Jigang Wang

The multi-relay wireless transfer (MR-WPT) system offers significant advantages in long-distance wireless power supply for equipment in complex spaces unsuitable for maintenance. While ensuring functionality, the simple structure will undoubtedly have a strong appeal in terms of stability and cost. This paper presents an extremely simple battery load adaptive wireless charging system. A novel n-coil MR-WPT system has been designed that is capable of achieving constant current output with a zero phase angle (CC-ZPA) and constant voltage output with a zero phase angle (CV-ZPA) at different frequencies. The port characteristics of the system were obtained after and a method for sensing the battery charge state at the transmitter (Tx) side was obtained. This is employed as a foundation for establishing the switching criterion for the two charging modes. The system uses the Tx’s current to determine the battery status and adjust the frequency to change the charging mode, it is the simplest structure as it does not require any communication and topology switching. Finally, a 6-coil prototype was used to validate the system with a maximum resonator transmission efficiency of 97%.

Random coil parameters and corresponding coupling relationships make it complex to analyze and understand the circuit characters of general domino wireless power transfer (DWPT) systems. This paper proposes a novel method called the coefficient decomposition model (CDM) to solve general D-WPT systems. First, without any coil self-inductance and mutual inductance constraints, the D-WPT transmission equation is obtained based on KVL, and the self-inductance of each coil is normalized; the load factor is extracted using the determinant transformation, which creates the conditions for constructing the load-independent unique port characteristics; finally, the current expression of each coil is derived based on Cramer's law, and the coefficient determinant is expanded several times to obtain the current expression and expand the coefficient determinant several times to get the hidden term in the expression. Based on the CDM, the input-output characteristics of general D-WPT systems can be obtained, and formulate approaches to attain constant current (CC)/voltage (CV) output, zero-phase-angle (ZPA), CC-ZPA/CV-ZPA, and maximum power transfer. This model suits D-WPT systems with various center axes, non-center axes, different coil combinations, and arbitrarily varying spacings.

Data-driven methods have been systematically studied for identifying GIS insulation defects, while the imbalanced samples pose a great challenge to the methodology. Although existing imbalanced learning strategies (i.e. re-sampling or re-weighting) yield improved minority diagnosis performance, the robustness of the representation feature is damaged. Therefore, this paper proposes a novel bilateral branching networks (BBN) with cost-sensitive Res2Net (BBN-Ence-Res2Net) to improve minority diagnosis performance while enhancing the representation feature robustness of feature extractor. Specifically, BBN bridges normal and rebalanced learning via cumulative learning strategy, thus adaptively increasing the learning power for minorities while preserving better generalization representation features, especially for majorities, in normal learning. Moreover, Res2Net-18 with multi-scale feature learning is proposed as the feature extractor backbone and regulated by class-based effective sample number cost-sensitive mechanism in rebalanced learning, considering the marginal effect of sample information. Eventually, multiple sets of comparative experiments are implemented based on the GIS PD experimental platform to confirm the validity of the proposed model. The results indicate that the proposed model is superior to other methods, reaching 98.27%, 95.68%, 0.9554, 96.19% and 92.83% for Acc, Avg-acc, G-mean, F1-score and MIoU, respectively.

Domino wireless power transfer (D-WPT) has excellent advantages in long-distance wireless power supply for equipment in complex spaces unsuitable for maintenance. A simple receiver (Rx) will reduce the failure rate and system maintenance costs. This paper proposes a D-WPT system with battery State of Charge (SOC) perception at the transmitter (Tx) and adaptive adjustment of constant voltage (CV) and constant current (CC) output. Specifically, we construct a two-port network circuit model for the n-coil D-WPT system and propose the SOC identification method that uses the voltage and current of the Tx as input. Then, an adaptive output regulation strategy is constructed that uses a hybrid topology at the Tx to achieve the system’s CV or CC output. The system does not require Rx and Tx communication. Finality, an experimental platform with 6 coils and a 0.5m transmission distance was established to verify the system’s effectiveness. The results show that the system can accurately identify the change in the battery load charging state and adaptively switch between CC and CV. The maximum system efficiency is 70%.

Multireceiver wireless power transfer (multi-Rx WPT) technology is a promising method for charging multiple electronic devices simultaneously. However, the cross-coupling effect between Rxs not only leads to the correlation of output characteristics of each load, but also reduces the power transfer capability and efficiency of WPT systems. This article proposes a method to achieve constant current (CC) output with cross-coupling compensation in multi-Rx WPT systems, thus avoiding tedious parameter measurement and communication between Rxs. The cross-coupling can be compensated adaptively, and the output can be controlled independently and not affected by the fluctuation of system parameters. A closed-loop control strategy based on the phase shift control method of an active rectifier is carried out, which can continuously regulate the load current while maintaining cross-coupling compensation. An optimized phase synchronization auxiliary circuit for multi-Rx WPT systems is presented, which shares the power transfer coils and is not affected by the cross-coupling effect. A 70-W two-load experimental prototype is built to verify the theoretical analysis, and the maximum efficiency reaches 88.1%.

There are few studies on space-driven gear systems in the existing literature. In this paper, a spacedriven two-stage spur gear system is taken as the research object, and a 10 DOF dynamic model is established. A nonlinear dynamic response analysis was performed. The backlash was introduced into the dynamic model, and the time-varying stiffness was corrected to make the theoretical model closer to reality. By comparing two kinds of dynamic response curves with and without return difference, it was illustrated that the influence of return difference on dynamic transmission error in a gear system. The results obtained in this paper provide a reference and basis for subsequent research.

Simultaneously charging multiple devices is a unique advantage of wireless power transfer (WPT). However, it is challenging to analyze and optimize the performance of a multireceiver (multi-Rx) WPT system due to the intricate coupling among coils. In this article, a method for adaptive decoupling between Rxs is proposed for electronic device charging, which can achieve independent regulation of each output in a multi-Rx WPT system. A simple feedback control avoids tedious system structure and position parameters’ measurement and calculation. With a fixed driving frequency, the number of Rx is not limited. It is analyzed that the quadrature phenomenon of the transmitter (Tx) and Rx currents can be used as a sign that the cross-coupling is fully compensated. A phase synchronization channel is established that shares power transfer coils. An microprogrammed control unit-(MCU) free feedback loop is constructed in Rx to control the variable switched capacitor and finally realize adaptive cross-coupling compensation. A two-load experimental prototype was built to verify the theoretical analysis. Compared with the system without adaptive decoupling between Rxs, the measured system efficiency and output power of the proposed system were improved by 69% and 14.3 times, respectively.