Method and control unit for operating a fuel cell system, and fuel cell system

Abstract

The invention relates to a method for operating a fuel cell system (200). The fuel cell system (200) comprises at least one fuel cell module (100) designed as a proton exchange membrane fuel cell. The method comprises a step of determining at least one parameter (247) of the fuel cell system (200) during a starting operation of the fuel cell system (200) using at least one sensor signal (235) that represents a pressure on an anode side of the at least one fuel cell module (100) and / or a pressure on a cathode side of the at least one fuel cell module (100). During the starting operation, firstly a feed of hydrogen and subsequently a feed of air to the at least one fuel cell module (100) is activated. The method also comprises a step of generating an operating parameter (249) for the operation of the fuel cell system (200) by evaluating the at least one determined parameter (247) using an evaluation rule. The operating parameter (249) represents an internal state of tightness of the fuel cell system (200).

Description

[0001] Description

[0002] title

[0003] Method and control device for operating a fuel cell system and fuel cell system

[0004] State of the art

[0005] The present invention relates to a method, to a corresponding control device and to a corresponding computer program product.

[0006] PEM fuel cell stacks (PEM = Proton Exchange Membrane) or polymer electrolyte fuel cell stacks comprise many individual cells that are stacked and held together by a clamping system. The core element of an individual cell is a membrane that separates the media on the anode and cathode sides. The tightness of the membrane can deteriorate over time due to a variety of damage mechanisms, such as mechanical, chemical, and thermal. During shutdown or maintenance, internal and external leakage between the anode and cathode can be checked using standard leakage tests, such as pressure drop tests. A complex single-cell voltage monitor (CVM = Cell Voltage Monitor) can detect membrane damage during operation.By choosing a special operating strategy, damage to the membrane during operation can be minimized.

[0007] Disclosure of the invention

[0008] Against this background, the approach presented here presents a method, a control unit that uses this method, and finally a corresponding computer program product according to the main claims. Advantageous embodiments emerge from the respective subclaims and the following description.

[0009] According to embodiments, leak detection in fuel cell modules can advantageously be implemented by evaluating the pressure curve during start-up. Thus, for example, a potentially dangerous leak can be identified using on-board diagnostics during each start-up or so-called protected start of the system. Complex individual cell voltage monitoring, for example, can be omitted. During protected start, the pressure in the stack or in each fuel cell module is lower than during operation. A negative pressure compared to the environment is also possible. The anode side and the cathode side are preferably filled with hydrogen. In particular, timely leak detection can prevent the membrane's tightness from critically deteriorating due to a variety of damage mechanisms, e.g., mechanical, chemical, or thermal.For example, by detecting in time that a membrane is about to leak, high efficiency can be maintained because oxygen in the air and hydrogen cannot pass to the other side and react directly without generating electricity. This also prevents the leakage from becoming too high locally at one point, which could otherwise lead to significant heating at that location.

[0010] A method for operating a fuel cell system is presented, wherein the fuel cell system has at least one fuel cell module designed as a proton exchange membrane fuel cell, the method comprising the following steps:

[0011] Determining at least one characteristic of the fuel cell system during a start-up process of the fuel cell system using at least one sensor signal that represents a pressure on an anode side of the at least one fuel cell module and / or a pressure on a cathode side of the at least one fuel cell module, wherein during the start-up process, first a hydrogen supply and then an air supply to the at least one fuel cell module are switched on; and generating an operating parameter for operating the fuel cell system by evaluating the at least one determined characteristic using an evaluation rule, wherein the operating parameter represents an internal tightness state of the fuel cell system.

[0012] A plurality of fuel cell modules designed as proton exchange membrane fuel cells can also be referred to as a fuel cell stack or, more precisely, a PEM fuel cell stack (PEM = Proton Exchange Membrane), a proton exchange membrane fuel cell stack, or a polymer electrolyte fuel cell stack. Such fuel cell stacks comprise many individual fuel cell modules or individual cells, which can be stacked and held together, for example, by a clamping system. Each individual cell can have a membrane that separates the media on the anode and cathode sides. Between the individual cells or individual cells,

[0013] So-called bipolar plates can be arranged in fuel cell modules, through which cooling water can flow in order to dissipate heat loss from the chemical reaction. The method can also comprise a step of controlling the operation of the fuel cell system using the operating parameter. The fuel cell system can comprise at least one pressure sensor and additionally or alternatively at least one blocking device or device for controlling the air supply and additionally or alternatively at least one device for controlling the hydrogen supply. The at least one blocking device can be designed to close off the air supply to the fuel cell stack or each fuel cell module from the environment when switched off by means of flaps or valves, so that a closed space is formed. The at least one pressure sensor can be arranged on the anode side and additionally or alternatively on the cathode side.The at least one pressure sensor can be designed to measure a pressure curve during the start of the fuel cell system.

