摘要
提出一种基于多Agent的变压器故障诊断模型,该模型包含3个诊断Agent、1个管理Agent和1个融合Agent。各诊断Agent的建立以NB、SB和TAN 3种贝叶斯分类器算法为基础,以所获取的变压器油中溶解气体数据为依据。由管理Agent调节和控制,达到各诊断Agent协商诊断的效果。由融合Agent根据管理Agent对诊断Agent的控制情况、诊断Agent发送的诊断结果、各个诊断Agent的诊断概率和诊断Agent发送结果的次数等因素进行融合,给出最终的变压器故障类型。实例验证了该模型的高效性。
A diagnosis model based on multi-Agent system is proposed for transformer faults,including three diagnosis Agents,a management Agent and a fusion Agent. According to the obtained data of dissolved gas in transformer oil,the diagnosis Agent is established based on three Bayesian classifier algorithms：NB,SB and TAN. These diagnosis Agents are regulated and controlled by the management Agent to achieve a coordinated diagnosis effect. The fusion Agent will determine the final transformer fault type by fusing the control situation of management Agent on diagnosis Agents,the diagnosis results of all diagnosis Agents and the diagnosis probability and number of each diagnosis-Agent. Case study verifies the high efficiency of the proposed model.
出处
《电力自动化设备》
EI
CSCD
北大核心
2011年第1期23-27,共5页
Electric Power Automation Equipment
基金
河北省自然科学基金项目(E2009001392)
中央高校基本科研业务费专项资金资助项目~~
