A novel combinatorial optimization algorithm for energy management strategy of plug-in hybrid electric vehicle

https://doi.org/10.1016/j.jfranklin.2017.08.020Get rights and content

Highlights

Abstract

Optimization design of energy management strategy (EMS) for plug-in hybrid electric vehicle (PHEV), which significantly affects the vehicle performance on fuel economy and pollutant emission, has always been a focal issue. Of various EMSs, rule-based strategies are dominant in practical applications due to their relatively low computational burden, but to obtain the optimum control parameters precisely and efficiently remains an unsolved problem. In this paper, a novel combinatorial algorithm utilizing the historical data from remote monitoring platform is proposed for the EMS optimization of PHEV. Firstly, the historical driving data are processed, and then a table which records different conditions at different time is built for reflecting the future PHEV operation schedule. Based on the historical data, a combinatorial algorithm which combines the advantages of genetic algorithm (GA) with enhanced ant colony algorithm (EACA) is proposed to optimize the control parameters. The principle of algorithm transformation from GA to EACA is when the objective function value is smaller than the default value after five generations of changing continuously in GA optimization process, and then the control parameter combinations can be regarded as the pheromone for EACA. Results show that the combinatorial algorithm successfully overcomes the low solution precision by GA and the slow resolving speed by EACA. The energy consumption of PHEV on a specific bus route can be reduced greatly by the proposed method, and it can provide a theoretical guidance for practical applications.

Introduction

Nowadays, the dwindling fossil fuel supplies and the severe pollution caused by motor vehicles have aroused grave concern. To address this issue, the international automotive industry has actively been probing new solutions in the past decades. In this process, electric vehicle (EV), hybrid electric vehicle (HEV) and plug-in hybrid electric vehicle (PHEV) have made great progress [1], [2], [3]. PHEV, being equipped with high-capacity battery, has become a promising new mode of automobiles which can travel long distances and operate in complex hybrid running conditions. Plug-in hybrid electric bus (PHEB) has emerged as an important way to lower fuel consumption in urban transit area [4], [6]. However, the performance of PHEB is heavily dependent on energy management strategy (EMS), which in turn depends on advanced optimization algorithm.

To achieve excellent fuel economy of a PHEV, numerous significant research efforts on optimal energy management have been carried out in the past years. Dynamic programming (DP) is used to find the global optimal strategy in known conditions [7], [8]. Peng et al proposed a method to improve the performance of the rule-based energy management based on the results given by a DP algorithm [7]. Considering the influence of air conditioning in the bus on fuel consumption and electric consumption, an optimized EMS based on corrected DP is proposed in [9]. However, due to the highly complex computation of DP, and the prerequisite that the future driving cycle is already known to achieve the optimal results, it is difficult to apply DP online directly. An optimal control strategy based on Pontryagin's Minimum Principle (PMP) is a promising solution because it provides a much simpler solution for EMS of PHEVs [10], [12]. Hou et al proposes an optimized EMS based on the approximate PMP algorithm for parallel PHEV, and an optimal solution close to the DP strategy is obtained, which proves the applicability of PMP to energy management problem [11]. Although PMP has relatively small computational complexity, real-time computing is still difficult to achieve. As mentioned above, driving condition information is critical to EMS in PHEV. Model predictive control (MPC) offers a predictive scheme that future cycle information can be incorporated into various EMSs [5], [13]. Based on the current position of the vehicle, driving direction and geographic information, Zeng et al proposed an energy management strategy based on MPC which considers slope information [14]. Considering the coupling relationship between the parameters of vehicle transmission system and the EMS, convex optimization is used to optimize both of them at the same time, and the optimization target of traditional system parameters is the same as that of the optimization goal of EMS, thus obtaining optimization of the two aspects at the same time [15], [16], [17], [18]. For energy management problem of PHEV, the commonly used convex optimization method is to convert the complex nonlinear problem into semi-definite problem, which can reduce the computing time to a certain extent, but its real-time applications are still very difficult. Different from the traditional state-feedback or output feedback control scheme [19], [20], [21], the key point of the energy management strategy of PHEV is the optimization of energy flow in hybrid powertrain. During the PHEV controller design procedure, the optimization result of the energy consumption should be concerned about more than the dynamic control performance. In this study, the transient control performance is also not taken into account.

