Method and control device for operating a fuel cell system

Abstract

The present invention relates to a method of operating a fuel cell system (1) comprising a fuel cell stack (2) having a number of fuel cells passing through cooling passages which are supplied with coolant via a cooling circuit (3) using a coolant pump (4). According to the invention, during low temperature or sub-freezing start-up, the temperature of the fuel cells in the fuel cell stack (2) is indirectly detected via the pressure difference (Δp) of the coolant across the fuel cell stack (2) and the speed (n) of the coolant pump (4) is controlled in dependence on the indirectly detected temperature. The invention further relates to a control device for a fuel cell system (1).

Description

[Technical Field]

[0001] The present invention relates to a method for operating a fuel cell system, particularly during low temperature or sub-zero start-up, which method is therefore particularly suitable for operating a fuel cell system in mobile applications.

[0002] Furthermore, the invention relates to a control device arranged to carry out the steps of the method. [Background technology]

[0003] Fuel cells convert fuel, such as hydrogen, and oxygen into electrical energy, heat, and water. To increase power output, multiple fuel cells are typically connected to form a fuel cell stack, with reactant gases supplied through supply passages passing through the stack. Heat generated by the electrochemical process in the fuel cells is removed using a cooling circuit, typically a vehicle cooler in mobile applications, which is then exhausted to the ambient environment through one of the vehicle's coolers. The coolant in the cooling circuit is pumped through a coolant supply passage passing through the fuel cell stack using a coolant pump integrated into the cooling circuit. A directional control valve can be integrated into the cooling circuit to bypass the cooler. Cooler bypass can be advantageous, for example, during startup, when the fuel cell stack should be heated as quickly as possible to avoid the accumulation of water and / or ice, which could delay or even prevent startup, especially when the ambient temperature is below 0°C. However, the risk of icing can only be eliminated if the coolant is reliably heated above 0°C before entering the fuel cell stack.

[0004] To heat the coolant during sub-zero start-up, the heat generated in the fuel cell in conjunction with the electrochemical reaction can be used. Alternatively, the coolant can be heated externally. However, in both cases, the start-up process is time-consuming. In addition, since the fuel cell must always be cooled below 0°C, measures must be taken to increase the ice resistance of the fuel cell, for example, by using an ice buffer in the fuel cell and / or a heater in the fuel cell system.

[0005] During sub-zero startup, the coolant volumetric flow rate must be sufficiently high to avoid localized peak temperatures, so-called "hot spots," and excessive temperature differences between the coolant's inlet and outlet temperatures. At the same time, the coolant volumetric flow rate must be sufficiently low to prevent excessive temperature drops at the inlet to the fuel cell stack, and therefore freezing. The coolant volumetric flow rate is controlled via the pump speed of the coolant pump. This pump speed is typically controlled depending on the coolant temperature at the inlet and outlet of the fuel cell stack. However, due to the high viscosity of the coolant during sub-zero startup, the change in coolant temperature lags significantly in time relative to the temperature change in the fuel cell. Summary of the Invention [Problem to be solved by the invention]

[0006] The present invention therefore addresses the problem of providing a method for operating a fuel cell stack that allows the temperature in the fuel cell stack to be detected as quickly and reliably as possible in the event of low temperature or sub-zero start-up, in order to be able to adjust the speed of the coolant pump and therefore the volumetric flow rate of the coolant through the fuel cell stack in dependence on the temperature. [Means for solving the problem]

[0007] To achieve the above object, a method is proposed having the features of claim 1. Advantageous developments of the invention can be read from the dependent claims. Furthermore, a control device for carrying out the method is provided.

[0008] A method for operating a fuel cell system including a fuel cell stack having a plurality of fuel cells through which cooling passages pass is proposed. In the method, the cooling passages are supplied with coolant via a cooling circuit using a coolant pump. According to the invention, during cold or sub-zero start-up, the temperature of the fuel cells in the fuel cell stack is indirectly detected via the coolant pressure difference across the fuel cell stack. The speed of the coolant pump is controlled depending on the indirectly detected temperature.

[0009] Therefore, the quantity used to detect the temperature in the fuel cell is no longer the temperature of the coolant at the inlet and outlet of the fuel cell stack, but the coolant pressure or the coolant pressure difference across the fuel cell stack. Since the viscosity of the coolant flowing through the cooling passages of the fuel cell stack changes with the temperature in the fuel cell stack, the temperature in the fuel cell stack can be inferred from the pressure difference between the coolant pressure at the inlet of the fuel cell stack and the coolant pressure at the outlet of the fuel cell stack.

