Optimized Design and Testing of Enhanced Heat Transfer Secondary Micro-Channels on the Surface of Fuel Cell Bipolar Plate Flow Channels

Abstract

Air-cooled proton exchange membrane fuel cells (PEMFCs) offer advantages such as light weight, compact size, and simple structure, and have been widely used in fields such as portable electronics, drones, and new energy electric vehicles. However, due to the influence of air convective cooling efficiency, air-cooled PEMFC can only operate at low power to avoid overheating. To improve the air-cooling efficiency and the maximum output power of PEMFCs, a new enhanced cooling structure has been proposed, which adds secondary micro-channels on the surface of the original bipolar plate flow channels. Thermal simulation analysis was conducted for flow channels with and without an array of micro-channels on the surface. Through orthogonal simulation experiments, the optimal geometric parameters for the secondary micro-channels were determined. The simulation results show that for flow channels with optimized secondary micro-channels, the maximum temperature at the center plane of the MEA is reduced by approximately 10 °C, the thermal resistance of heat transfer in the channel decreases by about 21.2%, and the experimental results on heat transfer in the channel indicate that the maximum heat flux density increases by approximately 22.5%. Finally, performance tests were conducted on air-cooled PEMFC stacks with and without enhanced cooling secondary micro-channels. The test results show that the fuel cell stack with enhanced cooling secondary micro-channels exhibits a temperature reduction of approximately 14 °C at a current density of 0.5 A/cm², a maximum output power increase of about 27%, and improved voltage uniformity across individual cells, demonstrating the effectiveness of the enhanced cooling secondary micro-channel structure.

1. Introduction

With the increasing demand for new energy, renewable energy technologies and devices have attracted growing attention from many countries. One such technology is the proton exchange membrane fuel cell (PEMFC), which can directly convert chemical energy into electrical energy. However, it is important to note that natural hydrogen resources are scarce, and the production of grey hydrogen from fossil fuels leads to significant carbon emissions. In light of global decarbonization efforts, the rapid development of hydrogen energy, particularly green hydrogen, has become increasingly urgent. Green hydrogen, produced via water electrolysis using renewable energy sources, provides a sustainable and carbon-free alternative to traditional fossil-based hydrogen production. While renewable energy sources such as solar and wind power are widely recognized as efficient substitutes for fossil fuels, their large-scale deployment is often constrained by their intermittent and variable nature. This intermittency poses challenges for grid stability and energy supply reliability. Hydrogen energy offers a promising solution to these issues by serving as an energy carrier that enables long-term energy storage and on-demand power generation. In particular, water electrolysis not only provides a clean hydrogen production method but also contributes to mitigating the fluctuation of renewable energy sources, thereby enhancing the overall efficiency and feasibility of a renewable energy-based power system [1,2,3].

Proton exchange membrane fuel cells play a significant role in the practical application of hydrogen energy. They feature pollution-free reaction products and low noise levels, and, unlike conventional thermal engines, they are not restricted by the Carnot cycle, resulting in high energy utilization efficiency. As a promising energy conversion device, PEMFCs have attracted considerable attention [4,5,6]. In particular, air-cooled fuel cells, due to their simple structure and lightweight design, have gained widespread interest and are primarily used in portable electronic devices and unmanned aerial vehicles [7,8,9].

Although proton exchange membrane fuel cells have shown great potential in clean energy applications, their short lifespan remains a major obstacle to commercialization. Over prolonged operation, PEMFC performance gradually deteriorates due to complex operating conditions and component aging, ultimately reaching the lowest acceptable efficiency threshold. The primary degradation mechanisms include catalyst degradation, membrane thinning, electrode flooding, and gas diffusion layer deterioration, all of which contribute to voltage loss and reduced power output. To address these challenges, prognostics and health management (PHM) strategies have been proposed to enable real-time monitoring, early fault detection, and predictive maintenance of fuel cells. By analyzing operational data and degradation patterns, PHM can help extend PEMFC lifespan and enhance system reliability. For instance, recent studies on fuel cell life prediction considering the recovery phenomenon of reversible voltage loss have demonstrated the potential of advanced monitoring techniques in mitigating performance degradation [10].

However, the heat transfer in air-cooled fuel cells is limited by the low specific heat capacity and heat transfer coefficient of air, leading to weak cooling performance. At high power densities, the cooling performance of traditional air-cooled PEMFC is significantly limited due to the low specific heat capacity and heat transfer coefficient of air. This limitation leads to inefficient heat dissipation, non-uniform temperature distribution, and localized overheating, which in turn accelerates membrane dehydration, decreases proton conductivity, and degrades overall fuel cell performance. As heat accumulates within the stack, excessive drying of the proton exchange membrane (PEM) reduces ion transport efficiency and increases internal resistance, ultimately shortening the PEMFC’s lifespan [11]. Since the cathode flow channel of the bipolar plate is a crucial pathway for heat dissipation in air-cooled PEMFCs, enhancing the heat transfer performance should be considered by optimizing the cooling channel structure.

