A fault diagnostic system based on Petri nets and gray relational analysis for train–ground wireless communication systems

Abstract

The system structure of train–ground wireless communication systems (TWCSs) is extremely complicated due to the use of fault tolerant technology to improve their performance. This complex structure raises several challenges in fault diagnosis for TWCSs, such as epistemic uncertainty, dynamic fault behaviors, and common cause failure (CCF). A fault diagnostic system is proposed to deal with these challenges based on Petri nets and gray relational analysis in this paper. Specifically, the fuzzy analytic hierarchy process is used to evaluate the failure data of components to handle epistemic uncertainty. Furthermore, the dynamic fault tree of TWCSs is established and converted into a generalized stochastic Petri net to calculate several reliability parameters used for fault diagnosis. Besides, a β factor model is employed to resolve the problem of CCF in TWCSs. In addition, Birnbaum importance measure (BIM), risk achievement worth (RAW) and test cost are considered comprehensively to obtain the optimal diagnostic sequence using an improved gray relational analysis. Finally, a numerical example is presented to demonstrate the efficiency of the proposed fault diagnostic system.

A reliability analysis technique for complex systems with retaining fuzzy information

“…The conversion from dynamic and static gates to corresponding GSPN structures can be found in our previous work. 39,40 Step2: Based on the interconnection between the gates in the DFT model, the 𝐺𝐷 is constructed by linking all the subnets 𝑆 using directed arcs. Step3: For the parts of the system failure behaviors that cannot be modeled by DFT, the GSPN model 𝐺𝐶 = (𝑃 𝐺𝐶 , 𝑇 𝐺𝐶 , 𝐴 𝐺𝐶 , 𝑀 0,𝐺𝐶 , 𝑅 𝐺𝐶 ) is construct directly.…”

“…The basic steps of combinational modeling method are summarized as follows: Each static or dynamic gate $$S \in D$$ and its corresponding GSPN structure are represented as a subnet in $$GD$$, denoted by $$S = (P_S, T_S, A_S, M_{0,S}, R_S)$$. The conversion from dynamic and static gates to corresponding GSPN structures can be found in our previous work 39,40 Based on the interconnection between the gates in the DFT model, the $$GD$$ is constructed by linking all the subnets S using directed arcs. For the parts of the system failure behaviors that cannot be modeled by DFT, the GSPN model $$GC = (P_{GC}, T_{GC}, A_{GC}, M_{0,GC}, R_{GC})$$ is construct directly.…”

A reliability modeling method for the system subject to common cause failures and competing failures

“…Generalized stochastic Petri net (GSPN) is an extended Petri net (PN), and could model the stochastic dependencies between the components in the complex system. 13,14 Besides, GSPN inherits the advantages of PN and has a detailed description of system status changes, which is suitable for the competing failure modeling. In this paper, GSPN is adopted as a modeling method to evaluate the effect of competing failure in the system.…”

Performance Evaluation of Complex Equipment Considering Resume Information

“…The following research on equipment design and management considers resume information, but there is no quantitative test data involved in the performance evaluation and the testability of the test data itself is still taken as an index. The performance evaluation method in light of the qualitative knowledge is mainly based on FMEA (failure mode and effects analysis) [ 21 ], fault tree analysis [ 22 ], finite element analysis [ 23 ], Petri net [ 24 ] and AHP (analytic hierarchy process).…”