Evaluating Faults with CBR - an Expert System

by Fernando Henao, Jaime Amaya, Rodrigo Jaramillo and Farmer Monterrosa, Interconexion Electrica S.A. E.S.P. (ISA), Colombia


Thanks to Interconeción Eléctrica S.A. E.S.P ISA
for sharing their new
developments related

In Colombia, the electric power transmission companies have transmission control centers from which they remotely operate all substation equipment. When a fault occurs, transmission control center dispatchers have to analyze power system faults and restore the transmission network as soon as possible, using information from protection relays, fault recorders, sequence of events analyzers (SOE), and operational guidelines. Also, Transmission Control Center Dispatchers must rely heavily on past experiences with faults in order characterize new faults.
Expert systems are powerful tools of Artificial Intelligence (AI) that emulate human reasoning to solve problems. Therefore, this methodology was used in the project because expert systems perform similarly to a transmission control center dispatcher in real time.

Recently, expert systems have taken on relevance in some fields of electric power engineering. One of the most common fields studied has been the fault diagnosis of power systems. Many of these studies related to expert systems are used in power systems control centers to aid control center dispatchers in recognizing and characterizing faults in real time. The dispatcher´s responses are complicated by the large amount of information that must be analyzed at one time. This difficulty is called “human cognitive barrier.”

Commonly, these studies have taken into account expert systems for fault analysis power system using the signal operation from relay´s information and CB´s operation. However, in many transmission control centers, complete information is not available, or time - download information can take a long time, because this activity is done manually by the dispatcher.
ISA has been developing a project to automatize the fault diagnosis of power systems called “Diagnostico Automatico de Eventos (DAE)” since 2007. To complete this project, it was necessary to create an expert system to find the fault source.

Therefore, this article presents an expert system using Case-Based Reasoning (CBR) to evaluate and characterize faults in Transmission Lines. This CBR is represented by rules, which were created by taking only fault oscillographic records into account. These fault records come from relays or fault recorder devices installed in substations.

To avoid problems in relation to downloading the fault recorders in real time, ISA has developed an additional software to download records automatically in real time called “Sistema Automatico de Gestión de Equipos de Subestaciones” (SAGES).

This software allows the control center dispatcher to concentrate only on fault analyses and not on retrieving data systems from computer storage. The expert system rules were created based on real faults that occurred in ISA’s power transmission system. Initially, the expert system had been tested with cases of transmission lines’ faults caused by open conductors. The rules for this case were created by expert engineers with many years of experience in power system fault analysis.


Protecting your electrical assets? today and tomorrow