[0010] The use of the pressure difference as a quantity for indirectly detecting the temperature in the fuel cell has the advantage that the pressure difference changes, i.e. decreases, substantially inversely proportionally to the temperature increase in the fuel cell, and precisely without a time delay, and therefore the temperature can be detected more quickly and reliably.

[0011] In this way, the proposed method allows for improved regulation of the coolant pump speed during low-temperature or sub-zero start-up. For example, the risk of icing due to excessive coolant volume flow can be reduced. The reduced risk of icing also reduces the need for measures to make the fuel cell more resistant to ice, thereby reducing costs. Furthermore, a faster sub-zero start-up can be achieved, since the coolant volume flow does not need to be reduced to avoid excessive temperature differences. Furthermore, leakage due to temperature differences can be eliminated, thereby extending the life of the fuel cell system.

[0012] Taking the pressure difference into account as a quantity for indirectly detecting the temperature change in the fuel cell stack has the further advantage that the pressure difference can be detected relatively easily and cheaply using a simple pressure sensor.

[0013] Therefore, to detect the coolant pressure difference across the fuel cell stack, it is preferable to use multiple pressure sensors or differential pressure sensors to measure the coolant pressure at the inlet and outlet of the fuel cell stack. Pressure sensors or differential pressure sensors are much cheaper to procure than, for example, volumetric flow sensors. A first pressure sensor can be located at the inlet of the fuel cell stack, and a second pressure sensor at the outlet. These pressure sensors can also be combined with a temperature sensor, thereby using a pressure and temperature sensor. Since a temperature sensor is usually already present, the number of sensors can be kept low in this way.

[0014] Furthermore, it is proposed to continuously or gradually increase the coolant pump speed during cold or sub-zero start-up to maintain the coolant pressure difference across the fuel cell stack substantially constant or within a predetermined range. Increasing the coolant pump speed leads to rapid heating of the coolant and therefore the fuel cell. The pump speed can be increased continuously or gradually or in steps. Gradual increases have the advantage that the pump speed does not increase too much during cold or sub-zero start-up, preventing excessive pump speed. To this end, in a first step, the pump speed is increased until a predetermined maximum pressure difference, such as 10 mbar, can be measured. The pump speed then remains constant for a short time, allowing the coolant to warm up and again reducing the coolant pressure difference across the fuel cell stack. When the pressure difference reaches a predetermined minimum pressure difference, such as 5 mbar, in a second step, the pump speed is increased again until the pressure difference reaches a new predetermined maximum value. This can be repeated until the coolant temperature reaches its target temperature and the cold or sub-zero start-up is complete. If the pump speed is gradually increased, the coolant pressure difference across the fuel cell stack in particular exhibits a sawtooth-like progression within a certain range.

[0015] On the other hand, if the pressure difference of the coolant across the fuel cell stack shows an upward trend and / or exceeds a predetermined maximum value, the speed of the coolant pump must be reduced, since such a trend in the pressure difference indicates an excessive volumetric flow rate and there is a risk of icing, since in the case of an excessive volumetric flow rate the fuel cell would be cooled rather than heated by the coolant.

[0016] Therefore, it is proposed as an alternative measure that the coolant pump speed is reduced during low temperature or sub-zero start-up if the coolant pressure difference across the fuel cell stack exhibits an upward trend and / or exceeds a predetermined maximum value. In this way, ice buildup caused by the coolant volume flow can be reliably avoided.

[0017] According to another preferred operating strategy, the coolant pump speed is kept constant during cold or sub-zero start-up in order to reduce the coolant pressure difference across the fuel cell stack. This corresponds to normal heating of the fuel cell. At the same time, the risk of icing due to excessive coolant volume flow is therefore reduced. If the pressure difference increases, countermeasures can also be taken by reducing the coolant pump speed.

[0018] In a further development of the invention, it is proposed to determine a first threshold value defining an initial minimum pressure difference of the coolant across the fuel cell stack, and to increase the speed of the coolant pump until the threshold value is reached in preparation for a cold or sub-zero start. This measure ensures that the available measurement range is utilized during a cold or sub-zero start. This is because the temperature of the fuel cell rises rapidly during a cold or sub-zero start, which results in a very steep temperature transition, and therefore a very steep transition in the pressure difference. Therefore, the initial pressure difference must be selected to be sufficiently large. The speed of the coolant pump is therefore increased until the threshold value is reached. To prevent the resulting high volumetric flow rate from causing ice formation in the inlet region of the fuel cell, the speed must also not be selected too high. Therefore, in preparation for a cold or sub-zero start, the threshold value must be reached but not significantly exceeded. In this way, a coolant volumetric flow rate large enough to prevent ice formation while preventing the formation of "hot spots" is achieved.

