Enhancing Heat Removal and H₂O Retention in Passive Air-Cooled Polymer Electrolyte Membrane Fuel Cells by Altering Flow Field Geometry

Abstract

This numerical study presents six three-dimensional (3D) cathode flow field designs for a passive air-cooled polymer electrolyte membrane (PEM) fuel cell to enhance heat removal and H₂O retention. The data collected are evaluated in terms of water content, average temperature, and current flux density. The proposed cathode flow field designs are a straight baseline channel (Design 1), converging channel (Design 2), diverging channel (Design 3), straight channel with cylindrical pin fins (Design 4), trapezium cross-section channel (Design 5), and semi-circle cross-section channel (Design 6). The lowest cell temperature value of 56.67 °C was obtained for Design 2, while a noticeable water retention improvement of 6.5% was achieved in a semi-circle cathode flow field (Design 5) compared to the baseline channel. However, the current flux density shows a reduction of 0.1% to 1.2%. Nevertheless, those values are relatively small compared to the improvement in the durability of the fuel cell due to heat reduction. Although the modifications to the cathode flow field resulted in only minor improvements, ongoing advancements in fuel cell technology have the potential to make our energy landscape more sustainable. These advancements can help reduce emissions, increase efficiency, integrate renewable energy sources, enhance energy security, and support the transition to a hydrogen-based economy.

1. Introduction

The growing need for electronics, including telecommunications devices, military intelligence apparatus like unmanned aerial vehicles, medical devices, and drones, has prompted researchers in the field to investigate readily available alternatives for power sources that align with sustainability goals. Because of significant and ongoing technological and industrial development, almost everything around us now depends on electricity. Therefore, scientists have begun searching for methods that enable various devices to produce clean electrical energy. Fuel cells, which convert existing chemical energy into electrical energy, are the leading solutions. With the interaction of liquid hydrogen with liquid oxygen, this conversion produces water in addition to electrical energy.

Fuel cells can be classified into different types depending on the electrolyte materials they use, such as polymer electrolyte membrane (PEM) fuel cells [1], alkaline fuel cells [2], molten carbonate fuel cells (MCFC) [3], phosphoric acid fuel cells [4], and solid oxide fuel cells (SOFC) [5]. PEM fuel cells have several unique benefits. They can promptly adapt to fluctuations in power and capacity requirements. Furthermore, they initialize quickly, have a prolonged lifespan, operate at low temperatures, and demonstrate less self-discharge. Notably, they require a temperature range of 50 °C to 80 °C, and any increase in temperature should be avoided to prevent the drying of the outer coating.

The most prominent and best techniques for cooling PEM liquid [6,7] include air cooling [8,9], phase-change cooling [10], the use of heat spreaders [11], and heat pipe cooling [12,13]. In recent years, several air-cooled PEM fuel cells have been proposed, deployed, and tested. In a PEM fuel cell system, heat transmission is crucial to cell cooling. Excess heat builds according to the reaction rate, and if the temperature exceeds a recommended limit, its performance is severely reduced [14,15]. The geometry of the cooling channel is a major factor in enhancing heat removal and H₂O retention. Lasbet et al. [16] showed that during convection, channel geometry affects the trajectory of water removal in addition to heat removal. Similarly, Ying et al. [17] demonstrated that changes in the channel geometry significantly influence the performance of the passive air-cooled PEM fuel cell. Jeong et al. [18] studied the cathode open area ratio of a cell and discovered that the performance of the cell decreased when the ratio increased. This was predominantly due to greater plane resistance and higher activation loss. Since membrane dehydration caused by an excessive supply of dry air is a common issue limiting PEM fuel cell performance in PEMs [19,20,21], in the absence of external humidification or under dynamic operating conditions, water transport behavior plays an important role in the design and functioning of air-cooled fuel cells [22,23].

