Method for determining anode integrity during fuel cell vehicle operation

Abstract

Procedure comprising: Identifying an anode leak during vehicle operation while a fuel cell is not under stable load via a leak test that compares a current generated by the fuel cell with a current predicted for a hydrogen flow to the fuel cell, wherein the power generated by the fuel cell maintains vehicle power, comprising - Calculation of the predicted current to be generated based on the hydrogen flow via electrochemical conversion of hydrogen in the fuel cell stack, - storing the predicted current and the generated current in a data buffer, wherein the predicted current includes a running average of the predicted current calculated from the stored predicted current data corresponding only to the time at which a purge valve is closed, and the generated current includes a running average of the generated current calculated from the stored generated current data corresponding only to the time at which the purge valve is closed, and - Identification of anode leak based on generated current compared to predicted current.

Description

[0001] The present invention relates to methods for performing anode leak tests in hydrogen fuel cell vehicles.

[0002] In fuel cell vehicles, hydrogen leak tests can be performed to determine anode integrity. Hydrogen can be used as a fuel source for fuel cells connected together to form a fuel cell stack. Hydrogen is introduced into the fuel cell stack on the anode side, while air is introduced on the cathode side. The fuel cell stack generates electrical power in response to the electrochemical conversion of hydrogen and oxygen to water, which can then be used to power various onboard devices in addition to the vehicle itself.

[0003] Current approaches to anode leak detection include performing an anode leak test (ALT) while the vehicle is operating at low fuel cell power, such as during very heavy traffic or while the vehicle is idling at a stoplight. Other approaches to detecting a hydrogen leak on the anode side of the fuel cell stack may rely on allowing vehicle power to drop without adding power to the vehicle to perform the anode leak test. However, one problem with such leak tests, which rely on identifying periods of low power, is that infrequent testing may occur depending on the type of operation a vehicle experiences.For example, US 8 524 405 B2 discloses performing an anode leak test during a vehicle shutdown process, while US 7 942 035 B2 performs a leak test in a fuel cell vehicle only when a zero-load requirement is met. An alternative approach increases the frequency of anode leak tests by providing additional power to the vehicle while the anode leak test is taking place. However, there are still periods in the vehicle cycle when anode leak tests cannot be performed (e.g., when driving under high load and / or at low battery state of charge, etc.). When fuel cell vehicles are configured with a power source for supplying additional power, they may include additional equipment such as an electric motor, which increases the vehicle cost. Further prior art includes US 2007 / 0 207 355 A1 and US 6 815 107 B2.

[0004] The inventors have recognized the above problems. Therefore, it is an object of the invention to propose improved methods, or at least alternative methods to the prior art.

[0005] This problem is solved by methods according to independent claims 1, 9 and 15.

[0006] Disclosed are methods for identifying an anode leak during vehicle operation. In one described embodiment, an anode leak test is performed while the vehicle is operating at a load by comparing a current generated by the fuel cell with a current predicted for hydrogen flow to the fuel cell, wherein the hydrogen flow to the fuel cell maintains vehicle power during the anode leak test. When configured with this arrangement, the method further enables reducing vehicle power in response to identifying the anode leak while still providing sufficient power to operate the vehicle via hydrogen flow.As described, adjusting vehicle power occurs in response to the actuation of a tank valve associated with a fuel storage tank included onboard the vehicle for storing hydrogen fuel. Closing the tank valve may result in a reduction in hydrogen fuel flow, which reduces vehicle power during operation. The advantage of the disclosed method is that the anode leak test can be performed during vehicle operation, e.g., while the vehicle is operating and traveling along a road. Another advantage of the disclosed methods is that anode leak tests can be performed at a higher frequency during operation for determining the degradation status of the fuel cell system, essentially in real time.In this way, the technical result is achieved that the performance of the anode leak test can be extended to allow for more frequent testing of anode leaks during vehicle operation, e.g., while the vehicle is driving on the road, during more vehicle operating conditions.

[0007] In one example, the method may include a statistical comparison including a two-sample Student's test that considers an average current and an envelope thereof to determine a range of the difference between the current produced by the hydrogen fuel cell and the current predicted for hydrogen flow into the fuel cell. The advantage of a statistical comparison based on a current and / or power produced thereby during operation is that an estimate of the size of the leak may be generated based on the range of the identified difference. Thus, in some embodiments, the method further includes determining a size of the leak during operation in response to identifying the anode leak while maintaining vehicle power based on the flow of hydrogen to the fuel cell.The advantage of incorporating such methods is the inclusion of alternative vehicle operating modes in the presence of an anode leak. For simplicity, the system is described here in terms of the hydrogen fuel cell, although the described methods can also be included in hybrid vehicles configured to provide auxiliary power while the hydrogen fuel cell is shut down.

[0008] The above advantages and other advantages and features of the present description are readily apparent from the following detailed description, taken alone or in conjunction with the accompanying drawings. Of course, the above summary is provided to introduce, in a simplified form, a selection of concepts further described in the detailed description. It is not intended to identify principal or essential features of the claimed subject matter, the scope of which is clearly defined by the claims following the detailed description. Furthermore, the claimed subject matter is not limited to implementations that solve any disadvantages mentioned above or in any part of this disclosure.

[0009] The advantages described herein will become more apparent from the following description and the embodiments of the invention in the drawings, in which: Fig. 1 illustrates an exemplary fuel cell stack system in accordance with an embodiment of the present description; Fig. 2 shows an exemplary flowchart illustrating a routine for identifying an anode leak during vehicle operation; Fig. 3 schematically illustrates exemplary data collection cycles for identifying the presence and extent of an anode leak during vehicle operation; and Fig. 4 schematically illustrates an exemplary operating sequence in which the anode leak test is performed in accordance with the present description during vehicle operation; Fig. 5 illustrates a block diagram for setting first and second predetermined pulse width modulated values ​​as used in accordance with first and second level anode leak tests; Fig. 6 is a block diagram for performing a first level anode leak test; Fig. 7 is a block diagram for performing a second level anode leak test; and Fig. 8 is a schematic block diagram to show a method for adjusting engine operations in response to the anode leak test.

[0010] The following description relates to a method for detecting anode leaks in a hydrogen fuel cell system during vehicle operation. For simplicity, the methods are described herein with respect to a hydrogen fuel cell vehicle that may be configured to utilize the described methods to identify leaks during vehicle operation and to adjust vehicle operations in response to the identified leaks. For this reason, Fig. 1 illustrates an exemplary fuel cell stack system included in a fuel cell vehicle propulsion system. In some cases, the fuel cell vehicle propulsion system may also be included in a hybrid vehicle that includes an alternative power source, such as an electric motor, that provides supplemental power during vehicle operation. Fig. 2 and 3-4 further illustrate an exemplary routine and schematic operating sequences for performing the anode leak test during vehicle operation. The described methods included making statistical comparisons, e.g., using a two-sample Student's t-test (sometimes referred to as the two-sample T 2 test or the multidimensional two-sample test), based on data inputs collected during vehicle operation. Thus, Fig. 3 schematically illustrates an exemplary data collection cycle that can be used to identify anode leaks during Fig. 4 shows an exemplary operating sequence that can utilize statistical comparison to demonstrate a method for identifying leaks while maintaining the vehicle based on hydrogen flow during operation. Fig. 5-7 are included to illustrate exemplary idle leak tests that may be performed in response to the described anode leak test. As described herein, an idle leak test may represent a confirmation test used to identify a leak in response to the described real-time anode leak test. Because a confirmation test based on a stable idle is included, the methods may also rely on adjusting vehicle operations in response to the statistical anode leak test to perform the confirmation leak test. For this reason, Fig. 8 is a schematic block diagram illustrating a method for adjusting engine operations in response to the anode leak test.

