Statistical approaches for analysis of failure data in power cables

Chengke Zhou, Xiaoping Jing, Zeyang Tang, Wei Jiang, Babakalli Alkali, Wenjun Zhou, Jianhui Yu

    Research output: Contribution to conferencePaperpeer-review

    Abstract

    Power cables are subject to electrical, thermal, mechanical, and environmental stresses on a constant basis when in service. These stresses, and occasionally problems resulting from inadequate installation and maintenance practices, lead to insulation degradation or defects, for example, excessive sheath circulating current, partial discharge activity and increasing dielectric loss can result in premature cable failure causing unplanned outages. The aim of this paper is to analyse data in connection with cable failures, model the failure trend and identify the stress factors leading to cable defects or degradation. This is to support asset managers to optimally manage and utilise the cable assets.
    Two statistical models are considered here: the Weibull model and the Crow-AMSAA model have been contrasted when fitted to cable failure data collected from the high voltage cable network in Wuhan City, China. The paper shows that the Weibull distribution, which models the age-to-failure data against the cable population, provides a more accurate account of the failure trend when the data
    sample is small. The paper also identifies the use of Proportional Hazard Model (PHM) for cable failure data analysis when further operational or condition monitoring data are available. The use of PHM can help to quantify the influencing factors leading to cable defects or accelerated insulation degradation. Results obtained from using the PHM could provide adequate information that could support asset managers make optimum decisions on cable maintenance and replacement times.
    Original languageEnglish
    Number of pages9
    Publication statusPublished - Aug 2012

    Keywords

    • cable
    • failure
    • statistics
    • power cable
    • cable failure mechanism
    • ageing
    • insulation
    • life estimation

    Fingerprint

    Dive into the research topics of 'Statistical approaches for analysis of failure data in power cables'. Together they form a unique fingerprint.

    Cite this