IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 67, NO. 1, JANUARY 2018 315 Power Split Strategy Optimization of a Plug-in Parallel Hybrid Electric Vehicle Nicolas Denis , Member, IEEE, Maxime R. Dubois , Member, IEEE, Jo˜ ao Pedro F. Trov˜ ao , Senior Member, IEEE, and Alain Desrochers Abstract—Hybrid electric vehicles (HEV), plug-in HEV (PHEV) need an energy management system (EMS) to ensure good fuel economy while maintaining battery state-of-charge (SOC) within a safe range. The EMS is in charge of the power split decision be- tween the engine and the electrical motor. For a PHEV, the optimal power split scenario will depend on the driving cycle, initial SOC, and trip length. Heavy computation and accurate knowledge of the future trip are required to find the optimal power split control and this represents a significant difficulty for the development of an EMS. The aim of this paper is to propose a genetic algorithm (GA) that optimizes the power split control parameters for a given driv- ing cycle in a relatively short computation time, thus, overcoming the problem of heavy computation. The methodology consists in 1) defining the control laws and their associated control pa- rameters based on the observation of optimality obtained by dynamic programming; and 2) developing a GA that will be able to compute the near-optimal values of these parameters in a short time and for a given driving cycle. It is demonstrated that the GA provides short computational burden and near-optimality for a wide variety of driving cycles. It then offers a promising tool for a future real-time implementation. Index Terms—Energy management system (EMS), genetic algo- rithm (GA), plug-in hybrid electric vehicles (PHEV), power split strategy, three-wheel electric vehicle (EV). I. INTRODUCTION S INCE the end of the XXth century, the automotive industry is facing the challenge of reducing fuel use and emissions caused by transportation. A promising solution is, of course, ve- hicle electrification. Since the release of the Toyota Prius, which is the first modern commercial hybrid electric vehicle, many car Manuscript received March 11, 2017; revised July 23, 2017; accepted Septem- ber 7, 2017. Date of publication September 26, 2017; date of current version January 15, 2018. This work was supported in part by the Automotive Partner- ship Canada program and in part by the Canada Research Chairs Program. The review of this paper was coordinated by the Guest Editors of the VPPC 2016 Special Section. (Corresponding author: Jo˜ ao Pedro F. Trov˜ ao.) N. Denis is with Challenergy Inc., Tokyo 131-0031, Japan (e-mail: nicolas. denis@challenergy.com). M. R. Dubois is with the Department of Electrical and Computation Engi- neering, Universit´ e de Sherbrooke, Sherbrooke, QC J1K 2R1, Canada (e-mail: Maxime.Dubois@USherbrooke.ca). J. P. F. Trov˜ ao is with the Department of Electrical Engineering and Computer Engineering, Universit´ e de Sherbrooke, Sherbrooke, QC J1K 2R1, Canada, and also with the Institute for Systems and Computers Engineering at Coimbra, Coimbra 3030-290, Portugal (e-mail: Joao.Trovao@USherbrooke.ca). A. Desrochers is with the Department of Mechanical Engineering, Universit´ e de Sherbrooke, Sherbrooke, QC J1K 2R1, Canada (e-mail: Alain.Desrochers@ USherbrooke.ca). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TVT.2017.2756049 manufacturers have launched hybrid electric vehicles (HEV), plug-in hybrid electric vehicles (PHEV), and pure electric ve- hicles (EVs). EVs only have one electrical motor (EM) and no internal combustion engine (ICE). As a consequence, they do not consume fuel while running. However, the main restriction toward a wide-spread use of this technology is the high cost of the battery for an acceptable driving range. Hybrid technology, used in both HEVs and PHEVs, over- comes these drawbacks by combining an ICE and an EM in the powertrain. This allows a significant increase in the driv- ing range and a better fuel economy compared to conventional vehicles. Therefore, PHEVs are considered as long-term tech- nologies by some authors [1], [2] as they promise a better fuel economy compared to HEV. Hybrid powertrain can be made with various configurations: parallel, series, mild hybrid, series- parallel. The parallel hybrid configuration has the advantage of a downsized EM and drive compared to the series configuration. In this paper, only the parallel hybrid configuration with one EM and one ICE will be studied. PHEVs include a high-level controller, named energy management system (EMS), which minimizes the vehicle fuel consumption and manages the bat- tery state-of-charge (SOC) by accurately assigning the required power level to be provided by the EM and the ICE. In the parallel hybrid configuration, a key control parameter is the percentage of the demanded power to be supplied by the ICE, hereafter called the power split ratio (PSR). The controller must then es- tablish if the ICE should be off (PSR of 0%), if the required power should be split between the EM and the ICE (PSR be- tween 0% and 100%) or if the ICE should be overdriven (PSR over 100%), with the EM operated in generator mode. Control strategies are usually classified into rule-based and optimization-based categories for both HEV [3] and PHEV [4]. The rule-based strategies are based on deterministic [5], [6] or fuzzy rules [7]–[9]. Generally, these rule-based strategies make the PHEV run in pure electric mode like an EV dur- ing the first part of the trip, and then impose the use of the ICE to perform charge sustaining (CS) operation like an HEV. Rule-based strategies are widely used in industry because of their simplicity and reliability; however, they exhibit a lack of optimality because they only rely on the current driving con- ditions [4]. Optimization-based strategies use a mathematical approach to determine the optimal PSR. Global optimization has been particularly investigated as it is able to find the optimal control sequence of the two motors over a complete prede- fined driving cycle. A very common global optimization tool is 0018-9545 © 2017 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications standards/publications/rights/index.html for more information.