融合网联信息的混合动力汽车工况预测与能量管理优化研究

It is hard to further improve the adaptability to driving cycles and fuel economy for hybrid electric vehicles (HEVs) without integrating vehicle networking information into the energy management. HEVs can access more traffic environment information in the networked environment, and their energy management involving multi-source information fusion is a multivariable, strong constraint optimization problem. It is an urgent issue to investigate how to integrate multi-source information into the energy management to improve the vehicle adaptability and economy due to the highly uncertainty for the vehicle operating conditions. To this end, this project presents a novel equivalent consumption minimization strategy (ECMS) based energy management method for HEVs in the networked environment. First, the data are extracted by constructing a traffic network model under urban conditions and then the driving cycle prediction model based on Bayesian network is established by fusing the internet information of vehicles. Second, the operation energy characteristics of HEVs are analyzed to reveal the impact mechanism of multi-energy state on equivalent factor (EF). Based on this, the energy spectrum is also constructed for HEVs over the prediction horizon to estimate the equivalent factor boundary of ECMS. Finally, a dynamic optimization method for EF is developed based on the predicted demand energy, which can be applied into ECMS energy management to coordinately optimize the gear decision and torque distribution over the prediction horizon. This project offers a solid theoretical foundation and technical support for improving the fuel economy as well as adjustability to driving cycles.

传统未融合网联信息的混合动力汽车能量管理对于提升工况适应性和整车经济性程度有限。网联环境下混合动力汽车可获取更多的交通环境信息，其能量管理涉及多源信息融合，是多变量、强约束优化问题。由于实际工况呈现高度不确定性，如何在能量管理中融合多源信息以提升车辆工况适应性和经济性是亟待解决的关键问题。本项目针对网联环境下混合动力汽车，采用等效消耗最小化方法（ECMS）实施能量管理。首先，构建城市工况交通路网模型以提取数据，建立融合网联信息的贝叶斯网络型工况预测模型；其次，分析混合动力汽车运行能量特征，揭示多能量状态对等效因子的影响机理；据此构建混合动力汽车预测时域能量图谱，并估计ECMS等效因子边界；最后，提出基于预测需求能量的等效因子动态优化方法，进而在预测时域采用ECMS优化能量分配，实现挡位决策与转矩分配的协同优化。本项目为提升网联环境下混合动力汽车燃油经济性和工况适应性提供理论基础和技术支持。

智能网联新能源汽车是今后汽车产业的重点方向。网联环境下，车辆可以获取更多的交通信息，有助于提升混合动力汽车工况预测精度与能量管理性能。本项目围绕融合网联信息的混合动力汽车工况预测与能量管理优化开展研究，针对现有能量管理方法难以适应复杂工况，节能效果提升程度有限的关键问题，开展了融合网联信息的混合动力汽车行驶工况预测、混合动力汽车预测时域内ECMS等效因子边界估计、混合动力汽车预测时域ECMS能量管理协同优化等工作。所取得的主要成果：.1. 针对网联信息对工况预测精度影响机制不清晰的问题，通过分析不同网联信息对预测精度的影响，提出了融合网联信息的贝叶斯网络型工况预测模型，实现了不确定车速预测。结果表明，多车场景下的预测结果优于两车场景下的预测结果。.2. 阐明了能量变化对等效因子的影响作用，估计了等效因子边界，并提出了基于预测信息的等效因子自适应调整方法，为设计ECMS能量管理方法奠定了基础。.3. 基于预测车速信息，提出了滚动时域内基于ECMS的能量管理方法，协同优化挡位决策与转矩分配，提升了算法工况适应性和计算效率。.4. 提出了基于能量变化的混合动力汽车自适应能量管理方法，并设计了基于能量的等效因子自适应调整方法，实现了能量管理优化的自适应分配。结果表明，与常数等效因子方法相比，根据能量变化来动态调整ECMS的等效因子，在燃油经济性和电池SOC平衡方面获得了明显优势。.上述研究成果解决了混合动力汽车能量管理算法工况适应性较差和节能效果受限的问题，对于提升混合动力汽车经济性具有重要的学术价值和工程应用前景。

内容获取失败，请点击重试