A deep learning-based cooperative fault diagnosis method is proposed in this study. The proposed method aims to improve the reliability of contemporary fault diagnosis methods for the multipermanent magnet synchronous motor (PMSM) drive system of four-wheel-independent-drive electric vehicles (4WID-EVs). Specifically, a novel 1-D signal to 2-D RGB image conversion method is proposed to transform the fault diagnosis task into an image classification problem, and the impressive image classification capability of the convolutional neural network (CNN) is utilized to achieve a highly accurate fault diagnosis function. Following this, CNNs of various architectures are trained using the constructed RGB image dataset, and the fault diagnosis accuracy under different architectures is compared. In addition, the adaptability of the trained network is increased by utilizing two adaptive factors to ensure that the trained CNN can be compatible with all untrained combinations of angular speed and load torque. Furthermore, by combining the designed single fault diagnosis method with directed-graph theory, a comprehensive evaluation criteria-based cooperative diagnosis method is proposed, which further improves the accuracy significantly. The effectiveness of the proposed fault diagnosis method is demonstrated via adequate comparison.

The interleaved parallel topology of the dc–dc converter can increase the load capacity and reduce the ripple. However, the multiphase control system requires many current sensors, which increases the software/hardware cost. To solve this problem, this article proposes a dual time-scale model predictive control (DMPC) algorithm based on a single current sensor. First, the current estimator replaces current sensors to measure the unbalanced branch currents and reduce hardware complexity. Then, a DMPC algorithm is proposed to reduce the parallel computational complexity of the controller. Furthermore, a parameter identification module dynamically updates the current estimator and controller to further increase the system’s accuracy. This method reduces the software/hardware complexity while ensuring control effectiveness. The validity of the proposed algorithm is verified on a three-phase buck converter.

Dynamic wireless power transfer (DWPT) technology provides a reliable and safe approach for charging electric vehicles. However, the mobility of electric vehicles results in fluctuations in the dynamic coupling coefficient, which can reduce the stability and transmission power of the system. To improve the DWPT system’s dynamic performance, this paper proposes the adaptive tube based model predictive control algorithm (ATMPC) to control the DWPT system. The ATMPC algorithm adds a non-linear autoregressive neural network (NAR)-based tube to a two-layer MPC. The algorithm can dynamically plan the expectation trajectory according to the fluctuation of mutual sensing of the DWPT system and adaptively tighten or loosen the tube for constraints, thus improving the dynamic performance. In addition, the algorithm also deals the hardware environment delay problem by the predictive compensation. Finally, the proposed algorithm is verified on an 85kHz, 1.5-kW DWPT experimental platform.

Dynamic wireless power transfer (DWPT) technology provides a more convenient and safe charging method for electric vehicles. Power density α and transmission efficiency η are the key indexes of DWPT systems. This paper proposes a two-stage optimization method with α-η as the optimization objective and the Coupling & Resonant part as the optimization part. The method decouples the parametric and structural design problems, comprehensively optimizes α and η of the system, and finally obtains the global optimal solution. A 30 kW, 85 kHz DWPT optimization system is constructed as the application case to verify the method’s effectiveness. The experimental results show that the system designed by the two-stage optimization method can achieve the expected performance. The system efficiency reaches 96.4%, and the power density reaches 2.65 kW/kg.

The implementation of dynamic wireless charging (DWC) infrastructure on highways effectively utilizes the renewable energy while providing electric vehicles (EVs) with a more convenient continuous charging solution, significantly enhancing the traffic efficiency of EVs. However, the coupled system of microgrid (MG) and DWC highways encounters numerous challenges arising from various uncertainties, including the variability of MG generation, the randomness of charging loads, the unpredictability of EV driving behavior, and the diversity of charging demands. To address the above problems, this article proposes a multiscale collaborative optimization (MSCO) strategy with microcharging characteristics, which balances the demands of MG, road operators, and EV users, thereby enhancing the overall performance of the coupled system. The proposed strategy is implemented through the collaborative operation of the cloud-based controller and EV-side controllers. The cloud-based controller generates a power allocation plan that balances the needs of the three stakeholders through multiobjective (M.O.) optimization algorithms. This plan is implemented by EV-side controllers using a tube-based model predictive control (TMPC) method. Furthermore, the proposed hybrid surrogate-metaheuristic optimization (HSMO) algorithm significantly enhances the computational efficiency for long-term cloud optimization problems, ensuring the coupled system maintains stable operation at optimal working points. The effectiveness and superiority of the MSCO strategy are validated through multiscenario case analyses on the cloud-based platform, demonstrating a 22.4% increase in MG utilization rate, a 46.16% reduction in DWC power losses, and a 9.25% improvement in the median EV charging power.