Research And Realization On Remote Collaboration Fault Diagnosis System Based On Grid

Due to its degree of vessel large-sized, automatic, high-speed constantly improved, complex equipments usually need multiple inspect measures and diagnosis methods, which can form integrated and merged diagnosis strategy and system. So we can reduce the uncertainty in the fault diagnosis. Correspondingly, tekecommunications, collaboration and intelligence has become an important development direction and inevitable trend in modern fault diagnosis field. RCFD (Remote Collaboration Fault Diagnosis) is an important technical measures to guarantee the stability running of complex instrument. Now, diagnosis method has transform into using intelligent technical constructing ES (Expert System) from traditional data measuring and experiment estimating. Because there is little relative complex fault sample and much complex fault situation, single intelligent diagnosis system can not meet the requirement of fault diagnosis, which impact the implement of complex instrument situation inspecting. So it is necessary to establish integrated intelligent distribute grid fault diagnosis system. How to adapt to this change of diagnosis model, it has not established the completed integrated frame and applied system in the fault diagnosis field.Grid is becoming the basic information supporting architecture platform for next generation scientific research, Grid Service Platform, as the better carrier to realize resource sharing and collaboration work, provides a new way to realize RCFD, embodies a wide area distribute resource sharing diagnosis idea of the combination of focus manager and distribute compute of RCFD system, realizes the fault diagnosis requirement of rapid response client.RCFD based on grid realize society resource sharing and integrating, which support integrated method of group collaboration running and management, is an important complement and development of RCFD theory. Finally, it provides the theory basis and method supporting for grid technique applying in RCFD. This dissertation belongs to new research field of multi-subject crossing and multi-technique integrated, it can widen grid technique application area in our homeland and accelerate multi-subject crossing and integration by implying this subject. This dissertation, combined The National Natural Science Fund Project, conducted a comprehensive study in theory, system architecture, resource modeling, task management and resource schedule, knowledge requirement and representation method and so on of RCDF, whose mainly content and fruit as follows:1) The composition, characteristics and advantages of the grid-based remote collaboration fault diagnosis are analyzed by using the principles of intelligent diagnosis, remote diagnosis and grid-related. Appropriate diagnosis of remote collaboration grid architecture is established and its working principle, system architecture and functional structure is described so that the system has the intelligence, openness, scalability and strong features as well as provides a computing platform for more complicated application of fault diagnosis in future.2) The diagnosis resources service package model framework based on the XML Schema and WSRF is introducesd and the fault diagnosis resources are classified. Then, useing the UML Unified Modeling Language, the diagnosis resources are modeled. And the diagnosis resources template to generate metadata is established, using XML Schema standard to describe the diagnosis resources. Eventually the fault diagnosis resources is packaged into web services-resources structure of grid resources services based on WSRF, which shields the complexity and heterogeneous nature of fault diagnosis resources, and supports resources package level for solving the sharing of resources and collaborations in grid environment.3) The functional structure and workflow of the diagnosis task management and resource scheduling subsystem is analyzed. the market economy principles is introduced to make a resource management model. And the model structure, step timing, the transaction process of fault diagnosis resources, as well as the corresponding resource scheduling algorithm are given, the.simulating experiments are done to verify the rationality of model.4) The object-oriented approach based on metadata is chosen to represent the diagnosis knowledge. The improved virtual method is usesd to integrate the metadata for heterogeneous, and the unity of diagnosis knowledge is realized. On this basis, according to the forms and distribution of knowledge in remote collaborative fault diagnosis system, the knowledge map of the fault diagnosis based on metadata is used to manage the diagnosis knowledge and a unified method to describe the knowledge map is put forward. And the process of building a knowledge map is explained through the establishment of mathematical models.5) Finally, in the case of automatic super finishing line for ball bearing ring grinding, the requirement analysis for the goals and tasks of remote collaborative fault diagnosis are based. As the foregoing sections are the main contents of the study, the role and the function structure of the prototype system is established, with an emphasis on the portlet in which the analysis and design of components are carried out.