With the rapid advancement of urbanization, the operating conditions of rail transit vehicles have become increasingly complex and stochastic. Consequently, conventional energy management strategies based on fixed control rules often fail to achieve optimal performance. This study proposes an adaptive energy management strategy for hydrogen-powered hybrid systems in rail transit, informed by real-time recognition of operating condition features. The proposed method classifies the diverse load conditions into three representative modes—low-speed braking, economic cruising, and high-speed operation—based on distinctive dynamic characteristics. A hierarchical and coupled global sensitivity analysis framework, incorporating an online particle swarm optimization algorithm, is developed to determine the optimal multi-objective weighting under varying conditions. This enables real-time adaptive multi-objective energy management tailored to dynamic scenarios. Comparative results demonstrate that, relative to fixed weight strategies, the proposed method significantly reduces fluctuations in the state of charge (SOC) of the lithium battery while maintaining comparable fuel cell efficiency and hydrogen consumption. Moreover, all three key performance metrics exhibit substantial improvements compared to conventional ECMS.

The growing scarcity of resources and the prevalence of environmental contamination has led to increased demand for energy trading and transportation. While integrated energy system (IES), as a crucial component in multi-energy coupling and energy conservation and emission reduction, are confronted with significant challenges. In particular, hydrogen energy systems face difficulties in temporal and spatial coupling, wherein pricing mechanisms remain decoupled from geographical delivery constraints and transportation logistics, and supply-demand optimization operates within fragmented frameworks lacking coordinated decision-making. In light of the aforementioned challenges, this paper proposes a trading method of IES considering hydrogen energy trading and transportation. First, this paper establishes an electricity-heat-hydrogen IES model and a transportation network based on the geographical information between IES and hydrogen refueling stations (HRSs). Then, considering that the hydrogen energy transaction between IES and HRSs is affected by hydrogen energy transportation time, distance and price, a two-stage optimization method based on non-cooperative game is proposed. In this game, both IES and HRSs engage in energy transactions with the objective of maximizing revenue. Finally, the effectiveness of the proposed method is verified by the case studies. The results show that the optimal scheduling method considering hydrogen transport can achieve an economic hydrogen trading and transport solution to complete the hydrogen supply to HRSs. The multi-energy transaction was profitable at $1140.07. Compared to conventional fixed-pricing models, the integrated approach achieves significant improvements in system performance, with an 18.57 % reduction in operating costs, a 3.2 % increase in energy utilization efficiency, and up to 22.3 % enhancement in hydrogen trading revenue.

Owing to its advantages of low voltage stress and a wide voltage conversion range, the symmetric three-level buck-boost converter has attracted considerable interest. However, significant switching losses under conventional hard-switching operation limit its power density and efficiency. To address this issue, this article proposes a unified trapezoidal current mode (TZCM) modulation scheme that enables zero-voltage switching (ZVS) operation in buck, buck-boost, and boost modes without explicit mode switching. The trapezoidal current waveform is formed by coordinating the 2-1-0 level sequence of the left bridge with the 0-1-2 level sequence of the right bridge. Based on the analytical relationship between the duty cycles and the output voltage, the operating mode is implicitly defined by the sum of duty cycles, allowing smooth transition among different voltage conversion regions. The proposed method not only achieves ZVS, but also eliminates the need for mode switching across the entire voltage conversion range. Finally, experimental results of the symmetric three-level buck-boost converter validate the effectiveness and feasibility of the proposed TZCM modulation method, achieving an efficiency of 98.54% at 1 kW, compared with 94.83% under continuous conduction mode operation.

