Accurate estimation of the state of health (SOH) of lithium-ion batteries is an important determinant of their safe and stable operation. In this paper, a method for the SOH estimation of lithium-ion batteries based on the least squares support vector machine error compensation model (LSSVM-ECM) is proposed. This method achieves a combination of an empirical degradation model and a data-driven method. Battery degradation can be divided into overall trends and local differences, where the former can be described by an empirical degradation model (EDM) established by the historical data of the battery capacity, while the latter can be mapped by a least squares support vector machine (LSSVM). An LSSVM-ECM is established, where the input is the time interval of the equal charging voltage rising (DV_DT) and the output is the fitting error of the EDM, which represents the local difference of the capacity degradation to dynamically compensate the prediction results of the EDM that represents the global trend in terms of the capacity degradation. Validations are carried out with battery data provided by Oxford and NASA datasets. Results show that the proposed method has a high prediction accuracy and a strong robustness.

准确估计锂离子电池的健康状态（SOH）是其安全稳定运行的重要决定因素。本文提出了一种基于最小二乘支持向量机误差补偿模型(LSSVM-ECM)的锂离子电池SOH估计方法。该方法实现了经验退化模型和数据驱动方法的结合。电池退化可以分为整体趋势和局部差异，前者可以用电池容量历史数据建立的经验退化模型（EDM）来描述，而后者可以用最小二乘支持向量机映射（ LSSVM）。建立LSSVM-ECM，其中输入为等充电电压上升的时间间隔（DV_DT），输出为EDM的拟合误差，它代表了容量退化的局部差异，以动态补偿代表容量退化全局趋势的 EDM 的预测结果。使用牛津和 NASA 数据集提供的电池数据进行验证。结果表明，该方法具有较高的预测精度和较强的鲁棒性。