[0014] According to embodiments, in particular an internal leakage of fuel cell modules between the anode and cathode, as well as additional external leaks of the cathode, can be detected without complex individual cell voltage monitoring and independently of the operating strategy by means of on-board diagnostics, for example so that maintenance or servicing can be planned at an early stage.

[0015] According to one embodiment, in the determining step, a pressure rise rate in response to the activation of the hydrogen supply can be determined as a characteristic variable. The evaluation rule used in the generating step can include a comparison of the pressure rise rate with a predefined reference value. The operating parameter generated in the generating step can represent an internal leak if the pressure rise rate falls below the reference value.

[0016] In the determination step, a pressure curve on the cathode side after the hydrogen supply is switched on and before the air supply is switched on can also be determined as a parameter. The evaluation rule used in the generation step can include a comparison of the pressure curve with a predefined threshold value. The operating parameter generated in the generation step can represent an internal leak if the pressure curve exceeds the threshold value.

[0017] Furthermore, in the determining step, a time period from the activation of the air supply until a target pressure is reached on the cathode side can be determined as a parameter. The evaluation rule used in the generating step can include a comparison of the time period with a predefined reference time period. The operating parameter generated in the generating step can represent an internal leak if the time period falls short of the reference time period.

[0018] In addition, pressure and / or temperature curves on the anode side, on the cathode side, in the hydrogen supply, and additionally or alternatively in a cooling water supply during the start-up process can be determined as parameters in the determination step. The evaluation rule used in the generation step can include a classification of the parameters into different classes using machine learning. The operating parameter generated in the generation step can represent an internal leak if at least one of the parameters falls into a class representing a leak. The evaluation rule can include a machine learning algorithm, more precisely, so-called supervised learning.

[0019] In the determination step, pressure and / or temperature curves on the anode side, on the cathode side, in the hydrogen supply, and additionally or alternatively in a cooling water supply during the start-up process can also be determined as parameters. The evaluation rule used in the generation step can monitor the parameters for unusual behavior patterns using machine learning. The operating parameter generated in the generation step can represent an internal leak if at least one of the parameters exhibits an unusual behavior pattern. The evaluation rule can use a machine learning algorithm, more precisely, so-called unsupervised learning.

[0020] The parameters mentioned here can be processed individually or in any combination within the method. All relevant embodiments offer the advantage that a leak can be detected easily, early, accurately, and reliably.

[0021] The method may also include a step of reading the at least one sensor signal via an interface of at least one pressure sensor of the fuel cell system. This allows the measured values ​​for determining the parameters to be obtained in a simple manner.

[0022] The approach presented here further provides a device designed to carry out or implement the steps of a variant of a method presented here in corresponding devices. This embodiment of the invention in the form of a device also allows the problem underlying the invention to be solved quickly and efficiently.

[0023] In this case, a device can be understood as an electrical device that processes sensor signals and outputs control and / or data signals depending on them. The device can have an interface, which can be implemented in hardware and / or software. In a hardware implementation, the interfaces can, for example, be part of a so-called system ASIC, which contains a wide variety of functions of the device. However, it is also possible for the interfaces to be separate integrated circuits or to consist at least partially of discrete components. In a software implementation, the interfaces can be software modules that are present, for example, on a microcontroller alongside other software modules.

[0024] A fuel cell system is also presented, comprising the following features: at least one fuel cell module designed as a proton exchange membrane fuel cell; and an embodiment of a control unit mentioned herein.

[0025] The fuel cell system can have at least one pressure sensor and, additionally or alternatively, at least one blocking device or device for controlling the air supply and, additionally or alternatively, at least one device for controlling the hydrogen supply. The control unit can be connected in a signal-transmitting manner to the at least one pressure sensor and, additionally or alternatively, to the at least one blocking device and, additionally or alternatively, to the at least one control device.

[0026] The approach presented here further provides a control unit configured to perform or implement the steps of a variant of a method presented here in corresponding devices. This embodiment of the invention in the form of a control unit also allows the problem underlying the invention to be solved quickly and efficiently.