[0011] Fig. 1 illustrates an exemplary fuel cell recirculation system 100 in accordance with one embodiment of the present description. The system 100 may be implemented in a hydrogen fuel cell vehicle, or in some embodiments, may be implemented in an electric vehicle or any other such hybrid fuel cell-based device that uses electrical power generated by a hydrogen fuel cell to power various devices while providing vehicle power via a flow of hydrogen from a fuel cell tank onboard the vehicle. When implemented in a hybrid vehicle, an alternative energy source may also be present to supplement vehicle power when fuel cell power is reduced or shut down during vehicle operation.

[0012] A first fluid stream (or cathode stream) consisting of air is supplied to a mass air flow (MAF) sensor 102. The air passing through the MAF sensor 102 is atmospheric air. The MAF sensor 102 measures the amount of airflow in the fluid stream. An air compressor 104 pressurizes the air stream and delivers the air stream to a fuel cell stack 106. The fuel cell stack 106 includes a first inlet 107 for receiving the air stream. A humidifier (not shown) may be added to the system 100 to add water vapor to the air stream. If the air contains a high water content, a humidity sensor (not shown) may be included, for example, to measure the moisture content of the wet air. The water may be used to ensure that membranes (not shown) in the fuel cell stack 106 remain humidified to provide optimal operation of the fuel cell stack 106.

[0013] A fuel storage tank (or fuel storage supply) 108 provides a supply fuel stream (or anode stream) in the form of hydrogen. The supply anode stream comprises compressed hydrogen. Although compressed hydrogen may be used in the system 100, any hydrogen fuel source may be implemented in the system 100. For example, instead of the compressed gas, liquid hydrogen, hydrogen stored in various chemicals such as sodium borohydride or alanates, or hydrogen stored in metal hydrides may be used.

[0014] A tank valve 110 controls the flow of supply hydrogen. A pressure regulator device 112 regulates the flow of supply hydrogen. The pressure regulator device 112 may include any type of device used to control the pressure or flow of a fluid. The pressure regulator device 112 may be implemented, for example, as a pressure regulator or as a variable or multi-stage ejector. The pressure regulator device 112 is configured to combine the supply anode stream (e.g., hydrogen received from the tank 108) with the unused anode stream (e.g., recirculated hydrogen from the fuel cell stack 106) to generate an input anode stream (or stack hydrogen). A humidifier (not shown) may be provided to add water vapor to the input anode stream.The humidified water vapor in the input anode stream may be included to ensure that the membranes in the fuel cell stack 106 remain humidified to provide optimal operation of the fuel cell stack 106.

[0015] The pressure regulator device 112 controls the flow of the input anode stream to the fuel cell stack 106. The fuel cell stack 106 includes a second inlet 109 configured to receive the input anode stream from the pressure regulator device 112. The fuel cell stack 106 generates a stack stream in response to the electrochemical conversion of the hydrogen from the input anode stream and the oxygen from the air in the first fluid stream.

[0016] Various electrical devices 120 are coupled to the fuel cell stack 106 to consume this power to operate. If the system 100 is used with a vehicle, the devices 120 may include a motor or multiple vehicle electrical components, each consuming power to function for a specific purpose. For example, these devices 120 may be associated with, but need not be limited to, a vehicle powertrain, occupant heating and cooling, interior / exterior lighting, entertainment devices, and power windows. The specific types of devices 120 implemented in the vehicle may vary based on the vehicle contents, the type of motor used, and the specific type of fuel cell stack implemented. A current sensor 122 measures the stack current generated by the fuel cell stack 106.Current sensor 122 sends the measured current values ​​to a controller 124. Controller 124 sends control values ​​to pressure regulator device 112 to control the flow of the input anode stream to fuel cell stack 106. A pressure sensor 125 is coupled between pressure regulator device 112 and fuel cell stack 106 to measure the pressure of the hydrogen before the hydrogen is delivered to fuel cell stack 106. Pressure sensor 125 sends a feedback signal (e.g., "pressure") indicating the pressure of the hydrogen in the input anode stream.

[0017] Excess amounts of air and hydrogen may be supplied to the fuel cell stack 106 to increase the operational robustness of the fuel cell stack 106. The fuel cell stack may vent unused hydrogen into the unused anode stream. An outlet 111 of the fuel cell stack 106 is configured to vent the unused anode stream. In addition to hydrogen, the unused anode stream may contain various impurities such as nitrogen and water, both in liquid and vapor form.

[0018] The fuel cell stack 106 includes an outlet 115 configured to vent excess air. The fuel cell stack 106 includes an outlet 116 configured to deliver coolant in the form of deionized water / ethylene glycol (DIWEG) or another suitable coolant in response to heat generated due to the chemical processing of the oxygen and hydrogen (e.g., from the air stream). A cooling interface 138 may receive the DIWEG from the fuel cell stack 106. The cooling interface 138 may also provide stack coolant to an inlet 113 of the fuel cell stack 106.

[0019] The pressure control valve 112 is configured to increase or decrease the pressure of the input anode flow into the stack 106 in response to control values ​​sent by the controller 124. The pressure control valve 112 is configured to receive the supply anode flow from the tank valve 110 at a constant pressure. The pressure control valve 112 can receive the unused anode flow from the fuel cell stack 106 (or from a purge assembly not shown) at variable flow rates.

[0020] As noted above, the pressure regulator device 112 may be implemented as any pressure regulator device commonly known in the art. Generally speaking, such pressure regulator devices adjust the pressure of a fluid in response to control values ​​(or control signals) with various electrically based characteristics. Such characteristics may include pulse-width modulated (PWM) values, analog values, or digital values ​​depending on the particular type of pressure regulator device implemented. In one example, the pressure regulator device 112 may be implemented as a pulsating valve or a pulsating injector that regulates the pressure of the fluid delivered to the hydrogen fuel cell. The advantage of incorporating a pulsating injector is that the fuel flow into the fuel cell stack 106 can be inferred during operation.The controller 124 may send control values ​​as PWM-based values ​​to control the pressure regulator to increase or decrease the flow of the input anode stream into the fuel cell stack 106. In another example, the pressure regulator device 112 may be implemented as an ejector to provide recirculation of the unused anode stream. The ejector may be implemented with a solenoid-controlled needle actuator (not shown). One such solenoid-controlled needle actuator is disclosed in U.S. Patent 7,943,260, entitled "System and Method for Recirculating Unused Fuel in Fuel Cell Application," to Brighton et al., which is incorporated herein by reference in its entirety. The solenoid may move the needle in the ejector in response to the control values, which are analog, thereby enabling an increase or decrease in the pressure of the input anode stream delivered to the fuel cell stack 106.The movement of the needle is inherently variable to allow different flow rates of the unused anode stream to be combined with the supply anode stream from tank 108. Such a variable characteristic allows the ejector to adjust the total flow and pressure of the input anode stream delivered to fuel cell stack 106. In yet another example, a combination of pulsating injectors for regulating the pressure of the fluid and ejectors for providing recirculation of the unused anode stream may also be included.

