algorithm,and the test model of selection and optimization was established. Finally, we collected vibration signal of the spindle bearing system and completed fault detection through adopting optimized Morlet wavelet to remove the early fault signal and strengthen the characteristic. This method in this paper not only can reduce the fault detection time but also can shorten the repair time, which can significantly contribute to reducing the technical requirements for maintenance and supply certain fault information according to the maintenance situation, raise the reliability of CNC machine performing mission, enhance the maintainability of CNC machine.