基于BDI智能体的配电网监控信息批处理方法  

A Method for Batch Processing of Distribution Network Monitoring Information Based on Belief-Desire-Intention Agents

作　　者：屈志坚[1] 刘明光[1] 刘靖[2] 王健[3] 杨罡[1] 王林[3] 

机构地区：[1]北京交通大学电气工程学院,北京市海淀区100044 [2]国家能源快堆工程研发实验中心(中国原子能科学研究院),北京市房山区102413 [3]华东交通大学电气工程学院,江西省南昌市330013

出　　处：《电网技术》2012年第3期1-8,共8页Power System Technology

基　　金：国家自然科学基金资助项目(51007023);江西省教育厅科技基金(GJJ09223;GJJ10138)的资助~~

摘　　要：针对配电网监控中量测点多、变化快,造成关键信息传输产生较大延迟的问题,提出了批量信息的信念–愿望–意图(belief-desire-intention,BDI)智能处理方法。通过建立BDI智能体的五元组模型,对多智能体进行交互分析,构建了公用信息的智能体联盟,综合运用模型自动映射和消息事件触发(message event trigger,MET)的技巧,从业务流和信息流角度进行智能化原型系统实现。利用10 kV配电网监控的工程算例进行了交互与性能实验,结果表明:文中方法的处理效率比传统2 s轮询方法提高了约70%。In allusion to the long time-delay during the transmission of key information due to numerous measuring points in distribution network monitoring and quick variation of measured information, an intelligent belief-desire-intention （BDI） batch processing method is proposed. By means of establishing a quintuple model of BDI agents, the interactive analysis on the agents are performed and an agent alliance of common information is constructed; then synthetically applying the skills in model auto-mapping and message event trigger （MET）, the prototype of the intelligent system is implemented in the viewpoint of business stream and information stream. Taking the monitoring system for 10kV distribution network as example, the interaction and performance tests are carried out, and test results show that the processing efficiency of the proposed method is potentiated by 70% than traditional 2s polling method.

关 键 词：智能监控技术 多智能体系统 公用信息模型 映 射技术 消息事件触发 

分 类 号：TM73[电气工程—电力系统及自动化]