Optimization of Cooling Channel Structure of Bipolar Plate for Proton Exchange Membrane Fuel Cells Based on CFD Analysis

Abstract

The working temperature affects the performance of PEMFC, so a reasonable and efficient cooling channel is necessary to control the working temperature in an efficient area. In this study, the channel structure of the bipolar plate for PEMFC is analyzed using the FLUENT simulation calculation method. The influence of cell size and cooling water flow direction on cell temperature distribution is analyzed, including an examination of the channel ridge width, depth, and aspect ratio of the bipolar plate. After comparing and analyzing three ridge width sizes (0.5 mm, 1.5 mm and 2 mm) in the paper, it was found that a ridge width of 2 mm had the best heat transfer performance. And it was found that a groove depth of 0.5 mm had the best heat transfer performance when comparing three groove depth dimensions (0.5 mm, 1 mm and 1.5 mm). The aspect ratio size parameters had almost no effect on the maximum and average temperatures of the electric stacks, while the relative flow direction of cooling water had a great influence on the temperature distribution of the bipolar plate.

1. Introduction

In recent decades, many scholars around the world have been conducting research on the development of new energy sources to address issues, such as energy shortages and environmental pollution [1]. Due to the advantages of high energy and cleanliness of hydrogen energy, it has become the “ideal clean energy of the 21st century”. In the utilization technology of hydrogen energy, the research on hydrogen fuel cell engines is one of the key directions [2]. Since the first introduction of the fuel cell principle in the last century to the invention of the first fuel cell and the subsequent development of its name [3], fuel cell technology continuously developed and improved in the following decades [4]. It was not until recent decades that real breakthroughs were made in technology, especially through the research on proton exchange membrane fuel cells, which have a wide range of applications, and have promoted the rapid development of fuel cell technology [5]. Due to its high energy efficiency, green environmental protection, low operating temperature, and fast start-up, PEMFC is widely used in fields such as automotive power and portable power sources [6]. In addition, its volume and efficiency are directly related, making it widely used in low-power generation equipment [7,8]. If the heat generated by PEMFC is not discharged in a timely manner, it will cause excessive battery temperatures, leading to a significant decrease in performance and even irreversible damage [9]. In addition, excessive temperature differences in the battery can cause uneven heating of the membrane, shortening the battery life [10]. Therefore, the cooling problem of fuel cells is also one of the key technical issues in the development of battery technology. Coolant is an effective way of thermal management for fuel cells. When researchers investigate cooling channels, their main focus lies in the uniformity of temperature distribution and pressure drop across different cooling plate configurations. Previous studies have already conducted extensive research on cooling plates.

For instance, Ebrahim Afshari and his team [11] conducted a study on the coolant flow and heat transfer in the cooling plates of proton exchange membrane fuel cells. They evaluated and compared the performance of four different coolant flow fields, including simple parallel channels and serpentine channels with one and two parallel channels, based on the maximum surface temperature, uniform temperature distribution, and pressure drop. The results indicated that, compared to the serpentine model, the straight flow field model exhibited a higher temperature uniformity index, lower pressure drop, and lower temperature difference. Furthermore, Choi et al. [12] studied the performance of six different cooling plates, including parallel flow fields, serpentine flow fields, and their improved structures. They found that the improved serpentine flow field exhibited lower peak surface temperature, while the improved parallel flow field showed better temperature distribution uniformity compared to other parallel flow fields. The study was conducted using a constant thermal flux. Asghari et al. [13] investigated the influence of different operating conditions on the performance of a serpentine cooling flow field. They observed that, at high flow rates, the cooling plate exhibited a more uniform temperature distribution, whereas, at low flow rates, the cooling channels had a lower pressure drop. The experimental validation of their numerical model involved comparing the temperature distribution curves of the cooling plate obtained from numerical simulations with those obtained from actual experiments. Sasmito et al. [14] conducted a simulation study on six different cooling flow fields, namely parallel, serpentine, oblique-fins, coiled, parallel-serpentine, and a novel hybrid parallel-serpentine–oblique-fins channel. The research results showed that the novel hybrid channel exhibited optimal temperature uniformity and pressure drop, demonstrating superior performance in thermal water management. In Li et al.’s numerical study of the thermal performance of traditional straight channels and wavy channels in cooling plates [15], they found that the wavy channels exhibit better cooling performance compared to the straight channels. Increasing the amplitude or reducing the wavelength further enhances the cooling performance, but it also leads to an increase in pressure drop. Therefore, they proposed two novel variable-wavelength wavy flow channels. These variable-wavelength flow channels show improved cooling performance compared to constant-wavelength flow channels while significantly reducing the pressure drop. In the field of heat transfer, different analytical perspectives can be crucial for research. Mohamed et al. [16] conducted a three-dimensional numerical analysis on enhanced heat transfer using nanofluids in triangular fin micro-heat exchangers. They investigated thermal and flow fields, heatlines, overall heat transfer coefficient, thermal efficiency, and thermal performance factor (TPF). The conclusion drawn from their study was that utilizing CNT nanofluids and triangular fins can significantly enhance the performance of the heat exchanger.