[0019] The first threshold value can be, for example, 50 mbar, in which case the coolant is circulated in such a way that the heat generated in the fuel cell is well dispersed, while at the same time, the low coolant flow rate prevents icing in the inlet region of the fuel cell.

[0020] It is further preferred that two additional thresholds are determined, defining the range of coolant pressure differential across the fuel cell stack that must be achieved during cold or sub-zero start-up. This means that at the end of cold or sub-zero start-up, a predetermined range of pressure differential must be achieved, with the lower limit defined by the second threshold and the upper limit defined by the third threshold. The second threshold, which may be 5 mbar for example, ensures that the coolant volume flow rate is large enough to circulate the coolant. The larger third threshold, which may be 15 mbar for example, is a reference amount that allows adaptation of the first threshold, as explained below. The same applies to the second threshold.

[0021] This is because, according to a preferred embodiment of the present invention, if the coolant pressure difference across the fuel cell stack at the end of a cold or sub-freezing start-up is outside the range of pressure differences to be achieved, the first threshold is increased or decreased. This means that, to achieve the predetermined range, the first threshold is increased if it is below the second threshold and decreased if it is above the third threshold. This means that the first threshold is adapted accordingly. This adaptation allows optimal adjustment of the initial speed of the coolant pump and therefore the volumetric coolant flow rate through the fuel cell stack at the end of a cold or sub-freezing start-up depending on the coolant pressure difference across the fuel cell stack and therefore depending on the temperature.

[0022] Because the viscosity of the coolant is strongly dependent on temperature or because the resulting pressure difference varies widely, the first threshold value may be different for each initial temperature. Therefore, in particular, a first threshold value is defined for each initial temperature. This threshold value can then be stored in the control device. Furthermore, when determining the first threshold value, other dependencies can be taken into account, such as the shutdown duration, which determines whether the fuel cell needs to be fully or partially thermostated.

[0023] Preferably, the pressure difference profile during or after cold or sub-zero start-up is evaluated, and if the profile temporarily stagnates, the third threshold is lowered, so that it approaches the second threshold. In a pressure profile diagram, the stagnation profile can be recognized as a plateau, which indicates an ice formation and melting process. This can be considered a sign that the initial speed of the coolant pump was too high. Lowering the third threshold adapts the first threshold, thus reducing the initial speed of the coolant pump, thereby minimizing the risk of ice formation during the next cold or sub-zero start-up.

[0024] It is further preferred that at least one threshold value, in particular all threshold values, are stored in a control device configured to carry out the steps of the method, which allows the method to be largely automated.

[0025] Furthermore, a control device for a fuel cell system is proposed, which is configured to carry out the steps of the method according to the invention. In particular, the control device can be used to realize the various operating strategies described above. For this purpose, at least one threshold value can be stored in the control device. Furthermore, the control device can be used to evaluate the progress of the coolant pressure difference across the fuel cell stack. The control device obtains the necessary measurement values ​​from a pressure sensor or a differential pressure sensor. Depending on the result of the evaluation, the control device can be used to control the coolant pump to increase, decrease, or keep the coolant pump speed constant. [Brief explanation of the drawings]

[0026] [Figure 1] 1 is a schematic diagram of a mobile fuel cell system suitable for carrying out the method according to the present invention; [Figure 2] FIG. 2 is a schematic diagram of the media supply of a fuel cell stack of a fuel cell system. [Figure 3] This figure shows the changes in the coolant pressure difference Δp across the fuel cell stack over time t, the coolant temperature TEin at the inlet and TAus at the outlet of the fuel cell stack, the coolant volume flow rate Q, the coolant pump rotation speed n, and the current I. [Figure 4] 1 is a block circuit diagram illustrating the steps of a first method according to the present invention; [Figure 5] FIG. 4 is a block circuit diagram illustrating the steps of a second method according to the present invention. [Figure 6] FIG. 10 illustrates a preferred progression of coolant pressure differential Δp across a fuel cell stack during cold or sub-freezing start-up. [Figure 7] 1 is a block circuit diagram illustrating the steps of the method according to the invention; [Figure 8] 1 is a block circuit diagram illustrating the steps of the method according to the invention; [Figure 9] 1 is a block circuit diagram illustrating the steps of the method according to the invention; [Figure 10]FIG. 10 shows another preferred progression of coolant pressure differential Δp across the fuel cell stack including pump speed during cold or sub-freezing start-up. [Figure 11] FIG. 10 is a diagram showing a temperature-dependent pressure difference threshold. DETAILED DESCRIPTION OF THE INVENTION

[0027] The present invention and its advantages will be explained in detail below with reference to the accompanying drawings.