To enhance the heat dissipation capability of flow channels and optimize fuel cell performance, extensive research has been conducted by scholars both domestically and internationally. Zhang et al. [12] designed a staggered fin manifold microchannel structure, and the results showed that compared to traditional rectangular microchannels, these channels exhibited superior heat transfer characteristics, reducing the maximum temperature by 5–10 K. Wang et al. [13] proposed a complex microchannel heat sink with fan-shaped grooves and triangular truncated ribs to enhance heat dissipation for high heat flux electronic devices. Zhang et al. [14] introduced fins to optimize the heat dissipation performance of microchannel heat sinks, addressing the low heat transfer efficiency caused by thermal boundary layers. Using CFD analysis, they examined the structural parameters and arrangement of internal fins, and the optimized model reduced the maximum and average temperatures by 6.67% and 6.75%, respectively. Nie et al. [15] improved the heat transfer performance of microchannels by embedding triangular ribs on the sidewalls. Their study found that isosceles or equilateral triangular ribs performed better than right-angled triangular ribs, with the maximum performance evaluation criterion (PEC) reaching 2.35 for equilateral ribs and 2.1 for isosceles ribs, while the maximum pressure drop was only 0.41 kPa higher than that of obstacle-free microchannels. Rahimi-Esbo et al. [16] proposed a novel metallic bipolar plate design incorporating baffles in the cooling flow field to allow fluid to pass through more uniformly, enhancing temperature uniformity within the fuel cell and optimizing overall performance. Yu et al. [17] designed an air-cooled PEMFC stack with a bipolar plate featuring concave–convex dual flow channels and investigated its heat dissipation performance. At a current density of 0.4 A/cm², when the airflow speed increased from 1 m/s to 1.5 m/s, the maximum temperature of the original bipolar plate dropped by 16.5%. Zhang et al. [18] proposed a novel elliptical dimple cooling (EDC) channel structure. Through numerical simulations, they compared the heat transfer characteristics of EDC, circular dimple cooling (CDC), and smooth cooling (SC) channels. The results showed that the narrow surface of the elliptical dimples enhanced fluid flow and created vortices that facilitated heat exchange. The EDC structure with a depth of 0.1 mm and a length of 0.6 mm outperformed SC, improving heat transfer efficiency by 10.6%. Zhang et al. [19] developed a cooling flow field characterized by waveform interweaving. Using 3D multiphase CFD analysis, they assessed how the cooling rate influenced temperature distribution. Their findings highlighted that the interleaved waveform pattern on both the cathode and anode resulted in more uniform water distribution across the membrane electrode assembly (MEA). Additionally, the reduced size of the cathode channels promoted more efficient heat dissipation, enhancing convective heat transfer. Despite the increased pressure drop, this cooling field exhibited improved temperature uniformity. Afshari et al. [20] compared a sawtooth-shaped water-cooled fuel cell cooling channel with a straight channel. The results indicated that in the sawtooth-shaped channel model, the maximum surface temperature, surface temperature difference, and temperature uniformity index were reduced by approximately 5%, 23%, and 8%, respectively, compared to the straight channel model. These findings highlight that designing an effective cooling structure for bipolar plate flow channels is a crucial measure to enhance the heat dissipation efficiency of air-cooled PEMFCs. Peng et al. [21] designed cooling channels with circular dimples. Compared with smooth channels, the cooling performance of dimpled channels improved by 10%, while pressure loss was reduced by 13% compared to wavy channels. It is evident that the design of heat dissipation structures for bipolar plate flow channels is a crucial measure to enhance the cooling efficiency of air-cooled PEMFCs.

2. Principle and Structure Design

Air-cooled fuel cell stacks are typically composed of many individual cells stacked together, as the output power of a single cell is too low to meet the demands of practical applications. However, as the output power of the stack increases, a significant amount of heat is generated. To ensure stable operation of the air-cooled fuel cell stack, effective cooling must be applied to maintain the temperature within the required range [22]. The operating temperature of air-cooled stacks is generally around 50–80 °C. When the temperature exceeds the specified limit, catalytic reactions are inhibited, leading to a significant decline in performance, and potentially affecting the lifespan of the membrane electrode assembly (MEA) [23]. During operation, electrochemical reactions occur within the MEA of the air-cooled PEMFC, with the generated electrical energy being output to the external circuit, accompanied by the production of heat. Figure 1 illustrates the structure and principle of the air-cooled PEMFC. The cathode flow channels of the air-cooled PEMFC are open and continuous, with air being supplied directly to the cathode and used for cooling via a fan located outside the fuel cell stack.

During the operation of an air-cooled PEMFC, electrochemical reactions occur within the membrane electrode assembly (MEA), generating electrical energy that is delivered to an external circuit while also producing reaction heat. The power output P of the PEMFC can be expressed by the following equation [24]:

$$P=V_{\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}}\times I\times A$$

(1)

where $V_{\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}}$ is the voltage of a single cell, $I$ is the current density, $A$ is the active area.

However, due to ohmic losses and polarization effects, a significant amount of heat is generated during PEMFC operation. This heat mainly originates from Joule heating $Q_{\mathrm{o}\mathrm{h}\mathrm{m}\mathrm{i}\mathrm{c}}$ and electrochemical reaction heat $Q_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}$:

$$Q_{\mathrm{o}\mathrm{h}\mathrm{m}\mathrm{i}\mathrm{c}}=I^{2}R$$

(2)

$$Q_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}=∆H-∆G$$

(3)

where $R$ is the resistance of the membrane and electrodes, $∆H$ is the enthalpy change in the reaction, $∆G$ is the Gibbs free energy change.