The above-mentioned research illustrates the significant importance of the coolant flow field design for thermal management in PEMFCs as well as the feasibility of various structures in enhancing cooling plate performance. However, while complex coolant flow field designs undoubtedly offer better performance, they also come with higher design, optimization, and production costs. Therefore, it is necessary to conduct detailed research on typical coolant flow fields. The design of coolant flow fields should strike a balance between temperature distribution uniformity and pressure drop. In this study, we focused on straight channel cooling passages, which are characterized by lower costs and good performance. We investigated the performance of straight channel cooling plates for PEMFCs under different operating conditions (varying thermal flux conditions), aiming to explore straight channel structures that can achieve both of the aforementioned performance aspects. The results aim to serve as a reference for the development of simple-structure and high-performance cooling plates for industrial PEMFC production.

2. Research on the Cooling Channel Structure of Single Battery

2.1. Model Description

2.1.1. Geometric Modeling

The PEMFC stack exhibits a symmetrical structure on the packaging, consisting of a membrane electrode (MEA) and a bipolar plate. The membrane electrode consists of a proton exchange membrane, a cathode catalytic layer, and a cathode gas diffusion layer. Therefore, this model adopts a symmetrical structure. As shown in Figure 1, from top to bottom are half-membrane electrodes, bipolar plates with gas and cooling channels, and half-membrane electrodes, respectively. Moreover, the cooling channel structure is shown in Figure 2, and the basic parameters of the battery are shown in

Table 1. Shape parameters of batteries.

Structure	Size/mm
Cooling runner	2 × 1
Cooling water inflow section	10 × 10 × 2
H2/O2 Runner	1 × 0.5
H2/O2 Infusion section	10 × 1 × 0.5
Semi-membrane electrode section	158 × 120 × 1
Cooling plate thickness	2
Cell thickness per layer	5

.

2.1.2. Theoretical Model

This model assumes the following properties:

(1)

No distinguishability is considered between proton exchange membranes, gas diffusion layers, and reaction catalytic layers, thus treating membrane electrodes as a whole part.

(2)

The gas velocity and temperature at the entrance of the stack are constant.

(3)

Radiation heat transfer is not considered during the process.

(4)

The reaction process and mechanism of H₂ and O₂ are not considered during the process.

(5)

Membrane electrodes are treated as constant heat flux heat sources for fuel cells.

The fuel cell model is divided into two parts: fluid and solid. There are four conservation equations used in the fluid part, including the mass conservation equation, momentum conservation equation, and energy conservation equation [17]:

Mass conservation equation:

$$\frac{\partial \rho }{\partial t}+\nabla \left(\rho u\right)=0$$

(1)

Momentum conservation equation:

$$\rho u\nabla u=-\nabla p+\left[\mu \right(\nabla u+{(\nabla u)}^T)-2/3 \mu \nabla u]$$

(2)

in which p represents static pressure, and ρ is the fluid density, respectively

Energy conservation equation:

$$\frac{\partial \left(\rho T\right)}{\partial t}+\nabla \left(\rho uT\right)=\nabla \left(\frac{k}{c_p}gradT\right)+S_T$$

(3)

in which dp is the specific heat capacity of the fluid, U is the temperature, and k is the heat transfer coefficient of the fluid; S_T is the internal heat source of the fluid and heat which come from the mechanical energy of the fluid due to viscosity.

The control equation for the solid part is

$$\nabla {\mathrm{N}}_T=\nabla \left(-{\mathsf{\lambda }}_{\mathrm{s}}\nabla T\right)=\mathrm{Q}$$

(4)

in which N, λ_s, and Q represent heat flux, thermal conductivity, and heat source, respectively.