Many studies have investigated the operating current of a PEM fuel cell, which also influences the temperature distribution in a cell stack. For example, Shahsavari et al. [24] used a simplified thermal fuel cell model to analyze the behavior of an air-cooled PEM fuel cell. Renau et al. [25] conducted a study on the operating current, applying a phosphoric acid-doped polybenzimidazole membrane to a UAV at a high temperature in preparation for a high-amplitude trip, successfully maintaining the cell at 160 °C with air cooling. In another study, Meyer et al. [26] monitored the temperature and current flux density in a passive air-cooled cell utilizing a PCB (printed circuit board) sensor plate device. They observed a significantly nonuniform current density and temperature, especially at high current density values. The cell temperature increased during the experiment, so they applied a neutron imaging technique to evaluate water film thickness dispensation. After 900 s of testing, the cell was almost dry, causing the variation in cell temperature and current density dispensation [27,28].

Based on the above literature and the relevant authors’ informative insight and competence, limited studies were reported on the influence of cathode flow field geometry on the performance of PEM fuel cells. This study seeks to address this gap by delving into the novel designs of cathode-side flow channels and their impact on crucial parameters such as heat removal, water retention, and current flux density within the PEM fuel cell. By systematically investigating these aspects, this research aims to contribute valuable insights into optimizing PEM fuel cell performance and advancing the development of sustainable energy technologies.

2. Numerical Methodology

2.1. Concept Development

This research focused on changing the flow field geometry of the cathode-side channel in a passive air-cooled PEM fuel cell to enhance heat removal and manage H₂O retention within the cell for better performance and durability. This quantitative study obtained results through simulations using certain boundary conditions set for the designs. Relevant data and research on passive air-cooled PEM fuel cells were collected to establish a clear direction for designing and simulating 3D fuel cell geometry. Overall, the aspects that influenced the 3D design and simulation of the PEM fuel cell are as follows:

1.

Dimensions of the simulated fuel cell geometry;

2.

Assumptions applied to the design;

3.

Boundary conditions for the simulation process.

Utilizing the data gathered, we visualized the 3D design and simulation results using SolidWorks and ANSYS Fluent 19.1 simulations to evaluate the research findings, adhering to the goal of enhancing heat removal and H₂O retention within the PEM fuel cell.

2.2. Modeling the Conventional Air-Cooled PEM Fuel Cell

In this study, the first design (Design 1) was based on a conventional PEM fuel cell, which employs a parallel flow field. The main purpose of developing this model was to validate our simulation approach and initial water content, temperature, and current flux density magnitude values based on previous studies. We created the model using SolidWorks 2019 software, as shown in Figure 1. The dimensions of the conventional fuel cell model are also listed in

Table 1. Fuel cell baseline model dimensions [29].

Description	Value (mm)
Length	18
Width	2
Membrane thickness	0.03
Anode/Cathode CL thicknesses	0.03/0.03
Anode/Cathode MPL thicknesses	0.03/0.03
Anode/Cathode GDL thicknesses	0.5/0.28
Bipolar plate thickness	1.05
Flow channel inlet area	1 × 1

.

2.3. Cathode Flow Field Design 2 (Converging Channel)

Considering the design of a conventional cathode flow field and our research objectives, we devised a cathode flow field design (Design 2), shown in Figure 2. We modified the conventional flow field, making the inlet wider than the outlet and decreasing the channel width, i.e., converging the flow field. The inlet width of 1.5 mm and outlet width of 0.8 mm resulted in a convergence angle of 2.23°.

Our goal with this design was to increase flow velocity, lowering pressure in the flow channel. This modification was based on the continuity principle, which states that a mass flow rate must remain constant in a closed system. Hence, when the flow area decreases, the velocity increases to maintain a constant mass flow rate. The increase in velocity promotes the efficient removal of excess water to prevent flooding. This improvement in water management is essential to maintaining adequate hydration in the membrane and catalyst layers. Additionally, the higher gas velocity due to the reduced flow area can significantly improve convective heat transfer within the flow channel. On the other hand, a decrease in pressure may help in preventing the fuel cell from dehydrating by slowing down the removal of water vapor, potentially increasing the humidity within the cell.

Additionally, a pressure reduction may cause localized hotspots or inefficient temperature control within the fuel cell. Thus, the higher velocity combined with the reduced channel width may lead to more efficient heat dissipation, which has advantages for controlling temperature and preventing localized overheating.

