Reseatch On CNC Machine Tool Wear State Diagnosis Technology Based On Cloud Theory And SVM

CNC machine tool failures are common problems that seriously affect the quality and accuracy of the workpiece in the CNC machining process. Tool failure resulting in downtime will also affect the processing efficiency. At present the popularity of CNC machine tools extensively used in the of domestic SMEs generally do not have the intelligent diagnostic capabilities of the tool wear state. To improve the quality of processing and the efficiency of the production process, it is urgent to monitor tool wear during CNC machine processing for such enterprises.The main objective of this paper is to solve the intelligent diagnosis of tool wear state in such enterprises. In the research subject prior to this paper has completed the work of gathering and processing the tool monitoring data, and on its basis the paper continues to focus on the problems and analysis and research work related to them was launched.In the different small and medium-sized enterprises, the number of CNC machine tools needed to be monitored are quite different. Based on the application requirements and characteristics of the application for CNC machine tools in SMEs, a program that the architecture of embedded modules and server use C/S structure and the server uses distributed systems architecture is proposed. In the program the size of the distributed servers to be deployed change according to the number of CNC machine tools needed to monitor. The program has the security and stability, scalability, strong, economical and practical features.At present, intelligent diagnosis of tool wear is mostly used methods such as artificial neural network (ANN). ANN has a great relationship with the user’s experience and other factors, and easily falls into local optimal solution. Therefore, ANN is not easy to promote is not easy to promote. Combined the advantage of the cloud theory and support vector machine (SVM), cloud-SVM model was proposed. The model retains the cloud theory’s advantages that is the unity of randomness and fuzziness, but also has the advantage of support vector machine to solve high dimensional problems by introducing the kernel function cleverly and to avoid the local optimal solution, and thus the operability and replicability are more stronger. The model was verified through simulation the EUNITE electricity peak prediction task, compared with the results of SVM the accuracy of the model was better, and the model is practical.Tool wear monitoring system has a large amount of real-time data. For this feature, an integration database structure with relational database and NoSQL database is proposed. The model of the integration database of MySQL and Cassandra was Designed and the data access integration layer was described in detail.