LANGAGE: an expert system for diagnosis in a real-time context

The operators at Hydro-Quebec's nine regional control centers (RCCs) must react promptly as soon as an equipment failure occurs, and must analyze the alarm messages displayed by the real-time (SCADA) system to diagnose the fault. To alleviate this complex task, Hydro-Quebec has developed the LANGAGE expert system to perform a continuous analysis of alarm messages, automatically detect the application of the protection or restoration control, and then produce a concise, real-time diagnosis identifying the origin and consequences of the fault. Thus, diagnosis allows RCC operators to complete their analysis in a shorter period of time. The authors describe the need for these diagnoses and the application of (AI) techniques to meet this need. They also discuss the constraints and problems experienced with the use of AI techniques in a real-time context.<
>