Accurately predicting the remaining useful life (RUL) of lithium-ion batteries is of significance for improving the safety of working environment and the reliability of equipment. To improve the stability and accuracy of RUL prediction, a battery RUL prediction method based on the combination of denoising technology and hybrid data-driven model is proposed. First, the original data is decomposed by variational mode decomposition, and the noise components are filtered by the analysis of correlation. The residual error is combined with the components which have a strong correlation to complete the sequence reconstruction process. Second, with the combination of Tent chaotic mapping, sine cosine algorithm and Levy flight strategy, the sparrow search algorithm (SSA) is optimized, and the optimal weight threshold of extreme learning machine (ELM) is obtained. Finally, the improved SSA-ELM model is trained by using the smoothed denoised data, and the RUL prediction is completed. The NASA data sets are used to verify the effectiveness of the proposed method. Experimental results show that the average absolute error and root mean square error of the prediction result obtained using this method are controlled within 1.58% and 2.14%, respectively, indicating that this method has a high robustness and a high prediction accuracy. Therefore, the proposed method can be applied to battery RUL prediction.

Aimed at the problem that the converter current ripple and electromagnetic interference (EMI) noise will increase due to the increasing switching frequency of a three-level neutral point clamped (NPC) converter, a variable switching frequency modulation strategy based on current ripple prediction is proposed to reduce the current ripple, harmonic noise and EMI noise of the three-level NPC converter. According to the requirements of current ripple, the switching cycle and sampling cycle are calculated to synthesize the latest switching cycle and form feedback, so as to realize variable frequency modulation. The random cycle is distributed around the expected cycle, so that the harmonic noise and electromagnetic noise of the converter are more evenly distributed in a wide frequency band. As a result, the electromagnetic noise of the converter is reduced, and the output inductance current ripple is improved. The relevant simulation and experimental results verified that compared with those under the traditional modulation strategy, the common mode noise was reduced by about 20 dB/µV under the proposed modulation strategy, the differential mode noise was reduced by about 10 dB/µV, and the amplitude of output inductance current ripple was also reduced accordingly.

In order to improve the DC voltage gain and reduce the electrical stress, a novel high voltage gain DC-DC converter based on Z-source is proposed. Theoretically, the ratio of output voltage to input voltage can reach (2-D)/(1-2D). Compared with the traditional diode capacitor filter Z-source DC-DC converter, the proposed topology can provide a higher DC voltage gain at the same duty cycle. It has lower voltage stress and inductance current stress when the DC voltage gain is the same. In addition, the input port and output port of the proposed DC-DC converter share the common ground, which helps to reduce the electromagnetic interference of the system. On this basis, the steady-state principle and characteristics of the proposed DC-DC converter are introduced, and the parameter design and theoretical efficiency calculation are also carried out. Finally, an experimental prototype with a power level of 200 W was fabricated, and experimental results proved the feasibility and superiority of the proposed circuit topology.

A direct control strategy is proposed for a suspension winding DC excitation double-inding bearingless flux switching permanent magnet motor (BFSPMM) with 12/10 pole U-type stator core. First, the influence of rotor eccentricity on the mathematical model of the motor is analyzed, and a double-winding BFSPMM mathematical model under the condition of eccentricity is constructed. Then, a mathematical model of the torque of double-winding bearingless flux switching motor is derived according to the principle of electromechanical energy conversion, and the direct torque control based on space vector pulse width modulation (SVPWM) is constructed. Finally, the virtual displacement method is used to obtain the mathematical model of suspension force, and the voltage vector synthesized by SVPWM is used to precisely control the flux of suspension winding, thus realizing the direct control of suspension force. Experimental results show that the rotor can be suspended stably, and the suspension force and torque can be controlled independently, indicating that the system has good dynamic and static characteristics.

To maximumly protect the medium-voltage motor in a back-to-back medium-voltage motor driving system without transformer based on modular multilevel converter (MMC) from the influence of asymmetric grid faults and switching actions, a control strategy for minimizing the common-mode voltage of the front-end transformerless grid-connected MMC is designed. The common-mode voltage caused by the asymmetric grid fault can be canceled by the MMC counterpart voltage, and the switching ripples caused by the switching action of the MMC can be suppressed by arranging the arm-voltage pulses end-to-end. In addition, the influence of MMC common-mode voltage suppression on the single-phase power deviation is analyzed, and the feedforward control is proposed accordingly. Tests were carried out using a grid-connected MMC prototype system, and experimental results verified that the maximum common-mode voltage of the grid-connected system under severe asymmetric grid conditions can be reduced to 1/3N of its original value by the proposed control strategy, where N is the per-arm submodule number. Meanwhile, the unity power factor, constant DC voltage, and balanced single-phase power were also realized.

