基于短时工况预测的混动拖拉机能量管理策略研究  

Research on Energy Management Strategy of Hybrid Tractor Based on Short-time Working Condition Prediction

作　　者：南祥 洪达 缪浩楠 唐小康 解玄 Nan Xiang;Hong Da;Miu Hao-nan;Tang Xiao-kang;Xie Xuan(School of Automotive and Traffic Engineering,Jiangsu University,Jiangsu Zhenjiang 212000)

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212000

出　　处：《内燃机与配件》2024年第21期1-4,共4页Internal Combustion Engine & Parts

基　　金：江苏省高等学校大学生创新创业训练计划省级重点项目(202310299038Z)。

摘　　要：针对油电混动拖拉机利用DP算法无法实现实时寻优的问题,为了实时将DP算法与拖拉机能量管理策略联系起来,利用模型预测的控制策略,设计一种基于车速预测并通过加载一段预测视界t_(p)的确定性时间-速度序列,并利用DP算法求解全局优化的优点求解预测视界的最佳控制方式。本文利用MATLAB/Simulink搭建油电混动拖拉机的DP算法仿真模型,根据预测速度计算出预测视界长度内的需求功率序列,利用动态规划算法求解了预测视界长度内的需求功率序列,进而去控制最佳的发动机燃油消耗。仿真结果表明:相比基于规则的EMS,提升了5.18%的油耗,进一步验证了该工况预测方法的有效性和自适应性,效果略差于全局最优的动态规划算法,但是解决了动态规划算法无法在线优化的问题。Aiming at the problem that DP algorithm cannot realize real-time optimization of hybrid tractor,in order to link DP algorithm with tractor energy management strategy in real time,a deterministic time-velocity sequence based on vehicle speed prediction and loading a segment of forecast horizon t_(p) is designed by using the control strategy of model prediction.The optimal control mode of prediction horizon is solved by using the advantage of DP algorithm to solve global optimization.In this paper,MATLAB/Simulink is used to build a simulation model of DP algorithm for the hybrid tractor.The required power sequence within the predicted horizon length is calculated according to the predicted speed,and the required power sequence within the predicted horizon length is solved by dynamic programming algorithm,so as to control the optimal engine fuel consumption.The simulation results show that compared with the rule-based EMS,the fuel consumption is improved by 5.18%,which further verifies the effectiveness and adaptability of the working condition prediction method.The effect is slightly worse than that of the globally optimal dynamic programming algorithm,but it solves the problem that the dynamic programming algorithm cannot be optimized online.

关 键 词：油电混动拖拉机 DP算法 神经网络 车速预测 能量管理 

分 类 号：S219.0[农业科学—农业机械化工程]