To decelerate the degradation of the proton exchange membrane fuel cell (PEMFC) system for hydrogen locomotives during long-term operation in plateau environments, this article proposes an adaptive health-conscious operation strategy. First, a degradation mechanism model based on membrane electrode assembly (MEA) decay is established to accurately simulate the degradation of the PEMFC system for hydrogen locomotives in a plateau environment. The electrochemical surface area (ECSA) and membrane thickness are then selected as the key degradation indices to extract the health state of the stack. Combined with the net power of the system, a comprehensive evaluation model is constructed to capture the tradeoff relationship between the state of health (SOH) and performance output. By analyzing the operation characteristics under different altitudes and load currents, an optimal health-conscious operation region is identified. A decoupling sliding mode control (SMC) approach based on a disturbance observer is designed to reliably track this optimal operation region. Comparative experiments on a hardware-in-the-loop (HIL) platform demonstrate that the proposed strategy effectively extends the service life while maintaining high net power output of the PEMFC system for hydrogen locomotives under plateau conditions.

With the continuous acceleration of global urbanization, fossil energy is gradually depleting. Hydrogen energy is characterized by rich sources and low-carbon environmental protection. The use of hydrogen energy and fuel cells in rail transit is also one of the aspects to meet the country to promote green and low-carbon energy transformation. To ultimately enhance the overall economy of the fuel cell hybrid system in urban electric multiple units (EMUs), this paper would put forward an energy management strategy by focusing on the essence of minimum value, the minimum system hydrogen consumption as the objective, and then optimize the overall model of the system under the conditions of the state variable (lithium battery state of charge) and the control variable (the output power of fuel cell). Experimental results indicate that all performance metrics including the total amount of hydrogen consumed by the system, average efficiency of fuel cells have been improved substantially compared with the performance of the energy management strategy which is based on the state machine.

The harsh environment and high-power and high-dynamic traction load along the high-altitude railways increase the failure risk and the control difficulty of the proton exchange membrane fuel cell (PEMFC) system for locomotives, which can even lead to failure shutdown. In this article, a fault-tolerant control method is proposed to deal with the control problem of concurrent failure for air supply system under high-altitude environment. Based on the close energy interaction between the system and environment, the PEMFC system model for high altitude is established. Combined with safe constraints, the operation characteristics in the case of failure under high-altitude environment are analyzed. Considering the complex state coupling between fuel cell and its auxiliary system, a strong tracking extended Kalman filter is adopted to realize fault reconstruction and monitoring. The induced mechanism of high-altitude environment and complex traction load to the fault of PEMFC system is analyzed. To realize fast response of controller under fault conditions, an improved fast terminal sliding mode control method is adopted. The comparative experiments on the hardware-in-the-loop platform show that the proposed fault-tolerant control method can effectively guarantee the stable operation of PEMFC system at high altitude, suppress the net power fluctuation, and enhance the safety of the PEMFC system.

To address the difficulty faced by traditional energy management methods in balancing system operation economy and durability when controlling large-scale cluster systems, as well as challenges such as poor power source consistency, this article proposes a hierarchical and domain-partitioned coordinated control method (HDP-CCM) for electric multiple units (EMUs). The proposed method consists of two orthogonal dimensions. The domain partitioning mechanism in the horizontal dimension divided power sources with relatively large performance differences into multiple relatively independent logical control domains. The vertical dimension decouples the output power of the fuel cell and the battery system through a system-domain module-level hierarchical architecture, and controls the power output of each domain and the power sources within the domain based on the domain partition results to achieve active consistency management of power source performance and the system's economic optimization. The effectiveness of the proposed method is validated through hardware-in-the-loop (HIL) testing. Results demonstrate that compared to the equivalent consumption minimization strategy (ECMS) and the hierarchical-only coordinated control method (H-CCM), the proposed method achieves superior performance in both economic efficiency and durability, effectively maintaining power source performance consistency. Notably, the method maintains robust performance under system fault conditions.