2.4. Cathode Flow Field Design 3 (Diverging Channel)

Considering the conventional cathode flow field design and our research objectives, we created a second design modification to the cathode flow field, as shown in Figure 3. In Design 3, we applied a similar modification to the conventional cathode flow field as in Design 2, but diverging instead of converging, widening the flow field from the inlet to the outlet by 2.23°.

In Design 3, the objective is to reduce the gas velocity, increasing the pressure within the flow channel. In terms of water retention, the reduction in velocity may help in preventing excessive water removal, minimizing the risk of dehydration in the membrane and catalyst layers. As for heat removal, the convective heat transfer inside the flow channel may be impacted by the lower gas velocity due to the larger flow area, increasing the temperature of the cell.

2.5. Cathode Flow Field Design 4 (Straight Channel with Cylindrical Pin Fins)

Considering the conventional cathode flow field design and our research objectives, we created cathode flow field Design 4, as shown in Figure 4.

This cathode flow field incorporates cylindrical pin fins of 0.5 mm in diameter along its centerline. In terms of heat removal, this modification aims to distribute the heat more evenly within the cell, preventing the development of localized, harmful hotspots. For water management, this design may alter the flow pattern within the cell, transporting water at varying rates. As a result, areas with lower flow rates allow for increased water retention in these regions.

2.6. Cathode Flow Field Design 5 (Trapezium Channel)

Considering the conventional cathode flow field design and our research objectives, our fifth design of the cathode flow field is shown in Figure 5. In Design 5, the cathode flow field cross-section is changed from rectangular to trapezoidal to enhance heat removal and water retention. The intent of this design was to increase the area in contact with the environment for better heat removal while maintaining the area in contact with the membrane. In terms of water retention, a trapezoidal flow field has significant potential for creating pockets or channels to keep water away from the catalyst layers. This geometry alters the flow dynamics of the water within the cell, causing water to be directed away from the catalyst layers and decreasing the likelihood of flooding within the cell.

2.7. Cathode Flow Field Design 6 (Semicircle Channel)

The last design for the cathode flow field is shown in Figure 6. In Design 6, the cross-section of the cathode flow field is now semicircular. In terms of heat removal, a change in the flow field cross-sectional geometry influences the flow pattern within the flow field, and a semicircular cross-section channel yields different fluid dynamics compared to a square cross-section. The semicircular shape provides smoother pathways for water drainage without the presence of pockets to accumulate water in the cell, reducing the risk of flooding and improving water management.

2.8. Numerical Solver and Boundary Conditions

The three-dimensional fluid flow and heat transfer in the PEM fuel cell was modelled by means of the resolution of the Navier–Stokes transport equations using ANSYS Fluent CFD solver. More details on the equations used to govern the different parts of the fuel cell can be found in reference [30]. The applied initial and boundary conditions are listed in

Table 2. Boundary conditions assigned to the model [29].

Operating Condition	Value
Inlet temperature	$25°$C
Set-point outlet temperature	60 °C
Inlet pressure (anode and cathode)	101.325 kPa
Relative humidity (anode)	35%
Relative humidity (cathode)	10%
Inlet air velocity	2.7 m/s
Voltage	1.0 V

.

2.9. Mesh Independence Study

Initially, we selected an element size of 0.0003 mm in the meshing process to optimize accuracy and computational time and resources. In the software, we set the mesh physics preference to computational fluid dynamics (CFD) because it specializes in analyzing fluid flow and heat transfer behavior. Subsequently, the element size was reduced, and the average temperature and water content were monitored for denser mesh. Figure 7 and Figure 8 show the average temperature and water content versus the mesh size, respectively. From these graphs, it can be deduced that the best number of elements for simulating the PEM fuel cell model is 126,000, which is obtained from an element size of 0.0001 mm.

The average temperature and water content changes after increasing the mesh size beyond this number were insignificant. In contrast, the simulation computational time increased with the number of elements; hence, 0.0001 mm was the optimal value. Accordingly, Figure 9 shows the finalized mesh for the model.

