Multi-objective design optimization of battery thermal management system for electric vehicles

https://doi.org/10.1016/j.applthermaleng.2021.117235Get rights and content

Highlights

  • Genetic programming shows reliable surrogate modelling.

  • Properties of liquid-based battery thermal management system are analyzed.

  • Cooling plate design and inlet coolant velocity are the main factors affect the temperature uniformity and pressure drop.

  • The pressure drop has been significantly optimized with the proposed optimization framework.

Abstract

Lithium batteries are commonly used as the primary power storage unit for electric vehicles, and their performance is sensitive to temperature. Thus, the battery thermal management system is crucially needed to allow the EVs to work safely and efficiently. This paper mainly focuses on the performance analysis and design optimization of the battery thermal management system with a U-shaped cooling channel. A Computational fluid dynamics model of a battery thermal management system is built to study the battery temperature distribution and pressure distribution. Through the establishment of the genetic programming model, sensitivity analysis and parameter interaction analysis are carried out to analyze the influence of cooling plate thickness, cooling plate wall thickness, inlet coolant temperature and flow velocity on the comprehensive performance of the battery thermal management system. A new surrogate-assisted multi-objective optimization scheme is proposed by introducing an integrated AI system that includes a surrogate battery model built with genetic programming (GP) and a design optimizer driven by the second-generation non-dominated sorting genetic algorithm (NSGA-II). Results show that the inlet coolant temperature has the most significant influence on the rise of battery temperature (59.87%) but has no influence on the pressure drop. The structural parameters of the cooling plate and the velocity of the inlet coolant have apparent effects on the uniformity of the battery temperature distribution and the pressure drop. The battery thermal management system achieves an ideal comprehensive performance when the thickness of the cooling plate is 4.50 mm, the thickness of the cooling plate wall is 1.49 mm, the inlet coolant temperature is 298.15 K, and the inlet coolant velocity is 0.29 m/s. Under such optimized parameter settings, the max temperature rise of the battery reduces from 7.72 K to 7.69 K, the standard deviation of the temperature distribution 2.54 K (a drop of 0.02 K), and the pressure drop decrease from 1022.1 Pa to 436.43 Pa (decrease by 57.3%). Such results have guiding significance for the design of the battery thermal management system with a U-shaped channel and the application of genetic programming in system performance analysis and optimization.

Introduction

The exhaust emissions of traditional fuel vehicles have caused severe environmental pollution problems and led to global climate changes, endangering the survival and development of humanity. Therefore, the development of new energy vehicles is of great significance. The electric vehicle (EV), which applies power batteries as the power source, is ideal for private and public transportation [1], [2]. The use of electric energy can realize a low or even zero load emission rate, which is suitable for environment protection.

The extensive use of EVs has brought a series of challenges to its power storage components. Thanks to the high energy density, low self-discharge rate, long lifespan etc., lithium-ion batteries (LIBs) are generally considered ideal power storage units in EVs [3], [4], [5]. LIBs pack commonly consists of lots of single cells in series and parallel. LIBs pack performance is crucial to guarantee the safety and durability of EVs [6], [7]. Charging and discharging of the LIBs involve many electric-chemical reactions sensitive to temperature, e.g., LIBs pack ought to run in a proper temperature range (20–40 °C) [8]. The charge and discharge processes generate a large amount of heat, making the LIBs overheat or even combust [9]. Except for overheat, the non-uniformity of heat distribution among single cells is also a critical problem needed to be focused on because the inconsistencies highly restrict the battery pack performance among individual cells, such as the failure of some individual cells that cause a significant voltage drop in the pack [9]. The temperature difference among individual cell should be as slight as possible [10]. Therefore, a reliable and effective battery thermal management system (BTMS) is vital for EVs safety [11], [12].

The typical cooling methods developed for BTMS include air cooling [13], liquid cooling [14], phase change materials (PCM) [14], heat pipe [15], [16], and thermoelectric cooling [17]. Fewer researches have been conducted on thermoelectric cooling [18]. Air cooling-based BTMS is no longer suitable for EVs with high safety requirements and endurance due to its low heat conductivity rate [10]. Though the PCM cooling system owns an excellent cooling performance, the high cost and the heavy mass of the system cannot be ignored [19]. The utilization of heat pipe cooling needs other cooling strategies to complete the condensation section, which leads to higher cost [20]. Compared with the above, liquid cooling has a wide range of applications in battery thermal management systems.

Liquid cooling has been widely used in EVs for the following reasons [18]: 1) the coolant thermal performance (e.g., heat capacity, heat conductivity rate) of liquids is better than that of the air; 2) liquid cooling system owns a more compact structure, and lower cost than the PCM cooling system; and 3) liquid cooling system is more reliable than the heat pipe cooling system. Malik et al. studied the cooling performance of the BTMS with different coolant temperature at four selected discharge rates [21]. Liang et al. focused on the effect of the coolant temperature [22] and the ambient temperature [23] on the performance of the BTMS for serially connected battery module. Al-Zareer et al. claim that the application of the coolants with high thermal conductivity can obtain an ideal cooling performance [24]. Wang et al. investigated the effect of individual cell arrangement (serial and parallel) and the coolant flow rate on the BTMS performance. They found that parallel cooling mode can improve the uniformity of temperature distribution, and the increase of the coolant flow rate can bring a high cooling performance [25]. Lai et al. focused on the compact and lightweight design of the liquid cooling system [26]. Hong et al. believed that a two-phase refrigerant is an ideal coolant employed in a traditional liquid cooling system [27].

Except for coolant categories, coolant flow rate, the initial temperature of the coolant, single-cell arrangement, the cooling plate is also the vital component of the BTMS. A reasonable design of the cooling channel can result in high cooling performance at a lower cost. Huo et al. found that liquid cooling plates (LCPs) with straight sub-channel show high heat dissipation efficiency [28]. Srinivaas et al. proposed that using a split-channel can improve the thermal uniformity and the utilization of the convergence channel decrease the temperature rise [29]. Jarrett and Kim designed LCPs with serpentine sub-channels. Their result showed that such LCPs could restrain the temperature rise while increasing the thermal non-uniformity [30]. More similar researches on LCPs have been conducted [31], [32].

For the battery pack built with cylindrical batteries, it is not suitable to apply traditional flat LCPs because of the shape mismatch. The shape of the cooling plate needs a special design to achieve optimal contact with the cylindrical battery. Bamdezh et al. designed a circular cylindrical cooling plate [33]. The cooling plate surrounded the cylindrical surface of the battery, and the coolant flows from the bottom of the circular cylinder and along the axial direction [33]. Wang et al. developed a cooling plate with a curved surface, which is in contact with part of the cylindrical surface of the battery. They claimed that the design of the cooling plate owned the most contribution to the cooling efficiency of the BTMS [34]. Zhao et al. proposed a serpentine channel that can increase the contact area between the cooling plate and the cylindrical battery surface. It was found that using such a design can improve the cooling performance and the uniformity of the battery temperature distribution [18]. A similar design scheme of the cooling plate was proposed by Xie et al. [35]. However, such a scheme result in a high value of the pressure drop of the BTMS, which led to a high energy consumption [36]. Following Zhao et al., Li et al. revise the design of the cooling plate by introducing a U-shaped scheme. They found that compared with the serpentine channel, using a U-shape channel can reduce the overall pressure drop of the system while ensuring heat dissipation performance and thermal uniformity.

