Thermal Behavior of an Air Cooled PEM Fuel Cell Stack I ...
Thermal Behavior of an Air Cooled PEM Fuel Cell Stack I. Tolja, M. Abdallahb, F. Barbira,b a
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split, R. Boskovica bb, 21000 Split, Croatia b UNIDO-International Centre for Hydrogen Energy Technologies Sabri Ulker Sk. 38/4, Cevizlibag, Zeytinburnu, 34015 Istanbul, Turkey
Thermal behavior of a 1 kW, air cooled PEM fuel cell stack was analyzed and simulated. Operation of the stack under various scenarios was simulated such as start-up from room temperature, steady state at full and half load and sudden changes of load (from full to half load and reverse), under constant or variable flow rate of the cooling air. In variable power load operation, the stack temperature is not constant. One interesting finding of this analysis is that liquid water occurs at the outlet whenever the stack temperature falls below 60°C, although the nominal operating conditions were selected such that there is no liquid water at the outlet at steady state operation. The results also indicate that this particular stack has insufficient heat exchange area with cooling air resulting in relatively large cooling air flow rates and low temperatures. Introduction Fuel cells stacks require active cooling in order to maintain the desired temperature. The most common temperature control strategies for an air-cooled fuel cell are on-off or multi-step flow rate control of the cooling air fan. The objective of this study was to analyze the effects of different cooling strategies on stack temperature, particularly at start-up, shut down and immediately after sudden load changes. An additional objective was to find out if the heat from the fuel cell stack can be used to release hydrogen form an adjacent metal hydride storage. Stack Description The stack used in this study (Figure 1) has been developed by Proton Energy Systems (1). The stack involves many innovative features particularly in flow configuration, cooling and stack clamping (1-3). It consists of 60 cells 65 cm2 active area each. The cell active area is of a rectangular shape, long and narrow (aspect ratio 10:1) to allow heat conduction to the edge of the cells where the fins are extended for cooling (Figure 2). The bipolar plates including the fins are made of molded polymer/graphite mix (made by SGL Carbon) with thermal conductivity of 20 W/mK. Cooling of the stack is with air that flows along the longer edge of the cells where the fins are located. Instead of the tie-roads the metallic shrouds that snap on the end plates are used to keep the stack compressed (3). The cells are compressed with five polyurethane springs distributed throughout the center line of the active area. The shrouds also form the passage for
cooling air. The cooling fan is installed directly on the shrouds. The stainless steel end plates also serve as the bus plates. The MEAs are made by 3M . These are 7-layer MEAs which include catalyzed membrane and two carbon paper gas diffusion layers with integrated gaskets. The active part of the stack therefore consists of only two repetitive components – 60 MEAs squeezed between 61 collector plates. Besides these repetitive components the rest of the stack is made of only 13 additional parts (three end plates, two shrouds, five polyurethane springs and 4 fittings).
Figure 1. Proton Energy System’s 1 kW fuel cell stack.
Figure 2. Cooling air passage between the fins at the tip of the bipolar plates. The stack physical properties are shown in Table I.
TABLE I. Stack Physical Properties. Property Number of cells Cell active area Cell aspect ratio L:W Dimensions Volume Mass Specific heat Heat exchange area Cooling medium
Value 60 65 cm2 10:1 31x23.6x9.6 cm 7l 17 kg 0.59 kJ kg-1 K-1 0.9378 m2 Air
The stack is designed to generate 1 kW power output at relatively high voltage of >0.7 V/cell. This is because of relatively poor thermal conductivity of the bipolar plates which at lower voltages would cause large temperature gradients between the edge and the center of the cell. The stack was tested at Connecticut Global Fuel Cell Center, University of Connecticut (1). The resulting polarization curve is shown in Figure 3 (1). The stack operating conditions during testing are shown in Table II. These operation conditions were selected such that the product water is just sufficient to saturate the cathode gas at the exit, i.e. there should be no liquid water present at the exit. TABLE II. Stack Operational Conditions. Property Nominal power Nominal current Nominal voltage Air stoichiometry Hydrogen stoichiometry Operating temperature Air and hydrogen inlet temperature Operating pressure Pressure drop (cathode) Humidity of reactant gases Cooling control
Value 1 kW 23 A 43.5 V 2 1.2 60 °C 20 °C Ambient 20 kPa Unhumidified On/off
Figure 3. Stack polarization curve. Steady State Calculations With known operating conditions and known stack geometry it is possible to calculate the heat fluxes coming in and out of the stack at steady state. The resulting stack heat balance at full power (60 °C) is shown in Table III. Because the cell potential is at 0.725 V the stack efficiency is close to 50% and the heat generation rate is about the same as the electricity generation rate. The heat is removed from the stack by excess air and evaporated product water (408 – 54 = 354 W), by radiation and convection to the surrounding air (90 W) and the remaining 600 W must be removed by the cooling air to avoid overheating. Table III. Stack Energy Balance. Electricity generated Enthalpy of the gases going in: Heat generated Enthalpy of gases leaving the stack Heat dissipation Heat to be removed by cooling air
1000 W 54 W 1045 W 408 W 90 W 601 W
i
The heat removed by the cooling air Q (W) is:
i
i
Q = m c p ( TAir ,out − TAir ,in )
[1]
where: i
- m is cooling air mass flow rate ( kgs −1 ) - c p is specific heat of air ( Jkg −1K −1 ) - TAir ,out temperature of cooling air at stack outlet (K) - TAir ,in temperature of cooling air at stack inlet (K) The same amount of heat must be transferred from the stack to the cooling air.
