The Design And Implement Of Fault Diagnosis System Based On Embedded System

Emergency communication system is a necessary communication system which put into operation to protect rescue, relief when large human errors or natural disasters occurred. Its requirement of system reliability and safety performance is very high. However, the emergency communication system is a vulnerable system, and the system application environment can not be expected, so that its system performance may be affected by adverse external environment, strong interference and other factors. All these factors cause that the fault incidence of hardware, software and links of the emergency communication system maintain a high level. In addition, the isomerization of the network and the diversification of its equipment and the comprehensive of its services and the non-linear features of system failure all make the diagnosis of the system fault more difficult. At present, the fault diagnosis system of emergency communication system faces with many problems, including low reliability, fault symptom acquisition mode being single, relying to the monitoring center and the objects monitored too much, and taking up the hardware and software, bandwidth and other resources of monitoring centers and the objects monitored heavily.To solve the problem mentioned above, fault diagnosis system based on embedded system is adopted in this thesis to replace the relevant function host, with which the system reliability and flexibility are improved as in the case of no using the resource of the relevant function host. In addition, the system hardware is designed in accordance with PCI standard, and the system with such features as power supply facilitate and diversity, easy installation and replacement, and applicable to field operations and so on. The results of this thesis include the following aspects:(1) Based on studying the characteristics of emergency communication system failure and combined with hierarchical and distributed multi-domain fault diagnosis architecture, the fault diagnosis system based on embedded system is adopted. The system hardware platform is designed in accordance with PCI standard, the CPU of which is Intel XScale of 404MHZ frequency and software platform is embedded Linux operating system. Through the system, it is played not only the advantages of embedded system such as high reliability, not to take up system resources and so on, but also the benefits of the hierarchical and distributed multi-domain fault diagnosis architecture. It is improved the resource-sharing capabilities, interoperability, real-time and survivability, and overall performance of the fault diagnosis system.(2) In order to improve the intelligence, scalability and diagnosis rates of the diagnosis system, expert system is adopted for fault diagnosis in this thesis which is suitable for embedded system. And the fault tree analysis is used for analyzing and mining the faults of computer software and hardware, switches, and link in emergency communication system in depth. Thus, the logic, effectiveness and correctness of the fault analysis are the enhanced. On this basis, a knowledge representation method based on object-oriented and production is adopted, which is not only beneficial to expressing the dynamic changes of object attributes but also to the realization of a parallel diagnosis. And it is established the rule base covering system hardware and software, switches, and link fault and designed the parallel inference mechanism and interactive interface of expert system in accordance with standards CLIPS strictly.(3) After reduced the CLIPS and ported to ARM-Linux environment, a multi-inference engine fault diagnosis expert system based on embedded system is developed with CLIPS by its expanding programming mechanisms. And the system is carried out in the simulation platform to validate. The experimental results show that the goals of system design and the target of improving the overall performance are achieved in system.