基于多Agent及Petri网的变压器故障诊断系统

将多Agent技术引入变压器故障综合诊断系统，采用模糊Petri网技术实现建模．该模型中每个Agent代表一种智能知识或智能处理方法，利用Petri网进行知识表示及推理的矩阵运算算法以描述故障征兆与故障的关系，进而达到多Agent之间的协同工作，并利用协调综合Agent给出诊断结果．运用所提出的诊断系统，不但可以充分利用各种故障信息资源，而且可对采用不同诊断方法时的诊断结果进行协调综合，并且在故障信息不全面的情况下也可以做出较准确的诊断结论．实例运算也验证了该诊断系统的快速性、协调性和准确性．

Fault Diagnostic System for Power Transformer Based on Multi-Agent and Petri Net

Multi-agent technique was introduced into synthetic fault diagnostic system for power transformer, and a multi-agent model based on Petri net was established, where each agent served as an intelligent knowledge body or intelligent disposal method. By means of Petri net, the matrix operation algorithm of the knowledge representation and inference was presented, the relationship between the symptom and the fault was described, and the cooperation work of multi-agent was fulfilled. Thus the diagnostic result was deduced by the coordination and synthesis agent. The presented system is effective in coordination and synthesis among multi diagnostic results, especially for the case of inadequate fault information. The efficiency, harmonization and accuracy of the presented diagnostic system are verified by some practical examples.