SOC estimation algorithm of power lithium battery based on AFSA-BP neural network

• Published in: Journal of engineering (Stevenage, England) Vol. 2020; no. 13; pp. 535 - 539
• Main Authors: Wang, Qiuxia, Wu, Peizhou, Lian, Jialing
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 01.07.2020
• Subjects: accurate battery models; AFSA algorithm; AFSA‐BP algorithm; AFSA‐BP neural network; artificial fish swarm algorithm‐back propagation neural network structure; backpropagation; iron phosphate power battery; neural nets; optimising BP neural network; power engineering computing; power lithium battery; secondary cells; SOC estimation algorithm; The 3rd Asian Conference on Artificial Intelligence Technology (ACAIT 2019); voltage 48.0 V; voltage 48.0 V; artificial fish swarm algorithm-back propagation neural network structure; power engineering computing; SOC estimation algorithm; AFSA algorithm; AFSA-BP neural network; accurate battery models; AFSA-BP algorithm; optimising BP neural network; power lithium battery; backpropagation; secondary cells; iron phosphate power battery; neural nets
• ISSN: 2051-3305; 2051-3305
• DOI: 10.1049/joe.2019.1214

The non-linear characteristic of power lithium battery restricts the establishment of accurate battery models. To overcome this problem and estimate the battery state of charge (SOC) more accurately, the artificial fish swarm algorithm-back propagation (AFSA-BP) neural network structure was designed based on AFSA and BP neural network theory. According to the test parameters of power lithium battery, the related mathematical model was established. The flow charts of optimising BP neural network with AFSA algorithm and estimating SOC value by AFSA-BP algorithm are given. The specific implementation steps are elaborated. Using the 48 V, 50 Ah lithium iron phosphate (LiFePO4) power battery as experimental object, through the periodic charging and discharging experiments and software simulation, the correctness, validity and accuracy of the application of AFSA-BP neural network in estimating SOC value of the power lithium battery are verified.