In summary, the total heat flux can be expressed as

$$Q_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}=Q_{\mathrm{o}\mathrm{h}\mathrm{m}\mathrm{i}\mathrm{c}}{+Q}_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}{-Q}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}{-Q}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}$$

(4)

where $Q_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}$ is the heat removed by cooling air through the flow channels, $Q_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}$ is the heat dissipated via thermal conduction through the bipolar plates and other structural components. If the heat flux is not effectively dissipated, a rise in local temperature may lead to low power output.

Therefore, it can be intuitively observed that one of the ways to reduce heat is to increase the amount of heat carried away by the cooling air in the bipolar plate channels. Since the cathode flow channels of the bipolar plate serve as the primary heat dissipation pathway, optimizing the channel structure to enhance heat transfer is crucial.

Currently, a structure with secondary grooves has been applied in some plate heat exchangers, as shown in Figure 2. This design offers advantages such as simplicity in structure and ease of manufacturing. Therefore, a new enhanced heat dissipation structure, which incorporates secondary micro-grooves on the surface of the air-cooled PEMFC bipolar plate flow channels, is proposed. This design ensures good compatibility with the original flow channel structure of the bipolar plate. The presence of secondary grooves increases the heat transfer area, enhances heat exchange efficiency, and can also help reduce the overall weight of the air-cooled PEMFC to some extent.

Based on this, to enhance the heat dissipation capability of the air-cooled PEMFC bipolar plate, a novel secondary micro-channel structure has been designed, as shown in Figure 3. The newly designed secondary micro-channels are located between the sidewalls and the bottom surface of the traditional rectangular straight channels. The structure is simple, and the presence of the secondary micro-channels significantly increases the forced convective heat transfer area of the air within the flow channel, thereby improving the heat dissipation performance. This paper focuses on enhancing heat transfer by designing a novel bipolar plate cathode flow channel structure with secondary grooves. A single-channel heat transfer simulation model is established to compare the heat transfer performance of the novel cathode flow channel with that of a traditional one and to analyze the mechanism by which secondary grooves enhance heat transfer. An orthogonal experiment is conducted to analyze and optimize the geometric parameters of the secondary grooves, determining the optimal parameters to further improve heat transfer performance. Finally, experiments and tests are carried out to validate the designed secondary groove flow channel’s effectiveness in enhancing heat dissipation and its overall performance in air-cooled PEMFCs.

3. Comparison of Heat Transfer in Secondary Micro-Channels Flow Field Structure

3.1. Heat Transfer Simulation Model and Boundary Conditions

This study primarily analyzes the heat transfer characteristics of the conventional cathode flow channel and the novel secondary micro-channel cathode flow channel in air-cooled fuel cell bipolar plates. Based on the heat dissipation characteristics of air-cooled fuel cells and simplified calculations, a single flow channel is used as the computational model due to symmetry and periodicity conditions, as shown in Figure 4. Figure 4a shows the conventional flow field (CF), while Figure 4b illustrates the secondary micro-channel flow field (SMF). Here, L represents the length of the flow channel, L = 60 mm; denote the width and depth of the conventional rectangular flow channel, W = 1.5 mm, H = 1.8 mm; the width and depth of the secondary micro-channels are w = h = 0.2 mm; s represents the spacing between the secondary micro-channels, set at s = 0.5 mm.

The following assumptions are made for the model: (1) During the operation of the air-cooled PEMFC, heat is primarily generated from the membrane electrode assembly (MEA), which is considered a uniform heat source. (2) The electrochemical reaction process of the fuel cell is not considered; only airflow and heat transfer in the cathode flow channel are analyzed. (3) The anode flow channel is neglected due to the low velocity of hydrogen flow on the anode side. (4) The airflow is assumed to be steady, laminar, and incompressible (since the Reynolds number in this study is low, <1000). (5) The flow is fully developed, and both the fluid and solid are assumed to have constant physical properties. (6) Radiation heat transfer is neglected. Based on these assumptions, the governing equations for analyzing fluid flow and heat transfer include the continuity equation, momentum equation, and energy equation.

Continuity equation:

$$\nabla ·\overrightarrow{u}=0$$

(5)

Momentum equation:

$$\rho \left(\overrightarrow{u}·\nabla \right)\overrightarrow{u}=-\nabla p+\mu {\nabla }^2\overrightarrow{u}$$

(6)

Energy equation:

$$\rho C_p\left(\overrightarrow{u}·\nabla \mathrm{T}\right)=k_f{\nabla }^{2}T$$

(7)

Energy equation for the solid domain:

$$\nabla \cdot \left(k_s\nabla T\right)=0$$

(8)

The boundary conditions for the single-channel model are as follows: Periodic boundary conditions are applied to the top and bottom surfaces of the model, while symmetric boundary conditions are applied to the side surfaces. For the inlet boundary conditions, it is assumed that the air velocity and temperature are uniform, with the inlet velocity specified as $u=u_{in}$ and the temperature as $T=T_{in}$. For the outlet boundary conditions, since it is exposed to the environment, the external atmospheric pressure is specified as $P=P_0$. The MEA is considered a heat source with a constant and uniform heat power of ${Q=Q}_0$. The specific simulation parameters are shown in

Table 1. Simulation parameters.