The operational energy consumption of the hydrogen-powered train is affected by both the load power profile and power-sharing mechanism of the hybrid power sources, which includes the fuel cell and battery. However, the existing energy-saving operation methods of traditional trains only optimize the load power profile, but fail to achieve the collaborative optimization of it and the power-sharing mechanism. Moreover, the state of power (SOP) of battery varies with the state of charge and operation temperature, which influences the performance of acceleration and braking energy recovery of the train. Thus, the battery SOP needs to be considered into the source-load collaborative optimization framework to ensure the accessibility of the planned operation trajectory and the efficient recovery of braking energy. However, it is also ignored in the literature. Thus, this article first derives the maximum traction power and absorbable braking power of the motor load with the battery SOP. On this basis, it proposes the novel source-load coordination control method for fuel cell trains considering the battery SOP to reduce the global operation energy consumption throughout the entire operating range. Experimental result verifies that the proposed method ensures the executability of the planned trajectories and significantly reduces the overall operation energy consumption by up to about 29% under multiple test conditions.

The harsh environment and high-power and high dynamic traction load along the plateau railway led to severe energy loss, limited output capacity and even stack halt of proton exchange membrane fuel cell (PEMFC) systems. To this end, an energy-saving optimization control (ESOC) method is proposed to simultaneously enhance energy recovery and output capacity of PEMFC systems. Based on the close energy interaction between the system and environment, a plateau PEMFC system model is established, and an expander is integrated to achieve the conversion of gas waste heat energy. The contradictory characteristics between the energy recovery capacity and output capacity of the system in the plateau environment are analyzed. Because maximum net power and maximum energy recovery cannot be achieved at the same oxygen excess ratio (OER), a dual-objective optimization method is proposed to determine the optimal OER while strictly adhering to safety constraints. To realize rapid response under complex traction loads, the OER is dynamically adjusted through an active disturbance rejection controller. The comparative experiments on the hardware-in-the loop platform show that the proposed ESOC method can effectively ensure the stable operation of PEMFC systems in plateau areas, and achieve energy recovery and output capacity improvement.

For the multi-module fuel cell hybrid power systems (FCHPSs) deployed in hydrogen-powered trains (HPTs), this article develops a condition-perception-driven hierarchical adaptive energy management strategy (EMS), aiming to improve energy management adaptability, global optimality, and dynamic coordination among heterogeneous power sources under complex cross-regional operating scenarios. At the top layer, a sliding-window scheme is introduced to extract train operating condition features. In the offline stage, an enhanced Gaussian mixture model (GMM) clustering method is employed to generate high-fidelity training data for random forest (RF) model training, followed by online deployment for real-time operating-condition recognition. At the middle layer, a comprehensive value-loss function is constructed by jointly considering system energy efficiency, hydrogen consumption (HC), battery state of charge (SOC), and power source performance degradation. Coupled with a regime-mapped weight optimization framework (RMWOF), the optimal weight parameters of the objective function for diverse operating conditions are derived and deployed online in a dynamic manner, thereby enabling self-adaptive system optimization. Underpinned by a multi-agent distributed communication network, the bottom layer enables consensus-based dynamic energy regulation of multi-module fuel cell system through lightweight intra-domain communication. Hardware-in-the-loop (HIL) platform test results demonstrate that, compared to conventional strategies and optimization strategies that neglect condition variations, the proposed strategy can perceive condition features and map them to optimization parameters, thereby further enhancing the overall system performance and significantly improving operating condition adaptability.

With the integration of new energy, the uncertainty of new energy output will have a serious impact on the operational economy and reliability of the traction power supply system. In order to cope with prolonged extreme weather, achieve long-term stable energy supply for the system, and improve energy utilization efficiency, this paper proposes a capacity allocation method for traction power supply system considering wind-photovoltaic-hydrogen-storage access under extreme weather conditions. Firstly, a capacity allocation model was established to minimize the annual investment cost, operating cost, and carbon trading cost of the system. Then, extreme weather scenarios are obtained through the analysis of extreme power data measurement indicators. Finally, the effectiveness of the relevant schemes was tested through numerical simulations, and the results indicate that the proposed capacity allocation method has improved the consumption capacity of renewable energy, enhanced the economic efficiency of the system, and promoted the green development of rail transit.

