基于随机模型预测控制的插电式混合动力汽车 多目标能量管理策略

为了改善插电式混合动力汽车的燃油消耗和排放, 开展多目标随机模型预测控制策略的研究. 首先, 建立适用于模型预测的多元线性回归的发动机和电池模型, 建立融合燃油消耗和排放的多目标价值函数的模型预测控制, 随后, 基于随机驾驶员模型未来时刻的车速, 结合交通信息并利用动态规划(DP)算法进行参考电荷状态(SOC)优化, 进而建立多目标随机模型预测控制策略. 最后, 通过与DP, MPC等策略进行对比验证, 及给出两组不同权值进行多目标控制效果分析. 结果表明, 该策略的燃油消耗和排放最接近DP的控制效果, 且设置不同权重值可获得相应的控制目标, 说明该策略对提升燃油消耗和排放的多目标性能的有效性.

Multi-objective energy management strategy based on stochastic model predictive control for a plug-in hybrid electric vehicle

In order to improve the fuel consumption and emissions of plug-in hybrid electric vehicles, a multi-objective stochastic model predictive control strategy was studied. First of all, the engine and battery models of multiple linear regression suitable for model prediction are established, and the model predictive control of multi-objective value function integrating fuel consumption and emission is also established, then based on the random driver model of future time speed, combined with the traffic information and by using the dynamic programming (DP) algorithm to optimize the reference state of charge (SOC). Then the multi-objective stochastic model predictive control strategy is established. Finally, through the comparison and verification with the DP, MPC and other strategies, two groups of different weights are given to analyze the effect of multi-objective control. The results show that the fuel consumption and the emissions of this strategy are the closest to the control effect of the DP, and the corresponding control objectives can be obtained by setting different weight values, indicating the effectiveness of this strategy in improving the multi-objective performance of the fuel consumption and emissions.