提出了基于行为的多代理故障诊断方法,给出了系统的实现策略。与传统的人工智能方法不同,基于行为的人工智能方法采用了“自下而上”的推理方法,首先,对诊断对象的故障行为进行分解,然后,在多代理技术的基础上,动态构建故障诊断系统,对不同部件的故障行为分而治之,并且根据部件故障行为的不同特点采用各异的诊断方法,增强了诊断系统的稳定性,提高了系统的诊断精度。 A multi-agent fault diagnosis method based on behavior is proposed and the strategy of system implementation is given. Different from the traditional artificial intelligence method, the behavior-based artificial intelligence method uses a “bottom-up” reasoning method. First, the fault behavior of the diagnostic object is decomposed. Then, based on the multi-agent technology, the fault is dynamically built The diagnosis system divides and rules the fault behaviors of different components, and adopts different diagnosis methods according to the different characteristics of component fault behaviors, which enhances the stability of the diagnosis system and improves the diagnostic accuracy of the system.