2.1.3. Boundary Conditions

The boundary conditions used in the simulation are as follows:

• Inlet boundary:

$$\begin{array}{l}{q}_w=1.966\times {10}^{-3}\mathrm{kg}/\mathrm{s}\\ v_{H_2}=v_{O_2}=0.02 \mathrm{m}/\mathrm{s}\end{array}$$

where $q_w$ is the inlet mass flow rate, and $v_{H_2}$ and $v_{O_2}$ are the inlet velocities of anode hydrogen and cathode oxygen, respectively.

2.

Inlet temperature:

$$\begin{array}{l}{T}_w=60 °\mathrm{C}\\ T_{H_2}=T_{O_2}=60 °\mathrm{C}\end{array}$$

$T_w$, $T_{H_2}$, and $T_{O_2}$ are inlet temperatures of water, hydrogen, and oxygen, respectively.

3.

Outlet boundary:

$$P_w=P_{H_2}=P_{O_2}=1 \mathrm{bar}$$

$P_w$, $P_{H_2}$, and $P_{O_2}$ are the outlet pressures of water, hydrogen, and oxygen, respectively.

4.

Reynolds number:

The fluid flow is laminar, and the Reynolds number of cooling water is 930.8.

2.1.4. Numerical Procedures

The computational fluid dynamics (CFD) software Fluent was used to solve the conservation equations, employing a structured hexahedral grid type. Figure 3 shows the mesh diagram of the cooling channel with a ridge width of 2 mm. The cooling plate consists of 16 layers of mesh, the cooling channel has 8 layers of mesh, and both the anode and cathode channels have 4 layers of mesh.

Figure 4 represents the mesh independence verification. It demonstrates the relationship between the cooling channel inlet–outlet temperature difference, pressure drop, and maximum temperature of the cooling plate with varying mesh numbers, indicating the simulation’s dependence on mesh refinement. From the figure, it can be observed that, except for the pressure drop at the cooling channel inlet–outlet, the other parameters remain relatively constant. Once the mesh number exceeds 532,640, the pressure drop no longer changes significantly. Therefore, this study employs a mesh partitioning approach with a total of 532,640 grid cells.

2.1.5. Verification of Simulation Results

Comparing the simulation data with the experiments conducted by Asghari et al. [13], it can be observed from Figure 5 that, as the coolant inlet flow rate increases, the temperature difference between the coolant outlet and inlet gradually decreases. This trend is consistent with the experimental results. However, the simulation curve deviates slightly from the experimental values, which can be attributed to the different operating conditions and configurations used in the simulation and their experiments.

2.2. The Influence of Cooling Channel Ridge Width on Cooling Efficiency

2.2.1. Cooling Channel Geometry

The optimized flow channel structure can effectively improve the heat transfer capacity of the cooling water, thereby enhancing the heat dissipation capacity of the stack [18]. In order to investigate the effect of channel ridge width on battery cooling efficiency, three forms of cooling channels were designed in this study. The ridge widths of the cooling channel are 0.5 mm, 1.5 mm, and 2 mm, respectively, the cross-section of the channel is 2 × 1 mm², and the single battery area is 158 × 120 mm². Figure 6 shows the geometric structure of the three cooling channels.

2.2.2. Result Analysis and Discussion

Figure 7 shows the temperature distribution contour plots of the fuel cell at a heat source heat flux density of 3 × 10⁶ W/m³. The temperature distribution of the fuel cell is similar for all three structures. From the perspective of the lowest temperature, the lowest temperature of the fuel cell increases with the increase in ridge width. From the contour plots, it can be observed that after the cooling water enters the channels, it absorbs a significant amount of heat, causing a rapid decrease in the fuel cell temperature. As the cooling water flows, its heat transfer capacity gradually decreases, absorbing less heat, and, consequently, the rate of temperature reduction of the fuel cell slows down.

Figure 8 shows the relationship between temperature and heat flux density for cells with different channel ridge widths. The maximum temperature, average temperature, and maximum temperature difference of the battery increase linearly with the increase of the battery heat source. In terms of maximum temperature and maximum temperature difference, there are significant temperature differences among the three types of battery structures. Among them, the channel structure with a ridge width of 2 mm has a better heat transfer effect. The battery with this structure has the lowest maximum temperature, average temperature, and maximum temperature difference, while the battery with a ridge width of 0.5 mm is most prone to local overheating and temperature non-uniformity.