Multi-stack fuel cell (FC) hybrid power systems (MSFCHPSs) are high-order nonlinear systems with inherently complex nonlinear characteristics due to their integration of multiple FC stacks, power electronic converters, and nonlinear loads. Thus, MSFCHPSs are very vulnerable to destabilization by transient disturbances, which exceeds the capabilities of traditional small signal stability analysis. To explore the destabilization mechanism of MSFCHPSs, this article proposes a comprehensive large-signal stability (LSS) analysis framework integrating a full-order nonlinear model, including all main circuits, complete control loops, and FC aging effects, with virtual inertia-based control. The nonlinear system is transformed into a Takagi-Sugeno (T-S) fuzzy representation, enabling domain of attraction (DOA) estimation via Lyapunov theory and linear matrix inequalities (LMIs) with low computational burden and reduced conservatism. The method quantitatively assesses the influence of power distribution, virtual inertia parameters, circuit elements, control gains, and FC aging on LSS, identifying dominant stability factors. A hardware-in-the-loop (HIL) platform is developed to experimentally validate the proposed approach. Results show strong agreement between estimated and actual stability boundaries, confirming the method's accuracy and practical applicability for MSFCHPS design and operation.

To address the renewable energy source (RES) and railway load uncertainties problem and to reduce the safety cost and energy consumption associated with coping with the risk, this article proposes a risk-aware multitimescale dispatch strategy considering risk for a railway energy hub (REH) integrating photovoltaic (PV) power generation and energy storage system (ESS). The global Sharpe ratio (GSR) is used to quantify the risk to establish a day-ahead energy-economic dispatch model. Considering the stochastic characteristics of intraday prediction, the fluctuation entropy is used as a risk-aware parameter to guide the model predictive control (MPC) adaptive optimization objective, and the intraday time-varying optimal dispatch model is established. Based on the actual railway data, the simulation results show that the proposed strategy can effectively reduce the energy consumption of the grid by 39.88% and improve the dispatch flexibility of the model to cope with fluctuations.

Mitigating voltage imbalances in series or flying capacitors is critical for the reliable operation of multilevel DC-DC converters. Conventional voltage balancing techniques rely on individual voltage measurements for each capacitor, requiring multiple sensors that increase system cost and hardware complexity. To provide a more cost-effective solution, this article proposes a novel voltage balancing method that utilizes a pre-existing inductor current sensor to balance all series and flying capacitor voltages, significantly reducing the number of voltage sensors. The proposed method leverages special inductor current sampling and decoupling techniques to calculate capacitor voltage values, enabling voltage balancing for all capacitors with only a single current sensor. An alternating modulation strategy is introduced to address the case where the decoupling method fails when the duty cycle equals 0.5. The key contribution is a novel voltage balancing method that eliminates the need for additional voltage sensors, and it can be seamlessly integrated into a conventional control system. Additionally, the accuracy requirements of the proposed method for the sensor are also analyzed, and the results show that this method can tolerate certain errors, and common current sensors can meet the requirements. Experimental results from a five-level DC-DC converter prototype validate the effectiveness of the proposed inductor current decoupling-based voltage balancing method.

Typically, the triangular current mode (TCM) and the near critical conduction mode (near-CRM) are widely used to achieve zero-voltage switching (ZVS). However, When the duty cycle of the conventional modulation method is equal to 0.5, the inductor current ripple is close to 0, and it is not possible to realize soft switching based on TCM or near-CRM mode. To address this issue, a quadrilateral current mode is proposed in this paper for soft switching implementation. Moreover, the proposed mode can serve as a transition process for bidirectional power flow in a bidirectional three-level DC-DC converter, ensuring that all switches achieve ZVS during the bidirectional operation. Meanwhile, a phase shift control scheme is adopted in this paper to enable the quadrilateral current mode while maintaining ZVS of all switches. The validity and feasibility of the proposed scheme to implement ZVS is verified by a 2 KW prototype bidirectional three-level DC-DC converter. The results demonstrate its practical performance.

