The estimation of the state of health (SOH) of lithium-ion batteries (LIBs) plays an important role in ensuring the safe and stable operation of LIB management systems. In order to more accurately predict SOH, a model based on a sparse self-attentive transformer (SSAT) with multitimescale feature fusion is proposed. The SSAT follows an encoder-decoder structure construction, and the model inputs are the extracted health indicators and SOH sequences. The encoder stacks three cross-stage partial (CSP)-ProbSparse attention self-attention blocks, between every two CSP-ProbSparse attention blocks, connections are made by dilated causal convolution and max-pooling layers to obtain exponential growth of the sensory field. All the feature maps output from the self-attention blocks are integrated by multiscale feature fusion, and finally, the appropriate feature dimensions are fed to the decoder through a transition layer to obtain the estimation of SOH. Numerous comparative and ablation experiments have demonstrated that the SSAT model achieves superior performance in a wide range of situations.

Three phase PWM rectifiers have excellent performance when the grid voltages are balanced, and it is widely adopted in the industry. However, the asymmetrical loads or grid faults will lead to the unbalanced grid voltages. In addition, the negative sequence (NS) voltages and currents in the unbalanced grid result in the fluctuation of the DC bus voltage. In this paper, the intrinsic mechanism between the voltage ripple amplitude and the NS currents is revealed. Furthermore, this paper proposes a data-driven adaptive current control method. By shaping the NS currents, the corresponding amplitudes of the 2nd-order harmonic voltage on the DC bus are observed. Based on the observation data, the reference currents that can minimize the DC bus voltage ripple are adaptively derived online. As a result, the DC bus voltage ripple can be suppressed effectively. In addition, the proposed method does not need the NS decomposition or the parameters such as the grid voltages and inductances. Therefore, it has a strong robustness against parameter mismatch. Finally, the accuracy of the theoretical analysis and the effectiveness of the proposed method are validated by experiments.

For an electric vehicle charging system, its control performance and reliability are affected by the health and aging of dc bus capacitor in the three-phase pulsewidth modulation (PWM) rectifier. To monitor the status of capacitor and then adjust the operation mode in time, a dc bus capacitance identification method based on the frequency domain data-driven approach is proposed. It has advantages such as high accuracy and high applicability. It uses all the data in one cycle for identification and is less affected by noise. There is no need for iterative calculation, and the identification speed is fast. Compared with the neural network identification method, the proposed method does not need to be trained under various working conditions and overcomes the disadvantage of limited application conditions. The bandpass filter is used to extract the capacitor voltage and current components, which can be used in the case of unknown loads. Transforming the time-domain model into frequency-domain model, the identification stability can be improved. In this article, the experimental platform of a three-phase PWM rectifier is built. The experiment results have verified the accuracy of the proposed method and its applicability in different loads.