By comparing the maximum temperature, average temperature, and maximum temperature differences of cells with different ridge widths, we conclude that the optimal cooling channel form is a structure with a ridge width of 2 mm.

2.3. The Influence of Cooling Channel Depth on Cooling Effectiveness

2.3.1. Cooling Channel Geometry

Based on the thickness of the electrode plate, three different depths of cooling channels were designed, with channel depths of 0.5 mm, 1 mm, and 1.5 mm, respectively. Figure 9 shows a partially enlarged view of the channel inlet section. The cooling channel ridge width of this structure is 2 mm, and the battery area is 158 × 120 mm².

2.3.2. Result Analysis and Discussion

Figure 10 shows the temperature distribution contour plots for different cell structures with a heat source of 4 × 10⁶ W/m³. The temperature distribution in the contour plots is similar, and the cell temperature gradually increases with the flow direction of the cooling water. From the perspective of the lowest temperature, the cells with a channel depth of 0.5 mm, 1 mm, and 1.5 mm have the lowest temperatures of 63.8 °C, 63.2 °C, and 62.7 °C, respectively. Moreover, the lowest temperature decreases as the channel depth increases.

Figure 11 shows the temperature variation plots for three different cell structures under various heat source heat flux densities. From Figure 8a–c, it can be observed that the cell with a channel depth of 0.5 mm exhibits better cooling performance, while the difference between the depths of 1 mm and 1.5 mm is minimal, with only a slight variation in temperature uniformity. When the heat source heat flux density is 1 × 10⁶ W/m³, both the 1 mm and 1.5 mm deep cell structures have the same maximum temperature difference, which is 3.3 °C. However, as the heat flux density increases, the difference between the two structures gradually widens.

In conclusion, the channel depth has a certain impact on the cooling effectiveness of the fuel cell. When comparing three aspects—the highest temperature, average temperature, and maximum temperature difference of the fuel cell—the most effective cooling performance is achieved with a channel depth of 0.5 mm.

2.4. The Influence of Length–Width Ratio on Cooling Performance of Electric Stacks

2.4.1. Geometry of Cooling Channel

In addition to the direct influence of the cooling channel structure on the fuel cell temperature distribution, the fuel cell itself may also impact the cooling effectiveness. To further explore the factors influencing the cooling effectiveness of the fuel cell, this section focuses on the study of the fuel cell’s dimensions. While maintaining a consistent cell area of 158 × 120 mm², we conducted a comparative study by varying the aspect ratio of the fuel cell to analyze its impact on the temperature distribution. We designed four different fuel cell structures with aspect ratios of 3.40:1, 1.75:1, 1.32:1, and 0.95:1. The cooling channel structures for all four configurations have a channel spacing of 2 mm and a channel depth of 1 mm. Figure 12 illustrates the structures of the four cooling channels.

2.4.2. Result Analysis and Discussion

Figure 13 shows the temperature distribution contour plots of different fuel cell aspect ratios at a heat source heat flux density of 3 × 10⁶ W/m³. From the figure, it can be observed that the surface temperature distribution of all four fuel cells is quite similar, with the cell temperature gradually increasing along the flow direction of the cooling water. As the fuel cell aspect ratio decreases, the minimum temperature of the cell slightly increases. However, the difference in the maximum temperature of the fuel cells is not significant, and the effect of the aspect ratio on the highest temperature of the fuel cell surface is relatively minimal compared to its impact on the lowest temperature.

Figure 14 shows the temperature variation plots of fuel cells with different aspect ratios. It can be observed that the temperature curves of the four different aspect ratios almost overlap, with only a slight difference in the maximum temperature difference for the larger aspect ratio structure at high heat flux density. The temperature of the fuel cells under the four aspect ratios varies only with the heat flux density, and there is no difference in the temperature distribution of the fuel cells for different aspect ratios at the same heat source heat flux density. Therefore, it can be concluded that the aspect ratio does not significantly affect the temperature distribution of the fuel cell under the given conditions.

From the above analysis, it can be basically concluded that the length–width ratio of the size parameter of the stack has little effect on the temperature distribution of the stack.

3. Research on the Flow Direction of Cooling Water in Double Layer PEMFCs

In order to simulate and compare the impact of the relative direction of cooling water flow in the stack on the cooling effect, a double-layer fuel cell was used for research to reduce computational costs. The simulation calculation conditions are the same as for a single battery. By keeping the cooling water flow rate constant, the heat flux densities of the battery are set at 1 × 10⁶ W/m³, 2 × 10⁶ W/m³, 3 × 10⁶ W/m³, 4 × 10⁶ W/m³ and 5 × 10⁶ W/m³, to compare the temperature distribution of co-flow and counter-flow cells.

