The wireless charging technology for electric vehicles has been widely studied and applied owing to its advantages of small footprint, convenience and flexibility, low maintenance cost and strong interaction with power grid. A bidirectional wireless charging system for electric vehicles based on the dual phase shift (DPS) control strategy is proposed. Through the DPS control strategy, the phase shift angle on the primary and secondary sides is changed to change the output power. By changing the phase angle difference between voltages on the primary and secondary sides, the energy flow direction is changed, and the function of “peak shaving and valley filling” is realized eventually. First, the architecture of the bidirectional wireless charging system for electric vehicles based on the DPS control strategy is presented, and its working mode and principle are analyzed. Second, the control strategy for the bidirectional wireless charging system is described in detail, and the output power and direction of the system are adjusted by changing the phase shift angle on the primary and secondary sides, as well as the phase angle difference between voltages on the primary and secondary sides. Third, the working principle for the system and the control method to realize bidirectional charging are analyzed. Finally, a simulation model was built in MATLAB, and experimental verification was carried out. Results show that the designed bidirectional wireless charging system can achieve bidirectional energy transmission, satisfying control performance and good symmetry of positive and negative charging.

Due to the rank deficiency of identification models, parameter coupling exists in the online identification of a permanent magnet synchronous motor (PMSM). To realize the parameter decoupling identification, an adaptive parameter identification algorithm based on the affine projection algorithm is proposed. By injecting instantaneous flux-weakening current signal on the d axis, the online decoupling identification of parameters such as the stator resistance, inductance and rotor flux linkage of a surface mounted PMSM is realized. Considering the nonlinear voltage drop caused by the dead-zone time of an inverter, turn-on and turn-off delay of switches, freewheeling diode saturated voltage drop, etc., a parameter identification model which is not affected by nonlinear voltage drop is constructed. Experimental results demonstrate that the proposed method can improve the identification accuracy of stator resistance, inductance and rotor flux linkage while speeding up the convergence rate of online identification.

To address the issue that the conventional model predictive torque control algorithm for a permanent magnet synchronous motor relies on position sensors, a model predictive torque control approach based on a dualsliding-mode observer is proposed. First, to accomplish the sensorless control, the real-time speed in the prediction model is substituted by the observed speed by coupling the back electromotive force sliding-mode observer with the optimized quadrature phase-locked loop. Second, the stator flux sliding-mode observer is used to observe the stator flux, and the observed flux introduces prediction error through feedback correction to obtain the prediction torque and flux with higher accuracy, thus improving the prediction speed and accuracy of the model predictive torque control system. By contrasting the MATLAB/Simulink simulation and experimental platform with the conventional model predictive torque control approach incorporating position sensors, the efficacy and feasibility of the proposed method are confirmed.

The high-power traction inverter allows for a lower switching frequency. To improve the output performance of a system, synchronous space vector pulse width modulation (SVPWM) strategies are usually used in the medium-speed range, and four synchronous modulation strategies with different performances can be flexibly implemented based on the number and positions of sampling points, as well as the sampling point vector action order. Under this background, the basic principle of four synchronous SVPWM strategies is summarized at first, with a focus on the smooth and shock-free switching method between different frequency-division and different modulation strategies. Second, to address the issue of fundamental phase discontinuity between the boundary sampling strategy and other modulation strategies during the switching, a method is proposed to split the original sampling sector and further reassemble into a new sector, which not only achieves phase continuity but also makes the setting of carrier wave ratio more flexible. Finally, the feasibility of the proposed method was verified by simulation and experimental results.

Due to the influence of microgravity in space, the water management of fuel cells in space is quite different from that on the ground. To meet the application requirements of space reliability and lightweight fuel cell stacks, the non-flow through water management technology is applied to space fuel cell stacks, thereby significantly optimizing the stack structure and reducing the corresponding size and weight. In this paper, the principle for the non-flow through technology, influencing factors of non-flow through, stack structures and research status of the non-flow through technology are reviewed, providing reference for the design of space fuel cell stacks and research on non-flow through water management.

