Improved K-means clustering-genetic backpropagation modeling for online state-of-charge estimation of lithium-ion batteries adaptive to low-temperature conditions.

Hai, Nan; Wang, Shunli; Huang, Qi; Xie, Yanxin; Fernandez, Carlos

Abstract

Accurate state-of-charge (SOC) estimation of lithium-ion batteries (LIBs) in low temperatures is significant to maximize their performance and application. An improved K-means clustering-genetic backpropagation (KMC-GBP) algorithm consisting of five innovative parts is established to achieve the goal. Specifically, an improved KMCG algorithm that redefines adaptation functions and introduces the K operator into genetic manipulation. Then, an improved genetic algorithm with dynamic search, matching a high selection rate to high adaptation, and the crossover and mutation probability led by the K factor will be adjusted based on individual behavior, is proposed to avoid falling into the local optimum. Alternatively, an adaptive elitist presentation strategy and an improved substitution strategy are induced. Moreover, a further performance comparison of variable algorithms is made under different working conditions at variable temperatures to prove the effectiveness. The experimental results showed that the maximum error of the IKMC-GBP reached 0.356%, 0.373%, and 0.380% at −5°C, −15°C, and −35°C under BBDST. Similarly, it reached 0.267%, 0.022%, 0.004% at −10°C, −20°C and −30°C under DST.

Citation

HAI, N., WANG, S., HUANG, Q., XIE, Y. and FERNANDEZ, C. 2024. Improved K-means clustering-genetic backpropagation modeling for online state-of-charge estimation of lithium-ion batteries adaptive to low-temperature conditions. Journal of energy storage [online], 99(B), article number 113399. Available from: https://doi.org/10.1016/j.est.2024.113399

• Journal Article Type: Article
• Acceptance Date: Aug 13, 2024
• Online Publication Date: Aug 21, 2024
• Publication Date: Oct 10, 2024
• Deposit Date: Aug 23, 2024
• Publicly Available Date: Aug 22, 2025
• Journal: Journal of energy storage
• Print ISSN: 2352-152X
• Electronic ISSN: 2352-1538
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 99
• Issue: B
• Article Number: 113399
• DOI: https://doi.org/10.1016/j.est.2024.113399
• Keywords: K-means; Genetic algorithm; Backpropagation; State of charge; Lithium-ion batteries
• Public URL: https://rgu-repository.worktribe.com/output/2440712