[0028] 1 is used to generate electric drive energy. To this end, the fuel cell system 1 comprises a fuel cell stack 2 having an anode 10 and a cathode 23. During operation of the system, the anode 10 is supplied with a fuel, e.g., hydrogen, via an anode passage 11, and the cathode 23 is supplied with air as an oxygen source via a cathode passage 24.

[0029] The fuel is stored in a tank 12, which can be shut off by a shutoff valve 13. Downstream of the shutoff valve 13, a heat exchanger 14 for regulating the fuel temperature and a pressure regulator 15 for controlling the pressure are located in the anode path 11. Furthermore, an injection pump 16 and a blower 18 located in the recirculation path 17 are provided, which can be used to recirculate the fuel leaving the fuel cell stack 2. Because the outflowing fuel may contain liquid water, the fuel is fed to a water separator 19 before recirculation, which separates the liquid water from the gas and collects it in a container 20. When the container 20 is full, a drain valve 21 is opened and the container 20 is emptied. Over time, nitrogen diffusing from the cathode side to the anode side accumulates in the fuel, so the anode region is periodically purged. For this purpose, a purge valve 22 is opened. The amount released through the purge valve is replaced with fresh fuel from the tank 12.

[0030] Air is taken from the ambient environment and supplied to an air compressor 26 for compression via an air filter 25 arranged in the cathode path 24. Since the air is heated during this process, a heat exchanger 27 is further provided to cool the air. When the system is shut down, air flow into the fuel cell stack 2 can be prevented by shut-off valves 28 on the inlet and outlet sides. Exhaust gas flowing out from the fuel cell stack 2 is again released into the ambient environment via an exhaust path 29. Furthermore, a bypass path 30 in which a bypass valve 31 is arranged is provided to bypass the fuel cell stack 2.

[0031] In addition to electrical energy, the operation of the fuel cell stack 2 also generates heat. Therefore, the fuel cell stack 2 is connected to a cooling circuit 3 incorporating a coolant pump 4. The circulating coolant transfers the captured heat to a cooler 6, which may in particular be the vehicle's main cooler. To bypass the cooler 6, a cooler bypass 7 is provided, which is opened via a directional control valve 8. Pressure sensors 5 are integrated into the cooling circuit 3 in the region of the inlet 2.1 to the fuel cell stack 2 and in the region of the outlet 2.2 from the fuel cell stack 2. These pressure sensors measure the coolant pressure in the region of the inlet 2.1 and outlet 2.2, so that the coolant pressure difference across the fuel cell stack 2 can be determined on the basis of these measurements (see also FIG. 2). Instead of the two pressure sensors 5, a differential pressure sensor (not shown) can also be used.

[0032] As shown exemplarily in Figure 3, during cold or sub-freezing start-up, the coolant pressure difference Δp across the fuel cell stack 2 decreases for a constant rotation speed n of the coolant pump 4. This progression is in the opposite direction to the coolant temperature, since the viscosity of the coolant decreases with increasing temperature. In Figure 3, the coolant temperature T Ein and the coolant temperature T in the area of ​​outlet 2.2 Aus, the temperature change is indirectly readable from the coolant volumetric flow rate Q or the pressure difference Δp. The present invention makes use of this relationship. Since the pressure difference Δp can be detected using a simple pressure sensor (see pressure sensor 5 in Figures 1 and 2), which is much more affordable than a volumetric flow sensor, the present invention focuses on the pressure difference Δp to indirectly detect the temperature change in the fuel cell stack 2.

