Scholarlink

<p>Series hybrid vehicles under variable-duty operating conditions face rapid power-demand changes, frequent load fluctuations, and transitions between low-speed and high-power operation, which make real-time energy scheduling difficult. To coordinate adaptive power allocation and efficient engine–generator operation, this study proposes a two-layer energy management framework for series hybrid special-purpose vehicles. At the lower layer, an optimization-oriented coordinated controller maps the generation-power command to the optimal engine speed and generator torque based on the composite efficiency map of the engine–generator unit. At the upper layer, a Twin-Delayed Deep Deterministic Policy Gradient (TD3) scheduler determines the desired generation power according to vehicle speed, acceleration, power demand, and battery state of charge. The framework is modeled and trained in MATLAB/Simulink and validated through simulations and Speedgoat-based hardware-in-the-loop tests. Results show that the proposed strategy limits the speed-tracking error within 3&#xa0;km/h, maintains stable SOC regulation under the tested scenarios, and reduces equivalent fuel consumption by approximately 7.5% compared with the rule-based strategy under the NEDC cycle. It also shows a 7.6% deviation from the DP benchmark, indicating near-optimal fuel economy with real-time implementability under variable-duty conditions.</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

Energy Management of Series Hybrid Special-Purpose Vehicles Using a Two-Layer Architecture with Coordinated Engine–Generator Control and TD3 Algorithm

  • 期刊论文

  • 发表日期：

    2026年8月3日

查看全文
  • Zi Qiang Luo,
  • Hui Jing,
  • Jin Xia Yang,
  • Xiong Chao Mo,
  • Hou Jia Qin,
  • Tao Wang,
  • Cong Li

摘要

Series hybrid vehicles under variable-duty operating conditions face rapid power-demand changes, frequent load fluctuations, and transitions between low-speed and high-power operation, which make real-time energy scheduling difficult. To coordinate adaptive power allocation and efficient engine–generator operation, this study proposes a two-layer energy management framework for series hybrid special-purpose vehicles. At the lower layer, an optimization-oriented coordinated controller maps the generation-power command to the optimal engine speed and generator torque based on the composite efficiency map of the engine–generator unit. At the upper layer, a Twin-Delayed Deep Deterministic Policy Gradient (TD3) scheduler determines the desired generation power according to vehicle speed, acceleration, power demand, and battery state of charge. The framework is modeled and trained in MATLAB/Simulink and validated through simulations and Speedgoat-based hardware-in-the-loop tests. Results show that the proposed strategy limits the speed-tracking error within 3 km/h, maintains stable SOC regulation under the tested scenarios, and reduces equivalent fuel consumption by approximately 7.5% compared with the rule-based strategy under the NEDC cycle. It also shows a 7.6% deviation from the DP benchmark, indicating near-optimal fuel economy with real-time implementability under variable-duty conditions.