FHI: A Fault Intensity-based Hierarchical Association Analysis Model for Mining Fault Database of Railway OCS

Qian, Kaiyi; Yu, Long; Liu, Yigu

doi:10.1109/iske47853.2019.9170406

Cited by 3 publications

(1 citation statement)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To mine association rules by algorithm, the thresholds of FI and confidence are needed, which are set as (6, 80%) both for DB1 and DB2. The number of rules returned is 82 for DB1 and 66 for DB2, which are shown as the failure mode relation network [27] in Figure 6. In Figure 6, each arrow represents one association rule, and the nodes on both sides of arrow are input event and output event respectively.…”

Section: Comparison Of Fault Treementioning

confidence: 99%

Fault Tree Construction Model Based on Association Analysis for Railway Overhead Contact System

Qian¹,

Yu²,

Gao³

2020

IJCIS

Self Cite

View full text Add to dashboard Cite

The overhead contact system (OCS) is the power source of electrified railway, it is very important to evaluate the system status accurately to maintain its safe and stable operation. At present, fault tree analysis is the main method of reliability assessment for OCS. Existent methods to establish the fault tree divide the system into functional modules and require professional knowledge, which limits the failure cause and makes the result subjective. Besides, the fault tree constructed by traditional method cannot reflect the difference between the same system under different working conditions. In this paper, we applied association analysis to OCS for the first time and proposed a novel fault tree construction model based on characteristics of OCS. Firstly, hierarchical structure and method of multi-dimensional partition (MDP) is proposed to flexibly transform the failure records of OCS into transaction database, which can reduce the sparsity of data. Secondly, the failure mode intensity is defined to replace the support, which can eliminate the influence by complex layout of railway and make the mining result more reasonable. Adjustment factor and new pruning strategy are proposed for mining association rules among levels. Thirdly, the transformation from rules into fault tree structure is presented and fault tree construction model based on association rules is proposed. At last, a case study is conducted based on real failure records database of two local railway administrations of China sharing the same OCS design and the comparison between fault trees constructed by our model and traditional method verifies that the proposed model can build the adaptable fault tree in a more objective way without the limitation of modules.

show abstract

Section: Comparison Of Fault Treementioning

confidence: 99%