Predicting future capacities and remaining useful life (RUL) with uncertainty quantification is a key but challenging issue in the applications of battery health diagnosis and management. This article applies advanced machine-learning techniques to achieve effective future capacities and RUL prediction for lithium-ion (Li-ion) batteries with reliable uncertainty management. To be specific, after using the empirical mode decomposition (EMD) method, the original battery capacity data is decomposed into some intrinsic mode functions (IMFs) and a residual. Then, the long short-term memory (LSTM) submodel is applied to estimate the residual while the Gaussian process regression (GPR) submodel is utilized to fit the IMFs with the uncertainty level. Consequently, both the long-term dependence of capacity and uncertainty quantification caused by the capacity regenerations can be captured directly and simultaneously. Experimental aging data from different batteries are deployed to evaluate the performance of proposed LSTM+GPR model in comparison with the solo GPR, solo LSTM, GPR+EMD, and LSTM+EMD models. Illustrative results demonstrate the combined LSTM+GPR model outperforms other counterparts and is capable of achieving accurate results for both 1-step and multistep ahead capacity predictions. Even predicting the RUL at the early battery cycle stage, the proposed data-driven approach still presents good adaptability and reliable uncertainty quantification for battery health diagnosis.

In conventional equalizers, the facts of bulky size and high cost are widespread. Particularly, the zero-switching loss and zero-voltage gap (ZVG) between cells are difficult to implement due to the high-frequency hard switching and the voltage drop across power devices. To overcome these difficulties, a direct cell-to-cell battery equalizer based on quasi-resonant LC converter (QRLCC) and boost dc-dc converter (BDDC) is proposed. The QRLCC is employed to gain zero-current switching, leading to a reduction of power losses. The BDDC is employed to enhance the equalization voltage gap for large balancing current and ZVG between cells. Moreover, through controlling the duty cycle of the BDDC, the topology can online adaptively regulate the equalization current according to the voltage difference, which not only effectively prevents overequalization but also abridges the overall balancing time. Instead of a dedicated equalizer for each cell, only one balancing converter is employed and shared by all cells, reducing the size and implementation cost. Simulation and experimental results show the proposed scheme exhibits outstanding balancing performance, and the energy conversion efficiency is higher than 98%. The validity of the proposed equalizer is further verified by a quantitative and systematic comparison with the existing active balancing methods.

This paper proposes an automatic any-cells-to-any-cells battery equalizer, which merges the forward and flyback converters through a common multiwinding transformer. The windings of the transformer are divided into two groups, which have opposite polarities. The principles of the proposed equalizer are that the equalization in one group is achieved based on forward conversion and the balancing between the two different groups is based on flyback conversion, by which the magnetic energy stored in the transformer can be automatically reset without using additional demagnetizing circuits. Moreover, only one MOSFET and one primary winding are required for each cell, resulting in smaller size and lower cost. One pair of complementary control signals is employed for all MOSFETs, and energy can be automatically and directly delivered from any high-voltage cells to any low-voltage cells without the requirement of cell monitoring circuits, thereby leading to a high balancing efficiency and speed. The proposed topology can achieve the global equalization for a long battery string through connecting the secondary sides of transformers without the need of additional components for the equalization among modules, which also overcomes the mismatching problem of multiple windings. The validity of the proposed equalizer is verified through experiments, and the balancing efficiency can reach up to 89.4% over a wide range of conditions.