As a key component of energy structure transformation, integrated energy systems effectively realise the synergistic operation of multiple heterogeneous energy sources. However, the existing studies on the optimal operation of integrated energy systems have not taken into account the system structure improvement and model refinement, which makes it difficult to fully grasp the complex interaction mechanisms and optimisation potentials within the system. In order to address this limitation, this paper proposes an optimal scheduling method for the integrated electricity-gas energy system that takes into account the dynamic characteristics of the gas network. Firstly, a two-stage P2G model and a gas network dynamic model are established by considering the two-stage P2G operation characteristics and the gas dynamic transmission characteristics. Then, the optimal scheduling model of the integrated electric-gas energy system including the two-stage P2G, gas turbine and other key equipments is constructed. Finally, the effectiveness and superiority of the proposed method in improving system energy efficiency and reducing costs are verified through simulation.

For enhancing the economy and durability of the multi-stack fuel cells system (MFCS) under the long-term cycle conditions of hydrogen electric multiple units (HEMU), a multi-objective power management strategy is proposed considering the stack performance consistency. Firstly, an aging model of proton exchange membrane fuel cell (PEMFC) is established based on the steady-state and dynamic electrochemical surface area (ECSA) model. To update the output characteristic curves and hydrogen consumption curves of PEMFCs in real-time, an online estimation method is employed based on an improved particle filter (PF) algorithm. Building upon this, an adaptive multi-objective optimal control model is established, incorporating MFCS performance consistency and real-time hydrogen consumption, to balance the operational economic efficiency and aging performance consistency of MFCS. To solve the optimal power distribution problem, the variable-order adaptive Legendre-Gauss-Radau orthogonal collocation method is applied, utilizing the GPOPS toolbox. The research findings demonstrate that the proposed method significantly reduces hydrogen consumption compared to the equalization distribution method, mitigates power fluctuations in poorly durable stacks, and promotes convergence of stack aging, effectively extending the system's lifespan.

To extend the lifespan of hybrid locomotive multistack fuel cell systems (MFCSs) and ensure the consistency of multiple fuel cells (FCs), this article proposes an energy management strategy with integrated optimization of system service life (IOSL-EMS). The strategy first unifies the hydrogen consumption (HC) of the FC and the equivalent hydrogen consumption of the battery into quadratic polynomial frameworks. It then considers the service life consistency of multiple FCs and the battery state of charge (SOC) constraint as additional hydrogen consumption, constructing the comprehensive equivalent hydrogen consumption (CEHC) of the system. By analyzing the incremental form of CEHC, the quadratic programming method is used to minimize the CEHC increment, effectively allocating load power between the FCs and the battery. To verify the effectiveness of IOSL-EMS, a hardware-in-the-loop platform was built, and tests were conducted. Results show that IOSL-EMS effectively controls the service life consistency of multiple FCs while ensuring system economy. Compared to the existing rule-based state machine control strategy and optimization-based equivalent consumption minimization strategy, IOSL-EMS significantly extends the service life of the MFCS and reduces the equivalent hydrogen consumption of the hybrid locomotive. In addition, the proposed strategy has similar control capabilities to dynamic programming (DP).

For back-to-back railway energy routers (BTB-RER), renewable energy utilization, regenerative braking energy (RBE) recycling, and traction inter-phase power balance are all achieved through two back-to-back (BTB) inverters, resulting in high hardware capacity requirements and system costs. This paper proposes a novel star-connected structure of an interphase-bridging inverter (IBI) and BTB inverters. This star-connected structure leverages not only the advantage of the smaller hardware rating of IBI but also the flexible controllability of BTB inverters, thereby reducing system costs while realizing similar functionality. Firstly, the power transfer principles and operational modes of the star-connected structure are analyzed to reveal the power complementarity relationship between the IBI and the BTB inverters. Subsequently, in response to the characteristics of multi-functionality and multi-directional energy flow, a power optimization control strategy is proposed to maximize the utilization of photovoltaic (PV) and RBE while enhancing power quality as much as possible. Finally, the superiority of the proposed star-connected structure is validated via hardware-in-the-loop experiments, and the measured data shows a 15% reduction in both inverter and transformer capacity requirements.