The operational reliability of rail vehicle pantograph systems is evaluated by transforming T-S multistate fault trees into dynamic Bayesian networks (DBNs), which take into account system multistability, long-lasting operation, dynamic failure, and maintenance recovery. The T-S multistate fault tree structure is constructed by the content validity ratio and content validity index; the T-S gate rule expressing causal uncertainty is constructed by using fuzzy theory and dependent uncertain ordered weighted averaging expert scoring, and finally, the pantograph T-S multistate fault tree is transformed into a DBN model characterizing the dynamic interaction and time dependence of the system. The dynamic evolution laws of reliability of a pantograph system in maintenance and maintenance-free states over time are inferred, compared and analyzed. The results show that the system availability of a pantograph system decreases continuously during 720 days of operation. The system availability without maintenance decreases to 0.881, and the system availability with maintenance is 0.952. The reliability of a pantograph system can be effectively ensured with maintenance during the operation period; the sensitivity analysis is performed by changing the failure rate of the equipment to 120% or 80%; the fall indicator, the electrical control box, and the elevating bow motor are the weak links in the system, and the impact of fault escalation on the reliability of a pantograph system is analyzed. It is then verified that the system reliability can be further improved by using a preventive maintenance strategy, and the steady-state reliability can be gradually reached, which is about 0.9968, providing a reference for the maintenance of a pantograph system.

As a key component of rail vehicle operation, the running condition of the wheelset significantly affects the operational safety of track vehicles. The wheel diameter, flange thickness, and flange height are key dimensional parameters of the wheelset, which directly influence the correct position of wheelsets on the track, and the train needs to be continuously monitored during the passenger operation. A prediction model for the key parameters of the wheelset is established based on LSTM (long short-term memory) neural network, and real measured data of wheelsets from the Shanghai Metro vehicles are selected. The predicted results of the model are compared and analyzed, and the results show that the LSTM-based prediction model for key parameters of wheelsets performs well, with the mean absolute percentage errors (MAPEs) for wheel diameter, flange thickness, and flange height being 0.08%, 0.42%, and 0.44%, respectively, for the left wheel and 0.07%, 0.35%, and 0.44%, respectively, for the right wheel. The prediction model for the train wheelset parameters established in this paper has a good prediction accuracy. By predicting the key parameters of the wheelset, the faults and causes of the wheelset can be found and determined, which is helpful for engineers to overhaul the wheelset faults, make maintenance plans, and perform preventive maintenance.

Brake systems are subjected to various factors such as wear and fatigue over a long period of time. They bring a great challenge to the reliability analysis of the braking system. In this paper, the hyperellipsoid model is used and the hyperellipsoid dynamic Bayesian network model is constructed. The model synthesizes the strengths of the dynamic Bayesian network with the precise constraint capability of the superellipsoid model, offering a novel methodology for the reliability analysis of the braking system. In order to verify the validity of the super ellipsoidal dynamic Bayesian network model, this paper applies it to the reliability analysis of the braking system and compares it with the traditional interval Bayesian network. The results show that the hyperellipsoidal dynamic Bayesian network model enhances the range accuracy of the probability intervals by 54.06%. This solves the conservatism problem of interval Bayesian networks. This suggests that the method improves the accuracy of brake system reliability analysis.

Based on the current research status of fault prediction in rail transit reliability, this paper proposes a cumulative failure rate prediction method for key components of subway vehicles based on the RF-CNN-LSTM model. The article describes the prediction method based on cumulative failure rate data and takes the subway EDCU as an example of cumulative failure rate prediction. Three models, ARIMA, MLP, and LSTM, are introduced and compared with the RF–CNN–LSTM model by R² and adjusted R² index. The results show that the RF–CNN–LSTM model can predict the failure rate of the underground door controller well, with accuracy rates of 99.78% and 97.88%. Based on the prediction results, the cumulative failure rate of the EDCU peaks in about 10 years at 4.5% and 10.6%, respectively; the maintenance strategies can be adjusted through the actual situation of the EDCU to reduce maintenance costs and optimize maintenance plans.

The reliability of traction systems in urban rail transit vehicles is critical to safe and efficient operations. However, existing reliability assessment methods face challenges such as state space explosion, computational complexity, and difficulties in modeling fault interdependencies and imperfect maintenance. This paper proposes an integrated framework that combines dynamic fault trees with fault dependency models and Monte Carlo simulation. The method captures fault dependencies through functional dependency gates incorporating fault impact factors, models component degradation using a two-parameter Weibull distribution, and accounts for imperfect maintenance through a age reduction factor. The simulation efficiency of three random number generators—LCG, MT, and PCG64—was compared. A Monte Carlo simulation of 100,000 trials under maintenance-free conditions yielded a system MTBF of 29.9 months. A mode significance analysis identified the pantograph control unit and the traction control unit as the most critical components. Under targeted maintenance based on component criticality, the MTBF increased to 39.77 months—a 33% improvement—while the peak failure rate was maintained at approximately 3%. The MT generator exhibited the fastest convergence, achieving stability within 1% after 3000 iterations. This framework provides a practical foundation for optimizing preventive maintenance strategies for urban rail transit systems, and sensitivity analysis confirmed the robustness of weak link identification to parameter variations. This method is applicable to other complex systems with interdependent failures and multiple maintenance schedules.