Rail Train Door System Hidden Danger Identification Based on Extended Time and Probability Petri Net

摘要

This paper proposes a hidden danger inference approach to solve the rail train door fault predicting problem. Based on Extended Time and Probability Petri Net (ETPPN) the door action process model is built and the fault reasoning process is realized by TC-PPN which is the backwards inference of ETPPN. The time interval of every event occurrence and the history statistical probability are two information aspects for the search evidence of hidden fault. In terms of time interval, the hidden fault could be insulated with most other hidden faults, and the probability sequence could give the confidence level to search the relative hidden faults. And based on the train door system physical constitution, the hidden fault could be searched by TC-PPN step and the result satisfies the real maintenance strategy indeed. Overall, this algorithm based on ETPPN could give the accurate result for train door system hidden trouble identification, and is valuable to apply in train maintenance engineering.