i
Q = h ⋅ A ⋅ LMTD
[2]
where: - h is the average heat transfer coefficient (Wm −2 K −1 ) - A is the heat exchange area (m 2 ) - LMTD is logarithmic mean temperature difference between the stack and cooling air ( C) For a given stack and ambient conditions the only variables in the above equations are the coolant flow rate, the air outlet temperature and the stack temperature. Figure 4 shows the cooling air and stack temperatures as a function of the cooling air flow rate and also as a function of the stack heat transfer coefficient h ⋅ A .
100 100
hA = 15 hA = 20
80 80 temperature, °C
Temperature (°C)
90
hA = 30
70
Tair out
60 60
Tstack
50 40 40 30 20 20 10
00 00
20 20
40 40
60 60
80 80
.
100 100
120 120
flow rate (mCp) Specific Thermal Flux, mCp (kW/K)
Figure 4. Stack and cooling air steady state outlet temperatures as a function of cooling air flow rate and overall stack heat transfer coefficient.
It was noticed during testing of the stack that the air exit temperature could not reach above 30 °C, due to the high air flow rates, and the resulting stack temperature could not be maintained at desired 60 °C in spite of the high air flow rates. This indicates that the heat transfer ( h ⋅ A ) was insufficient (i.e., h ⋅ A
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split, R. Boskovica bb, 21000 Split, Croatia b UNIDO-International Centre for Hydrogen Energy Technologies Sabri Ulker Sk. 38/4, Cevizlibag, Zeytinburnu, 34015 Istanbul, Turkey
Thermal behavior of a 1 kW, air cooled PEM fuel cell stack was analyzed and simulated. Operation of the stack under various scenarios was simulated such as start-up from room temperature, steady state at full and half load and sudden changes of load (from full to half load and reverse), under constant or variable flow rate of the cooling air. In variable power load operation, the stack temperature is not constant. One interesting finding of this analysis is that liquid water occurs at the outlet whenever the stack temperature falls below 60°C, although the nominal operating conditions were selected such that there is no liquid water at the outlet at steady state operation. The results also indicate that this particular stack has insufficient heat exchange area with cooling air resulting in relatively large cooling air flow rates and low temperatures. Introduction Fuel cells stacks require active cooling in order to maintain the desired temperature. The most common temperature control strategies for an air-cooled fuel cell are on-off or multi-step flow rate control of the cooling air fan. The objective of this study was to analyze the effects of different cooling strategies on stack temperature, particularly at start-up, shut down and immediately after sudden load changes. An additional objective was to find out if the heat from the fuel cell stack can be used to release hydrogen form an adjacent metal hydride storage. Stack Description The stack used in this study (Figure 1) has been developed by Proton Energy Systems (1). The stack involves many innovative features particularly in flow configuration, cooling and stack clamping (1-3). It consists of 60 cells 65 cm2 active area each. The cell active area is of a rectangular shape, long and narrow (aspect ratio 10:1) to allow heat conduction to the edge of the cells where the fins are extended for cooling (Figure 2). The bipolar plates including the fins are made of molded polymer/graphite mix (made by SGL Carbon) with thermal conductivity of 20 W/mK. Cooling of the stack is with air that flows along the longer edge of the cells where the fins are located. Instead of the tie-roads the metallic shrouds that snap on the end plates are used to keep the stack compressed (3). The cells are compressed with five polyurethane springs distributed throughout the center line of the active area. The shrouds also form the passage for
cooling air. The cooling fan is installed directly on the shrouds. The stainless steel end plates also serve as the bus plates. The MEAs are made by 3M . These are 7-layer MEAs which include catalyzed membrane and two carbon paper gas diffusion layers with integrated gaskets. The active part of the stack therefore consists of only two repetitive components – 60 MEAs squeezed between 61 collector plates. Besides these repetitive components the rest of the stack is made of only 13 additional parts (three end plates, two shrouds, five polyurethane springs and 4 fittings).
