| The service-oriented manufacturing is an important developing direction of the traditional manufacturing which wants to extend and enhance the product value chains, and the fault diagnosis services are the key contents of it. In the fault diagnosis modes and fault diagnosis technologies field of general equipment and product, large number of studies were carried out, some academic and application achievements were obtained. But gear hobbing machine has very complex structure, has high failure rate, and requires high expertise to diagnose the fault, the researches of fault diagnosis mode and supporting technologies are very few although widely applications of it today. This study is to address the problem, to help the gear hobbing machine manufacturers solving the problem of valuing manufacturing and belittling service that widely exists in the industry. So, the research has very important scientific significance and application value.Aiming at the urgent demands of fault diagnosis new mode in gear hobbing machine industry, proposed a service-oriented fault diagnosis mode of gear hobbing machine. Firstly, analyzed its special fault diagnosis needs of gear hobbing machines, highly specialized knowledge dues to the complexity of structure and a large number of knowledge needs happening in different places dues to its widespread popularity; Secondly, based on the topology of the fault diagnosis grid, proposed a service-oriented fault diagnosis mode of gear hobbing machine which composed by resources layer, supporting technologies layer, application layer and users layer, which had features such as shared, collaborative, cheap, dynamic and convenience and etc; Thirdly, the implementation of the mode in grid was designed, the architecture and the services packaging methods were researched too; Finally, proposed the key supporting technologies of the mode.Put forward out the gear hobbing machine fault knowledge acquisition method based on multi-source information collection and the knowledge organization description methods based on ontology. In order to implement it, we focus our energies on the following five steps: the first step is collecting the base data according to collecting the gear hobbing machine’s status information, fault codes or manual entry the fault data; following step is treating the collected data of pre-processing and data fusion; third step is using the semi-structured knowledge description method, organizing and standardization describing the knowledge, to eliminate the fault knowledge in grammar, semantics ambiguity; forth step is obtaining the knowledge based on KDD, to extract the fault knowledge; the last step is using the modeling method of ontology, built the fault diagnosis ontology, described the fault diagnosis knowledge as XML, to solve the fault knowledge transfer and applications in grid environments.Based on the grid characteristics of distributed and hierarchical features, put forward out the fault diagnosis method named SPF with flexible components. First, in order to solve the registration, distribution and access problems in grid, we proposed a resource and services invocation method based on grid services registration; secondly, aiming at the fault diagnosis grid has the characteristics of power law, proposed the strategy of selecting the maximizing degree node as the router and resources center. About the knowledge discovery problems, used two different search methods, that is, to discover the core link using Chord ring and to search the external link using flooding method; finally, proposed a fault diagnosis method of SPF tree based on flexible component comprehending the researches above, the SPF tree constituted by some virtual branches of flexible components based on iterative algorithm, it only expand part of branches while fault searching, so it has the characteristics of high efficiency and fast speed, additional, SPF tree has the characteristics that the expression is very intuitive and visual, has the ability of learning the new knowledge automatically also.On the basis of the urgent demands of Chongqing Machine Tool Group, using the fault diagnosis mode, knowledge treating method and the grid serving technologies proposed in the paper, the prototype system of gear hobbing machine’s fault diagnosis was developed, and analyzed the trials of it, and the preliminary application proved that the system acts as a positive actions in the fault diagnosis. |
