Web-based Bearing Fault Diagnosis System Development

As a rolling bearing is one of the most widely used parts in the machinery, which possesses high efficiency, lower friction, convenient assemblage and lubricating easily advantages. Therefore rolling bearing is frequently used as a key part in rotating machines. At present, most of rolling bearing fault diagnosis systems can be only used for simple diagnosis. It is difficult to satisfy requirement of accurate, intelligent diagnosis and development of enterprise's information. Developing a diagnosis system which can find out fault timely and accurately to avoid loss is both inevitable and necessary. On the background of the project of Science and Technology Foundation of KMUST, The Development of Bearing Fault Diagnosis System Based on Web, The research work of this dissertation is carried out. Surrounding the central subject, Bearing Fault Diagnosis Prototype System Based on Web, profound researches have been done in correlative areas. Moreover, an expert system-BFDS has been designed and implemented, based on these researches.In this dissertation, the research work begins with fault diagnosing problems of rolling bearing. Based on constructing involved concepts of equipment fault diagnosis, trend in development of rolling bearing fault-diagnosing has been expounded; Then founded on the description of fundamental form of the bearing abnormity, its diagnostic methods have been introduced respectively and explained vibration diagnosis detailedly. On the background of vibratory mechanism of rolling bearing, formulations of the fault characteristic frequency and parameters of time-domain and frequency-domain have been given.Acquirement of fault symptoms is the base of fault diagnosis. According to types of characteristic signals, extracting method of rolling bearing symptom has been studied in this dissertation. Specially, resonance demodulation method of symptom extraction for rolling bearing has been researched. Meanwhile, discusses implementation of resonance demodulation in code.To the problems of traditional Expert System and Neural Networks, which are incapable of learning, weak reasoning, inefficient, slow training speed and etc. A kind of integrated intelligent diagnosis system is introduced in this dissertation. The system thinks of combination of production rule and neural networks technology. It is propitious to overcome the defects which correlate with traditional Expert System and Neural Networks. To possess remote collaboration and adapt for need of enterprise informationization, a diagnosing model based remote collaboration is presented. Then, the diagnosing mode is designed using NetMeeting Component.Design and implementation of BFDS has been done in the final part of the dissertation, where the OOP thought has been permeated in whole the designing process. Based on requirements, developing environment of software has been selected accordingly. Foundations of BFDS have been marked out detail implementations of every functional module have been carried out, including explanation of key technologies. Finally, a Software of Experiment Data Acquirement Based on LabView has been developed, which is used to validate BFDS.