Figure 1. Proton Energy System’s 1 kW fuel cell stack.
Figure 2. Cooling air passage between the fins at the tip of the bipolar plates. The stack physical properties are shown in Table I.
TABLE I. Stack Physical Properties. Property Number of cells Cell active area Cell aspect ratio L:W Dimensions Volume Mass Specific heat Heat exchange area Cooling medium
Value 60 65 cm2 10:1 31x23.6x9.6 cm 7l 17 kg 0.59 kJ kg-1 K-1 0.9378 m2 Air
The stack is designed to generate 1 kW power output at relatively high voltage of >0.7 V/cell. This is because of relatively poor thermal conductivity of the bipolar plates which at lower voltages would cause large temperature gradients between the edge and the center of the cell. The stack was tested at Connecticut Global Fuel Cell Center, University of Connecticut (1). The resulting polarization curve is shown in Figure 3 (1). The stack operating conditions during testing are shown in Table II. These operation conditions were selected such that the product water is just sufficient to saturate the cathode gas at the exit, i.e. there should be no liquid water present at the exit. TABLE II. Stack Operational Conditions. Property Nominal power Nominal current Nominal voltage Air stoichiometry Hydrogen stoichiometry Operating temperature Air and hydrogen inlet temperature Operating pressure Pressure drop (cathode) Humidity of reactant gases Cooling control
Value 1 kW 23 A 43.5 V 2 1.2 60 °C 20 °C Ambient 20 kPa Unhumidified On/off
Figure 3. Stack polarization curve. Steady State Calculations With known operating conditions and known stack geometry it is possible to calculate the heat fluxes coming in and out of the stack at steady state. The resulting stack heat balance at full power (60 °C) is shown in Table III. Because the cell potential is at 0.725 V the stack efficiency is close to 50% and the heat generation rate is about the same as the electricity generation rate. The heat is removed from the stack by excess air and evaporated product water (408 – 54 = 354 W), by radiation and convection to the surrounding air (90 W) and the remaining 600 W must be removed by the cooling air to avoid overheating. Table III. Stack Energy Balance. Electricity generated Enthalpy of the gases going in: Heat generated Enthalpy of gases leaving the stack Heat dissipation Heat to be removed by cooling air
1000 W 54 W 1045 W 408 W 90 W 601 W
i
The heat removed by the cooling air Q (W) is:
i
i
Q = m c p ( TAir ,out − TAir ,in )
[1]
where: i
- m is cooling air mass flow rate ( kgs −1 ) - c p is specific heat of air ( Jkg −1K −1 ) - TAir ,out temperature of cooling air at stack outlet (K) - TAir ,in temperature of cooling air at stack inlet (K) The same amount of heat must be transferred from the stack to the cooling air.
i
Q = h ⋅ A ⋅ LMTD
[2]
where: - h is the average heat transfer coefficient (Wm −2 K −1 ) - A is the heat exchange area (m 2 ) - LMTD is logarithmic mean temperature difference between the stack and cooling air ( C) For a given stack and ambient conditions the only variables in the above equations are the coolant flow rate, the air outlet temperature and the stack temperature. Figure 4 shows the cooling air and stack temperatures as a function of the cooling air flow rate and also as a function of the stack heat transfer coefficient h ⋅ A .
100 100
hA = 15 hA = 20
80 80 temperature, °C
Temperature (°C)
90
hA = 30
70
Tair out
60 60
Tstack
50 40 40 30 20 20 10
00 00
20 20
40 40
60 60
80 80
.
100 100
120 120
flow rate (mCp) Specific Thermal Flux, mCp (kW/K)
Figure 4. Stack and cooling air steady state outlet temperatures as a function of cooling air flow rate and overall stack heat transfer coefficient.
It was noticed during testing of the stack that the air exit temperature could not reach above 30 °C, due to the high air flow rates, and the resulting stack temperature could not be maintained at desired 60 °C in spite of the high air flow rates. This indicates that the heat transfer ( h ⋅ A ) was insufficient (i.e., h ⋅ A