[0021] In some embodiments, the system 100 may further include an energy storage device 140 configured to utilize or consume an energy source other than the fuel cell stack 106. For example, in some cases, the energy storage device 140 may be a battery that supplements or replaces the fuel cell power during vehicle operation. That is, the energy storage device 140 may supplement all or a portion of the fuel cell power during vehicle operation. Thus, the vehicle may be referred to as a hybrid fuel cell vehicle or, in some cases, a hybrid electric vehicle when used in a vehicle with a propulsion system such as the one in Fig. 1, an alternative power source is present. As another example, the system 100 may be operated to drive the energy storage device 140, which in turn may provide a generator function, to convert the output into electrical energy, where the electrical energy may be stored for later use in the system 100. Other embodiments may include, for example, one or more rechargeable batteries, fuel cells, and / or capacitors. In such examples, electrical energy may be temporarily converted to chemical or potential energy for storage. The vehicle propulsion system may be configured to transition between two or more of the operating modes described herein in response to operating conditions. In addition, the energy storage device 140 may periodically receive electrical energy from a power source external to the vehicle (e.g.,is not part of the vehicle). As a non-limiting example, in other embodiments, the propulsion system of system 100 may be configured as a plug-in hybrid vehicle, where electrical energy may be supplied to the energy storage device from a power source via an electrical energy transmission cable. The controller 124 may identify and / or control the amount of electrical energy stored in the energy storage device 140, if present, which may be referred to as the state of charge (SOC). Thus, it should be appreciated that to recharge energy storage devices from a power source that is not part of the vehicle (e.g.,Charging via a cable, via a wireless connection, where the energy storage device can receive electrical energy from a power source via electromagnetic induction and / or via radio waves and / or via electromagnetic resonance), any suitable approach may be used. Incorporating an alternative energy source via the energy storage device 140 enables vehicle propulsion using a power source other than the fuel cell stack 106. In this way, the hybrid electric vehicle described with reference to the propulsion system of the vehicle system 100 may be configured, in some modes of operation, to utilize the secondary form of energy (e.g., electrical energy) to provide power for uninterrupted operation of the vehicle.

[0022] Now moving on to a description of the procedures, Fig. 2-4 to illustrate exemplary routines for identifying an anode leak during vehicle operation via a leak test configured to compare a current generated by the fuel cell with a current predicted for a hydrogen flow to the fuel cell, wherein the flow of hydrogen to the fuel cell maintains vehicle performance. For example, Fig. 2 is an exemplary flowchart illustrating a routine for identifying an anode leak during vehicle operation. Fig. 3 and Fig. 4 then schematically illustrate exemplary data collection and operating sequences for performing the anode leak test in accordance with the description during vehicle operation.

[0023] Fig. 2 illustrates an exemplary leak detection method 200 configured to identify leaks by comparing the current generated by the fuel cell with a current predicted for hydrogen flow to the fuel cell. When a vehicle is operating in the presence of a leak in the fuel cell system 100, an increased flow of hydrogen may be directed to the fuel cell to maintain vehicle operations, e.g., by supplying an amount of fuel to the fuel cell based on a vehicle operator's demand in the presence of the leak. In other words, the total hydrogen flow may be increased because a portion of the hydrogen in the fuel stack is reacted to generate a current that provides vehicle power, while the remaining portion of the hydrogen is removed from the fuel cell recirculation system 100, e.g.,into an engine compartment within the vehicle.

[0024] At 202, the leak detection method 200 includes reading a sensor in the fuel cell recirculation system. For example, a system and / or tank pressure, stack current, stack pressure, etc., may be determined while monitoring conditions within the fuel cell recirculation system during operation. Monitoring system parameters and / or vehicle operations may provide indicators of system usability that may be processed by the controller 124 to determine whether degradation has occurred by identifying leaks within the fuel cell system.

[0025] At 204, method 200 includes determining a hydrogen flow to the fuel cell stack 106. Determining the hydrogen flow delivered to the fuel cell stack enables a prediction of the potential current to be generated based on the hydrogen flow via electrochemical conversion of hydrogen in the fuel cell stack. For this reason, at 206, method 200 further includes calculating a current that can be supported based on the determined hydrogen flow. As an example, pressure sensor 125 may be coupled between pressure regulator device 112 and fuel cell stack 106 and measure the pressure of the hydrogen prior to delivering the hydrogen to the fuel cell stack 106. Further, pressure sensor 125 may provide a feedback signal (e.g.,"Pressure") indicating the pressure of the hydrogen in the input anode stream to the controller 124, which processes the received feedback. The controller 124 can then calculate the hydrogen flow in the input anode stream and can further calculate the current supported by the hydrogen flow into the fuel cell stack. As described in more detail below, predicting the current supported by the hydrogen flow into the fuel cell stack enables the identification of an anode leak by comparing the actual current generated with the current estimated from the hydrogen flow into the input anode stream.

[0026] As described herein, the controller 124 may include a data storage buffer configured to store input data received during vehicle operation. The inclusion of a data storage buffer allows received data input to be stored for further computational analysis of the data by the described methods. As one example, the calculated hydrogen flow in the fuel cell stack may be stored in a data buffer, which is used to calculate a running average of the calculated current that this fuel flow supports. A separate buffer may also be included that stores the actual current generated, which may also be used to calculate a running average of the generated stack current.Comparing the two buffers then allows leaks to be identified during vehicle operation with a desired degree of confidence by performing a statistical analysis such as a two-sample t-test. In one specific example, an anode leak is identified based on the statistical comparison of the generated current and the predicted current for the hydrogen flow to the fuel cell, where the statistical comparison is a two-sample t-test.

[0027] In such an example, a two-sample t-test may be performed to compare the entire sets of predicted and generated currents to identify an anode leak. In general, the two-sample t-test can be used to determine whether the two sets of data are equal to each other. The two-sample t-test is set forth in "STATISTICS FOR ENGINEERING AND SCIENCE" by W. Mendenhall and T. Sincich, fourth edition, pp. 422-494, which is incorporated herein by reference. If the generated currents are represented by the mean and / or by the standard deviation of the sample size (e.g., N data points received during one data collection cycle), the controller 124 may perform a one-sample t-test. The one-sample t-test is also set forth in "STATISTICS FOR ENGINEERING AND SCIENCE" by W. Mendenhall and T. Sincich, as set forth above.It should be appreciated that the approach described here is more than the simple application of a well-known statistical test. Rather, it presents a procedure, combined with various hardware and software, that enables an improved technical result.

[0028] At 210, method 200 includes determining a purge state of the fuel cell recirculation system. Hydrogen is a flammable substance with an autoignition temperature that may be within a vehicle operating temperature. Precautions may be included to prevent the hydrogen content from becoming overpressurized. Thus, occasional purging of hydrogen from the fuel cell recirculation system may be included. Purging hydrogen from the system may include, for example, opening a purge valve. Opening a purge valve may introduce a leak for which data collection and data interpretation of another leak is difficult. Thus, the removal of a portion of hydrogen from the flow may not be easily quantified in the presence of the purge leak created by opening the purge valve.For this reason, no attempt is made to detect a possible leak in the fuel cell recirculation system 100 when purging occurs based on an open purge valve. If the purge state of the fuel cell recirculation system 100 is set to "on," the method 200 proceeds to box 212, where a data storage buffer is reset to begin a new data collection cycle. Thereafter, the method 200 continues as previously described. Alternatively, a trouble-free data collection cycle may occur, enabling reliable data collection and statistical processing, as described below, if a purge state of the fuel cell system is "off."

