Data-Driven Time Dependent Reliability Prediction Approach for Overhead Contact Lines Subjected to Imperfect Maintenance, Common Cause Failure and Lightning Strike

Authors:
Publisher website
Google Scholar
Abstract
To ensure the high-reliability operation of overhead contact lines (OCLs), this paper proposes a data-driven time dependent reliability prediction approach, incorporating the maintenance factor, common cause failure (CCF), and lightning strike into reliability modeling and prediction for multi-component OCLs. Firstly, a new lightning-related failure probability calculation method is developed, including probabilistic lightning modeling, lightning identification, tracking and association analysis (LITAA), lightning movement trajectory prediction, and electric geometric model-based failure rate calculation. Moreover, a Bayesian network-based reliability model is embedded to integrate historical maintenance information, predictive lightning-related failure rate and fault tree model into reliability assessment of OCLs, which effectively captures the dependencies between failure events and depicts the impact of CCF on OCL components. In this context, the reliability of OCL components and OCLs can be predicted, with update of lightning strike data. Finally, the actual maintenance and lightning-related failure records are implemented to validate the effectiveness of the proposed approach. The result analysis demonstrates that our approach can not only illustrate how the reliability can be predicted using the predictive lightning-related failure rate, simultaneously accounting for imperfect maintenance and CCF, but also enhance risk prevention against lightning strike and faulty OCL components.
Funding Information
  • National Natural Science Foundation of China (52177115)
  • National Key R&D Program of China (2021YFB2601500)
  • Natural Science Foundation of Shandong Province (ZR2023QF036)
  • Qilu University of Technology (2023RCKY054, 2023PX081)
  • National Natural Science Foundation of China (52107017)

This publication has 23 references indexed in Scilit: