Method for determining membrane protonic resistance of a fuel cell stack

Abstract

A method for determining membrane humidification by determining the membrane protonic resistance of a fuel cell stack at humidified conditions, and normalizing the base resistance of the fuel cell stack against the base resistance of a reference fuel cell stack.

Claims

1. A method for determining a normalized high frequency resistance of a fuel cell stack, said method comprising: starting up operation of the fuel cell stack; adjusting a cathode air compressor and a water vapor transfer (WVT) unit to increase cell membrane humidification, where the compressor and the WVT unit are adjusted by a controller including a processor and a memory; using a water buffer model, calculated in the controller, to estimate a membrane hydration state of the fuel cell stack at humidified conditions, where the water buffer model integrates product water output by the fuel cell stack; measuring a high frequency resistance of the fuel cell stack using a high frequency resistance (HFR) sensor electrically coupled to the fuel cell stack; identifying, using the controller, a base high frequency resistance of the fuel cell stack at humidified conditions from the high frequency resistance measured by the HFR sensor, including filtering and using data gathered for the base fuel cell stack high frequency resistance at humidified conditions once the water buffer model estimates that the membrane hydration state of the fuel cell stack is within 10% of saturation; and normalizing the base high frequency resistance of the fuel cell stack against a base high frequency resistance of a reference fuel cell stack to determine the normalized high frequency resistance of the fuel cell stack: and adjusting the cathode air compressor and the WVT unit, by the controller, to achieve a desired cell membrane humidification based on an estimated relative humidity calculated from the normalized high frequency resistance.

2. The method according to claim 1 further comprising intermittently bleeding the fuel cell stack to improve water management in the fuel cell stack.

3. The method according to claim 1 further comprising increasing a base stoichiometry of the fuel cell stack to higher than a nominal stoichiometry before identifying the base high frequency resistance of the fuel cell stack at humidified conditions so as to lower the temperature to achieve a high relative humidity and reduce the chance of flooding in anode and cathode flow channels of the fuel cell stack.

4. The method according to claim 3 wherein increasing a base stoichiometry of the fuel cell stack to higher than a nominal stoichiometry includes increasing the base stoichiometry of the fuel cell stack to twice as high as the nominal stoichiometry.

5. The method according of claim 1 wherein adjusting a cathode air compressor and a water vapor transfer (WVT) unit to increase cell membrane humidification includes increasing a relative humidity of cathode air entering the fuel cell stack to 80% relative humidity.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a schematic block diagram of a fuel cell system;

(2) FIGS. 2a-2c are plot diagrams of the HFR of fifty fuel cell stacks operated under dry conditions, wet conditions, and normalized, respectively; and

(3) FIG. 3 is a graph with relative humidity on the x-axis and HFR on the y-axis illustrating the relationship between HFR and relative humidity at three different stack current densities.

DETAILED DESCRIPTION OF THE EMBODIMENTS

(4) The following discussion of the embodiments of the invention directed to a method for determining base stack resistance for high frequency resistance-based relative humidity control is merely exemplary in nature, and is in no way intended to limit the invention or its applications or uses.

(5) FIG. 1 is a schematic block diagram of a fuel cell system 10 including a fuel cell stack 12. The fuel cell stack 12 receives hydrogen from a hydrogen source 14 on anode input line 16 and provides an anode exhaust gas on line 18. A bleed valve 26 is provided in the anode exhaust gas line 18 to allow anode exhaust gas to be exhausted as desired. A compressor 20 provides an airflow to the cathode side of the fuel cell stack 12 on cathode input line 22 through a water vapor transfer (WVT) unit 24 that humidifies the cathode input air. A cathode exhaust gas is output from the stack 12 on a cathode exhaust gas line 28. The exhaust gas line 28 directs the cathode exhaust to the WVT unit 24 to provide the humidity to humidify the cathode input air. A by-pass line 30 is provided around the WVT unit 24 to allow the cathode exhaust gas to by-pass the WVT unit 24. A by-pass valve 32 is provided in the by-pass line 30 and is controlled to selectively redirect the cathode exhaust gas through or around the WVT unit 24 to provide the desired amount of humidity to the cathode input air.