[0029] Further data collection allows additional data points to be stored in a data buffer in the controller 124, which are used to identify an anode leak during vehicle operation. At 214, the method 200 includes storing the generated stack current in a data buffer and further storing the stack current predicted based on the hydrogen flow in box 216. In one implementation, the method 200 may be configured to perform an anode leak test upon completion of a data set, which may be timed to occur upon reaching a full data buffer. In this arrangement, the data storage buffer advantageously acts as a timing device to enable periodic testing of anode integrity during operation.However, this is not limiting, and the timing of the anode leak test in other implementations may alternatively be performed based on other criteria. For example, a sudden increase in hydrogen flow without an increase in operator demand may indicate that a potential leak has occurred in the system. Thus, alternative configurations of the method may include analyzing a subset of the data stored in a data storage buffer, for example, in response to a spike in hydrogen flow during operation. However, for simplicity of description, the leak test is performed when the data buffer reaches full capacity, which in some cases may be specified by a threshold associated with a storage capacity of the buffer.In other words, the anode leak test may be performed when the computational memory associated with the stored data reaches a memory threshold. Alternatively, as indicated by a no response to a full buffer at 220, method 200 may continue to accumulate data used to identify a data leak in the fuel cell recirculation system while the buffer falls below the memory threshold.

[0030] At 230, the controller 124 may be configured to perform the anode leak test in response to the full buffer. If the anode leak test is to occur, the method 200 proceeds by performing the comparison of the predicted current data relative to the generated current data using the statistical method such as the two-sample t-test. In some cases, additional or alternative statistical tests may be performed to identify the presence of an anode leak in the fuel cell recirculation system. If no leak test is performed, the method may proceed to monitoring the fuel cell system during vehicle operation. Although this is Fig. 2 is not shown, if no leak test is performed while a data buffer is full, the buffer can be reset to allow for further collection of input data while monitoring anode integrity onboard the fuel cell vehicle. As described herein, the statistical comparison includes performing the two-sample t-test, which considers an average current and its envelope, to determine the magnitude of the difference between the generated current and the predicted current.

[0031] Because the example method 200 includes a statistical data comparison, the identification of an anode leak may be made based on a p-value associated with the statistical comparison. The p-value may represent the confidence level of obtaining a test result (e.g., the predicted current) based on the actual observed test result (e.g., the generated current). Further, the confidence level may refer to an envelope surrounding the running average collected during the data cycle. Thus, the comparison of method 200 may further include processing a data envelope associated with one or more of the predicted and generated currents, wherein the comparison further includes a t-test that determines anode integrity based on the data envelope.For this reason, the method 200 may further include identifying a leak based at least on a degree of difference between the predicted and generated currents and the data envelope associated therewith.

[0032] Continuing with a description of the p-value, a threshold is often used to indicate a predetermined level of confidence in the data or a level of confidence that the collected data is likely to differ from a reference curve. For example, the predetermined level of confidence may be represented by a p-value of 0.05 or 0.01, which is used as a threshold to identify statistical differences in the collected data or the presence of an anode leak. If the assisted estimated current subsequently exceeds the generated current, the vehicle may consume more fuel than is actually used to produce the vehicle's power. If this occurs, there may be a high probability of a leak in the fuel circulation system, e.g., in the anode loop of the fuel cell system.As indicated at 240, the leak may be identified by a p-value falling below the threshold indicating the leak. Upon detecting a leak, the method 200 proceeds to 242 by setting a flag indicating the presence of a leak and communicating the message to a vehicle operator, e.g., by illuminating a dashboard lamp communicating the anode leak. In this way, the method further includes communicating the anode leak to a vehicle operator during vehicle operation in response to identifying the presence of the anode leak. However, as described below, one or more vehicle adjustments may also be made along with the leak communication.If no leak is detected because the determined p-value exceeds the threshold, the method 200 continues by resetting the buffer to enable a new data collection cycle that determines the anode integrity on board the fuel cell vehicle.

[0033] For simplicity, Fig. 2 schematically illustrates a routine for identifying the leak via a comparison of the predicted current and the actual current generated. As described in more detail below, one or more operational adjustments may be made in response to the anode leak test and / or its results to continue vehicle operation if desired, particularly in the presence of a hydrogen leak on board the vehicle.

[0034] Since the methods include statistical comparisons between two sets of data in the presence of a data envelope, the processing in another example method may further include estimating an extent of the anode leak based on the distance a data point lies outside the data envelope. In this way, the predicted current exceeds the generated current when a leak is present, with the extent of the predicted current relative to the generated current being further used to determine the size of the leak. For example, a relatively high predicted current that lies outside the data envelope of the generated current, such asA relatively small predicted current, determined using the variance or noise in the generated current data for a data collection cycle, may indicate that a larger leak exists in the fuel cell system, whereas a relatively small predicted current that lies outside the generated current data envelope by a small amount may indicate that a smaller leak exists. Thus, there may also be a relative gradient based on the statistical comparison that can be advantageously used to enable operational adjustments such as limp home mode in the presence of a leak. Alternatively, in some cases, a small leak may allow temporary vehicle operation based on operator demand without degradation of performance. Fig. Figure 3 schematically illustrates an exemplary vehicle operating sequence to illustrate the collection of data to identify the presence of an anode leak during vehicle operation. For simplicity, the schematic data shown is enlarged to illustrate the comparisons in more detail.

[0035] The upper figure shows a time-vs-current diagram 300 containing an example graph of the current data (e.g., predicted and generated current) used to power a vehicle during operation. The middle figure shows ALT activity 350, showing active periods when anode leak tests are being performed. Following this, the lower figure shows ALT results 360, which are used to indicate the status of anode leaks within the system. Such results can be used to identify anode leaks in accordance with the methods described herein, further enabling operational adjustments based on the anode leaks during vehicle operation. Time increases from left to right.

[0036] The current diagram 300 shows an average generated current 302 that may be present, e.g., the stack current generated in response to the electrochemical conversion of the hydrogen from the input anode current and the oxygen from the air in the first fluid stream during vehicle operation. For simplicity, the noise of the average generated current 302 is shown as the generated current envelope 310, which may be defined by a range or variance 312 within the measurement data. Based on the average input data and the data envelopes thereof, the generated current window 314 is shown to represent a window associated with the average buffer data that provides a means for identifying anode leaks based on the comparisons to a predicted current 320 (shown without a data envelope for simplicity).If the predicted current 320 exceeds and falls outside the generated current window 314 during a data collection cycle, a potential leak may be present within the system. The diagram shown includes two such regions to illustrate how the predicted current exceeds the generated current when an anode leak is present.

[0037] The ALT activity graph 350 shows areas where the buffer has reached a storage capacity, thus triggering the statistical comparisons (e.g., two-sample t-test) used to identify anode leaks. A first exemplary active ALT 352 is identified. After the test is executed, the data collection cycle 354 may begin collecting a new data set and maintain a running average of the current data during the collection. Upon reaching a storage capacity, another anode leak test may be triggered, represented by a second active ALT. Thereafter, the process may continue in the same manner.

[0038] Returning to the example comparisons shown in the current diagram 300, a small leak 330 is represented by a portion of the predicted current that lies just outside the generated current window 314, but is detectable by a statistical comparison such as the two-sample t-test described herein. As indicated above, the described methods may identify the extent of the leak as a small leak based on the close distance 332 between the two curves within the region of the data collection cycle for which data was used to perform the test. In accordance with the described methods, such a comparison may trigger the first ALT leak 362 shown in the ALT results 360.

