铅酸蓄电池的内阻会随其运行时间增加而增大，从而使其容量下降并导致循环使用寿命减小。因此，对其使用寿命的准确评估预测将有助于提高变电站直流电源系统的持续供电能力和运行可靠性。LIBSVM支持向量机是遵循结构风险最小化原则发展的机器学习方法，将其用于蓄电池寿命预测，具有不依靠蓄电池详细数学模型建立其循环寿命预测模型的特点。基于此，在研究支持向量机的基本原理基础上，进一步研究利用LIBSVM支持向量机基于蓄电池健康状态、端电压和电池剩余容量的训练样本数据，建立反映电池容量与健康状态和端电压非线性映射的建模方法，并讨论基于交叉验证设计LIBSVM回归机最优参数的方法。实验结果表明，基于LIBSVM的铅酸蓄电池寿命预测模型具有较高的预测精度，该方法是切实可行的。

The internal resistance of the lead-acid battery increases as its operating time increases, which will result in the decrease in its capacity and a consequent reduction in service life. Therefore, the accurate assessment and prediction of its useful life is benefit for improving the ability of continuous power supply and operational reliability of substation DC power system. The support vector machine of LIBSVM is a machine learning method that follows the principle of structural risk minimization. It has the characteristic of using support vector machine to establish the predict model of the battery useful life without modeling the detailed mathematical model of battery. So based on studying the basic principle of support vector machine, the method that uses the support vector machine of LIBSVM to model the non-linear mapping between useful life and both the terminal voltage and state of health based on the training sample data of these three state variables is proposed. At the same time, the method of designing the optimal parameters for the regression machine of LIBSVM based on cross-validation is discussed. The experimental results show that the LIBSVM-based lead-acid battery life prediction model has high prediction accuracy. The feasibility of the proposed prediction method is verified as well.