An Accident Prediction and Diagnosis Method Based on Bayesian Network

Abstract:

Against the limitations of some traditional accident prediction and diagnosis methods, a new method based on Bayesian network is proposed on the basis of the Fault Tree Analysis method. Using this method, accident prediction and diagnosis can become easily, quickly and more accurately. The method includes mainly two steps: the first step, to establish Bayesian network topology; the second step, to determine conditional probability table for each node. In addition, to update the marginal probability of the leaf nodes, in order to get a more objective evaluation, changing the past methods which mainly rely on statistics or expert experiences to give the probability value, it proposes an evaluation method using Bayesian network to combine different expert opinions effectively. Finally, taking the gas explosion accident as an example, using GeNIe Software to model and reason, taking analysis of the experimental data, we strive to prove the validity of the model.

• Periodical: Applied Mechanics and Materials (Volumes 513-517)
• Pages: 1581-1585
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.513-517.1581
• Online since: February 2014
• Authors: Lin Shi*, Zhi Gang Li
• Keywords: Accident Diagnosis, Accident Prediction, Bayesian Network (BN), Conditional Probability Table, Genie

References

[1] Wang Hui, Shao Weihang, Analysis on mine gas risk by Fault Tree and its prevention measures[J]", Shanxi Meitan, 2011, 30(3)，pp.22-24.

[2] Ma Xiaodong, Lu Xu, Application of Disaster Tree Analysis Method in Gas Explosion Disaster [J], Coal Mining Technology, 2006, 10, 11(5), pp.70-72.

[3] Wang Changshen, Sun Yajun, Hang yuan, Application of Fault Tree Analysis to Risk Assessment of Potential Water-inrush Hazards in Coal Mining, Chinese Journal of Rock Mechanics and Engineering, 2009, 02, 28(5), pp.298-305.

[4] Dong Dingwen, Study on Coal Mine Intrinsic Safety Management Information System [D]", Xi'An: Xi, An University of Science and Technology, (2010).

[5] Liu Wensheng, Zeng Fengzhang, Applicatio n of Bayesian Networks in Safety Assessment of Coal Mine Production System[J], Industry and Mine Automation, 2008(1), pp.1-4.

[6] Cai Linqin, Multi-Agent Based Virtual Environment of Coal Mine [D], HeFei: University of Science and Technology of China, (2007).

[7] Zhou Zhongbao, Research on Methods and Application of Probabilistic Safety Assessment Based on Bayesian Networks [D], Changsha: National University of Defense Technology, (2006).

[8] Ding Songbin, Ru Yi, Airlines Flight Safety System Assessment Method Based on Bayesian Networks [J], Journal of Civil Aviation University of China, 2009, 06, 27(3).