Characterization of underground cable incipient failures from on-line monitoring of underground distribution power systems

Abstract

With the imminent re-regulation of the power industry and investor owned utilities becoming more cost conscious, the need for an on-line, non-destructive, incipient cable fault detection system is prevalent. With such an incipient fault detection system, utilities could schedule service on the failing cable and minimize the outages their customers would experience due to power failure. The long-term objective of this research is to develop an on-line, non-destructive, fault detection technique that could detect incipient cable failures in underground power distribution systems. As part of this research, this thesis presents: 1. The development of experimental setups to generate an extensive data library from:a.on-line monitoring of cable in TU Electric's underground power distribution systems; b. controlled experiments on field aged cable samples and on new intentionally damaged distribution cable sections at the TAMU Downed Conductor Test Facility. 2. Analysis of data recorded at TAMU and TU Electric. 3. The characterization of recorded data from on-line monitoring of underground distribution cable systems and staged cable experiments. 4. A comparison of the data from the TAMU cable experiments and data collected from TU Electric underground power systems. Analysis of field aged cable samples tested on various days at TAMU showed similarities in abnormal activities. Comparable characteristics were observed in the TAMU and TU Electric data, but with differing levels of magnitude and duration. Preliminary conclusions were presented for incipient underground cable behavior.