Bayesian network-based evaluation of transmission tower structural safety

To identify the key factors contributing to transmission tower failures and evaluate potential mitigation strategies, this study introduces a structural safety assessment framework utilizing Bayesian networks. The research begins with a systematic examination of diverse risk factors that may lead to structural damage in transmission towers, followed by the development of a three-tiered network model grounded in Bayesian causal relationships. The model’s prior and conditional probabilities are quantified through an integrated approach incorporating expert evaluations and fuzzy set theory. A connectivity tree is subsequently generated by constructing a belief map and optimizing the Build Constructive Tree (BuildCT) algorithm through element reduction. The proposed methodology employs both forward and backward inference mechanisms, leveraging real-time monitoring data to determine the probabilistic relationships between causative agents and structural consequences for maintenance purposes. Analytical results indicate that wind-induced conductor displacement presents the highest activation likelihood under certain weather scenarios, consistent with professional expertise. Furthermore, severe wind loads and ice accumulation are identified as primary contributors to conductor separation and structural deformation.