[0033] 4, a first possible operating strategy according to the method of the invention is shown. At the beginning of a low temperature or sub-zero start-up (step 100), the coolant pump 4 is first switched on (step 110). Subsequently, the coolant inlet temperature T Ein It is checked (step 120) whether the inlet temperature T is greater than 0°C. If this is the case ("yes"), the method can already be ended (step 130). If this is not the case ("no"), the pressure difference Δp of the coolant across the fuel cell stack 2 is sensed (step 140) and it is checked whether the pressure difference Δp decreases, i.e. whether the fuel cell is heating up. If this is the case ("yes"), the inlet temperature T Ein is measured anew, which means that steps 120 and 130 or 140 are repeated. If this is not the case ("No"), the pump speed n is reduced (step 150) before these steps are repeated.

[0034] In Fig. 5, a modified operating strategy is shown, in which steps 200 to 230 correspond to steps 100 to 130. In step 240, the pressure difference Δp of the coolant across the fuel cell stack 2 is likewise detected and evaluated. However, what is checked is whether the pressure difference Δp is constant. If this is the case ("yes"), the measured temperature T EinDepending on the value of the temperature difference Δp, step 220 and step 230 or step 240 are repeated. If this is not the case ("No"), it is checked in step 250 whether the pressure difference Δp decreases. If this is the case ("Yes"), the pump speed n is increased (step 260). If this is not the case ("No"), the pump speed n is decreased (step 270). Subsequently, steps 220 and 230 or the measured temperature T Ein Steps 220, 240 and 250 are repeated accordingly.

[0035] FIG. 6 shows the progression of the coolant pressure difference Δp across the fuel cell stack 2 during a cold or sub-zero start. The actual cold or sub-zero start first begins at time t1. Prior to this, there is a phase in which the cold or sub-zero start is prepared. During this preparation phase, the pressure difference Δp is set to a minimum value, specifically an initial value corresponding to threshold value S1. Then, during the subsequent cold or sub-zero start, the pressure difference Δp is set to a target value within the range defined by threshold values ​​S2 and S3. By determining threshold values ​​S1-S3, excessive volumetric flow rates and therefore icing are prevented. Meanwhile, a sufficient volumetric flow rate for circulating the coolant is still guaranteed.

[0036] 7 shows an exemplary possible operating strategy for the preparatory phase. At the beginning of the preparatory phase (step 300), the coolant pump 4 is first switched on (step 310). Subsequently, in step 320, the pressure difference Δp is detected and it is checked whether this pressure difference has reached the threshold value S1. If this is the case ("yes"), the preparatory phase can be ended and a low-temperature or sub-zero start can be initiated (step 330). If this is not the case ("no"), the pump speed n must be increased (step 340).

[0037] 8 shows another possible operating strategy for adapting the initial pump speed n depending on the temperature. According to the invention, the temperature is detected indirectly via the coolant pressure difference Δp across the fuel cell stack 2. With the start of adaptation (step 400), it is first checked whether a cold or sub-zero start has already been performed and terminated (step 410), since only then is the information required for adaptation present. Subsequently, it is checked whether the pressure difference Δp was below threshold value S2 at the end of the cold or sub-zero start (step 420). If the result of the check is positive ("yes"), threshold value S1 is increased in step 430 and adaptation is terminated (step 440). If the result of the check is negative ("no"), it is checked whether the pressure difference Δp was above threshold value S3 at the end of the cold or sub-zero start (step 450). If this is not the case ("no"), no adaptation needs to be performed and the method can be terminated with step 440. However, if threshold S3 is exceeded ("yes"), threshold S1 must be lowered (460), and only then can the adaptation be terminated (step 440).

[0038] FIG. 9 shows another possibility for adapting the initial pump speed n based on an evaluation of the course of the coolant pressure difference Δp across the fuel cell stack 2. This method exploits the fact that a temporarily stagnant course indicates icing. A stagnant course can be recognized as a plateau in an otherwise descending curve (see FIG. 3). With the start of adaptation (step 500), the method first checks whether a cold or subzero start has already been performed and completed (step 510), since only then is the information necessary for adaptation present. Subsequently, it is checked whether the course of the coolant pressure difference Δp indicates a plateau (step 520). If this is not the case ("No"), no adaptation is necessary and the method can be terminated (step 530). However, if the result of the check is positive ("Yes"), the threshold value S3 is lowered in step 540, and the adaptation is terminated in the subsequent step 530.