After determining the cooling plate shape, it is necessary to optimize the geometric design parameters to improve the BTMS performance. Nowadays, the equivalent models are commonly used in the performance analysis and optimal design of BTMS. Gan et al. developed a thermal equilvalent circuit model to analyze the designed BTMS properties [37]. In Li et al. research, surrogate model technology was applied to optimize the mini-channel liquid cooling system to reduce the temperature difference and the pressure drop [38]. Li et al. utilized kriging method for the optimization of a BTMS involving herringbone fins [39]. Liu et al. applied a radial basis function-based surrogate model to carry out the multi-objective optimization of the settings of BTMS design variables [40]. An optimization framework consists of support vector regression surrogated method and particle swarm optimization (PSO) is proposed in Tang et al. research [41]. Li et al. used the surrogate model and the second-generation non-dominated sorting genetic algorithm (NSGA-II) to carry out a multi-objective optimization process to redesign the geometric structure of the BTMS. They determined the optimal operating parameters values [10]. More optimization frameworks are proposed in [42], [43], [44], [45].

As can be found from the literature review, most of the existing work focus on building the surrogate models for BTMS optimization. They are either focusing on modelling or optimization, there is lack of study regards surrogate modelling and surrogate-assisted optimization as an integrated system. This paper proposed an integrated AI system for surrogate-assisted design optimization of BTMS. The work is conducted with two main contributions: 1) a new surrogate battery model is built with Genetic programming (GP), and its advantage in model accuracy and adaptiveness is demonstrated by a comparative study with a response surface model and an ANN model; and 2) the design parameters in the BTMS is optimized using NSGA-II, and a parametric and visual analysis is conducted to validate the optimization results.

The remainder of this paper is orgnised as follows. Section 2 states the main research problems of this paper. Section 3 introduces the parameter selection. Section 4 elaborates on the modelling and optimization methods. Section 5 presents the results and discussion. Section 6 shows the conclusion.

Section snippets

Problem statement

The energy storage system in the EVs contains thousands of individual batteries connected in series and parallels. Extensive heat is generated during the charging/discharging process. Untimely heat dissipation will lead to battery life degradation and potential safety hazards (e.g. thermal runaway, fire and explosion) [46], [47], [48].

Liquid cooling BTMSs are commonly used to maintain battery packs operating within a proper temperature range. In this research, a liquid cooling BTMS for

Research framework

Fig. 2 shows a flowchart of the proposed method. All simulation and numerical calculation are performed with such device: CPU Inter(R) Core(TM) i7-10700 K CPU of 3.80 GHz, RAM 16.0G. Firstly, the geometric design of the BTMS is carried out, and the initial conditions are given for CFD analysis. Secondly, 200 sample points are obtained by the design of experiment (DoE) method, and the corresponding output objectives are obtained through generated sample points. Some model evaluation methods are

Parameter selection

As above mentions, this research mainly studies the effect of the geometric design of the cold plate channel and the parameter settings on the liquid-cooling based BTMS performance (thermodynamic and fluid dynamic characteristics). In the CFD simulation and numerical analysis phase, the thickness of the cooling plate – t1, cooling plate wall thickness – t2, inlet coolant temperature – Tc and inlet coolant velocity – v are the design/input parameters. The temperature difference between the

Heat dissipation model

The complicated internal chemical reaction of LIBs causes to the complex inner heat generation mechanism. An accurate thermal model is established based on a proper calculation of the heat generation rate of LIBs, which is difficult to obtain accurately in actual application. Bernardi equation is an alternative approximation model which is popular in the heat generation rate calculation. The battery was assumed as a stable and uniform heat source [51], [52].Q=IVb(EO-U)+TEOTwhere Q, I, R, T, Vb

Sensitivity analysis

Based on the generated numerical models, sensitivity analysis is conducted to study the effect of the input variables on the change of output variables. The result is shown in Table 9. In such a geometric design scheme of the cooling plate, the TD is mostly affected by Tc, which is up to 59.87%. t1 has the second-highest contribution to the temperature drop, which is 19.94%. v and Tc have a similar influence on the heat dissipation efficiency of the system (10.37% and 9.82%, respectively). TSD

Conclusions

This research analyzes the performance of the designed BTMS and proposes a multidisciplinary multi-objective optimization problem of the BTMS using genetic programming. Computational fluid dynamics is applied to study the battery temperature distribution and pressure distribution of the battery. Based on the generated GP models, the effects of the thickness of the cooling plate, the thickness of the cooling plate wall, the temperature and velocity of coolant at the inlet on the performance of

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (61)

  • T.H. Tran et al.

