Research On The Theories And Technologies Of Reconfigurable Remote Fault Diagnosis System

The rapid expand of firm service region, with the shortening of service time demanded by consumer and the increase of morden manufacture device complexity, calls for new approving service measures to maintain customer satisfaction. Internet-based romote monitoring and fault diagnosing technology is an available approach to meet the requirements. Aiming to enhance the openness and versatility of remote fault diagnosis system (FDS) and starting with its software structure analysis, the cooperation method among diagnosis resources is given and some technologies of software and diagnosis resources reuse are studied. A novel structure model of remote FDS and its reconfiguration modes are presented. A reconfigurable architecture of remote FDS, which includes a web-based service platform managing and scheduling the diagnosis resources and devices, is given. With the platform, the heterogeneity of resources and devices can be shielded and specific FDS can be customized to meet requirements of different applications. A remote FDS can be described as a dynamic system in three-dimensions that are diagnosis resource, device and time. Its three-layer function model is described and its working model, which consists of multiple resources locateded in different places and serving for the fault device cooperatively, is detailed also. Using two-layer reconstruction of remote FDS, static reconfiguration and dynamic reconfiguration, the diagnosis resource share and software module reuse can be actualized. Accordingly, the diagnosis task will be decomposed and carried out in two layers. Reconfigurable remote FDS can be implemented in four levers as system software developing, specific FDS designing, fault diagnosing and information managing. Several reconfiguration technologies, which are the means to achieve the opening for devices and diagnosis resources of remote FDS, are studied. To realize the first layer reconstruction of FDS —static reconfiguration, a reusable structure model of diagnosis resource subsystem is presented and a novel approach to customize fault diagnosis resource subsystem based on software intelligent chip (SIC) is discussed. With this approach, different diagnosis resource subsystem can be designed and the configuration of their SICs can be specified to serve for various types of devices rapidly. Remote monitoring subsystem is an indispensable part of remote FDS, so its reconfiguration is an indispensable part of FDS static reconfiguration. To improve the flexibility and reusability of monitoring system, a data-centered architecture is proposed. Some key technologies to implement the remote monitoring system reconstruction platform are discussed. The steps to design specific monitoring system and its working process are described. To realize the second layer reconstruction of FDS —dynamic reconfiguration, the description models of diagnosis task, diagnosis resource and the collaborative FDS are given. And a reduction algorithm of choosing appropriate diagnosis resources to satisfy the diagnosis task is elaborated. Then, the collaboration mode among diagnosis resources and its realization method are presented. In addition, two data processing methods of fault diagnosis are studied. The learning vector quantization neural network has been integrated successfully with the case-based reasoning approach to reduce the case indexing space and to enhance the indexing efficiency. And a reduction algorithm based on decision-making capability has been applied to acquire concise rule knowledge by reducing the fault symptoms effectively. With VC++ language, a remote monitoring reconfiguration software system is developed with which various specific monitoring systems for different specific applications can be customized. And a prototype system of remote FDS reconfiguration platform is accomplished with which the diagnosis resource subsystem can be implemented using SIC-based software module reconfiguration method and the collaborative FDS can be constructed dynamically using diagnosis resources recomposition method.