In addition to the methods mentioned above, some other rule-based optimization algorithms have been proposed to solve the EMS problem of PHEV [22], [23], [24], [25], [26], [27], [28]. At the present stage, the optimal results for these algorithms can be obtained by offline operation for some fixed driving cycle, and then the optimization of PHEV energy management is applied to an online strategy. For a specific rule-based strategy, the combination of control parameters has great influence on fuel economy and dynamic performance. S. Bashash et al. set energy cost and battery longevity as the optimization target, used Particle Swarm Optimization (PSO) to optimize battery charging and then the optimizations with different electricity prices and mileages are discussed in detail [23]. Wu. et al. considered fuel consumption, exhaust emission, and manufacturing cost of HEV, using the weighting coefficient method to convert the multi-objective optimization problem into single objective optimization, and the PSO is used to obtain the optimal solution. The results show that the method can reduce the energy consumption effectively without sacrificing the dynamic performance [24]. Zhang et al. optimized the parameters of the power system components and the control strategy parameters of the plug-in series hybrid vehicle by using Genetic Algorithm (GA). Simulation results show that the fuel economy is increased by 8.97% and emissions have also been greatly reduced [25]. However, PSO and GA are easy to fall into local optimum and thus are not able to reach global optimization results. And for these methods, much redundant iteration is likely to occur in the later optimization process to search precision results, which leads to low efficiency. Pourhashemi et al. proposed a optimized EMS based on Ant Colony Optimization (ACO) for fuel cell hybrid electric vehicle. However, the speed of ACO is slow due to the lack of feedback pheromones [26].

Based on the above analyses, the rule-based optimization strategies are real-time operation EMS, but it is difficult to obtain desired results using a single optimization algorithm. Therefore, many researchers combine several algorithms to solve the energy management problem in recent years. Chen et al. proposed a power-split strategy based on GA and quadratic programming for PHEV. The optimized fuel economy show obvious improvements [27]. Hui proposed an adaptive Simulated Annealing-Genetic Algorithm (SA-GA) to improve the performance of the vehicle with improved fuel economy [28]. These methods solve the problem of falling into local optimum but convergence is still slow. In addition, these optimized EMS methods show their effectiveness only in the driving cycle preselected. For other driving cycles they do not necessarily give good energy economy.

With the rapid development of wireless network communication technology, remote monitoring technology in intelligent transportation system (ITS) has begun to prove its great potential for urban transportation management, which makes the traffic condition more easily available [29], [30], [31]. So in this paper, a representative PHEB with parallel hybrid powertrain is studied, and a novel combinatorial optimization algorithm, which combines GA with enhanced ant colony algorithm (EACA) is proposed. The proposed algorithm takes advantage of the fact that GA provides good optimization in the first part of iteration to generate groups of variables for the latter part, and then turn them into initial pheromones used in EACA, in which adaptive crossover operation are used to accelerate convergence and prevent local optimization. In this way, the redundancy operation in later period of GA is avoided, and the low-efficiency problem caused by the lack of initial pheromones of ACA is solved at the same time.

The paper is organized as follows. Models and preparative optimization parameters are described in Section 2. The optimization design for EMS including the processing of driving cycle data and the combinatorial optimization algorithm is proposed in Section 3. The optimization results and discussions are given in Section 4. Finally, the conclusions and future work are given in Section 5.

Section snippets

Configurations and models of PHEB

With the development of powertrain technology, there have been a variety of configurations of hybrid powertrain. In this paper, we focus on a typical parallel hybrid powertrain with automated mechanical transmission (AMT) [32], [33]. The diagram of this powertrain is shown in Fig. 1.

As seen from Fig. 1, for the coaxial parallel configuration, there are two power sources, engine and electric motor (EM). In order to decrease energy consumption, the combination of proper operating modes can be

Combinatorial optimization Algorithm-based EMS for PHEBs

In this section, a combinatorial optimization algorithm, which combines the advantage of GA and EACA, is proposed as the rule-based optimization algorithm for the EMS of PHEBs. The optimization framework of the abovementioned method applied to PHEBs is shown in Fig. 4 which can be divided into three main parts. The first part is the acquisition of driving cycles based on remote monitoring platform (RMP). The second part is the combinatorial algorithm for the optimization of control parameters.

Optimization results

In the optimization, the values of some parameters must be defined properly. According to the range of each parameter value which has been suggested in several papers and some tests for PHEB optimization, the parameter values adopted for the following optimizations are listed in Table 8.

Optimizations using GA-EACA are conducted on driving cycles of two different levels of congestion for 10 times respectively. The optimal solutions for each cycle are listed in Tables 9 and 10.