(6) An HFR sensor 36 measures the high frequency resistance (HFR) of the fuel cell stack 12 to determine the cell membrane humidification of the fuel cell stack 12. The HFR sensor 36 operates by measuring the ohmic resistance, or membrane protonic resistance, of the fuel cell stack 12. Membrane protonic resistance is a function of membrane humidification of the fuel cell stack 12, however, stack-to-stack variation in the HFR sensor measurements may be caused by differences in design parameters, such as compression and variability of parts, degradation of fuel cell stack components over the life of the stack 12, and measurement errors from the HFR sensor 36 itself. Therefore, HFR-based membrane humidification control systems must be able to adapt to these variations to ensure that the level of humidification of the fuel cell membrane is maintained at an appropriate level for the life of the fuel cell stack 12.

(7) The system 10 also includes a controller 34 that receives HFR measurement signals from the sensor 36 and controls the speed of the compressor 20, the injection of hydrogen from the hydrogen source 14, and the position of the by-pass valve 32 and the bleed valve 26.

(8) Variation in parts in the balance of plant (i.e., rest of plant) may also impact fuel cell stack humidification. For example, suppose one vehicle includes an inefficient WVT unit and a fuel cell stack with a low HFR sensor reading. This fuel cell stack will be running dryer than desired and the HFR sensor will not detect it. Alternatively, suppose another vehicle has a WVT unit with high efficiency and a fuel cell stack with a high HFR sensor reading. This fuel cell stack will be running wetter than desired and the HFR sensor will not detect it. In addition, the rest of plant items, such as valves, thermocouples, heat exchangers, pressure and delta-P transducers, etc., could also cause variations in HFR sensor measurements.

(9) FIGS. 2a-2c are plot diagrams of the HFR measurements of fifty fuel cell stacks operated under dry conditions, wet conditions, and normalized, respectively. FIGS. 2a-2c illustrate that variation tightens when the actual HFR is normalized with respect to stack HFR measurement under wet conditions. When operated under dry conditions and wet conditions, the stacks exhibit a variation in HFR measurements that may be quite large, as is illustrated in FIGS. 2a and 2b. FIG. 3 is a graph with RH on the x-axis and HFR on the y-axis that illustrates that as the level of cell membrane humidification approaches 100%, the HFR measurement of the cell membrane will converge to a known value, regardless of the variation in the HFR sensor measurements that may occur at lower RH values. Thus, for example, the variation in HFR of the stacks of FIG. 2 operated under dry conditions will be eliminated as the stacks converge to a known value upon cell membrane saturation. Thus, HFR “noise” is eliminated when the fuel cell stacks reach an RH value equal to or greater than 100%. This convergence to a known value upon membrane saturation is illustrated in the normalized plot diagram of FIG. 2. Accordingly, membrane protonic resistance can be illustrated using the following equation:
HFR=HFR.sup.base+f(RH)  (1)
Where HFR.sup.base is fuel cell membrane base stack resistance and f (RH) is resistance rise of the fuel cell membrane as a function of relative humidity.

(10) Going back to FIG. 2, the fuel cell stacks in the normalized plot diagram illustrate that once HFR “noise” is determined and subtracted for each fuel cell stack, which is described in detail below, the remaining membrane resistance for each of the fuel cell stacks is similar, e.g., around 90 miliohm-cm.sup.2. Thus, by removing HFR “noise” due to various contact resistances and other non-membrane resistances, fuel cell membrane humidification for each fuel cell stack can be calculated with a reasonable degree of accuracy. How the HFR “noise” is determined and filtered out is set forth in detail below.

(11) One way to address stack-to-stack HFR measurement variation due to HFR “noise” is to identify each fuel cell stack's membrane protonic resistance by determining base resistance (j) at humidified conditions (HFR.sub.j.sup.wet)and normalize it against the HFR measurement of some reference fuel cell stack, such as a fuel cell stack which is known to have accurate compression, and thus no appreciable contact resistance (i.e., a stack without “noise”). To identify base stack resistance at humidified conditions, the following options are available: (1) ensure the fuel cell stack comes with base stack resistance at humidified conditions identified during handover to module, i.e., prior to insertion in a vehicle, (2) identify base stack resistance at humidified conditions by running the stack wet and having the controller 34 auto-learn this prior to inserting the fuel cell stack in a vehicle (e.g., in certain situations, modules go through a “break-in” hydration step, thus auto-learning may be combined in this step with the value being stored in the memory of the controller 34), or (3) allow base stack resistance at humidified conditions to be learned by the controller 34 after the fuel cell stack 12 has been place in a vehicle, and determined during a cold and wet start-up. Additionally, if contact resistance changes appreciably over the life of the fuel cell stack 12, it may be necessary to determine base stack resistance at humidified conditions periodically while in a vehicle during a cold and wet start.

