基于数字孪生和PSO算法的混动车辆能量管理策略鲁棒优化

doi:10.3969/j.issn.1674-8484.2022.03.013

摘要/Abstract

摘要：

该文针对混动车辆能量管理策略开发任务,基于车辆数字孪生平台,提出了一种融合全局交叉验证和粒子群优化(PSO)的鲁棒优化算法,以获得高可靠性、适应性的能量管理策略。基于转鼓台架试验结果建立了某混动车辆数字孪生模型,定义了综合考虑车辆能量转换效率和电池剩余电量的控制效用指标,搭建了基于自适应神经模糊推理系统(ANFIS)能量管理控制器;利用粒子群鲁棒优化算法在JC08、WLTC、UDDS等国际常用行驶工况对控制器进行超参数优化,并基于硬件在环平台对优化结果进行了对比验证。结果表明：通过综合考虑训练工况和验证工况下的控制效用,粒子群鲁棒优化算法相比标准粒子群算法,能够提升11%以上的控制效用值,获得0.41%至27.92%的燃油经济性提升。

关键词: 混合动力汽车, 能量管理策略, 粒子群优化(PSO), 鲁棒优化, 交叉验证

Abstract:

A robust particle swarm optimization (PSO) scheme for the development of energy management strategy for hybrid vehicles was proposed based on digital twin. By incorporating global cross-validation with local particle swarm optimization, the proposed scheme aimed to achieve more reliable optimization for energy management. First, a digital twin model for the hybrid vehicle was built based on the chassis dynamometer test data and an adaptive neural fuzzy inference system (ANFIS) was then developed for real-time energy management. By introducing the concept of control utility, which evaluated the vehicle energy efficiency with a penalty factor of battery usage, the robust particle swarm optimisation scheme was deployed to optimize the hyper parameters of the ANFIS controller. The optimization performances were evaluated through experiment based on the hardware-in-the-loop testing platform under worldwide driving cycles including JC08, WLTC, and UDDS. Compared to conventional particle swarm optimisation, the proposed robust particle swarm optimization can achieve more than 11% higher control utility value in both learning cycles and testing cycles and improve the fuel economy by up to 27.92%.

Key words: hybrid electric vehicle, energy management strategy, particle swarm optimization (PSO), robust optimization, cross-validation

中图分类号:

U469.7

周泉, 张策腾飞, 李雁飞, 帅斌, 徐宏明. 基于数字孪生和PSO算法的混动车辆能量管理策略鲁棒优化[J]. 汽车安全与节能学报, 2022, 13(3): 517-525.

ZHOU Quan, ZHANG Cetengfei, LI Yanfei, SHUAI Bin, XU Hongming. Robust optimization of energy management strategy in hybrid vehicles based on digital twin and PSO algorithm[J]. Journal of Automotive Safety and Energy, 2022, 13(3): 517-525.

图/表 13

图1 混动车辆数字孪生系统

车辆质量	1 315 kg
车轮旋转半径	0.35 m
主减速比	9.655
电池额定电压	352 V
电池安全充放 / 电流	413 A / 16 A

表1 车辆参数

图2 基于自适应神经模糊推理的能量管理策略

图3 鲁棒粒子群优化算法

图4 折-集划分规则

图5 折内优化后的ANFIS模型输入输出关系比较

图6 基于ANFIS与动态规划的控制效果对比

图7 电池SOC轨迹追踪的绝对误差分布

	第1集	第2集	第3集	综合效用
第1折	0.199 2	0.273 6	0.225 7	0.199 2
第2折	0.172 1	0.278 6	0.206 7	0.172 1
第3折	0.179 3	0.274 4	0.226 7	0.179 3

表2 折内优化和验证条件下的效用矩阵

图8 高电量状态硬件在环试验结果

图9 电量保持阶段硬件在环试验结果

	$U$ _high	提升 / %	$U$ _main	提升 / %
鲁棒粒子群能优化	0.195 8	—	0.094 9	—
粒子群优化对照组1	0.190 6	2.66	0.094 6	0.32
粒子群优化对照组2	0.168 5	13.94	0.094 3	0.63

