Energy Informatics, Год журнала: 2024, Номер 7(1)

Опубликована: Март 14, 2024

Abstract This research addresses the issue of State Charge (SOC) prediction for electric vehicle batteries by employing a dynamic Kalman neural network model. The model is optimized using Genetic algorithm to adjust weights. Additionally, strategy involving support vector machines optimization proposed. involves preprocessing data, selecting appropriate kernel functions training, and merging results enhance stability Results indicated that Dynamic Neural Network (DGKNN) achieved minimum error percentage only 0.1529% when correction coefficient was set 0.7. DGKNN consistently exhibited lowest percentage, average absolute error, mean square root handling small, medium, large datasets. For instance, in small dataset, 0.1518, 0.0604. findings demonstrated proposed high real-time accuracy predicting battery SOC, enabling monitoring operating parameters. method this study can accurately predict state charge, extend life packs, improve performance vehicles. It has important significance promoting development industry.

Язык: Английский