Accurately predicting the remaining useful life (RUL) of lithium-ion batteries is critical to the safe and reliable operation of new energy vehicles. First, the research status of data-driven methods for predicting the RUL of lithium-ion batteries is analyzed in this paper, and the research progress in six commonly used data-driven methods is reviewed. Then, three problems existing in the practical applications of RUL prediction of lithium-ion batteries at present are summarized. At the same time, the issue of battery dataset collection is discussed comprehensively, and the importance of battery datasets to the development of data-driven methods is also elaborated upon. Finally, the development trend in the future is prospected.

In the phase field fracture method, the crack is represented by defining an order parameter, and the governing equation describing the crack development is obtained by using the minimum energy variational principle for the system. In this paper, the framework of phase field theory was established through the derivation of governing equations, including the establishment of phase field fracture damage evolution equation and boundary conditions based on the energy principle. Tensile and shear simulations were carried out on a plate with a notch, and the fracture path and support reaction were in good agreement with the results reported in the literature. The crack propagation of lithium-ion batteries with different diameters of silicon anode during charging was simulated using the phase field method. Results show that the large-diameter (400 nm) silicon particles quickly exhibit crack during charging due to the large tensile stress on their surfaces, and the crack propagates inwards. The fracture energy release rate has a great influence on the crack propagation, and a small energy release rate results in more crack initiation points and a high crack density. In comparison, internal fracture appears in small-diameter (100 nm) silicon particles during charging, and the fracture time is earlier than that of large-sized silicon particles.

To overcome the shortcomings of phase change materials (PCM) such as a low coefficient of heat conductivity, a novel axial extended corrugated fin is designed to improve the heat dissipation performance of lithium-ion batteries. Subsequently, the effects of the types of fins, convective heat transfer coefficient, number of fins, extended fin length and ambient temperature on the heat dissipation performance of batteries are studied by simulations using computational fluid dynamics (CFD). Results demonstrate that the introduction of corrugated fins can improve the heat dissipation performance of lithium-ion batteries. Compared with the traditional trapezoidal fins, the corrugated fins can further reduce the maximum temperature of batteries by 1.7 °C. In addition, the heat dissipation performance of batteries is improved by increasing the heat transfer coefficient, as well as the number and extended length of fins. With a trade-off between the cooling performance of the heat dissipation model and its weight, the maximum temperature and temperature difference of the lithium-ion battery are 39.9 °C and 4 °C respectively when the convective heat transfer coefficient, the number of fins and fin length separately take their optimal values of 35 W/m²K, 8 and 15 mm. Moreover, when the ambient temperature rises from 25 °C to 30 °C, the maximum temperature of the battery is 46 °C, which still remains within a safe range, indicating that the thermal management system presents good thermal stability.

The parameter identification accuracy of a lithium battery model is of significance for the development of a battery management system. To improve the accuracy and efficiency of parameter identification, a multi-step parameter identification method for the electrochemical-thermal coupling model of lithium battery is proposed in combination with the Arrhenius law, which uses the whale optimization algorithm improved by oppositional learning strategy and Levy mutation. In this method, constant-current data at 15 ℃, 25 ℃ and 35 ℃ are adopted for parameter identification, and constant-current and dynamic working conditions are used to verify the identification results. Simulation and experimental results show that the mean absolute percentage error of parameters is reduced by 7.2%, the maximum root mean square error of voltage predicted by the model is 26 mV and that of temperature is 0.4 ℃, indicating that the multi-step parameter identification method using the whale optimization algorithm based on oppositional learning strategy and Levy mutation can efficiently identify parameters.

The remaining useful life (RUL) prediction of lithium-ion batteries is an important part of battery health management. To improve the accuracy of RUL prediction, a battery RUL prediction method based on the fusion of source-domain battery iterations module (SIM) and long short-term memory (LSTM) neural network is proposed. First, multiple batteries with known degradation trends are used as source-domain batteries, and their full-life capacity degradation data is used to construct an SIM to obtain the optimal LSTM pre-training model. Second, the pre-training model is transferred to the target-domain and further fine-tuned using the target-domain training set. Finally, the fine-tuned LSTM pre-training model is used for the capacity prediction task of the target-domain battery and completes the RUL prediction. Three open-source datasets were used to validate the effectiveness of the algorithm. Results show that the RUL prediction error of this algorithm is less than 2 cycles for the same type of source-domain and target-domain batteries. For different battery types, the RUL prediction error is less than 10 cycles except for 20% of the starting points for prediction, indicating a high level of prediction accuracy.