基于聚类算法的蓄电池SOC模糊预测

Battery SOC fuzzy estimation based on fuzzy clustering

为了实现对蓄电池的准确在线估算,研究了利用蓄电池电动势、内阻与荷电状态(state of charge,SOC)之间的关系,设计了基于模糊C-均值聚类的模糊控制器。该控制器将模糊C-均值聚类方法与模糊控制系统有机结合,能有效地进行数据划分和构建模糊控制规则。实验表明,该方法将SOC预估误差控制在3%之内,很好地反映了铅酸蓄电池的能量状态。与现有的模糊预测控制器相比,准确度更高,具有一定的实用性。

In order to achieve an accurate online estimation of the battery, a research of the relationship between the battery electromotive force and internal resistance with state of charge （SOC） was carried on. a fuzzy controller which was based on the fuzzy C-means clustering was designed. By combining the fuzzy C-means clustering method and the fuzzy control system organically, this controller was effective to partition the original data and construct the fuzzy control rules of fuzzy controller. Experiments show that the relative prediction error can be controlled less than 3% and it can reflect the energy state of lead-acid battery effectively. Comparing with the existing fuzzy predictive controller, the accuracy was higher and has certain practicality.