[0027] In this context, a control unit can be understood as an electrical device that processes sensor signals and outputs control and / or data signals depending on them. The control unit can have an interface that can be implemented in hardware and / or software. In a hardware implementation, the interfaces can, for example, be part of a so-called system ASIC, which contains a wide variety of functions of the control unit. However, it is also possible for the interfaces to be separate integrated circuits or to consist at least partially of discrete components. In a software implementation, the interfaces can be software modules that are present, for example, on a microcontroller alongside other software modules.

[0028] Also advantageous is a computer program product with program code that can be stored on a machine-readable medium such as a semiconductor memory, a hard disk memory or an optical memory and is used to carry out the method according to one of the embodiments described above when the program product is executed on a computer or a device.

[0029] The approach presented here is explained in more detail below using the attached drawings. They show:

[0030] Fig. 1 is a schematic representation of a fuel cell module;

[0031] Fig. 2 is a schematic representation of an embodiment of a fuel cell system;

[0032] Fig. 3 is a schematic diagram of voltage and current waveforms during a starting process of an embodiment of a fuel cell system;

[0033] Fig. 4 is a schematic diagram of pressure curves during a start-up process of an embodiment of a fuel cell system; and

[0034] Fig. 5 is a flowchart of an embodiment of a method for operating a fuel cell system.

[0035] In the following description of advantageous embodiments of the present invention, the same or similar reference numerals are used for the elements shown in the various figures and having a similar effect, whereby a repeated description of these elements is omitted.

[0036] Fig. 1 shows a schematic representation of a fuel cell module 100. The fuel cell module 100 comprises a proton exchange membrane fuel cell. In other words, Fig. 1 schematically shows the structure of such a fuel cell. A fuel cell system comprises a plurality of fuel cell modules, such as the fuel cell module 100 shown here, which are arranged in a so-called fuel cell stack.

[0037] Such fuel cell stacks comprise many individual fuel cell modules or single cells, such as the fuel cell module 100 shown here, which are stacked and held together, for example, by a clamping system. Each individual cell or each fuel cell module 100 comprises a membrane 101 that separates the media on an anode side 103 or anode and a cathode side 105 or cathode. So-called bipolar plates 107 are arranged between adjacent individual cells or fuel cell modules, such as the fuel cell module 100, through which cooling water can flow to dissipate heat loss from the chemical reaction. Seals 109 are arranged between the membrane 101 and the bipolar plates 107 to seal the anode side 103 and the cathode side 105, respectively.

[0038] If the membrane is leaking, a pressure difference between the anode side 103 (anode) and the cathode side 105 (cathode) can cause hydrogen to flow from the anode side 103 to the cathode side 105, where it can react directly with the oxygen without generating electricity. This scenario is symbolically illustrated in Fig. 1 by an arrow X, which represents an internal leakage of hydrogen to the air.

[0039] Fig. 2 shows a schematic representation of an embodiment of a fuel cell system 200. The fuel cell system 200 comprises at least one fuel cell module 100 designed as a proton exchange membrane fuel cell, which corresponds to or is similar to the fuel cell module from Fig. 1. In the illustration of Fig. 2, the fuel cell system 200 comprises only four fuel cell modules 100 by way of example. The fuel cell system 200 can have any suitable, deviating number of fuel cell modules 100. A

[0040] A plurality of fuel cell modules 100 are arranged in the fuel cell system 200 in a so-called stack 210.

[0041] The fuel cell system 200 also comprises at least one pressure sensor 230 which is designed to detect a pressure on an anode side of the at least one fuel cell module 100 and / or a pressure on a cathode side of the at least one fuel cell module 100 and to provide it in the form of a sensor signal 235.

[0042] The fuel cell system 200 further comprises a control unit 240 configured to operate the fuel cell system 200 or to control the operation thereof. The control unit 240 is connected in a signal-transmitting manner to the at least one fuel cell module 100 or the stack 210, or more precisely, to at least one pressure sensor 230 associated with the at least one fuel cell module 100 or the stack 210.

[0043] For example, the control unit 240 comprises a reading device 244 which is designed to read the sensor signals 235 from the at least one pressure sensor 230 via an interface 241.