[0039] Likewise, a relatively large leak 340 is represented by a portion of the predicted current that lies outside the generated current window 314 by a greater distance than the small leak 330. As noted above, the described methods can identify the extent of the leak as a large leak based on the large distance 342 between the two curves. Such methods can trigger the second ALT leak 364 shown.

[0040] Fig. Figure 4 schematically illustrates an exemplary operating sequence in which the anode leak test is carried out during vehicle operation. Therein, the generated current 400 is shown in the upper diagram along with the predicted current 410. For simplicity, the schematic data is shown without data envelopes for clarity of illustration. However, the present statistical methods can still explain the data envelopes of the current curves when anode leaks are identified in the fuel cell system. The second diagram from the top illustrates the purge valve position 420, which may represent the open / close state of a purge valve during vehicle operation. Although the on / off state of the purge valve for system purging during the Fig. 4, the buffer of controller 124 may, in some cases, continue to collect data even after a system purge has occurred. When arranged in this manner, the data used to identify the presence of a leak may include contiguous data items that are close to each other but on different sides of the purge valve setting. In this way, the methods may, in some cases, advantageously enable uninterrupted data collection cycles for high-frequency leak detection, even in the presence of a purge.

[0041] The third diagram represents the ALT activity 430, which contains areas where the anode leak test is actively running to identify an anode leak, while the fourth diagram shows the ALT result 440, which graphically represents the result of the anode leak tests being run. Fig. Figure 4 shows the state of the ALT result in the on state (e.g., 1) or the off state (e.g., 0). For example, in some cases, the system state can be maintained as shown throughout the next data collection cycle. In other cases, the ALT result can indicate the measured p-value. Fig. 4, areas where an ALT flag has been set to 1 indicate excessive fuel usage, which may further indicate anode leaks and system quality degradation. Although in Fig. 4 does not show example engine settings, in some cases one or more operating settings may be made in response to the identification of an anode leak.

[0042] As an example, illustrate Fig. 5-7 show exemplary engine adjustments performed in response to an identified leak. Upon detection of a leak, vehicle adjustments may be made to confirm the presence of the leak. As one example, the confirmation test may be a pressure decay test in which the system is pressurized and further monitored to detect pressure changes therein. For example, WO 2008071402 entitled “LEAKAGE TEST IN A FUEL CELL SYSTEM,” which is incorporated herein by reference, may be included as an exemplary leak detection method. A drop in pressure may indicate the leak in the fuel cell recirculation system. As another example, an idle leak test may be performed, wherein operating adjustments are made to the fuel cell stack for a stable load. For example, US 7,942.035 entitled "ANODE LEAK TEST IMPLEMENTATION," which is incorporated herein by reference, may be included as an example of the leak detection method performed to determine anode integrity. As described and reproduced herein, the system 100 utilizes a series of anode leak tests that are generally performed when the load on the fuel cell stack 106 is stable. In some cases, the methods may include an alternative power source to supplement fuel cell power while the confirmation test is being performed, thus enabling the confirmation tests to be performed during vehicle operations. However, in other cases, operational adjustments may be made, for example, to perform the confirmation leak test at low fuel cell idle loads.

[0043] Operational settings executed in response to the anode leak test can be implemented in accordance with a functional safety feature protocol. For example, the additional tests and their vehicle settings can be implemented according to the functional safety standard ISO 26262, which can be used as a guide for the automotive product development phase. The guidelines of the ISO 26262 standard can range from the specification through the design, implementation, integration, verification, validation, and product phases. The ISO 26262 standard is an adaptation of the functional safety standard IEC 61508 for automotive electrical / electronic systems and defines functional safety for automotive equipment that is applicable throughout the entire life cycle of all electronic and electrical safety-related systems of the motor vehicle.

[0044] Upon reaching a stable load, a first-level anode leak test may be performed. In a first-level anode leak test, the controller 124 determines whether the fuel cell stack 106 has been reduced below a requested power, which in one example is a fuel cell idle state based on the amount of current generated by the fuel cell stack 106. In response to determining that the fuel cell stack 106 is in the idle state, the controller 124 controls the pressure regulator device 112 to control the pressure of the input anode current in response to various control values ​​at a first preselected pressure level. As mentioned above, such control value(s) may be PWM-based, analog, or digital.The pressure sensor 125 sends the actual pressure amount on the "Pressure" signal back to the controller 124 to determine whether the actual pressure amount is equivalent to the first preselected pressure level. The controller 124 measures and records the values ​​associated with driving the pressure regulator device 112 to determine whether these values ​​are equal to one or more predetermined control values. If the control values ​​are not equal to the first predetermined control values, the controller 124 may perform a second-level anode test in response to the fuel cell stack 106 being in an idle state. The second-level anode test may be similar to the first-level anode test, except that a second preselected pressure level is used instead of the first preselected pressure level. In addition, one or more second predetermined values ​​are set that correspond to the second preselected pressure level.The second preselected pressure level may correspond to a higher pressure level than the first preselected pressure level. In . Fig. 6 and Fig. 7, an exemplary first and second level anode leak test is shown in more detail. Before performing the first and second level anode leak tests, the system 100 is calibrated to define the predetermined control ranges for the first and second preselected pressure levels, respectively. Such calibration may be performed during an end-of-production (EOL) test while the vehicle is being manufactured or manufactured.

[0045] Fig. 5 illustrates a block diagram 500 for setting one or more first and second predetermined control values ​​as used in accordance with the first and second level anode leak tests, respectively.

[0046] In block 502, the pressure regulator device 112 is controlled (via the controller 124 with the control valves) to distribute hydrogen (e.g., in the input anode stream) at the first preselected pressure level. The first preselected pressure level generally corresponds to a low pressure level. The first preselected pressure level may correspond to the pressure difference between the anode and the cathode.

[0047] At block 504, the current sensor 122 measures the amount of current generated by the fuel cell stack 106 to determine whether the stack current equals a predetermined amount of current, which generally corresponds to the fact that the fuel cell stack 106 is in an idle state. In one example, a stack current of 3 A may correspond to the fact that the fuel cell stack 106 is in the idle state. The determined amount of stack current used to indicate whether the fuel cell stack 106 is in the idle state may vary from system to system based on the type of fuel cell stack implemented, various vehicle loads, and other fluctuations. If the measured stack current does not equal the predetermined amount of current, the graph 500 remains at block 504 until the predetermined amount of current is reached. If the measured stack current is equal to the predetermined amount of current, the graph 500 proceeds to block 506.

[0048] In block 506, a counter variable (e.g., i) is initialized.

[0049] In block 508, the controller 124 receives the corresponding control value, which is used to control the pressure regulator device 112 to achieve the first preselected pressure level reached in block 502. The controller 124 stores the control value in volatile memory.

[0050] In block 510, the counter variable i is incremented in response to storing the control value.

[0051] At block 512, chart 500 determines whether counter value i is equal to a predetermined sample size (e.g., N1). Generally speaking, system 100 experiences noise, which may cause a number of control values ​​to be measured while obtaining the first preselected pressure level. Because of such a condition, N1 is chosen to achieve a sample size that ensures a high degree of confidence. In one example, N1 may correspond to a sample size of 120 measured control values ​​used to achieve the first preselected pressure level. The sample size may vary based on the desired criteria of a particular implementation. If i is equal to N1, chart 500 proceeds to block 514. If i is not equal to N1, chart 500 proceeds to block 508 to obtain additional measurements.

