Scholarlink

<p>This research introduces an integrated approach that merges Fault Tree Analysis (FTA) with Bayesian Networks (BNs) to perform probabilistic failure assessment and degradation analysis of BB1-type centrifugal pumps employed in the oil sector. Initially, a detailed fault tree was constructed, and preliminary failure probabilities for key events were derived using empirical evidence and expert insights. This fault tree was subsequently converted into a BN, facilitating a more dynamic and accurate evaluation of failure scenarios and their underlying causes. In addition, degradation modeling was applied to track pump wear patterns under real-world operating conditions, examining how condition monitoring influences the precision of failure predictions. Sensitivity analysis highlighted misalignment, sealing system failures, and lubrication deficiencies as the primary factors impacting pump reliability. Overall, the finding confirms that the proposed hybrid method substantially improves the accuracy of failure forecasts and contributes to the formulation of more effective predictive maintenance plans.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A hybrid fault tree analysis and Bayesian network framework for probabilistic failure analysis and degradation modeling in BB1 centrifugal pumps

  • 期刊论文

  • 发表日期：

    2026年7月29日

查看全文
  • Mohammadreza Barouzeh,
  • Saeed Adibnazari,
  • Adel Maghsoudpour,
  • Shahram Etemadi Haghighi

摘要

This research introduces an integrated approach that merges Fault Tree Analysis (FTA) with Bayesian Networks (BNs) to perform probabilistic failure assessment and degradation analysis of BB1-type centrifugal pumps employed in the oil sector. Initially, a detailed fault tree was constructed, and preliminary failure probabilities for key events were derived using empirical evidence and expert insights. This fault tree was subsequently converted into a BN, facilitating a more dynamic and accurate evaluation of failure scenarios and their underlying causes. In addition, degradation modeling was applied to track pump wear patterns under real-world operating conditions, examining how condition monitoring influences the precision of failure predictions. Sensitivity analysis highlighted misalignment, sealing system failures, and lubrication deficiencies as the primary factors impacting pump reliability. Overall, the finding confirms that the proposed hybrid method substantially improves the accuracy of failure forecasts and contributes to the formulation of more effective predictive maintenance plans.