[0039] In order to avoid initially very high pump speeds n, instead of increasing the pump speed n continuously, a gradual or stepwise operating strategy can be selected, as exemplarily shown in FIG. 10. As the speed n gradually increases, the pressure difference Δp also increases. During the time that the speed n is kept constant, the pressure difference Δp decreases again. In this way, the pressure difference Δp is kept within a certain range, exhibiting a progression similar to the sawtooth shown in FIG. 10. This range is determined by a lower limit (Δp) which varies depending on the temperature, as exemplarily shown in FIG. 11. U ) and upper limit (Δp A ) can be determined by [Explanation of symbols]

[0040] 1. Fuel cell system 2. Fuel cell stack 2.1 Inlet (to fuel cell stack 2) 2.2 Outlet (from fuel cell stack 2) 3 Cooling circuit 4 refrigerant pump 5 Pressure Sensor 6 Cooler 7 Cooler Bypass 8. Directional Control Valve 10 anodes 11 Anode Path 12 Tank 13 Shut-off valve 14 Heat exchanger 15 Pressure Controller 16 Injection pump 17 Recirculation Path 18 Blower 19 Water separator 20 containers 21 Drain valve 22 Purge valve 23 Cathode 24 cathode path 25 Air Filter 26 Air Compressor 27 Heat exchanger 28 Shut-off valve 29 Exhaust route 30 Bypass Route 31 Bypass valve Q Refrigerant flow rate S1, S2, S3 thresholds T Ein Coolant temperature (in the area of ​​inlet 2.1) T Aus refrigerant temperature (in the area of ​​outlet 2.2) Δp pressure difference Δp U Lower limit Δp A upper limit

Claims

1. 1. A method for operating a fuel cell system (1) comprising a fuel cell stack (2) having a plurality of fuel cells passing through cooling passages to which a coolant is supplied via a cooling circuit (3) using a coolant pump (4), characterized in that during low temperature or sub-zero start-up, the temperature of the fuel cells in the fuel cell stack (2) is indirectly detected via a pressure difference (Δp) of the coolant across the fuel cell stack (2), and the speed (n) of the coolant pump (4) is controlled in dependence on the indirectly detected temperature.

2. 2. The method of claim 1, wherein the pressure (p1) of the coolant at the inlet (2.1) and outlet (2.2) of the fuel cell stack (2) is measured using a plurality of pressure sensors (5) or a differential pressure sensor to detect the pressure difference (Δp) of the coolant across the fuel cell stack (2).

3. 3. The method according to claim 1 or 2, characterized in that the rotational speed (n) of the coolant pump (4) is continuously or gradually increased during the low temperature or sub-zero start-up in order to keep the pressure difference (Δp) of the coolant across the fuel cell stack (2) substantially constant or within a predetermined range.

4. 3. The method according to claim 1 or 2, characterized in that the rotational speed (n) of the coolant pump (4) is kept constant during the low temperature or sub-zero start-up in order to reduce the pressure difference (Δp) of the coolant across the fuel cell stack (2).

5. 3. The method according to claim 1 or 2, characterized in that the rotational speed (n) of the coolant pump (4) is reduced during the low-temperature or sub-zero start-up if the pressure difference (Δp) of the coolant across the fuel cell stack (2) exhibits an upward trend and / or exceeds a predetermined maximum value.

6. 3. The method according to claim 1 or 2, characterized in that a first threshold (S1) defining an initial minimum pressure difference of the coolant across the fuel cell stack (2) is determined, and in preparation for a low temperature or sub-zero start, the rotational speed (n) of the coolant pump (4) is increased until the first threshold (S1) is reached.

7. 7. The method of claim 6, wherein two further thresholds (S2, S3) are determined that define a range of pressure differential of the coolant across the fuel cell stack (2) to be achieved during the low temperature or sub-zero start-up, the two further thresholds (S2, S3) being below the first threshold (S1).

8. 7. The method of claim 6, wherein the first threshold value (S1) is increased or decreased if the pressure difference (Δp) of the coolant across the fuel cell stack (2) at the end of the low temperature or sub-zero start-up is outside the range of pressure difference (Δp) to be achieved.

9. 8. The method according to claim 7, characterized in that the progress of the pressure difference during or after the cold or sub-zero start-up is evaluated, and if the progress temporarily stagnates, the third threshold (S3) of one of the two further thresholds (S2, S3) is lowered, so that the third threshold approaches the second threshold (S2) of the other of the two further thresholds (S2, S3).

10. 10. The method of claim 9, wherein at least one of the first threshold (S1), the second threshold (S2) and the third threshold is stored in a control device configured to perform the steps of the method.

11. A control device for a fuel cell system (1), configured to carry out the steps of the method according to claim 1 or 2.

Images