Design Of Remote And Real-Time Fault Diagnosis System For NC Machine Tools

NC machine tool is one of the important basic equipments in the modern industrial production. It plays an critical role in manufacturing industry, aviation and aerospace industry, national defense construction. It is also influencing national economy and comprehensive national strength. And its fault will be cost and risk, so it is important to diagnosis the NC machine tool fault. NC machine tool’s fault diagnosis technology is a synthetic application technique with signal acquisition, data processing, machine learning and software engineering. It is used to find faults in time, or even predict faults. This paper, Naive Bayes classification is used to fault recognition, and a remote and real-time fault diagnosis system for NC machine tools is built based on Python and websocket.The vibration data of bearing is applied to demonstrate the proposed algorithm scheme. Zero-mean-value and wavelet packet filtering is adopted to pre-process. And the feature vector is composed of time-domain, frequency-domain and time-frequency domain features. At last, Naive Bayes classification works out the bearing condition. Python and some scientific computing kits based on python was used for these algorithm programing.With the internal characteristics of mechanical structure, the parameters of sensors and signal conditional units (SCU) are chosen and set properly. Those hardware instruments are the first step of the whole fault diagnosis system.Software of the system was divided into three sections:data collector, server and client. Python is the main language to complete the fault diagnosis system due to its advantage of availability, simple and sufficiently rich third-party libraries. At the stage of data collector, VC++2010 is extended to improve the performance. Collector and server is based on the C/S framework, using TCP to transfer data and communicate. The acquired data is stored in MAT files. Other data is stored in database MariaDB. However, client and server works on the B/S framework. Tornado framework creates the web server and website, then user can access the website by browsers. Finally, in the front end of website, websocket in HTML5 enables process large amounts of data for real-time communication. In addition, VSG, Bootstrap, JQuery and Echarts is used for use interface friendly.Practical experiments show that the system running steadily, it can show the NC machine tool’s status in time. And it can meet the requirements of industrial application.