With the continuous development of locomotive technology, locomotive system and the on-board equipment have become more complex. Because of the increasing diversification of the type of locomotive failures as well as the fault reasons, the fault diagnosis system is put forward higher requirements. Fault diagnosis system, also shows a dynamic, distributed, multi-sector collaboration and uncertainty characteristics. Agent can make use of many parallel distributed processing technology and modular design idea; put a complex system into relatively independent subsystems Agent, by cooperation and competition between Agents to complete the diagnosis of complex problem solving. At the same time, taking into account the area of some fault has accumulated a wealth of fault cases; take the CBR (Case Based Reasoning) method will greatly improve the efficiency of case searching to improve the fault-solving ability. This paper studies for CBR-based fault diagnosis system design implementation on the Multi Agent platform.Firstly, this paper reviewed the history of development of the fault diagnosis system, provides an overview of Distributed Fault Diagnosis System Research. Then, introduced the basic theories of the expert system, CBR methods, multi-Agent technology .Based on these theories , proposed the model of CBR-based fault diagnosis system on multi-Agent platform, and given a detailed analysis of the coordination mechanism between the Multi Agent. For the case matching process, proposed matching in accordance with the fault level , thereby reducing the scope and improving the efficiency of the case matching .Using the nearest neighbor matching strategy , improves the accuracy of case matching. Finally, through the JADE platform, achieve the CBR-based fault diagnosis system on multi Agent platform. The Combination of the CBR method and the agent technology, making the system has the properties of collective intelligence, better robustness and fault tolerance. |