Expert System for Analysis of Electric Power System Harmonics, 1992 Engineering Conference Proceedings

The proliferation of both power electronics devices that cause harmonic voltages and currents along with the wide-spread use of harmonics-sensitive electronic equipment is creating a near-epidemic of "power quality" issues that are being addressed by both users and suppliers of electric energy. Numerous types, styles, and capabilities exist in the field of instrumentation and diagnostic equipment for the detection, recording, and analysis of the distorted waveform. The critical shortage is in experienced power system engineers who can interpret the data, diagnose the problems, and recommend practical solutions.

This paper reports on one effort to employ artificial intelligence (AI) in the form of a Harmonics Analysis Expert System (HAES). The resulting PC-based software program has captured the knowledge of expert power system engineers in a "rule base" which can be applied to information about the topology, power signature, operating practices, symptoms, equipment type and ratings, and chronology of system development to diagnose the power system for harmonics-caused problems. The goal is to accomplish the analysis with a less experienced engineer thereby multiplying the available nationwide pool of expertise. Additionally, repeated use of the system does result in impartation of knowledge from the system to the user.

The process of applying this AI approach involves: (1) training the user, (2) measurements (optional), (3) diagnosis, and (4) solution of the power system for harmonics disturbances to acceptable power quality. The emphasis of this paper will be on the diagnostic module (AI program) with a usage example presented.