Remote Fault Dignosis System For Injection Molding Machine

Remote fault diagnosis technology is a rising multi-disciplinary technology. Applying the remote fault diagnosis technology to mechanical equipments can not only improve the efficiency of fault diagnosis but also share diagnosis resources in wide range.At the same time, it fundamentally solves the passive situation that diagnostic staff is busy rushing. With the expanding industrial scale, the automation level of mechanical equipments has also been improved increasingly, which increases the diagnosis possibility of injection molding equipment. And due to the widely distributed users of injection molding machine, the fault diagnosis need to consume large amounts of human, financial and material resources. In order to improve the fault diagnosis efficiency of the injecting molding machine, improve the benefit of enterprises, solve the problem of the diagnosis staff lack, and reduce the investment of the manufacturers on the human, financial and material resources, a remote fault diagnosis system for injection molding is urgently needed to be developed.In this paper, firstly, the injection molding process is analyzed. And then according to the needs of the remote system, the overall framework of the remote fault diagnosis system for injection molding machine is introduced. B/S mode is chosen as the basis structure of the remote fault diagnosis system. And Visual Studio2008and SQL Server2005database are used as the tool for platform development. So that the functional development and interface design are conducted. The fault diagnosis method for injection molding machine in this paper is formed by combining phase-based PCA, wavelet analysis and case-based reasoning (CBR). In this method, phase-based PCA determines whether there is a fault by detecting the T2and SPE statistics overrun. After the wavelet analysis and reconstruction of the collected signal, fault diagnosis of wavelet packet compares the band energy value to the normal one. And then if there is a certain deviation in the same frequency band by the comparison, it can be considered that the system has a fault at that time. After the fault is detected by the phase-based PCA and wavelet analysis, case-based reasoning is used for fault diagnosis of injection molding machine. In the CBR, variable contribution features obtained by SPE variable contribution figures and the energy variation features of frequency band obtained by wavelet packet analysis are used as the diagnosis features of the injection molding machine. And fault case base is established by fully using the maintenance information, mechanism knowledge and expertise and expressed as hierarchy structure. And the nearest neighbor method is used for case retrieval. Through the above CBR, the fault diagnosis for injection molding machine can be implemented.Finally, remote fault diagnosis system is designed. System database is established and many basic functions are achieved, such as information management of the users, information management of the equipments, intelligent fault diagnosis, fault information submission, diagnosis feedback, etc. In the system, fault diagnosis includes two modes of intelligent diagnosis and expert diagnosis. In the intelligent diagnosis mode, the fault diagnosis method formed by combining the phase-based PCA, wavelet analysis and case-based reasoning is applied to the remote fault diagnosis system.