3.1. Model Settings

The structure of each part in the double-layer PEMFC model is shown in Figure 15a, and the cooling channel adopts a three-channel serpentine channel that facilitates heat transfer of cooling water, as shown in Figure 15b. The relative flow direction of the double-layer cooling water is also shown in Figure 15c,d. In the battery stack, except for the two end batteries, there are cooling channels on both sides of the battery, so the temperature curve analysis takes the temperature of the middle battery for comparison.

3.2. Analysis of Results and Discussion

Figure 16 shows the temperature distribution contour plots of the fuel cell at a heat flux density of 5 × 10⁶ W/m³ with two different cooling water flow directions. When the cooling water flows in the same direction as the fuel cell, the temperature distribution of each cell is nearly identical, and the fuel cell temperature gradually decreases along the flow direction of the cooling water.

However, when the cooling water flows in the opposite direction, the temperature distribution of each cell varies significantly due to the inconsistency in the cooling water flow direction. Specifically, the upper half of the fuel cell is more affected by the upper cooling water, resulting in lower temperatures at the inlet and higher temperatures at the outlet of the upper half of the cell. On the other hand, the temperature distribution of the lower half of the fuel cell is opposite to that of the upper half. The middle cell is located between the two layers of cooling water, resulting in a symmetrical temperature distribution, with lower temperatures on both sides and higher temperatures in the middle. Figure 17 shows the temperature variation plots of the fuel cell with cooling water flowing in the same and opposite directions. From Figure 14a, at the same heat flux density, the maximum temperature is higher with cooling water flowing in the same direction compared to the opposite direction. As the heat source heat flux density increases from 1 × 10⁶ W/m³ to 5 × 10⁶ W/m³, the maximum temperature of the fuel cell with cooling water flowing in the same direction increases from 64.1 °C to 79.1 °C, while the maximum temperature of the fuel cell with cooling water flowing in the opposite direction increases from 63.5 °C to 76.5 °C, resulting in a difference of 2.6 °C between the two, which is higher compared to the 0.6 °C difference observed at lower heat flux density. However, the average temperature of the fuel cell is lower when the cooling water flows in the same direction. Additionally, at a lower heat flux density, the average temperature of the fuel cells is almost the same for both flow directions. However, at a higher heat flux density, the average temperature of the fuel cell with cooling water flowing in the opposite direction is significantly higher than that with cooling water flowing in the same direction. Furthermore, the maximum temperature difference is lower with cooling water flowing in the opposite direction, and the temperature distribution of the fuel cell is more uniform. As the heat flux density increases from 1 × 10⁶ W/m³ to 5 × 10⁶ W/m³, the maximum temperature difference of the fuel cell with cooling water flowing in the opposite direction increases from 2 °C to 7 °C.

Thus, the cooling water flow direction of the double-layer stack has a significant impact on the cooling effect of the battery. When the cooling water flows in the opposite direction, the average temperature of the battery is higher, but when the cooling water flows in the opposite direction, the highest temperature and maximum temperature difference of the stack are lower.

4. Conclusions

(1)

The ridge width of the cooling channel has a significant effect on the overall cooling performance of the stack. Among the three cooling channels with channel ridge widths of 0.5 mm, 1.5 mm, and 2 mm, the structure with a channel ridge width of 2 mm exhibits the best heat transfer performance.

(2)

The depth of the cooling channel also plays a crucial role in the cooling performance of the stack. Among the structures with channel depths of 0.5 mm, 1 mm, and 1.5 mm, the structure with a channel depth of 0.5 mm shows the best cooling performance.

(3)

The aspect ratio of the stacking size parameters, while being an influencing factor, has a negligible influence on the temperature distribution of the stack.

(4)

The cooling water flow direction of the double-layer stack has a significant impact on the cooling performance. When the coolant flows in reverse, the average temperature of the stack is higher. However, when considering the highest temperature and maximum temperature difference, the cooling performance of the reverse flow configuration is better.

(5)

The spacing between adjacent cells in the stack also affects the overall cooling efficiency. A smaller spacing results in better cooling performance due to improved heat transfer between the cells.