2.10. Design Validation

To validate the numerical model, Figure 10 compares the temperature distribution on the outer surface of the GDL from experimental data by Lim et al. [29] with current simulations. The temperature distribution along the fuel cell from the simulation agrees well with a difference of less than 5%.

3. Results and Discussion

After completing the setup, we conducted ANSYS analyses, obtaining the contour plots of the temperature distribution, water content, and current flux density for each design. The current flux density was evaluated to determine how heat removal and water content affect the performance of the fuel cell.

Figure 11 shows the temperature distribution for all designs. Design 2 shows the most even temperature distribution along the flow field, while Design 3 shows the worst distribution, with the maximum temperature at a large area of the flow field. The improved temperature distribution in Design 2 is the result of increased velocity due to a converging channel profile. A faster flow in the channel increases the convective heat transfer coefficient and distributes heat more evenly across the cell, minimizing localized hotspots and improving heat removal. On the other hand, Design 5 and Design 6 are similar in that the temperature near the inlet is almost at a minimum and increases as the reactants flow toward the outlet, showing an uneven and undesirable distribution pattern. Next, Design 4 shows an interesting temperature distribution in which the temperature is low at areas around the pin fins. This is attributed to the improved heat removal due to the presence of extended surfaces.

Figure 12 depicts the water content for all proposed designs. Design 4 and Design 5 show good water content as they demonstrate the smallest areas of extreme water levels, indicating no flooding nor dehydration. The even water distribution in Design 4 may be the result of different flow patterns from the presence of fins. The fins serve as an obstacle, causing some regions to have a reduced flow and retain water, balancing out the water content in the cell. In contrast, Design 3 and Design 6 indicate a minimum water level for most of the channel, showing that these cells are at risk of dehydration and membrane damage. In Design 3, the low water level may be the result of reduced velocity caused by the diverging area, resulting in less reactant flowing through the flow field.

On the other hand, the low water content in Design 6 is the result of a smooth path for water drainage, with flooding avoided because of its circular cross-section design. Finally, Design 2 shows an interesting contour in which the water level at the inlet is at a maximum, with a large area of contact with the membrane, possibly resulting in flooding. The water level lowers as the reactants flow through it, owing to an increase in velocity. The water level continues to decrease, causing a rise in the temperature at the outlet of the channel. The same observations have been noted by Lee et al. [31].

The current flux density contours for all six designs are presented in Figure 13. Although the highest current is obtained by the baseline design (Design 1), it is clear that the Design 4 contour reveals the most evenly distributed current compared to the other designs. In contrast, Design 3 shows the lowest level of current. The remaining designs show sufficient current flux contours compared to the baseline design. A similar observation is made by monitoring the average current flux density in the cell as listed in

Table 3. Average results obtained from ANSYS simulations.

Channel Type	Average Temperature, (°C)	Average Water Content	Average Current Flux Density (A/m2)
Straight (Design 1)	57.43	4.07	169.72
Converging (Design 2)	56.67	3.95	169.14
Diverging (Design 3)	58.04	3.82	167.65
With Fins (Design 4)	59.04	3.85	169.03
Trapezoidal (Design 5)	57.13	3.89	169.53
Semicircular (Design 6)	57.49	3.81	168.63

. The average temperature value and water content are also discussed in the table. In terms of the average temperature of the cell, only the converging channel (Design 2) and trapezoidal channel (Design 5) showed reductions compared to the baseline geometry (Design 1). The average temperature value is in good agreement with the contour of the temperature distributions shown in Figure 11, in which Design 2, the converging channel, and Design 5, the trapezoidal channel, show more even distributions compared to the other designs. The remaining designs showed an increase in the average temperature value, indicating that the increase in the velocity of the reactants flowing through the flow field in the converging design and the increased surface area for convection in the trapezoidal design have positive impacts on temperature distribution. Although the current density of Design 2 is slightly reduced at 0.3% compared to the baseline design, the enhanced heat removal within the cell leads to less degradation of the fuel cell’s components over time, resulting in improved durability and a longer lifespan of the fuel cell.