Comparison results among different algorithms

In addition, GA,

Conclusions and future work

In this paper, a novel optimization framework based on RMP is proposed for energy management of PHEBs. In this approach, the driving cycle data can be processed according to the departure time of the buses on the same route, while a table which records different conditions at different time could be built to predict the future vehicle schedule. Based on this, a combinatorial optimization algorithm which combines the advantages of both GA and EACA could be fulfilled. Results show that the

Acknowledgments

The authors are very grateful to the China government by the support of this work through the National Natural Science Foundation of China (Grant nos. 51605243 and No. 51675293). And this work is also supported by the National Key Technology R&D Program of the Ministry of Science and Technology (Grant no. 2016YFB0101400) and the 1-class General Financial Grant from the China Postdoctoral Science Foundation (Grant no. 2016M590094).

References (45)

  • R.N. Fries et al.

    Meeting privacy challenges while advancing intelligent transportation systems

    Transp. Res. Part C

    (2012)
  • YangC. et al.

    Economical launching and accelerating control strategy for a single-shaft parallel hybrid electric bus

    Mech. Syst. Signal Process.

    (2016)
  • LiL. et al.

    Hybrid genetic algorithm-based optimization of powertrain and control parameters of plug-in hybrid electric bus

    J. Frankl. Inst.

    (2015)
  • W. Shabbir et al.

    Exclusive operation strategy for the supervisory control of series hybrid electric vehicles

    IEEE Trans. Control Syst. Technol.

    (2016)
  • C. Desai

    Design and Optimization of Hybrid Electric Vehicle Drivetrain and Control Strategy Parameters Using Evolutionary Algorithms (Master's Dissertation of Science)

    (2010)
  • T. Padovani et al.

    Implementation of an energy management strategy for hybrid electric vehicles including drivability constraints

    IEEE Trans. Veh. Technol.

    (2016)
  • G. Wu et al.

    Development and evaluation of an intelligent energy-management strategy for plug-in hybrid electric vehicles

    IEEE Trans. Intell. Transp. Syst.

    (2014)
  • HuX. et al.

    Advanced power-source integration in hybrid electric vehicles: multicriteria optimization approach

    IEEE Trans. Ind. Electron.

    (2015)
  • LiL. et al.

    Correctional DP-based energy management strategy of plug-in hybrid electric bus for city-bus-route

    IEEE Trans. Veh. Technol.

    (2015)
  • J. Gissing et al.

    Optimal control of series plug-in hybrid electric vehicles considering the cabin heat demand

    IEEE Trans. Control Syst. Technol.

    (2016)
  • ZouY. et al.

    Comparative study of dynamic programming and pontryagin's minimum principle on energy management for a parallel hybrid electric vehicle

    Energies

    (2013)
  • ZengX. et al.

    A Parallel hybrid electric vehicle energy management strategy using stochastic model predictive control with road grade preview

    IEEE Trans. Control Syst. Technol.

    (2015)
  • Cited by (37)

    • Investigation of novel intelligent energy management strategies for connected HEB considering global planning of fixed-route information

      2023, Energy
      Citation Excerpt :

      These provided many solutions for improving the optimization and adaptability of HEB in complex driving conditions [15–18]. Firstly, from the perspective of adaptive intelligent algorithms, an adaptive optimization-based algorithm has been widely applied to HEV energy management [19–21]. Yang et al. [22] studied an adaptive optimization-based intelligent strategy.

    • Meta-heuristics optimization in electric vehicles -an extensive review

      2022, Renewable and Sustainable Energy Reviews
      Citation Excerpt :

      The detailed optimization framework for the optimal grid integration of EVs is presented in Table 13 and the optimization framework for the adaptive and real-time EMS with fuzzy and neural networks is presented in Table 14. Table 13 is formulated based on the data from the references [94,102,107,128–137]. Table 14 is formulated based on the data from the references [91,98,105,107,108,121,124,138–143].

    • Critical review on structural architecture, energy control strategies and development process towards optimal energy management in hybrid vehicles

      2022, Renewable and Sustainable Energy Reviews
      Citation Excerpt :

      The complex structure of hybrid powertrains and various energy sources have to be taken into consideration while designing and developing EMS process in order to attain the maximum performance as shown in Fig. 1. With this context, the conceptual review of EMS concludes that HEV development fall under three main categories notably design, operating and control parameters [2,3]. The configuration of HEV is more complex than ever to find the optimum drivetrain structure, i.e., size and energy storage capacity of components in the vehicle.

    View all citing articles on Scopus
    View full text