基于GA-ANFIS的装甲车辆蓄电池SOH预测方法

SOH Prediction Method of Armored Vehicle Batteries Based on GA-ANFISZhu Yongli,Chang Tianqing,Liu Peng

摘要:

为保持装甲车辆的机动安全和运行可靠,提高其铅酸蓄电池健康状态的预测能力至关重要。本文将遗传算法与自适应模糊神经系统相结合,提出了一种基于GA-ANFIS的装甲车辆蓄电池SOH预测方法,着重分析了该方法的总体流程和训练过程。着眼装甲车辆的工作环境,在放电深度和输出能量的基础上,引入海拔和温度作为模型的输入。在Matlab的实验结果表明,GA-ANFIS相比ANFIS测试数据误差减小47.6%,四输入GA-ANFIS相比两输入GA-ANFIS测试数据误差减小51.2%,验证了方法的有效性。

Abstract:

In order to keep the armored vehicles safe and reliable, it is very important to improve the ability of predicting the state of health of lead-acid batteries. In this paper, a method of SOH prediction for armored vehicle batteries based on GA-ANFIS is proposed, which combines genetic algorithm with adaptiveSnetwork-based fuzzy inference system. The overall process and training process of the method are emphatically analyzed. Aiming at the working environment of armored vehicles, elevation and temperature are introduced as the input of the model on the basis of discharge depth and output energy. The experimental results in Matlab show that the error of GA-ANFIS is 47.6% less than that of ANFIS, and that of four-input GA-ANFIS is 51.2% less than that of two-input GA-ANFIS.