Technical gazette, Vol. 33 No. 1, 2026.
Original scientific paper
https://doi.org/10.17559/TV-20240824001939
A Bayesian Network-Based Risk Assessment Model for Gas Pipeline Intelligent Management Systems
ChunXiao Mei
; Hebei Gas Co., Ltd, Shijiazhuang, 050011, China
*
JianXin Tan
; Hebei Gas Co., Ltd, Shijiazhuang, 050011, China
Hao Li
; Hebei Gas Co., Ltd, Shijiazhuang, 050011, China
JingTao Chang
; Hebei Gas Co., Ltd, Shijiazhuang, 050011, China
YiFan Wang
; Hebei Gas Co., Ltd, Shijiazhuang, 050011, China
YiWei Lu
; Hebei Gas Co., Ltd, Shijiazhuang, 050011, China
* Corresponding author.
Abstract
To enhance the accuracy and efficiency of gas pipeline risk assessment within intelligent management systems, this study proposes a Bayesian network-based data modeling and risk assessment framework. Traditional risk assessment methods rely heavily on statistical analysis and expert judgment, often struggling with uncertainty and interdependencies between risk factors. In contrast, Bayesian networks effectively model complex probabilistic relationships, providing a more dynamic and adaptive risk evaluation approach. This study integrates a cloud-based uncertainty processing model with Bayesian inference to improve risk prediction. Experimental validation using real-world gas pipeline monitoring data demonstrates that the proposed method achieves a 98.8% accuracy rate, significantly outperforming conventional techniques. Additionally, the assessment period is reduced to approximately two months, enhancing real-time decision-making capabilities. These findings highlight the potential of Bayesian networks to transform gas pipeline safety management by offering precise, efficient, and adaptive risk assessment models. Future research will explore integrating real-time IoT sensor networks and machine learning-based anomaly detection to further optimize predictive capabilities.
Keywords
accident prediction; bayesian networks; gas pipeline; intelligent management; risk assessment
Hrčak ID:
342656
URI
Publication date:
31.12.2025.
Visits: 558 *