RBF neural network and modified pid controller based State of Charge determination for lead-acid batteries

• Published in: 2008 IEEE International Conference on Automation and Logistics pp. 769 - 774
• Main Authors: Yanqing Shen, Guangwei Li, Shanquan Zhou, Yinquan Hu, Xiang Yu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-09-2008
• Subjects: Adaptation model; Artificial neural networks; Batteries; Data models; Lead-Acid Batteries; Mathematical model; Modified PID Controller; Predictive models; Radial Basis Function Neural Network (RBFNN); State of Charge (SOC); System-on-a-chip
• ISBN: 9781424425020; 1424425026
• ISSN: 2161-8151; 2161-816X
• DOI: 10.1109/ICAL.2008.4636253

State of charge (SOC) determination is an increasingly important issue in battery energy storage system. Precise knowledge of SOC allows the controller to confidently use the battery packpsilas full operating range without fear of over- or under-charging cells. Taking into account of some transformed parameters like voltage and current, this paper describes a novel adaptive online approach to determinate SOC for lead-acid batteries by combining modified PID controller with RBFNN based terminal voltage evaluation model, which is used to simulate batterypsilas behavior while it is under load. Results of lab tests on physical cells, compared with model prediction, are presented. Results show that the ANN based terminal voltage evaluation model simulates battery system with great accuracy, and the prediction value of SOC simultaneously converges to the real value quickly within the error of plusmn1% as time goes on.