[0052] At block 514, a determination is made as to whether the pressure regulator device 112 should be controlled via the controller 124 with a different set of control values ​​to distribute hydrogen in the input anode stream at a second preselected pressure level. The second preselected pressure level is set to a pressure higher than the first preselected pressure level. If the second preselected pressure level has not been set, the diagram 500 returns to block 502. If the second preselected pressure level has been set, the method 500 proceeds to block 516.

[0053] In block 516, the controller 124 determines the first predetermined values ​​and stores the first predetermined values ​​in non-volatile memory (e.g., in an EEPROM). In one example, the first predetermined values ​​may correspond to a range of measured control values ​​for achieving the first preselected pressure level. Such a range may include the sample size N1. As noted above, in one example, the sample size may include 120 control value measurements. In this case, a range of 120 control values ​​may be stored in the non-volatile memory. In yet another example, the controller 124 may calculate an average and a standard deviation of the measurements comprising the sample size N1. In such an example, the average and / or standard deviation may be defined as one or more first predetermined values.

[0054] Block 502 is executed again to establish the second preselected pressure level. In block 502, the pressure regulator device 112 is controlled (via controller 124 with control values) to distribute hydrogen at the second preselected pressure level.

[0055] Block 504 is executed again to determine whether the stack current is equal to a predetermined current amount, which generally corresponds to the fact that the fuel cell stack 106 is in an idle state. If the measured stack current is not equal to the predetermined current amount, the graph 500 remains in block 504 until the predetermined current amount is reached. If the measured stack current is equal to the predetermined current amount, the graph 500 proceeds to block 506.

[0056] Block 506 is executed again to reinitialize i.

[0057] Block 508 is executed again so that the controller 124 receives the corresponding control value used to drive the pressure regulator device 112 to achieve the second preselected pressure level obtained in the re-executed block 502. The controller 124 stores the corresponding control value in non-volatile memory.

[0058] Block 510 is executed again to increment i in response to storing the control value.

[0059] Block 512 is executed again to determine whether the counter value i is equal to N1. As mentioned above, the system 100 experiences noise, which may require that a number of control values ​​be measured while obtaining the second preselected pressure level. Because of such a condition, N1 is chosen to achieve a sample size sufficient to establish a high degree of confidence. If i is not equal to N1, the chart 500 returns to block 508 to obtain additional measurements.

[0060] Block 514 is executed again to determine whether the second preselected pressure level control value has been established. The diagram proceeds to block 516.

[0061] In block 516, the controller 124 determines the second predetermined values ​​and stores the second predetermined values ​​in the non-volatile memory. In one example, the second predetermined control values ​​may correspond to a range of measured control values ​​for achieving the second preselected pressure level. Such a range may include the sample size N1. As mentioned above, in one example, the sample size may include 120 control value measurements. In this case, a range of 120 control values ​​may be stored in the non-volatile memory. In yet another example, the controller 124 may calculate an average and a standard deviation of the measurements comprising the sample size N1. In this case, the average and / or standard deviation may be defined as one or more first predetermined values.

[0062] Fig. 6 illustrates a block diagram 600 for performing the first-level anode leak test. In block 602, the current sensor 122 measures the current generated by the fuel cell stack 106 to determine whether the stack current equals the predetermined current. Such a condition indicates whether the fuel cell stack 106 has been reduced to idle or to a steady state. If the measured stack current does not equal the predetermined current, the first anode leak test cannot be performed. If the measured stack current does equal the predetermined current, the diagram 600 proceeds to block 604.

[0063] In block 604, a counter variable (e.g., j) is initialized.

[0064] In block 606, the controller 124 reads and stores a particular operating control value that is used to control the pressure regulator device 112 to adjust the pressure of the input anode stream to achieve the first preselected pressure level.

[0065] At block 608, the controller 124 determines whether the fuel cell stack 106 is coming out of the idle state. For example, the controller 124 may evaluate the current readings sent by the current sensor or read message activity on a multiplexed bus protocol (e.g., a control area network (CAN)) to determine what the driver is requesting (e.g., throttle position, brake position, etc.). If the fuel cell stack 106 is coming out of the idle state, the diagram 600 stops executing the first anode level test and suspends execution of the test.

[0066] In block 610, j is incremented.

[0067] At block 612, chart 600 determines whether counter variable j is equal to a predetermined sample size (e.g., N2). As mentioned above, based on the desired level of confidence in light of noise considerations, it may be necessary for controller 124 to obtain multiple readings of the control values. If counter variant j is equal to N2, chart 600 proceeds to block 614. If counter variant j is not equal to N2, chart 600 returns to block 606. Generally, blocks 606, 608, 610, and 612 may take seconds to execute (e.g., 5 seconds may correspond to the time necessary to obtain the PWM value measurements). The particular time required to execute blocks 606, 608, 610, and 612 may vary depending on the controller design, software, and the desired test confidence level.

[0068] In block 614, the controller 124 compares the operating control values ​​stored or measured from block 606 with the first predetermined control values ​​as established in the chart 500. For example, the controller 124 may compare the operating control values ​​of block 606 with the first predetermined control values ​​(e.g., the entire set of 120 measured values ​​of the first predetermined control values ​​obtained in block 506). In such an example, a two-sample t-test may be performed to compare the entire set of operating control values ​​of block 606 with the entire set of first predetermined control values. As described above, a one-sample t-test may be used to determine whether the two sets of data are equal to each other.

[0069] In block 616, the controller 124 determines whether the stored operating control values ​​of block 606 are equal to the first predetermined control values. If the stored operating control values ​​of block 616 are not equal to the second predetermined control values, the diagram 600 proceeds to block 618. If the stored operating control values ​​of block 606 are equal to the first predetermined control values, the diagram 600 proceeds to block 602. It will be appreciated that the controller 124 utilizes a number of statistically based methods to determine whether the values ​​are equal to each other. For example, the values ​​may be statistically equal to each other to be considered equal. Likewise, the values ​​may not be considered equal if the values ​​are not within the same predefined statistical range.

[0070] In block 618, the controller 124 sets a flag to indicate that the second anode leak test should be performed.

[0071] If the stored operating control values ​​of block 606 are different from the first predetermined control values, such a condition may generally indicate an anode leak in the system 100. For example, a leak may be inferred because the controller 124 is required to control the pressure regulator device 112 to operating control values ​​that are different from the control values ​​used to establish the first predetermined control values ​​to achieve the first preselected pressure level. The operating control values ​​necessary to maintain the hydrogen pressure at the first preselected pressure level are distinguishable as a fault condition or fault indicator if such control values ​​do not equal the first predetermined control values.

[0072] Fig. Figure 7 illustrates a block diagram 700 for executing the second-level anode leak test. The second-level anode leak test is generally executed when the controller 124 sets the flag indicating that the operating control values ​​are not within the first predetermined range of values, as noted in diagram 600.

[0073] In block 702, the controller 124 controls the pressure regulator device 112 to adjust the pressure of the hydrogen in the input anode stream to the second preselected pressure level.

[0074] At block 704, current sensor 122 measures the amount of current generated by fuel cell stack 106 to determine whether the stack current has decreased to the predetermined current level. If the measured stack current has not decreased to the predetermined current level, graph 700 waits for fuel cell stack 106 to enter the idle state. If the measured stack current has decreased to the predetermined current level, graph 700 proceeds to block 706.

[0075] In block 706, a counter variable (e.g., k) is initialized.

[0076] In block 708, the controller 124 reads and stores the determined operating control value used to control the pressure regulator device 112 to adjust the pressure of the input anode stream to achieve the second preselected pressure level.