Parameter	Value
$T_{in}$(K)	293.15
$u_{in}$(m/s)	3–4.5
$P_0\left(\mathrm{P}\mathrm{a}\right)$	0
$Q_0$ (W)	0.5

. The simulation was conducted on a computer with the following specifications: Intel Core i5-12400F, 6 cores, 3.0 GHz, and 32 GB RAM, the equipment was sourced from Intel (Atlanta, GA, USA) and Kingbank (Shenzhen, China). The software version used was COMSOL Multiphysics 6.0. The computational time for each case was approximately 1 h. The properties of air, such as density, specific heat capacity, viscosity, and thermal conductivity, were temperature-dependent and obtained from the COMSOL material library.

3.2. Mesh Scheme and Independence Verification

The heat transfer simulation was conducted using COMSOL Multiphysics 6.0. The physical field interfaces were set to “Heat Transfer in Solids and Fluids” and “Laminar Flow”, and the coupling was achieved using the non-isothermal flow method under steady-state conditions. The mesh generation was performed using the mesh module in COMSOL. Given the simplicity of the physical model and flow channel structure, an automatic mesh generation method controlled by the physics field was applied. To ensure computational accuracy, the mesh in the fluid domain was calibrated for fluid dynamics, resulting in smaller and more numerous mesh elements. The minimum element size was set to 0.02 mm, and the maximum element growth rate was 1.15. At the fluid–wall interface, boundary layer meshing was applied with two boundary layers and a stretching factor of 1.2 to improve the accuracy of heat transfer and flow characteristics near the walls.

To ensure result accuracy while improving computational efficiency, a mesh independence analysis was conducted. This helps evaluate the impact of different mesh configurations on the solution results, allowing us to determine an appropriate mesh density for efficient numerical computation. A traditional rectangular straight flow channel was used as a reference, and mesh independence verification was carried out by selecting mesh sizes of 3,968,520, 7,113,970, 11,658,580, 15,563,920, 21,840,640, and 31,619,890 elements. The highest temperature on the MEA surface and the pressure drop between the inlet and outlet of the flow channel were used as evaluation criteria, with an inlet airflow velocity of 4 m/s. The results of the mesh independence verification are shown in Figure 5. As the number of mesh elements increased, the computed values of the highest temperature on the MEA surface and the pressure drop between the inlet and outlet gradually stabilized. When the mesh size increased from 21,840,640 to 31,619,890 elements, the variation in the highest temperature on the MEA surface was only 0.05%, while the variation in the pressure drop was 0.6%. These changes were minimal, indicating that further mesh refinement would have a negligible effect on the numerical simulation results. Therefore, the final mesh size was determined to be 21,840,640 elements. The meshing results are shown in Figure 6.

3.3. Simulation Results Analysis

As shown in Figure 7, the surface temperature distribution of the bipolar plate and MEA for both flow channel models at an inlet velocity of 4 m/s is illustrated. It can be observed that the temperature is lower at the inlet for both channels and gradually increases from the inlet to the outlet. This is because the primary mode of heat transfer between the fluid and the wall is thermal conduction, which is mainly influenced by the temperature difference between the fluid and the solid. The inlet cooling air has a lower temperature, and as it flows through the channel, it absorbs significant heat, resulting in a temperature drop. However, as the air continues to flow towards the rear end of the channel, its temperature rises, reducing the temperature difference and consequently decreasing the heat transfer rate. Thus, the temperature at the rear end of the channel is higher than at the front end. Comparing the surface temperature distribution of the two flow channels, it is evident that from the inlet to the outlet, the surface temperature of the bipolar plate and MEA in the secondary micro-channels flow field is lower than that of the traditional straight flow channel, with the maximum temperature reduced by approximately 4 °C. To more intuitively illustrate the effect of the secondary micro-channels, the temperature distribution contour at the bottom cross-section of both flow channels was selected when the micro-channels depth H = 1.8 mm, as shown in Figure 8. As can be clearly seen in Figure 8b, the three lower-temperature strip-shaped regions correspond to the locations of the secondary micro-channels. The presence of these micro-channels increases the forced convective heat exchange area for air, converting part of the solid bipolar plate region into a fluid region. As a result, at the same position, the temperature of the fluid is lower than that of the solid region, creating a temperature gradient that enhances heat exchange. Consequently, compared with the traditional straight flow channel, the temperature of the bipolar plate and MEA in the secondary micro-channels flow field is significantly reduced.

To more clearly observe the effect of the secondary micro-channels, the velocity distribution at the mid-section of the flow channel is shown in Figure 9, where the left side represents the conventional straight channel, and the right side represents the channel with secondary micro-channels. As shown, the velocity near the solid walls of the conventional straight channel is almost zero, resulting in a thicker boundary layer. This increases thermal resistance in the convective heat transfer process and reduces overall heat transfer efficiency. In contrast, the channel with secondary micro-channels features multiple micro-channels on the bottom and sidewalls, which not only increase the airflow rate and enhance the velocity at the channel center but also alter the boundary layer near the original wall. A portion of the new boundary layer develops inside the secondary micro-channels, while another portion in contact with the original wall becomes thinner, ultimately enhancing heat transfer efficiency.

