Research on SOC fuzzy weighted algorithm based on GA-BP neural network and ampere integral method

• Published in: Journal of engineering (Stevenage, England) Vol. 2019; no. 15; pp. 576 - 580
• Main Authors: Guo, Na, Fang, Yu, Tian, Zhongli, Cao, Songyin
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 01.03.2019
• Subjects: ampere integral method; ampere integration method; backpropagation; charge measurement; computerised instrumentation; estimation theory; fuzzy set theory; fuzzy weighted SOC estimation model; GA‐BP neural network model; genetic algorithms; Jiangsu Annual Conference on Automation (JACA 2018); neural nets; secondary cells; SOC fuzzy weighting algorithm model; state of charge estimation model; voltage sampling accuracy; estimation theory; fuzzy set theory; computerised instrumentation; voltage sampling accuracy; SOC fuzzy weighting algorithm model; genetic algorithms; fuzzy weighted SOC estimation model; charge measurement; state of charge estimation model; ampere integration method; ampere integral method; backpropagation; secondary cells; GA-BP neural network model; neural nets
• ISSN: 2051-3305; 2051-3305
• DOI: 10.1049/joe.2018.9385

In view of the GA-BP neural network model for estimating the state of charge (SOC) of batteries be greatly influenced by the voltage sampling accuracy. Here, a SOC estimation model based on fuzzy weighting algorithm is proposed which is modified by combining GA-BP neural network with ampere integration method. In addition, the block diagram of the fuzzy weighted SOC estimation model and the specific design process of the model and the determination of input and output and the formulation of fuzzy rules are given here. Through comparing the simulation of SOC model based on GA-BP neural network with the simulation of SOC fuzzy weighting algorithm model based on GA-BP neural network and ampere integral method, it is concluded that the SOC fuzzy weighting algorithm studied here is superior to the SOC algorithm based on GA-BP neural network.