In multitrain subway systems, excessive rail potential and stray current (RP-SC) appear in the reflux system during the dynamic operations of over voltage protection device (OVPD) and drainage device (DD), which has not been clarified until now and has seriously affected the safe operation of the systems. To accurately evaluate the RP-SC with dynamic traction networks, the equivalent method of reflux system with mixed parameters is proposed, based on which a new simulation model that considers the dynamic characteristics of OVPD and DD is established. The control characteristics of OVPD and DD are analyzed in the established simulation model with multiple trains operating in the dynamic traction network. Dynamic RPSC are simulated in different cases, and are validated in the field tests at Guangzhou Metro Line 2. Results show that the RP-SC with dynamic traction networks can be effectively evaluated by the proposed model. The RP-SC are greatly influenced by the dynamic operations of OVPD and DD, and excessive RP-SC can be mitigated by adjusting the grounding resistances of the closed OVPD.

Effective machinery prognostics and health management play a crucial role in ensuring the safe and continuous operation of equipment, and satisfactory characteristics' expression of machine health status plays a key role in the ability to diagnose faults with high accuracy. At present, most methods based on signal processing and the shallow learning model rely on artificial feature extraction to identify the machine fault type. In practical applications, however, meaningful health management requires correct recognition of not only the health type but also the fault degree, if any occurs. Such recognition is useful for determining the priority level of mechanical maintenance and minimizing economic losses. Deep learning techniques, such as deep belief network (DBN), have demonstrated great potential in exploring characteristic information from machine status signals. In this paper, an end-to-end fault diagnosis model based on an adaptive DBN optimized by the Nesterov moment (NM) is proposed to extract deep representative features from rotating machinery and recognize bearing fault types and degrees simultaneously. Frequency-domain signals are inputted into the model for feature learning, and NM is introduced to the training process of the DBN model. Individual adaptive learning rate algorithms are then applied to optimize parameter updating. The performance of the proposed method is validated using a self-made bearing fault test platform, and the model is shown to achieve satisfactory convergence and a testing accuracy higher than those obtained from standard DBN and support vector machine.

Currently, the excessive reflux safety parameters widely exist in the multitrain subway systems, which cause a serious damage to the rail-side equipment and passengers. Nevertheless, the corresponding mechanism has not been clearly clarified. This paper elucidates the effect of crossing power restraint on reflux safety parameters. First, the calculation method of the power allocation in multitrain subway systems is proposed to analyze the crossing power among trains and traction substations. Second, a decomposition method of reflux safety parameters is proposed based on the power allocation to evaluate the impact of crossing power on reflux safety parameters. In the simulation and field tests based on Guangzhou Metro Line 2, the variation of the power distribution with multitrain is analyzed, and the effect of crossing power restraint on reflux safety parameters and power loss is clarified with the dynamic operation of multitrain. The results show that the restraint of crossing power can effectively realize the safe operation and energy saving of the system.

The dc traction power system is commonly adopted in multitrain subway systems. During the operation of multiple trains, different power distribution of the system directly affects the energy consumption and power supply safety. It is urgent for multitrain subway systems to optimize the operation parameters to ensure energy-saving and safety. In this article, nondominated sorting genetic algorithm II (NSGA-II) based on kernel density estimation (KDE) is proposed for collaborative optimization of total energy consumption and the maximum rail potential in the system. First, a simulation model for the dynamic operation of multiple trains with a parallel multiconductor traction network is established to realize the dynamic flow calculation with train–network coupling. Second, a dual-objective optimization model of the maximum rail potential and the total energy consumption is formulated to realize the energy-saving and power supply safety of the system. Finally, the NSGA-II algorithm based on KDE is proposed to handle the dual-objective optimization model. According to the dynamic simulations of Guangzhou Metro Line 2, the optimization method is verified. Results show that the total energy consumption is reduced by 22.2%, and the maximum rail potential is dropped by 40.3%, which effectively realizes the energy-saving and power supply safety.

Running rails used as a return conductor and ungrounded scheme have been widely adopted in DC traction power systems. Due to the longitudinal resistance of running rails and insulation resistance of rail-to-ground, there will be a potential rise between running rails and the ground when return current flows through the running rails, which is known as rail potential. At present, abnormal rise of rail potential exists widely in DC traction power systems. The present rail potential model still cannot simulate and explain the abnormal rail potential in the system. Based on the analysis of power distribution with multiple trains running in multiple sections, a dynamic simulation model of rail potential in the whole line is proposed. The dynamic distribution of rail potential and stray current in DC traction power systems when multiple trains run in multiple sections is analyzed, and the impact of traction current distribution on rail potential is evaluated. Simulation results show that the abnormal rise of rail potential during the dynamic operation of the system can be evaluated effectively.

Currently, power supply accidents and excessive energy consumption occur frequently in DC traction power systems of multitrain subway lines. Rail potential (RP) exceeds the standard limit frequently, energy consumption and fluctuation of the traction current are high, which are directly affected by the traction substation converter characteristic and train timetable. In this paper, multi-objective optimization of traction substation converter characteristic and train timetable is conducted to realize the safe and economic operation of subway systems. Firstly, a power flow calculation platform with multitrain running dynamically is established to work out the operating parameters of the system. Then, a multi-objective optimization model consisting of maximum RP, energy consumption and maximum traction current is designed. For the coordination of multiple goals to be optimized, an enhanced Non-Dominated Sorting Genetic Algorithm-III (NSGA-III) is proposed to obtain an optimal traction substation converter characteristic and train timetable. Finally, the multi-objective optimization is conducted with the data of Guangzhou Metro Line 2. Results indicate that three optimized objectives are dropped by 43.8%, 33.4% and 40.3% respectively based on the proposed optimization method, which will greatly enhance the energy conservation and power supply safety of subway systems.