	$U$ _high	提升 / %	$U$ _main	提升 / %
鲁棒粒子群能优化	0.195 8	—	0.094 9	—
粒子群优化对照组1	0.190 6	2.66	0.094 6	0.32
粒子群优化对照组2	0.168 5	13.94	0.094 3	0.63

表3 复合工况控制效用对比

	V_high / L	提升 / %	V_main / L	提升 / %
鲁棒粒子群能优化	2.883	—	4.599	—
粒子群优化对照组1	3.006	4.27	4.636	0.80
粒子群优化对照组2	3.688	27.92	4.618	0.41

表4 复合工况100 km油耗对比

参考文献 25

[1]	TAO Fei, ZHANG He, LIU Ang, et al. Digital twin in industry: State-of-the-art[J]. IEEE Trans Ind Info, 2019, 15(4): 2405-2415.
[2]	徐宏明, 周泉. 人工智能在发动机控制开发中的应用及前景[J]. 汽车安全与节能学报, 2021, 12(2): 150-162.
	XU Hongming, ZHOU Quan. Artificial intelligence technologies for engine control development: State-of-the-art review and outlook[J]. J Auto Safe Energ, 2021, 12(2): 150-162. (in Chinese)
[3]	International Energy Agency. World energy outlook 2020[R/OL]. (2020-10-13), https://www.iea.org/events/world-energy-outlook-2020.
[4]	新能源汽车产业发展规划(2021—2035年)[R]. 国务院办公厅,北京,2020.
	New energy vehicle industry development plan (2021-2035)[R]. General Office of the Chinese State Council, Beijing,2020. (in Chinese)
[5]	HUANG Yanjun, WANG Hong, Khajepour A, et al. A review of power management strategies and component sizing methods for hybrid vehicles[J]. Rene Sustain Energ Rev, 2018, 96: 132-144.
[6]	杨超, 杜雪龙, 王伟达, 等. 智能网联环境下的 PHEV 实时优化能量管理策略研究[J]. 汽车安全与节能学报, 2021, 12(2): 210-218.
	YANG Chao, DU Xuelong, WANG Weida, et al. Research on real-time optimization energy management strategy of PHEV under intelligent networking environment[J]. J Autom Safe Energ, 2021, 12(2): 210-218. (in Chinese)
[7]	PENG Jiankun, HE Hongwen, XIONG Rui. Rule based energy management strategy for a series-parallel plug-in hybrid electric bus optimized by dynamic programming[J]. Appl Energ, 2016, 185: 1633-1643. doi: 10.1016/j.apenergy.2015.12.031 URL
[8]	ZHOU Quan, ZHAO Dezong, SHUAI Bin, et al. Knowledge implementation and transfer with an adaptive learning network for real-time power management of the plug-in hybrid vehicle[J]. IEEE Trans Neur Net Learn Syst, 2021, 32(12): 5298-5308.
[9]	XIE Shaobo, HU Xiaosong, QI Shanwei, et al. An artificial neural network-enhanced energy management strategy for plug-in hybrid electric vehicles[J]. Energy, 2018, 163: 837-848. doi: 10.1016/j.energy.2018.08.139 URL
[10]	HUANG Yanjun, WANG Hong, Khajepour A, et al. Model predictive control power management strategies for HEVs: A review[J]. J Power Sources, 2017, 341: 91-106. doi: 10.1016/j.jpowsour.2016.11.106 URL
[11]	Ali A M, Ghanbar A, Soffker D. Optimal control of multi-source electric vehicles in real time using advisory dynamic programming[J]. IEEE Trans Vehi Tech, 2019, 68(11): 10394-10405.
[12]	Mura R, Utkin V, Onori S. Energy management design in hybrid electric vehicles: A novel optimality and stability framework[J]. IEEE Trans Contr Syst Tech, 2015, 23(4): 1307-1322. doi: 10.1109/TCST.2014.2363617 URL
[13]	LIU Teng, ZOU Yuan, LIU Dexing, et al. Reinforcement learning of adaptive energy management with transition probability for a hybrid electric tracked vehicle[J]. IEEE Trans Ind Elect, 2015, 62(12): 7837-7846. doi: 10.1109/TIE.2015.2475419 URL