    Experimental investigation on the feasibility of heat pipe cooling for HEV/EV lithium-ion battery

    Appl. Therm. Eng.

    (2014)
  • Y. Gan et al.

    A Numerical study on the performance of a thermal management system for a battery pack with cylindrical cells based on heat pipes

    Appl. Therm. Eng.

    (2020)
  • Y. Lyu et al.

    Electric vehicle battery thermal management system with thermoelectric cooling

    Energy Rep.

    (2019)
  • C. Zhao et al.

    Minimization of thermal non-uniformity in lithium-ion battery pack cooled by channeled liquid flow

    Int. J. Heat Mass Transf.

    (2019)
  • Y. Wei et al.

    Experimental investigation of a novel hybrid cooling method for lithium-ion batteries

    Appl. Therm. Eng.

    (2018)
  • X. Yuan et al.

    Experimental investigation on thermal performance of a battery liquid cooling structure coupled with heat pipe

    J. Energy Storage

    (2020)
  • M. Malik et al.

    Thermal and electrical performance evaluations of series connected li-ion batteries in a pack with liquid cooling

    Appl. Therm. Eng.

    (2018)
  • J. Liang et al.

    Thermal and electrochemical performance of a serially connected battery module using a heat pipe-based thermal management system under different coolant temperatures

    Energy

    (2019)
  • H. Wang et al.

    Cooling capacity of a novel modular liquid-cooled battery thermal management system for cylindrical lithium ion batteries

    Appl. Therm. Eng.

    (2020)
  • Y. Lai et al.

    A compact and lightweight liquid-cooled thermal management solution for cylindrical lithium-ion power battery pack

    Int. J. Heat Mass Transf.

    (2019)
  • S.H. Hong et al.

    Thermal performance of direct two-phase refrigerant cooling for lithium-ion batteries in electric vehicles

    Appl. Therm. Eng.

    (2020)
  • Y. Huo et al.

    Investigation of power battery thermal management by using mini-channel cold plate

    Energy Convers. Manag.

    (2015)
  • A. Jarrett et al.

    Design optimization of electric vehicle battery cooling plates for thermal performance

    J. Power Sources

    (2011)
  • J. Rostami et al.

    Optimization of conjugate heat transfer in wavy walls microchannels

    Appl. Therm. Eng.

    (2015)
  • M.A. Bamdezh et al.

    Role of foam anisotropy used in the phase-change composite material for the hybrid thermal management system of lithium-ion battery

    J. Energy Storage

    (2020)
  • J. Wang et al.

    Sensitivity analysis of factors influencing a heat pipe-based thermal management system for a battery module with cylindrical cells

    Appl. Therm. Eng.

    (2019)
  • L. Xie et al.

    Coupled prediction model of liquid-cooling based thermal management system for cylindrical lithium-ion module

    Appl. Therm. Eng.

    (2020)
  • J. Smith et al.

    Simulative method for determining the optimal operating conditions for a cooling plate for lithium-ion battery cell modules

    J. Power Sources

    (2014)
  • Y. Gan et al.

    Development of thermal equivalent circuit model of heat pipe-based thermal management system for a battery module with cylindrical cells

    Appl. Therm. Eng.

    (2020)
  • L. Cheng et al.

    Surrogate based multi-objective design optimization of lithium-ion battery air-cooled system in electric vehicles

    J. Energy Storage

    (2020)
  • Cited by (53)

    View all citing articles on Scopus
    View full text