Based on the above analysis, it is evident that the flow channel with secondary micro-channels exhibits significantly better heat transfer performance than the traditional straight channel structure, playing a crucial role in reducing the temperature of the membrane electrode assembly (MEA) heat source. Since the MEA is the primary heat source of the air-cooled PEMFC, a comparative analysis was conducted to evaluate the effectiveness of the secondary micro-channels flow field in cooling the MEA. This was conducted by examining the difference in the maximum temperature at the central plane of the MEA under the two flow channel structures. Furthermore, the total thermal resistance of the model was calculated based on the results of the heat transfer simulation model, which reflects the effectiveness of the heat transfer performance. The expression for the thermal resistance of the flow channel is given as follows:

$$R={\displaystyle \frac{T_{wmax}-T_{in}}{Q}}$$

(9)

where $T_{wmax}$ is the maximum temperature of the flow channel wall, $T_{in}$ is the inlet temperature, $Q$ is the heat power.

Due to the presence of secondary micro-channels, the cross-sectional area of the flow channel increases, inevitably leading to a higher internal pressure compared to the traditional straight flow channel. Therefore, it is necessary to compare the pressure drop variations between the two flow channel types. If the pressure drop in the novel secondary micro-channels flow field is significantly greater than that of the traditional straight channel, a comprehensive evaluation of the trade-off between heat transfer performance and pressure drop is required. Hence, the pressure drop between the inlet and outlet of the flow channel is calculated using the following formula:

$$\mathsf{\Delta }P=P_{in}-P_{out}$$

(10)

where $P_{in}$ is the inlet pressure, $P_{out}$ is the outlet pressure.

As shown in Figure 10, the line charts depict the variations in the maximum temperature at the center plane of the membrane electrode assembly (MEA), the thermal resistance, and the pressure drop of the flow channels under different inlet velocities for both flow channel types.

As shown in Figure 10a,b, at different air inlet velocities, the maximum temperature and thermal resistance of the MEA central plane are both lower when using the secondary micro-channels flow field. This demonstrates its significantly superior cooling performance compared to the traditional straight flow channel. Replacing the conventional design with the secondary micro-channels flow field enhances cooling efficiency, ensuring more stable fuel cell operation. Additionally, the secondary micro-channels flow field achieves the same cooling effect as the traditional straight channel at lower air inlet velocities. This reduces the power demand of the cooling fan, leading to lower overall system energy consumption. As shown in Figure 10c, the pressure drop across the inlet and outlet of the secondary micro-channels flow field is slightly higher than that of the traditional straight flow channel due to the increased cross-sectional area. However, data analysis shows that at different flow velocities, the pressure drop remains within 30 to 60 Pa, a relatively low range. At a flow velocity of 4.5 m/s, the pressure drop of the secondary micro-channels flow field is only 5.13 Pa higher than that of the straight flow channel. Given that the cathode flow channel of an air-cooled fuel cell is open to the external environment, this minor increase in pressure drop has a negligible impact on overall system performance.

3.4. Optimization of Secondary Micro-Channel Structural Parameters

The flow channel with secondary micro-channels plays a certain role in enhancing heat dissipation. To further improve the heat transfer performance of the channel, an optimization design of the secondary micro-channels parameters is carried out. In Figure 4b, the width (w), depth (h), and inter micro-channels distance (s) of the secondary micro-channels are selected as the main research parameters. These are taken as experimental factors, and their values are determined according to the actual structural dimensions of the bipolar plate channel, as shown in

Table 2. Different levels of experimental factors.

Item	Level Number	Parameter Name
		w (mm)	h (mm)	s (mm)
Parameter Values	1	0.1	0.2	0.4
	2	0.2	0.3	0.5
	3	0.3	0.4	0.6

. A three-factor, three-level orthogonal table was designed, with the maximum temperature at the center plane of the MEA selected as the objective function. The calculation results are shown in

Table 3. Orthogonal experimental simulation results.

Number	w (mm)	h (mm)	s (mm)	${\mathit{T}}_{\mathit{m}\mathit{a}\mathit{x}}$ (°C)
1	0.1	0.2	0.4	70.52
2	0.1	0.3	0.5	69.32
3	0.1	0.4	0.6	68.39
4	0.2	0.2	0.5	68.5
5	0.2	0.3	0.6	66.58
6	0.2	0.4	0.4	65.43
7	0.3	0.2	0.6	66.38
8	0.3	0.3	0.4	64.22
9	0.3	0.4	0.5	62.22

.

A range analysis of the experimental data in Table 3 was conducted to identify the parameter combination that minimizes the maximum temperature of the MEA. The results of the range analysis for the maximum temperature are shown in

Table 4. Range analysis results for maximum temperature.

Objective Function	Analysis Factors	Experimental Factors
		w	h	s
$T_{max}$	$k_1$	69.41	68.47	66.72
	$k_2$	66.84	66.71	66.68
	$k_3$	64.27	65.35	67.12
	$R$	5.173	3.12	0.44
Primary factor order	w > h > s
Optimal combination	w3h3s2

. As observed from the table, the width and depth of the secondary micro-channels are the primary factors influencing heat transfer performance, exerting the most significant impact on the maximum temperature. This is because the width and depth of the secondary micro-channels not only directly determine their cross-sectional area but also significantly affect the heat transfer surface area and the airflow rate through the channel. When the width and depth of the secondary micro-channels are optimized, the convective heat transfer effect of the fluid is significantly enhanced, thereby improving the overall heat dissipation performance of the channel. The inter micro-channels spacing has a relatively smaller impact. Based on the range analysis results, the optimal parameters that minimize the maximum temperature are determined as follows: w = 0.3 mm, h = 0.4 mm, s = 0.5 mm.

