SOC prediction of Volterra adaptive filter based on chaotic time series

• Published in: AIP advances Vol. 12; no. 11; pp. 115113 - 115113-9
• Main Authors: Yu, Zhilong, Na, Haipeng, Qi, Lihua, Li, Ran
• Format: Journal Article
• Language: English
• Published: Melville American Institute of Physics 01.11.2022; AIP Publishing LLC
• Subjects: Adaptive filters; Algorithms; Chaos theory; Electric potential; Electric vehicles; Error analysis; Lithium; Lithium-ion batteries; Multidimensional data; Parameters; Rechargeable batteries; Reconstruction; State of charge; Time series; Voltage
• ISSN: 2158-3226; 2158-3226
• DOI: 10.1063/5.0113098

This paper presents an SOC (State of Charge) prediction method based on a chaotic time series Volterra adaptive filter. This method first verifies the chaotic characteristics of the SOC time series of lithium-ion batteries and then implements the phase space reconstruction method to reorganize the voltage and current data into multi-dimensional data. Furthermore, we use the Volterra adaptive filter to predict the SOC of the lithium-ion battery pack on the reconstructed multi-dimensional data. This method can effectively solve the problem that the SOC estimation error increases due to the poor real-time performance of battery voltage, current, and other parameters of vehicular lithium-ion battery packs in electric vehicles obtained by the battery management system. According to the simulation and test results, the error of the SOC estimation algorithm studied in this paper is less than 0.6%, which indicates that the Volterra adaptive filter based on chaotic time series comprehensively uses the linear and non-linear parameter characteristics of lithium-ion batteries and recombines the voltage and current data into multi-dimensional data through phase space reconstruction technology. It can reveal the relationship between the battery SOC and other variables better, achieve higher precision lithium-ion battery SOC prediction, and have better realizability.