[0077] At block 710, the controller 124 determines whether the fuel cell stack 106 is coming out of the idle state. If the fuel cell stack 106 is coming out of the idle state, the diagram 700 proceeds to block 704. If the fuel cell stack 106 remains in the idle state, the diagram 700 proceeds to block 712.

[0078] In block 712, the counter variable k is incremented.

[0079] At block 714, the chart 700 determines whether the counter variable k is equal to a predetermined sample size (e.g., N3). If k is equal to N3, the chart 700 proceeds to block 716. If k is not equal to N3, the chart 700 returns to block 708.

[0080] In block 716, the controller 124 compares the operating control values ​​stored or measured in block 708 with the second predetermined control values ​​established in the chart 500. The controller 124 compares the operating control values ​​of block 708 with the second predetermined control values ​​in a manner similar to that disclosed in block 614.

[0081] In block 718, the controller 124 determines whether the stored operating control values ​​of block 718 are equal to the second predetermined control values. If the stored operating control values ​​of block 708 are not equal to the second predetermined control values, the diagram 700 proceeds to block 720. If the stored operating control values ​​of block 708 are equal to the second predetermined control values, the diagram 700 proceeds to block 724. As noted above, the controller 124 may use any number of statistically based methods to determine whether the values ​​are equal to each other. For example, the values ​​may be statistically equal to each other to be considered equal. Likewise, the values ​​may not be considered equal if the values ​​are not within a predetermined statistical range.

[0082] In block 720, the controller 124 calculates the worst-case leak size. The controller 124 may calculate the worst-case leak size based, for example, on f(PRESSURE, control input, and operating system parameters), where the function f depends on the system and controller design. In one example, the function may be described as: Leak size=−a+bH2Supply+c(μ)Panode−Pout where P OUT the pressure at the exit of the leak (cathode or environment), p H2Supply denotes the incoming pressure and P anodedenotes the pressure at the signal "pressure", u corresponds to one or more operating control values ​​obtained in block 708, and parameters a, b, and c are variables that depend on the system architecture and / or other design criteria associated with the fuel cell stack 106. µ is a control included to maintain a pressure, such as that described in M. Milacic, V. Booden, J. Grimes, and Bernd Maier, "Hydrogen leak detection method derived using DCOV methodology," SAE International Journal of Materials and Manufacturing, pp. 97-102, SAE, 2008, the entire contents of which are incorporated herein by reference. Diagram 600 contemplates that block 720 may also be executed in place of block 618 if diagram 700 is not implemented.

[0083] In block 722, the controller 124 sets a diagnostic trouble code that can be read by a service tool.

[0084] In block 724, chart 700 exits the second anode leak test.

[0085] The first and / or second anode leak tests may be applied to a fuel cell device generally arranged to generate power in response to the electrochemical conversion of hydrogen from the anode side and oxygen from the cathode side. Generally, system 100 contemplates that the first anode leak test may be performed to detect the presence of an anode leak without performing the second anode leak test to confirm the anode leak test performed during vehicle operation. As described in more detail above, the second anode leak test may also be performed at the second preselected pressure level higher than that of the first preselected test (e.g., at a high resolution) to confirm the results of the first anode leak test, with the anode leak test performed during vehicle operation.In some embodiments, an alternative power source may be included to power the vehicle using a supplemental power source, such as an engine, while the confirmation tests are performed.

[0086] In this way, the methods further include reducing vehicle power in response to identifying the anode leak while providing sufficient power to operate the vehicle. Thus, the methods may further rely on adjusting vehicle operations in response to the anode leak test performed during vehicle operation. Adjustments to engine operations in response to the leak test described herein may advantageously enable the confirmation leak test to be performed more reliably. For this reason, Fig. 8 is a schematic block diagram illustrating a method 800 for adjusting engine operations in response to the anode leak test.

[0087] Although Fig. 8 with respect to an idle leak test, the methods may, in some cases, be implemented based on a confirmatory test, which is a pressure decay test. Thus, operating adjustments may be made in response to the leak test to place the vehicle in the appropriate operating conditions for performing the pressure decay test. As an example, supplemental power may be provided onboard a hybrid vehicle while a pressure decay test is being performed. In the pressure decay test, the system is pressurized and further monitored to detect pressure changes after pressurization. Subsequently, a pressure decay after pressurization may be used to identify and confirm the leak in the fuel cell recirculation system. In this way, the confirmatory leak test may be an idle leak test and / or a pressure decay test.

[0088] At 802, method 800 includes adjusting engine operations to perform the confirmatory leak test. For example, after identifying a leak during operation using statistical methods, e.g., method 200, controller 124 may reduce vehicle power to a stable idle prior to performing the confirmatory test. Reducing vehicle power may include actuating 110 the tank valve to reduce hydrogen flow to fuel cell 106, where the reduced hydrogen flow, in some cases, provides the reduced vehicle power that operates the vehicle. As noted above, in some cases, supplemental power may be provided to meet an operator request while the idle leak test is being performed.At block 810, current sensor 122 measures the amount of current generated by fuel cell stack 106 to determine whether the stack current has decreased to the predetermined current level. If the measured stack current has not decreased to the predetermined current level, method 800 waits for fuel cell stack 106 to enter the idle state. If the measured stack current has decreased to the predetermined current level, method 800 proceeds to block 812.

[0089] At 812, the controller 124 makes settings to perform the confirmation test to confirm the presence of the leak. The confirmation test here is an open-circuit leak test (e.g., Fig.5-7), but the leak may instead also be confirmed by performing a pressure decay test. At 814, a flag or indicator may be set indicating that a leak has been detected. For example, the controller 124 may determine the extent of the leak to determine anode integrity and further communicate the leak to a vehicle operator. In some cases, communicating the leak to the vehicle operator may include illuminating a lamp on an instrument panel in response to the confirmation test indicating the presence of the leak. Thereafter, additional adjustments, e.g., based on ISO 26262, may be included to operate the vehicle in the presence of a hydrogen leak. As one example, the method 800 may include closing the tank valve to shut off hydrogen flow to the fuel cell in response to the leak.

[0090] The approach just described, together with various hardware and software, represents a method that enables an improved technical result for a hydrogen fuel cell vehicle. The method is described in terms of methods that include identifying an anode leak during vehicle operation based on a comparison of a current generated by the fuel cell with a current predicted for a hydrogen flow delivered to the fuel cell, wherein the hydrogen flow delivered to the fuel cell maintains vehicle operation in the presence of the leak; estimating a size of the leak and setting a flag that communicates the leak to a vehicle operator; and adjusting the hydrogen flow in response to the leak to adjust vehicle performance.The methods further include reducing the flow of hydrogen to the fuel cell in response to the leak and performing a confirmatory leak test in response to the reduced hydrogen flow, wherein the confirmatory leak test includes an idle leak test and / or a pressure decay test. Since maintaining vehicle operation is desirable for a vehicle operator, in some implementations the methods may also include switching vehicle power to an alternative power source to power the vehicle while reducing the flow of hydrogen to the fuel cell, for example, while performing the confirmatory test. If a hydrogen leak is subsequently identified onboard the vehicle, the methods further include shutting off the flow of hydrogen to the fuel cell for increased safety.Shutting off the hydrogen flow may occur in response to the anode leak and / or the confirmation leak test indicating the presence of the leak (e.g., by closing a tank valve).