A comparison between the optimized secondary micro-channels model and the traditional straight-flow channel model was conducted. At an inlet velocity of 4 m/s, the maximum temperature at the center plane of the MEA and the thermal resistance of the model were calculated, as shown in

Table 5. Optimization results for secondary micro-channels flow field.

	${\mathit{T}}_{\mathit{m}\mathit{a}\mathit{x}}$ (°C)	$\mathit{R}$ (°C/W)
Straight-flow channel	72.23	99.78
Optimized Secondary Micro-channel	62.22	78.64

. It is evident that, compared to the traditional straight-flow channel, the optimized secondary micro-channels flow field reduces the maximum temperature $T_{max}$ at the MEA center plane by approximately 10 °C and decreases the thermal resistance $R$ by about 21.2%, significantly enhancing the heat transfer performance.

3.5. Experimental Validation of Channel Heat Transfer

To validate the enhanced heat transfer performance of the optimized secondary micro-channels flow field, a heat transfer experimental platform was designed for comparative testing. The schematic of the heat transfer experimental system is shown in Figure 11. The test specimens were made of 316L stainless steel, with two pieces in total, each containing three channels on the top surface. One specimen featured a conventional straight channel structure, while the other incorporated the secondary micro-channels flow field design. During the experiment, a temperature-adjustable heating element was attached to the bottom of the test specimen and powered by a DC power supply. By adjusting the voltage and current of the power supply, the output heat power of the heating element could be precisely controlled. The test specimens were wrapped with Teflon insulation blocks, which effectively minimized heat exchange with the external environment due to Teflon’s extremely low thermal conductivity. Additionally, fiber insulation tape was applied to the top of the channel and the Teflon block to further reduce heat loss from the upper surface. Air was introduced from one side of the channel, and the airflow rate was precisely controlled using a flow meter. To monitor the temperature distribution on the sidewalls of the test specimens, three thermocouples were employed. The temperature data were recorded using a data acquisition instrument and then imported into a computer for further processing and analysis.

To evaluate the improvement in heat transfer performance of the secondary micro-channels flow field compared to the conventional straight channel, this experiment was conducted under controlled conditions, keeping the heat source and other variables constant while only varying the test specimens, namely, the secondary micro-channels flow field and the conventional straight channel. During the experiment, the heat source was set to a power of 1.5 W. Before introducing airflow, air was only supplied once the average temperature recorded by the thermocouples reached 50 °C. The inlet air temperature was maintained at 25 °C. The airflow rate was regulated using a flow meter connected to a gas cylinder. The adjustment knob on the flow meter allowed for precise control of the airflow. To investigate the heat transfer performance at different flow rates, the airflow was varied between 1 L/min and 2.5 L/min, divided into four levels. At each airflow level, the test was conducted for 10 min before switching to the next level. However, in practical experiments, since the heating element was positioned at the bottom of the test specimen, not all of the generated heat could be transferred to the channel, leading to inevitable heat loss. To more accurately assess the heat transfer performance of the two channel designs, three thermocouples were installed on the sidewall of the test specimens, with a spacing of 5 mm. Once the temperature stabilized, the average value of the last 10 temperature readings from the thermocouples was recorded as the steady-state temperature. The temperature gradient was then used to calculate the heat flux density. The heat flux density was determined using Fourier’s law, as expressed by the following equation:

$$q=-k_s\cdot {\displaystyle \frac{dT}{dx}}$$

(11)

where $q$ is the heat flux density, $k_s$ is the thermal conductivity of the test specimen, $dT$/$dx$ is the temperature gradient.

Since heat is transferred in a one-dimensional linear manner from the heating element through the test specimen, the temperature gradient is calculated using the following equation:

$${\displaystyle \frac{dT}{dx}}={\displaystyle \frac{1}{2}}\left({\displaystyle \frac{T_2-T_1}{\mathsf{\Delta }x}}+{\displaystyle \frac{T_3-T_2}{\mathsf{\Delta }x}}\right)$$

(12)

In the equation, $\mathsf{\Delta }x$ represents the distance between $T_2$ and $T_1$, as well as between $T_3$ and $T_2$, both of which are 5 mm in this case. $T_1$, $T_2$, and $T_3$ correspond to the temperatures measured by the thermocouples.

Once the system reaches a stable temperature, air is introduced. Data are recorded for 10 min at each airflow rate before switching to the next airflow setting. The variation in heat flux density for the two flow channels under different airflow rates is illustrated in Figure 12.