[14]	SHUAI Bin, ZHOU Quan, LI Ji, et al. Heuristic action execution for energy efficient charge-sustaining control of connected hybrid vehicles with model-free double Q-learning[J]. Appl Energ, 2020, 267, Paper ID: 114900.
[15]	XU Bin, HU Xiaosong, TANG Xiaolin, et al. Ensemble reinforcement learning-based supervisory control of hybrid electric vehicle for fuel economy improvement[J]. IEEE Trans Transport Electrificat, 2020, 6(2):717-727. doi: 10.1109/TTE.2020.2991079 URL
[16]	ZHOU Quan, LI Ji, SHUAI Bin, et al. Multi-step reinforcement learning for model-free predictive energy management of an electrified off-highway vehicle[J]. Appl Energ, 2019, 255: 113755. doi: 10.1016/j.apenergy.2019.113755 URL
[17]	ZHOU Quan, LI Yanfei, ZHAO Dezong, et al. Transferable representation modelling for real-time energy management of the plug-in hybrid vehicle based on k-fold fuzzy learning and Gaussian process regression[J]. Appl Energ, 2022, 305: 117853. doi: 10.1016/j.apenergy.2021.117853 URL
[18]	刘灵芝, 张冰战, 蒋通. 基于工况自适应的 PHEV 等效燃油最小策略[J]. 汽车安全与节能学报, 2020, 11(3): 371-378.
	LIU Lingzhi, ZHANG Bingzhan, JIANG Tong. Equivalent consumption minimization strategy for PHEV based on driving condition adaptation[J], 2020, 11 (3): 371-378. (in Chinese)
[19]	ZHOU Quan, ZHANG Yunfan, LI Ziyang, et al. Cyber-physical energy-saving control for hybrid aircraft-towing tractor based on online swarm intelligent programming[J]. IEEE Trans Ind Info, 2018, 14(9): 4149-4158.
[20]	HE Yinglong, ZHOU Quan, Makridis M, et al. Multiobjective co-optimization of cooperative adaptive cruise control and energy management strategy for PHEVs[J]. IEEE Trans Transport Electrificat, 2020, 6(1): 346-355. doi: 10.1109/TTE.2020.2974588 URL
[21]	LI Yangyang, WANG Shuqian, DUAN Xiongbo, et al. Multi-objective energy management for Atkinson cycle engine and series hybrid electric vehicle based on evolutionary NSGA-II algorithm using digital twins[J]. Energ Conv Manag, 2021, 230: 113788. doi: 10.1016/j.enconman.2020.113788 URL
[22]	张涛, 章佳莹, 王凌云, 等. 基于云模型粒子群算法的海岛交直流混合微网鲁棒优化运行[J]. 信息与控制, 2019, 48(4): 385-391. doi: 10.13976/j.cnki.xk.2019.8402
	ZHANG Tao, ZHANG Jiaying, WANG Lingyun, et al. Obust optimization operation of island AC-DC hybrid micro-grid based on cloud model particle swarm optimization[J]. Info Contr, 2019, 48(4): 385-39. (in Chinese)
[23]	TAO Jun, SUN Gang, WANG Xinyu, et al. Robust optimization for a wing at drag divergence Mach number based on an improved PSO algorithm[J]. Aero Sci Tech, 2019, 92: 653-667. doi: 10.1016/j.ast.2019.06.041 URL
[24]	MA Guoqi, WANG Binquan, GE Shuzhi. Robust optimal control of connected and automated vehicle platoons through improved particle swarm optimization[J]. Transport Res Part C: Emerg Tech, 2022, 135: 103488. doi: 10.1016/j.trc.2021.103488 URL
[25]	LI Ji, ZHOU Quan, HE Yinglong, et al. Driver-identified supervisory control system of hybrid electric vehicles based on spectrum-guided fuzzy feature extraction[J]. IEEE Trans Fuzz Syst, 2020, 28(11): 2691-2701.