融合遗传算法与动态规划的混合动力拖拉机等效燃油消耗最小策略优化方法

ECMS Optimization Method for Hybrid Tractor Based on Genetic Algorithm and Dynamic Programming

摘要:

针对混合动力拖拉机工况适应性差的问题,提出一种融合动态规划与遗传算法的分层优化策略。通过构建全局优化与瞬时控制协同决策框架,采用动态规划离线生成电池荷电状态(State of charge, SOC)全局参考轨迹,为等效燃油消耗最小策略(Equivalent consumption minimization strategy, ECMS)建立基准目标;利用遗传算法对初始等效因子进行多目标寻优,生成初始等效因子映射表,结合动态规划生成的SOC 轨迹,通过比例积分(Proportional-Integral, PI)控制实现等效因子的在线修正;通过全局轨迹跟踪与实时偏差调节的机制,缓解全局优化目标与瞬时工况的矛盾。结果表明,与传统ECMS 方法相比,通过遗传算法对等效因子进行优化可使能耗降低10.24% ,进一步引入SOC参考轨迹实时修正后,油耗额外降低1.59% ,验证了优化策略的有效性。

Abstract:

Aiming to address the issue of poor working condition adaptability in hybrid electric tractors by proposing a hierarchical optimization strategy that integrated dynamic programming and genetic algorithms, a collaborative decision-making framework combining global optimization and instantaneous control was established, where dynamic programming was employed to offline generate a global reference trajectory of battery SOC, providing a benchmark objective for the equivalent fuel consumption minimization strategy. When the actual working conditions differed in duration from those used in the dynamic programming for the reference SOC trajectory, a fitted approximate optimal SOC trajectory was adopted as the reference. Genetic algorithm was utilized to perform multi-objective optimization of initial equivalent factors, generating an initial equivalent factor map. A PI dynamic compensator was designed in combination with SOC tracking deviation to realize online correction of equivalent factors. The conflict between global optimization objectives and instantaneous working conditions was alleviated through a mechanism integrating global trajectory tracking and real-time deviation adjustment. Results demonstrated that compared with traditional ECMS methods, the genetic algorithm-optimized equivalent factors reduced energy consumption by 10.24%. After further introducing real-time SOC reference trajectory correction, an additional 1.59% fuel consumption reduction was achieved, which validated the effectiveness of the optimization strategy.