Multi-objective design optimization of battery thermal management system for electric vehicles
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].where 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)
- et al.
Electric vehicle battery technologies: from present state to future systems
Renew. Sustain. Energy Rev.
(2015) - et al.
Intelligent sizing of a series hybrid electric power-train system based on chaos-enhanced accelerated particle swarm optimization
Appl. Energy
(2017) - et al.
Thermal behavior study of discharging/charging cylindrical lithium-ion battery module cooled by channeled liquid flow
Int. J. Heat Mass Transf.
(2018) - et al.
A novel temperature-compensated model for power li-ion batteries with dual-particle-filter state of charge estimation
Appl. Energy
(2014) - et al.
Reciprocating air flow for li-ion battery thermal management to improve temperature uniformity
J. Power Sources
(2011) - et al.
An investigation of lithium-ion battery thermal management using paraffin/porous-graphite-matrix composite
J. Power Sources
(2015) - et al.
Thermal performance investigation of an air-cooled lithium-ion battery pack considering the inconsistency of battery cells
Appl. Therm. Eng.
(2019) - et al.
In-plane nonuniform temperature effects on the performance of a large-format lithium-ion pouch cell
Appl. Energy
(2016) - et al.
A theoretical and computational study of lithium-ion battery thermal management for electric vehicles using heat pipes
J. Power Sources
(2014) - et al.
Thermal management of cylindrical lithium-ion battery based on a liquid cooling method with half-helical duct
Appl. Therm. Eng.
(2019)