The active-clamped soft-switching inverter can realize the soft-switching of power devices, which is conducive to improving the power density and dynamic performance of the inverter. However, when overcurrent occurs, if the conventional cycle-by-cycle(CBC) current limit strategy( i.e., a strategy under which power devices will be blocked once overcurrent occurs) is adopted, the DC bus current will change its direction from flowing to the inverter bridge to flowing to the DC side. Due to the existence of a resonant inductor, both the DC bus current and the current flowing through the resonant inductor flow through the auxiliary switch, so there is high current stress on the auxiliary switch. In this paper, an improved CBC current limit strategy is proposed. By changing the switching state of the inverter bridge after the CBC current limit strategy is triggered, the DC bus current flowing to the DC side is reduced, thus significantly suppressing the current stress. In addition, the protection strategy was verified by an experiment of 3 kW active-clamped soft-switching inverter.

To address the difficulty in predicting the state-of-charge (SOC) of a Li-ion battery pack, an SOC prediction model based on kernel extreme learning machine (KELM) optimized by the improved sparrow search algorithm (ISSA) is proposed. First, Logistic chaotic mapping is introduced to improve the standard SSA and acquire the best population individuals. Second, the improved algorithm is used to optimize the kernel function parameter S and penalty coefficient C of KELM to create an ISSA-KELM prediction model. The simulation is carried out utilizing the historical data from an energystorage device, and the results predicted by ELM, KELM and ISSA-ISSA-KELM models were compared and analyzed. In addition, the robustness of the model was verified using data under other working conditions. Results show that the root mean square error and mean absolute error of predicted SOC decreased to 2.06% and 1.54%, respectively. The proposed model improved the prediction accuracy, and its convergence, generalization and robustness were also satisfying.

Aimed at new energy combined power supply systems such as photovoltaic and fuel cells, a non-isolated dual-input high step-up DC-DC converter is proposed. This converter is based on a dual-input Boost circuit, and the two input sources and output, as well as each switch tube, share a common ground. A diode capacitor network is introduced at the later stage to achieve high voltage gain and reduce the voltage stress of switching devices. The two input sources can supply power at the same time, and any one of them can supply power independently without adding extra switch tubes. In addition, the voltage gain can be further improved by expanding the booster unit to adapt to different application scenarios. The working principle for the converter and its extended circuit and the corresponding performance such as voltage gain characteristics and voltage stress of switching devices in three power supply modes are analyzed in detail, and its performance is compared with those of the existing similar converters. Finally, an experimental prototype was built to verify its feasibility.

The new generation of wide bandgap power semiconductors such as SiC and GaN are driving the rapid high-frequency, high-efficiency, and small volume development of power electronic equipment. However, they are also more likely to interfere with sensitive loads, affect wireless communication, and even endanger their own safety and reliable operation, which poses great pressure and challenges to the electromagnetic compatibility (EMC) performance of power electronic equipment. In recent years, the radiated frequency(RF) characteristics of power switches, wideband electromagnetic models of magnetic components, electromagnetic radiation mechanisms of switched mode power supplies, near-field characteristics of wireless power transmission(WPT), and the new designs of electromagnetic interference(EMI) filters have become current research hotspots and received continuous attention from academia and industry. The Journal of Power Supply has specially released the album "Electromagnetic Compatibility in Power Electronic Systems" to promote the exploration of difficult and hot issues in the field of EMC analysis and design of power electronic systems.

In view of the serious current harmonics in a doubly-fed induction generator (DFIG)-DC connection system and the large loss of a dual-voltage source inverter (VSI) connection system, a novel dual-converter connection system which connects a three-phase DFIG to DC microgrid is designed. First, the topology, pulse-width modulation (PWM) measurement and DFIG model of the dual-converter connection system are described in detail. Different from the traditional connection systems, the proposed connection system adopts an open-end winding structure and uses a three-bridge arm rectifier on each side of the stator winding. Considering that these arms are usually composed of insulated gate bipolar transistors (IGBTs), a diode is used instead of the IGBT in the rectifier to decrease the number of control switches and reduce the cost. Second, the control strategies for a stator-side converter (SSC) and a rotor-side converter(RSC) under the new topology are given. Third, a comparison with the DFIG-DC connection system and the dual-VSI connection system is performed through simulations, and results show the advantages of the proposed method in terms of current harmonic distortion, torque ripple and semiconductor loss. Finally, experimental verification was carried out on a 0.56 kW DFIG, and results also verified the advantages of this method in terms of loss, harmonics and torque ripple.