基于工况在线统计与电池能量随机存储优化的并联式混合动力系统能量管理策略研究

The real world driving conditions of vehicle is a stochastic dynamic process, the probability distribution of which was decided by the road environment, the intention of the driver, the vehicle status and other factors. Traditional energy management strategies of hybrid systems using a deterministic model to study this random process, which can only get an instantaneous optimal solution or globally optimal solution only applicable to a specific condition. These strategies were limited in practical applications. Considering the random characteristics of vehicle driving conditions, a statistical based decision-making and optimization method was used in this research. Energy management of hybrid system was simulated as a random inventory model which use battery as an energy storehouse, power access obey the real-turner conditions of statistical distribution of driving parameters. Energy distribution and management strategies were optimized to achieve the target of maximum expectations of energy-saving, based on online access of the statistical characteristics of vehicle driving conditions. Firstly, a quantitative analysis of energy flow relationship between multi-mode of parallel hybrid system was carried out, energy-saving modes and evaluation method of energy saving effects were extracted, which can be used as the value function for statistical optimization; Then, control objectives of SOC to achieve maximum energy-saving mode was determined by considering the statistical characteristics of driving parameters. Based on which, mode switching and power distribution law to achieve maximum expectation of energy saving were established using random storage optimization theory. Eventually, an energy management strategy which can adaptive the change of working conditions and can achieve maximum expectation of energy saving was established.

车辆实际工况是一个随机动态过程，其概率分布由道路环境、驾驶员意图、车辆状态等多种因素共同决定。传统的混合动力系统能量管理策略用确定性模型来研究这一随机过程，往往只能得到瞬时最优解或只适用于特定工况的全局最优解，在实际应用中受到极大限制。本研究针对车辆行驶过程的随机性特点，采用统计决策与优化的方法，将混合动力系统能量管理模拟为一个以电池为能量仓库，电能存取服从实车工况统计分布的随机存储模型，根据在线获取的车辆行驶工况统计特征，以系统获得最大节能期望值为目标进行能量的分配与管理。首先对并联式混合动力系统多模式能量流动关系进行定量分析，提取出各节能模式及节能效果定量评价方法，为统计优化提供价值函数；然后结合工况统计特征确定能使节能模式最大化实现的SOC控制目标，并根据随机存储优化理论获取节能期望值最大的模式切换与功率分配规律。最终形成一套能自适应工况变化、且能实现节能期望值最大化的能量管理策略。

车辆实际工况是一个动态过程，其统计特征由道路环境、驾驶员意图、车辆状态等多种因素共同决定。本研究针对车辆行驶过程的动态特征，采用统计决策与优化的方法，将混合动力系统能量管理模拟为一个以电池为能量仓库，电能存取服从实车工况统计分布的存储模型，根据在线获取的车辆行驶工况统计特征，以系统获得最大节能期望值为目标进行能量的分配与管理。研究首先对并联式混合动力系统各工况下系统效率与各主要部件效率及系统控制参数和状态参数之间的耦合关系进行分析，对并联式混合动力系统各工况以系统效率最佳为目标进行控制优化，仿真及测试结果显示，各工况平均节油率在2%左右；提出了基于电池能量等效燃油消耗率的节能评价方法，为系统优化及各工况间能量相互转换时的节能效果提供了定量评价手段；基于实时工况下制动能量统计特征的提取，以平均制动能量最大效率回收为目标优化SOC目标值，建立了并联式混合动力系统SOC目标值确定方法；提出了基于存储论的模式切换和功率分配策略，提出采用工况中各随机参数的统计平均值作为其随机性特征的表征，以提高算法实时性。仿真及试验显示，与传统的门限值策略相比发动机平均工作效率提升3%左右。本项目的研究内容和成果丰富了并联式混合动力系统能量管理策略，提出了一系列优化和管理思想，对车辆降低能耗、减少排放有较为重要的参考价值，有较好的应用前景。本项目所涉及的能量管理策略正应用到某型纯电动重卡动力总成开发中，现已在进行样车装配。项目发表论文8篇，其中EI已收录2篇，EI核心期刊录用2篇(有录用通知)。申请国家发明专利2项，培养硕士研究生6名(其中3人已毕业，3人在读)。