[0044] The control unit 240 comprises a determination device 246 and a generation device 248. The determination device 246 is designed to determine at least one characteristic variable 247 of the fuel cell system 200 during a start-up process of the fuel cell system 200 using the sensor signal 235, which represents the pressure on an anode side of the at least one fuel cell module 100 and / or the pressure on a cathode side of the at least one fuel cell module 100. During the start-up process, first a hydrogen supply and then an air supply to the at least one fuel cell module 100 are switched on. The generation device 248 is designed to generate an operating parameter 249 for operating the fuel cell system 100 by evaluating the at least one determined characteristic variable 147 using an evaluation rule.The operating parameter 249 represents an internal leakage state of the fuel cell system 200. According to one embodiment, the determination device 246 is configured to determine a pressure rise rate in response to the activation of the hydrogen supply as a parameter 247. The evaluation rule used by the generation device 248 comprises a comparison of the pressure rise rate with a predefined reference value. The generation device 248 is configured to generate an operating parameter 249 that represents an internal leak if the pressure rise rate falls below the reference value.

[0045] According to one embodiment, the determination device 246 is configured to determine a pressure profile on the cathode side after the hydrogen supply is switched on and before the air supply is switched on as a parameter 247. The evaluation rule used by the generation device 248 comprises a comparison of the pressure profile with a predefined threshold value. The generation device 248 is configured to generate an operating parameter 249 that represents an internal leak if the pressure profile exceeds the threshold value.

[0046] According to one embodiment, the determination device 246 is configured to determine a time period from the switching on of the air supply until a target pressure is reached on the cathode side as a parameter 247. The evaluation rule used by the generation device 248 comprises a comparison of the time period with a predefined reference time period. The generation device 248 is configured to generate an operating parameter 249 that represents an internal leak if the time period falls below the reference time period.

[0047] According to one exemplary embodiment, the determination device 246 is configured to determine pressure profiles and / or temperature profiles on the anode side, on the cathode side, in the hydrogen supply, and / or in a cooling water supply during the start-up process as characteristic variables 247. The evaluation rule used by the generation device 248 comprises classifying the characteristic variables into different classes using machine learning. The generation device 248 is configured to generate an operating parameter 249 that represents an internal leak if at least one of the characteristic variables 247 falls into a class representing a leak.

[0048] According to one exemplary embodiment, the determination device 246 is configured to determine pressure profiles and / or temperature profiles on the anode side, on the cathode side, in the hydrogen supply, and / or in a cooling water supply during the start-up process as characteristic variables 247. The evaluation rule used by the generation device 248 includes monitoring the characteristic variables for unusual behavior patterns using machine learning. The generation device 248 is configured to generate an operating parameter 249 that represents an internal leak if at least one of the characteristic variables 247 exhibits an unusual behavior pattern.

[0049] Fig. 3 shows a schematic diagram of voltage curves and current curves during a start-up process of an embodiment of a fuel cell system. The start-up process relates to the fuel cell system from one of the figures described herein or a similar fuel cell system. In the diagram, a first voltage curve V1, which represents a cell voltage limited to 0.85 V by way of example only, a second voltage curve V2, which represents a case in which the voltage can fall to -0.1 V by way of example only in the event of an insufficient air supply to a cell, a first current curve I1, which represents an actual current, and a second current curve I2, which represents a desired current value, are plotted over time t. Furthermore, an open circuit voltage Voc andOCV (open-circuit voltage), which is merely an example of 0.95 V, a first time t1 at which the hydrogen supply is switched on, and a second time t2 at which the air supply is switched on are plotted. Initially, the hydrogen supply or hydrogen supply is switched on at the first time t1. After a certain time, the air supply or air supply is switched on at the second time t2, and the fuel cell stack supplies current and voltage.

[0050] Fig. 4 shows a schematic diagram of pressure curves during a start-up process of an exemplary embodiment of a fuel cell system. The start-up process relates to the fuel cell system from one of the figures described herein or a similar fuel cell system. In the diagram, a first hydrogen pressure curve P1 a, which represents a hydrogen pressure on the anode side with internal tightness, a second hydrogen pressure curve P1 b, which represents a hydrogen pressure on the anode side with internal leakage, a first air pressure curve P2a, which represents an air pressure on the cathode side with internal tightness, and a second air pressure curve P2b, which represents an air pressure on the cathode side with internal leakage, are plotted against time t. Furthermore, a first time t1, at which the hydrogen supply is switched on, and a second time t2, at which the air supply is switched on, are shown.

