Scholarlink

<p>Improving energy efficiency and reducing emissions in hybrid electric vehicles is a critical global concern, driven by rising fuel costs and environmental regulations. Effective energy management strategies are essential for optimizing fuel usage, reducing engine wear, and improving overall vehicle performance. This research proposes a novel hybrid method that combines Recurrent Conditional Wasserstein Generative Adversarial Networks (RCWGAN) with the Harbor Seal Whiskers Optimization Algorithm (HSWOA), termed RCWGAN-HSWOA, to enhance the energy management system of Series Hybrid Electric Vehicles (SHEVs). The RCWGAN is used to predict the engine’s on/off state, while HSWOA optimizes energy usage by minimizing fuel consumption and reducing engine switching frequency to enhance drivability and comfort. The proposed method is put into practice on the MATLAB/Simulink platform and evaluated against benchmark models such as Q-Learning (QL), Deep Deterministic Policy Gradient (DDPG), and Multi-agent Deep Reinforcement Learning (MADRL). Results show that the RCWGAN-HSWOA achieves significantly lower fuel consumption at 17.5&#xa0;g, outperforming DDPG, QL, and MADRL respectively. The proposed approach outperformed the present strategy in terms of prediction accuracy, achieving the lowest Root Mean Square Error (RMSE) of 0.208%. Additionally, the proposed approach ensures smoother control transitions and improved system responsiveness. These outcomes validate the efficiency of the RCWGAN-HSWOA in improving the energy efficiency of SHEVs. In conclusion, the proposed strategy presents a promising solution for intelligent and fuel-efficient vehicle management systems, with potential implications for future green automotive technologies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Fuel efficiency enhancement in series hybrid electric vehicles through recurrent conditional wasserstein generative adversarial networks

  • 期刊论文

  • 发表日期：

    2025年6月11日

查看全文
  • S. S. Harish,
  • M. Ramesh Babu

摘要

Improving energy efficiency and reducing emissions in hybrid electric vehicles is a critical global concern, driven by rising fuel costs and environmental regulations. Effective energy management strategies are essential for optimizing fuel usage, reducing engine wear, and improving overall vehicle performance. This research proposes a novel hybrid method that combines Recurrent Conditional Wasserstein Generative Adversarial Networks (RCWGAN) with the Harbor Seal Whiskers Optimization Algorithm (HSWOA), termed RCWGAN-HSWOA, to enhance the energy management system of Series Hybrid Electric Vehicles (SHEVs). The RCWGAN is used to predict the engine’s on/off state, while HSWOA optimizes energy usage by minimizing fuel consumption and reducing engine switching frequency to enhance drivability and comfort. The proposed method is put into practice on the MATLAB/Simulink platform and evaluated against benchmark models such as Q-Learning (QL), Deep Deterministic Policy Gradient (DDPG), and Multi-agent Deep Reinforcement Learning (MADRL). Results show that the RCWGAN-HSWOA achieves significantly lower fuel consumption at 17.5 g, outperforming DDPG, QL, and MADRL respectively. The proposed approach outperformed the present strategy in terms of prediction accuracy, achieving the lowest Root Mean Square Error (RMSE) of 0.208%. Additionally, the proposed approach ensures smoother control transitions and improved system responsiveness. These outcomes validate the efficiency of the RCWGAN-HSWOA in improving the energy efficiency of SHEVs. In conclusion, the proposed strategy presents a promising solution for intelligent and fuel-efficient vehicle management systems, with potential implications for future green automotive technologies.