As shown in Figure 12, upon introducing air, the heat flux density of both flow channel workpieces significantly increases compared to the condition without airflow. This is because, in the absence of airflow, heat is transferred from the heating film at the bottom of the workpiece to its interior, leading to a certain level of heat flux density due to thermal conduction [25]. However, when cooling air is introduced, the heat flux density rises sharply due to the forced convection heat transfer occurring in the top flow channel. This causes the temperature at the top to decrease, increasing the temperature difference and consequently enhancing the heat flux density from the interior of the workpiece to the flow channel. In this experiment, as the airflow rate increases, meaning the air velocity within the flow channel rises, the heat flux density of both flow channels correspondingly increases. This is because higher airflow velocity enhances convective heat transfer at the solid surface, removing more heat and thus lowering the solid surface temperature. This increases the internal temperature gradient of the workpiece, leading to a higher heat flux density. However, as the airflow velocity continues to increase, the rate of heat flux density growth gradually diminishes. This is because the thermal conductivity of the workpiece material is significantly lower than the rate of forced convective heat transfer, thereby limiting the heat transfer rate. Comparison results indicate that the heat flux density of the workpiece with secondary micro-channels flow field is consistently higher than that of the traditional straight flow channel across different air inlet flow rates. Since all other conditions remain unchanged and the only variable is the flow channel type, this suggests that the internal heat flux density of the workpiece with secondary micro-channels is higher. This implies that under identical conditions, heat transfer is more efficient, significantly enhancing heat dissipation performance. Through calculations, the heat flux density of the secondary groove channel workpiece is increased by approximately 22.5% compared to the conventional straight channel workpiece. This indicates that the optimized secondary groove channel improves heat dissipation performance by about 22.5%, which is consistent with the heat transfer simulation results showing a reduction in thermal resistance of approximately 21.2%. This consistency validates the effectiveness of the secondary groove channel structure in enhancing heat transfer, providing strong support for the subsequent performance testing experiments of the air-cooled PEMFC stack.

4. Air-Cooled PEMFC Testing

4.1. Experimental Conditions and Methods

Based on the aforementioned study on the enhanced heat transfer of the secondary micro-channels flow field, an air-cooled fuel cell stack was independently designed and assembled to verify the beneficial effects of the secondary micro-channels cathode flow field on air-cooled fuel cells. A schematic diagram of the experimental setup is shown in Figure 13.

The air-cooled PEMFC stack used in the experiment consists of 15 membrane electrode assemblies (MEAs) with an active area of 37 cm², which were provided by Suzhou Qing dong Power Technology Co., Ltd. The stack is assembled with 16 bipolar plates, and the dimensions of the cathode flow channels are the same as those in the model described in the previous section. The secondary groove dimensions are based on the optimized parameters. The overall size of each bipolar plate is 120 mm × 60 mm × 3 mm, with 32 cathode flow channels, all manufactured using electrical discharge machining (EDM). The physical appearance of the two stacks is shown in Figure 14a, where the only difference is the cathode flow channel structure of the bipolar plates, while all other components remain the same. The testing experiments were conducted using the fuel cell test platform shown in Figure 14b. The ambient air temperature was approximately 25 °C, with a relative humidity of around 40%. The hydrogen inlet pressure was set to 50 kPa, and the airflow rate was automatically adjusted by a programmed controller via a fan. A solenoid valve installed at the hydrogen outlet was programmed to open every 20 s for 0.5 s to periodically purge the anode exhaust and water. Prior to the experiment, the fuel cell stack was activated. The polarization performance of the stack was then tested under a constant current mode. An electronic load was used to gradually apply current, with each set current value maintained for 5 min before recording the experimental data. The operation was halted once the stack temperature reached 60 °C.

4.2. Results and Discussion

The test results of the fuel cell stacks with conventional cathode flow channels and secondary micro-channels cathode flow field are shown in Figure 15a,b, which present the performance curves and the temperature variations in the stacks under different current densities. By analyzing both figures, it is evident that as the current density increases, the output power of the fuel cell stack also increases, accompanied by a rise in stack temperature. Notably, within 0.5 A/cm², the temperature of SMF-based stack is lower than that of CF-based stack. When the current density reaches 0.5 A/cm², the temperature of the conventional straight-channel stack reaches 58 °C. Further increasing the load would cause the temperature to exceed 60 °C, which led to the termination of the test at this point. Under these conditions, the maximum output power of the conventional stack is 162 W. In contrast, the stack with secondary micro-channels flow field exhibits superior heat dissipation due to the presence of secondary micro-channels. At the same current density of 0.5 A/cm², its temperature is only 44 °C, which is 14 °C lower than that of the conventional straight-channel stack, demonstrating a significant enhancement in cooling performance. As seen from the previous equations, introducing a secondary microchannel structure enhances convective heat dissipation and raises the $Q_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}$, thereby reducing the overall temperature and improving the performance of the air-cooled PEMFC. Consequently, the current density can be further increased to 0.7 A/cm² before the stack temperature reaches 60 °C. At this point, the maximum output power of the stack with secondary micro-channels flow field reaches 205 W, representing a 27% increase compared to the conventional straight-channel stack. This highlights a significant improvement in fuel cell performance.

The single-cell voltages of both fuel cell stacks under current densities ranging from 0.1 A/cm² to 0.5 A/cm² are shown in Figure 16. It can be observed that at all current densities, the single-cell voltage of the stack with secondary micro-channels flow field is consistently higher than that of the stack with conventional flow channels. At low current densities, the membrane electrode reaction occurs at a moderate rate, preventing the stack temperature from rising significantly. At this stage, both fuel cell stacks exhibit good uniformity in single-cell voltage. However, as the current density increases, the uniformity of the single-cell voltage in the conventional flow channel stack deteriorates at a current density of 0.5 A/cm². Specifically, the voltage of the central cells is lower than that of the cells at both ends of the stack. This phenomenon occurs because heat tends to accumulate more easily in the central region of the stack compared to the two ends, resulting in a higher central temperature and larger thermal gradients. In contrast, the stack with secondary micro-channels flow field maintains high single-cell voltage uniformity even at a current density of 0.5 A/cm². This indicates that the enhanced heat dissipation provided by the secondary micro-channels flow field improves the overall stability of the fuel cell stack.

