An on-line intelligent alarm analyzer for power systems based on temporal constraint network

The volume of alarm messages could be very large in modern large-scale power systems. Many intelligent methods have been developed for alarm processing in order to provide summarized and synthesized information instead of a flood of raw alarm data. The timestamps of alarms represent the temporal relationship among event occurrences. However, the temporal information has not yet been appropriately utilized in traditional intelligent methods. A temporal constraint network (TCN) is a kind of directed acyclic graph (DAG) which is of promise for the representation of temporal logics. Based on TCN, an on-line intelligent alarm analyzer with the temporal information of alarms taken into account is proposed for on-line operational environment. The advanced intelligent alarm analyzer is able to infer what events cause the reported alarms and to estimate when these events occurred, as well as to identify the abnormal or missing alarms. Finally, a case study is served for demonstrating the feasibility and efficiency of the proposed method.