Predictive Maintenance and Incipient Failure Detection


Project Scope:

Due to financial constraints being placed on power utility companies throughout the country, traditional preventive maintenance operations are being dismissed and utilities are becoming more failure driven. The failure of equipment in power distribution systems can have direct or indirect impact on the reliable delivery of power. Also certain failures can be catastrophic when they result in loss of service. For example, the failure of line pole equipment (bushings, insulators, arresters) can stress other parts of the system which decreases the overall reliability of the system and quality of power delivered to customers. The utility industry is greatly interested in low-cost, automated, real-time approaches which can assess the integrity of the system equipment or predict maintenance needs.

project62This research project involves the design and development of analytical signal processing techniques, new heuristic and probabilistic techniques, and an expert system method which will perform incipient fault detection and predictive maintenance for distribution systems. This system will reduce unscheduled down-time, increase quality up-time, and decrease overall maintenance costs. The incipient failure detection will focus on detecting and identifying incipient information extracted from substation measurements which indicate equipment deterioration. Various signal processing techniques will be investigated to determine the appropriate subset of information necessary to effectively detect and identify the incipient behavior.

Techniques will be developed to predict the remaining useful life of three types of equipment, transformers, insulators, and surge arresters. Mathematical techniques utilizing approximate and detailed models, advanced signal processing techniques, and acoustic, thermal and electrical monitoring will be employed to develop the transformer remaining life prediction scheme. Insulator and surge arrester prediction techniques will be developed which utilize signal processing techniques to analyze various equipment-related measurements for determination of remaining life.


Principal Investigator:                                     Dr. Karen L. Butler-Purry


Sponsor:                                                                                         Project Duration:

CAREER Award                                                                                 1996-2000

National Science Foundation

Research Assistants:

Peter Palmer                                                                 Hang Wang

Adedayo Kuforji                                                           Harini Sundaresan