The comparison of our research results with selected studies in the literature is shown in

Table 6. Comparison of research results with studies in the literature.

Study	Method	Result
This Study	Secondary microchannel structure in bipolar plate	Heat transfer performance improved by 22.5%, temperature reduced by 14 °C, and stack output power increased by 27%
Zhang et al. [12]	Staggered fin manifold microchannel structure	Reduced maximum temperature by 5–10 K
Wang et al. [13]	Microchannel heat sink with fan-shaped grooves and triangular truncated ribs	Improved heat dissipation for high heat flux electronic devices
Zhang et al. [14]	Microchannel heat sink with internal fins	Reduced maximum and average temperatures by 6.67% and 6.75%
Nie et al. [15]	Microchannels with embedded triangular ribs on sidewalls	Maximum PEC of 2.35 (equilateral ribs) and 2.1 (isosceles ribs); pressure drop increased by only 0.41 kPa
Rahimi-Esbo et al. [16]	Metallic bipolar plate with baffles in cooling flow field	Improved fuel cell temperature uniformity and overall performance
Yu et al. [17]	Air-cooled PEMFC stack with concave–convex dual flow channels	At 0.4 A/cm2, increasing airflow from 1 m/s to 1.5 m/s reduced max temperature by 16.5%
Zhang et al. [18]	Elliptical dimple cooling (EDC) channel structure	Improved heat transfer efficiency by 10.6% over smooth channels
Zhang et al. [19]	Waveform interweaving cooling flow field	More uniform water distribution in MEA; enhanced heat dissipation despite increased pressure drop
Afshari et al. [20]	Sawtooth-shaped water-cooled fuel cell cooling channel	Reduced max surface temperature, temperature difference, and uniformity index by 5%, 23%, and 8% compared to straight channels
Peng et al. [21]	Cooling channels with circular dimples	Improved cooling performance by 10%; pressure loss reduced by 13% compared to wavy channels

.

5. Conclusions

To enhance the heat dissipation capability of air-cooled PEMFCs and improve their operational performance, this study proposes a bipolar plate cathode flow field with a secondary microchannel structure. Based on simulation analysis and experimental testing, the conclusions are as follows:

(1) A novel bipolar plate cathode flow field with secondary micro-channels was designed, and a single-channel heat transfer simulation model was developed to compare the heat transfer performance of the secondary micro-channels cathode flow field with that of a conventional cathode flow field. The results indicate that the presence of secondary micro-channels increases the convective heat transfer area, improves the flow boundary layer, and enhances heat transfer efficiency.

(2) The optimal geometric parameters of the secondary micro-channels were determined through orthogonal simulation experiments. The simulation results indicate that, with the optimized secondary micro-channels parameters, the maximum temperature at the center plane of the membrane electrode assembly (MEA) is reduced by approximately 10 °C, and the thermal resistance of the flow field is decreased by about 21.2%. Furthermore, experimental results on heat transfer demonstrate that the heat flux density of the optimized secondary micro-channels flow field is enhanced by up to 22.5%, verifying the effectiveness of the secondary micro-channels structure in enhancing heat transfer.

(3) Performance tests were conducted on an air-cooled proton exchange membrane fuel cell (PEMFC) stack with and without enhanced heat dissipation secondary micro-channels. The experimental results show that the fuel cell stack incorporating enhanced heat dissipation secondary micro-channels achieves a 27% increase in maximum output power. At a current density of 0.5 A/cm², the stack temperature is reduced by 24%, and the uniformity of individual cell voltages is improved, confirming the effectiveness of the enhanced heat dissipation secondary micro-channels structure.

6. Outlook and Future Work

In summary, the positive feedback from the experiment results validates the effectiveness of the proposed secondary micro-groove bipolar plate flow channel structure in enhancing heat dissipation and improving power output for the air-cooled PEMFC stack from the perspective of heat transfer enhancement. However, it is well known that water management significantly impacts PEMFC output performance, as excessive water accumulation can lead to flooding, hinder gas diffusion, and accelerate performance degradation, particularly under high current densities [26]. While this study primarily focuses on enhancing heat dissipation by optimizing the flow channel structure, we acknowledge the critical role of water management in PEMFC operation. The influence of the secondary micro-groove structure on water removal has not been explicitly analyzed in this work. However, considering the close relationship between thermal management and water management, future research will explore the effects of secondary micro-grooves on both heat dissipation and water drainage to further optimize the performance of air-cooled PEMFCs. Furthermore, the influence of different bipolar plate materials on heat transfer characteristics can be explored, such as using high thermal conductivity composite materials to enhance cooling efficiency. Secondly, in experimental studies, higher-precision temperature sensors or infrared thermal imaging technology can be integrated to further optimize the measurement methods for flow and thermal fields. Additionally, multi-scale numerical simulation methods can be investigated to more accurately predict the impact of complex flow channel structures on heat transfer performance. Finally, in practical applications, the long-term stability of the optimized structure and its adaptability under different operating conditions should be further evaluated to promote the commercialization of air-cooled PEMFC technology.