Journal
ENERGY
Volume 208, Issue -, Pages -Publisher
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.energy.2020.118366
Keywords
Plug-in hybrid electric vehicles; Model predictive control; Dynamic programming; Neural network; Pontryagin's minimum principle
Categories
Funding
- National Natural Science Foundation of China [61763021, 51775063]
- National Key R&D Program of China [2018YFB0104900]
- EU [845102-HOEMEV-H2020-MSCA-IF-2018]
Ask authors/readers for more resources
Online optimal energy management of plug-in hybrid electric vehicles has been continually investigated for better fuel economy. This paper proposed a predictive energy management strategy based on multi neural networks for a multi-mode plug-in hybrid electric vehicle. To attain it, firstly, the offline optimal results prepared for knowledge learning are derived by dynamic programming and Pontryagin's minimum principle. Then, the mode recognition neural network is trained based on the optimal results of dynamic programming and the recurrent neural network is firstly exploited to realize online co-state estimation application. Consequently, the velocity prediction-based online model predictive control framework is established with the co-state correction and slacked constraints to solve the real-time optimal control sequence. A series of numerical simulation results validate that the optimal performance yielded from global optimal strategy can be exploited online to attain the satisfied cost reduction, compared with equivalent consumption minimum strategy, with the assistance of estimated real time co-state and slacked reference. In addition, the computation duration of proposed algorithm decreases by 23.40%, compared with conventional Pontryagin's minimum principle-based model predictive control scheme, thereby proving its online application potential. (C) 2020 Elsevier Ltd. All rights reserved.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available