All designs show a reduction in the average water content. The largest reduction was 6.5% in the semicircular channel (Design 6) compared to the baseline design, likely due to the smooth path created by the semicircular flow field, allowing more water to exit the cell. The smallest reduction, identified in Design 2, may be the result of the low pressure exerted on the water molecules from the increased velocity in the flow field. Here, if the pressure is decreased too much at the outlet, this could prevent enough water from being pushed out of the cell and leading to possible flooding at the inlet as shown in Figure 12.

Finally, the average current density across the cell, shown in Table 3, allows us to examine how changes in the temperature and water content affect the performance of the fuel cell. The new designs demonstrate a drop in fuel cell performance. Theoretically, an improvement in heat distribution results in an increase in cell performance [23]. Although the average temperature of the converging and trapezoidal designs showed some reductions, indicating better heat removal, the corresponding water content contours in Figure 12 appear to be unevenly distributed. The very high water levels at the inlet of the converging channel indicate flooding at the inlet, which decreases fuel cell performance. The minimal water level in the trapezoidal fuel cell suggests dehydration, which may be responsible for the slight decrease in the current flux density.

These results indicate that water content distribution and temperature are critical in fuel cell performance. Improving heat dissipation in a PEM fuel cell without considering water retention affects the fuel cell’s overall performance. The opposite is also true: an emphasis on water retention at the cost of considering heat dissipation negatively impacts fuel cell performance. As shown in the results of this study, Designs 4 and 6 exhibit even water contours, but because of increases in temperature, the average current flux density is slightly reduced in both cases. A PEM fuel cell must maintain a delicate balance between efficient heat removal and water retention to perform optimally and with durability.

Unfortunately, all the designs exhibited a decrease in power output compared to the baseline. This shows the complex balance required to design a successful PEM fuel cell flow field. The observed trade-offs between power output, water retention, and heat removal indicate that concentrating on only one factor may have undesirable effects on total performance.

To further improve flow field design, we suggest carrying out additional analyses and optimization. To achieve a more balanced performance and mitigate the observed trade-offs, additional factors and advanced modeling techniques should be considered. The obtained research data can help to gain a better insight into our current understanding of PEM fuel cell technology, highlighting the need for a comprehensive and sophisticated strategy to attain maximum efficiency in several parameters.

4. Conclusions

In this study, we examined six cathode flow field designs in PEM fuel cells using ANSYS Fluent simulations, focusing on temperature distribution, water content, and current heat flux. Our analysis provides valuable insights into how flow field geometry influences cell performance, crucial for enhancing PEM fuel cell efficiency and durability.

The temperature distribution analysis revealed significant differences among designs. Design 2 showed the most uniform temperature distribution due to increased velocity facilitating efficient heat transfer. Conversely, Design 3 exhibited poor distribution with elevated temperatures in a large area. Design 4 displayed low temperature points around the implemented pin fins due to improved heat removal. Designs 5 and 6 demonstrated uneven distributions, highlighting the need for optimized configurations for thermal management.

The water content analysis revealed diverse behaviors. Design 2 showed dynamic water levels, potentially leading to flooding, while Designs 3 and 6 indicated risks of dehydration. Designs 4 and 5 displayed favorable distribution, mitigating flooding or dehydration risks. The current flux density analysis identified Design 4 as the most productive in term of flux distribution, with Design 6 showing lower levels. Designs 2 and 5 exhibited a reduced average temperature, indicating improved heat removal, but faced challenges such as flooding or dehydration.

Based on the results provided, Design 4 appears to offer several advantages. It demonstrated an interesting temperature distribution with some areas of lower temperature, suggesting potential for effective heat dissipation. Additionally, it showed good water content distribution, indicating balanced water management without risks of flooding or dehydration. Therefore, considering its favorable temperature distribution and water content management, Design 4 could be recommended as a promising option for further investigation and potential implementation in PEM fuel cell applications. However, it is essential to conduct an additional analysis and validation to confirm its performance under various operating conditions and ensure its suitability for specific applications. Future research should explore additional factors and advanced techniques to achieve balanced performance. This study can contribute to advancing PEM fuel cell technology, guiding the development of more efficient flow field designs crucial for sustainable energy applications.