An online intelligent alarm-processing system based on abductive reasoning network

Large amounts of alarms coming into control center when faults occur have harassed operators for a long time, especially when there are false alarms, missing alarms or multiple faults. An online intelligent alarm-processing system based on abductive reasoning network (ARN) is developed to deal with the alarms automatically. Models and real-time data of the power system are collected from both supervisory control and data acquisition system / energy management system (SCADA/EMS) and fault information system, including primary and secondary models. First, when alarms arrive, rules reflecting cause-effect relationship between faults and alarms and between alarms themselves are established automatically, and the whole abductive reasoning network is constructed. Then, all the coming alarms are located on the network and traced back to the roots. Third, possible diagnoses are sifted and picked out according to their false alarm rate and missing alarm rate. At last, the diagnosis results are presented to the operators in several friendly manners. The system has been applied to the district control center of Jiaxing, China and is functioning well.