[0091] The advantageously described methods enable methods for determining anode integrity in hydrogen fuel cells during vehicle operation. As described, the methods include performing a leak test configured to compare a predicted current based on a hydrogen flow to the fuel cell with a current generated by the fuel cell, wherein the hydrogen flow to the fuel cell provides power to the vehicle during the leak test. Performing the anode leak test during vehicle operation includes, for example, performing the leak test when the vehicle is traveling on the highway under a substantially high engine load. Another advantage of the disclosed methods is that tests for anode leaks can be performed with increased frequency because such tests occur during vehicle operation.In this way, the methods enable the determination of the fuel cell system's degradation status essentially in real time, offering the attractive possibility of making adjustments based on the anode leak test during vehicle operation. The ability to perform anode leak tests during vehicle operation further extends system status checks to more, and in one example, virtually all, vehicle operating conditions, enabling more frequent testing for anode leaks over the useful life of the vehicle.

[0092] It is noted that the exemplary control and estimation routines included herein may be used with various engine and / or vehicle system configurations. The control methods and control routines disclosed herein may be stored as executable instructions in non-transitory memory. The specific routines described herein may represent one or more of any number of processing strategies, such as event-driven, interrupt-driven, multitasking, multithreading, and the like. Thus, various depicted activities, operations, or functions may be performed in the order depicted, executed in parallel, or, in some cases, omitted.Likewise, the order of processing is not necessarily required to achieve the features and advantages of the exemplary embodiments described herein, but is provided for ease of illustration and description. Depending on the particular strategy employed, one or more of the illustrated acts, operations, and / or functions may be performed repeatedly. Further, the described acts, operations, and / or functions may graphically represent code to be programmed into the non-transitory memory of the computer-readable storage medium in the engine control system.

[0093] It will be appreciated that the configurations and routines disclosed herein are exemplary in nature, and that these specific embodiments are not to be considered in a limiting sense, as numerous changes are possible. The subject matter of the present disclosure includes all novel and non-obvious combinations and subcombinations of the various systems and configurations, as well as other features, functions, and / or characteristics disclosed herein.

[0094] The following claims particularly point out certain combinations and sub-combinations considered novel and non-obvious. These claims may refer to "a" element or to "a first" element, or to the equivalent thereof. These claims are to be construed as including the inclusion of one or more such elements, but neither requiring nor excluding two or more such elements. Other combinations and sub-combinations of the disclosed features, functions, elements, and / or properties may be claimed by amending the present claims or by presenting new claims in this or a related application. Such claims, whether broader, more limited, the same, or different in scope than the original claims, are also to be construed as being within the subject matter of the present disclosure.

Claims

[1] Procedure comprising: Identifying an anode leak during vehicle operation while a fuel cell is not under stable load via a leak test that compares a current generated by the fuel cell with a current predicted for a hydrogen flow to the fuel cell, wherein the power generated by the fuel cell maintains vehicle power, comprising - Calculation of the predicted current to be generated based on the hydrogen flow via electrochemical conversion of hydrogen in the fuel cell stack, - storing the predicted current and the generated current in a data buffer, wherein the predicted current includes a running average of the predicted current calculated from the stored predicted current data corresponding only to the time at which a purge valve is closed, and the generated current includes a running average of the generated current calculated from the stored generated current data corresponding only to the time at which the purge valve is closed, and - Identification of anode leak based on generated current compared to predicted current. [2] The method of claim 1, further comprising reducing vehicle power in response to identifying the anode leak while providing sufficient power to operate the vehicle. [3] The method of claim 2, wherein reducing vehicle power includes actuating a tank valve to reduce hydrogen flow to the fuel cell, the reduced hydrogen flow providing the reduced vehicle power operating the vehicle. [4] A method according to any preceding claim, further comprising identifying the anode leak based on a statistical comparison of the generated current and the current predicted for the hydrogen flow to the fuel cell, wherein the statistical comparison is a two-sample t-test. [5] The method of claim 4, wherein the predicted current exceeds the current generated when a leak is present, the magnitude of the predicted current relative to the generated current being further used to determine the size of the leak. [6] The method of claim 5, wherein the hydrogen flow to the fuel cell is adjusted while providing power to the vehicle, wherein the hydrogen flow is adjusted to supply sufficient fuel to power the vehicle in the presence of the leak. [7] The method of claim 6, further comprising closing the tank valve to shut off hydrogen flow to the fuel cell in response to the leak. [8] The method of claim 7, wherein the operational adjustments for powering the vehicle using an alternative power source are made in response to the flow of hydrogen to the fuel cell. [9] A method for a hydrogen fuel cell, the method comprising: Determining the anode integrity during vehicle operation, wherein the fuel cell is under variable load, via a leak test configured to compare a current predicted based on a hydrogen flow to the fuel cell with a current generated by the fuel cell, wherein the predicted current comprises a running average of the predicted current calculated from predicted current data stored over time in a data buffer corresponding only to the time at which purging of the fuel cell is stopped, and wherein the generated current comprises a running average of the generated current calculated from generated current data stored over time in the data buffer corresponding only to the time at which purging of the fuel cell is stopped, wherein the hydrogen flow to the fuel cell powers a vehicle during the leak test,wherein a leak is identified in response to the generated current deviating from the predicted current by more than a threshold, and the predicted current is calculated based on the electrochemical conversion of the hydrogen feed to the fuel cell. [10] The method of claim 9, wherein the comparison includes processing a data envelope associated with one or more of the predicted and generated currents, the comparison further including a two-sample t-test determining anode integrity based on the data envelope. [11] The method of claim 10, further comprising estimating an extent of the leakage based on a degree of difference between the predicted and generated currents and the data envelope associated therewith. [12] The method of claim 11, further comprising adjusting the flow of hydrogen to the fuel cell in response to the leak and performing a confirmation leak test. [13] The method of claim 12, wherein the confirmation leak test is an idle leak test and / or a pressure decay test. [14] The method of claim 13, wherein the extent of the leak determines the anode integrity, wherein the anode integrity is further communicated to a vehicle operator. [15] A method for a hydrogen fuel cell, the method comprising: Identifying an anode leak during vehicle operation, with the fuel cell under variable load, based on a comparison of a current generated by the fuel cell with a current predicted for a hydrogen flow delivered to the fuel cell, wherein the predicted current includes a running average of the predicted current calculated from the stored predicted current data stored over time in a data buffer corresponding only to the time at which purging of the fuel cell is stopped, the generated current includes a running average of the generated current calculated from the stored generated current data stored over time in the data buffer corresponding only to the time at which purging of the fuel cell is stopped,the hydrogen flow delivered to the fuel cell maintains vehicle operation in the presence of the leak, wherein the predicted current is calculated based on the electrochemical conversion of the hydrogen flow delivered to the fuel cell, and the leak is identified in response to the generated current deviating from the predicted current by more than a threshold; estimating a size of the leak and setting a flag that communicates the leak to a vehicle operator; and adjusting the hydrogen flow in response to the leak to adjust vehicle performance. [16] The method of claim 15, further comprising reducing the flow of hydrogen to the fuel cell in response to the leak and performing a confirmation leak test in response to the reduced flow of hydrogen, the confirmation leak test comprising a pressure decay test. [17] The method of claim 16, further including switching vehicle power to an alternative power source for powering the vehicle while hydrogen flow to the fuel cell is reduced. [18] The method of claim 16 or 17, further comprising shutting off the flow of hydrogen to the fuel cell in response to the confirmation leak test indicating the presence of the leak.

Images