(12) Once it is determined that the fuel cell stack's base resistance at humidified conditions needs to be determined, which may occur periodically throughout the life of the fuel cell stack 12, the first step is for the controller 34 to determine the temperature of the fuel cell stack 12 at key on to ensure that the stack 12 is not warm, as saturation of the stack membranes is part of the auto-learning process and is more readily achieved utilizing a cold and wet start. Calibration and implementation may dictate having a trigger based on ambient temperature and thus no active diagnostics. It will be readily apparent to those skilled in the art that various methods for determining that the temperature is suitable to implement the determination of the base resistance of the fuel cell stack at humidified conditions are available without departing from the scope of the present invention.

(13) Once the controller 34 determines that the fuel cell stack 12 is cold at key on, the level of relative humidity exhausted from the fuel cell stack 12 on the cathode exhaust line 28 is set to greater than or equal to 100%. The base stoichiometry set-point is also increased, where base stoichiometry may be twice as high as nominal stoichiometry, or more if at low stack power. Inlet RH on the cathode air input line 22 is set to approximately 80% to fully utilize the WVT unit 24. Increasing the stoichiometery lowers the temperature set-point to achieve high cell membrane humidification of the fuel cell stack 12 and also reduces the chance of flooding in the anode and cathode flow channels of the fuel cell stack 12. Determination of the hydration state of the membrane electrode assembly is discussed in detail below.

(14) Once the cathode inlet RH on the cathode air input line 22 and the outlet RH on the cathode exhaust line 28 are set to the desired values, intermittent bleeds of the fuel cell stack 12 occur, preferably by bleeding to emissions, to maximize bleed velocity to improve anode water management. The increased stoichiometry operates to lower the temperature set-point, as discussed above, and fully opening any radiator by-pass valve will provide maximum cooling. The length of time the fuel cell stack 12 is kept at an outlet RH greater than or equal to 100% will depend on the cooling capacity of the fuel cell system 10. Hence, auto-learning of the base resistance of the fuel cell stack 12 may be more effective in cold conditions rather than warm, as saturation is more readily achieved when the fuel cell stack 12 is cold.

(15) A water buffer model in the controller 34 integrates the product water and estimates the fuel cell membrane electrode assembly (MEA) and diffusion media hydration state of the stack 12. When the model estimates that the MEA is close to saturation, which is a calibratable estimation, the auto-learning operation of the controller 34 is triggered. During auto-learning, the HFR of the fuel cell stack 12 is measured by the controller 34 using the HFR sensor 36 for a period of time, such as for a few seconds. This data is filtered as HFR.sub.j.sup.wet and is then saved in a non-volatile memory of the controller 34. During subsequent operation, until the next HFR.sub.j.sup.wet update, the HFR sensor 36 measurements are normalized by the controller 34 by comparing the measurement of HFR.sub.j.sup.wet, to the HFR measurement of a reference fuel cell stack's base resistance at humidified conditions, and using the following equation:
HFR.sub.j.sup.norm(t)=HFR.sub.j.sup.raw(t)−HFR.sub.j.sup.wet+HFR.sub.ref.sup.wet  (2)
Where HFR.sub.j.sup.norm(t) is HFR normalized, HFR.sub.j.sup.raw(t) is measured HFR including “noise”, HFR.sub.j.sup.wet is the HFR after the membrane of the fuel cell stack 12 has been fully saturated (i.e., HFR is a measure of membrane protonic resistance), and HFR.sub.ref.sup.wet is the HFR of the reference stack operating with a fully saturated membrane. Thus, the key to HFR based RH control using the controller 34 is not to feedback on the absolute value of HFR, but the HFR measurement in reference to a base HFR of the fuel cell stack during a humidified state, which is then compared to the HFR of a reference stack base HFR during a humidified state.

(16) A reconditioning mode of the fuel cell stack 12 could be used to generate the wet conditions necessary for auto-learning of the controller 34 to take place, and thus would replace the steps leading up to the auto-learning of the controller 34 described above, as the fuel cell stack membranes are fully saturated during the reconditioning mode. In addition, employing this strategy periodically during the life of the fuel cell stack 12 may enable the controller 34 to detect and filter out changing RH versus HFR over the life of the fuel cell stack 12, or changing operation of the rest of plant over time, for example, water vapor transfer efficiency, bypass valve operation, etc.

(17) The foregoing discussion discloses and describes merely exemplary embodiments of the present invention. One skilled in the art will readily recognize from such discussion and from the accompanying drawings and claims that various changes, modifications and variations can be made therein without departing from the spirit and scope of the invention as defined in the following claims.