[0051] In a fuel cell stack with internal tightness or no internal leaks, the first hydrogen pressure curve P1a and the first air pressure curve P2a are established. After the fuel cell supply or hydrogen feed is switched on at the first time t1, hydrogen flows into the anode volume of the stack. Since there is no oxygen on the anode or cathode sides, no power is generated and therefore no hydrogen is consumed. The pressure on the anode side rises relatively quickly. This is shown by the first hydrogen pressure curve P1a. Initially, no or only a slight change in pressure is measured on the cathode side. A pressure increase only occurs after the air supply or air feed is switched on at the second time t2. This is shown by the first air pressure curve P2a.

[0052] In a fuel cell stack with internal leakage, the second hydrogen pressure curve P1 b and the second air pressure curve P2 b occur. After the fuel cell supply or hydrogen feed is switched on at the first time t1, hydrogen flows into the anode volume of the stack. Since there is no oxygen on the anode or cathode side, no power is generated and therefore no hydrogen is consumed. Since the internal leakage allows hydrogen to flow from the anode side to the cathode side, the pressure increase in the case of the second hydrogen pressure curve P1 b is slower than in the first hydrogen pressure curve P1 a, since the larger volume of the anode and cathode now has to be filled. In addition, a pressure increase can already be measured on the cathode side between the first time t1 and the second time t2, as hydrogen flows to the cathode side. This can be read off the second air pressure curve P2 b.After the air supply is switched on at the second time t2, the pressure on the cathode side continues to rise until the target pressure is reached. This can also be seen from the second air pressure curve P2b.

[0053] Fig. 5 shows a flowchart of an embodiment of a method 540 for operating a fuel cell system. The method 540 for operating can be executed to operate a fuel cell system or to control the operation of a fuel cell system. The method 540 for operating can be carried out in connection with the fuel cell system from one of the figures described herein or a similar fuel cell system. The steps of the method 540 for operating can be executed using a control unit from one of the figures described herein or a similar control unit.

[0054] The operating method 540 comprises a determining step 546 and a generating step 548. In determining step 546, at least one characteristic of the fuel cell system is determined during a start-up process of the fuel cell system using at least one sensor signal that represents a pressure on an anode side of the at least one fuel cell module and / or a pressure on a cathode side of the at least one fuel cell module. During the start-up process, first a hydrogen supply and then an air supply to the at least one fuel cell module are switched on. Subsequently, in generating step 548, an operating parameter for operating the fuel cell system is generated by evaluating the at least one determined characteristic using an evaluation rule. The operating parameter represents an internal tightness state of the fuel cell system.

[0055] According to one exemplary embodiment, the operating method 540 also includes a step 544 of reading in the at least one sensor signal via an interface of at least one pressure sensor of the fuel cell system. With reference to the figures described herein, one exemplary embodiment is summarized below and, in other words, briefly explained.

[0056] To generate the operating parameter 249, the control unit 240 and / or the method 540 can evaluate the following parameters 247 individually or in any combination for diagnosis and evaluation: a) The rate of pressure increase during or after the H2 supply or hydrogen supply is switched on at the first time t1: a reduced rate of pressure increase indicates an increased internal leak. See also the second hydrogen pressure curve P1b in Fig. 4. b) The pressure curve on the cathode side after the H2 supply or hydrogen supply is switched on at the first time t1 and before the air supply or air supply is switched on at the second time t2: If the pressure curve rises here, for example, above a threshold, this is an indication of an internal leak. See also the second air pressure curve P2b in Fig. 4. c) The time period from the time the air supply is switched on to the first time t1: a reduced rate of pressure increase indicates an increased internal leak.Air supply at the second time t2 until the target pressure is reached on the cathode side: If the target pressure is reached too quickly, this is an indication of an internal leak. See also the second air pressure curve P2b in Fig. 4. d) Evaluation of the pressure curves of the anode, H2 supply, cathode, cooling water as well as the temperatures of the anode, H2 supply, cathode, cooling water in the start-up phase or during the start-up process and classification into good / bad, e.g. using machine learning or machine learning, in particular supervised learning. e) Evaluation of the pressure curves of the anode, H2 supply, cathode, cooling water as well as the temperatures of the anode, H2 supply, cathode, cooling water in the start-up phase or during the start-up process and monitoring for unusual behavior, e.g. using machine learning or machine learning, in particular unsupervised learning.

[0057] Relevant influencing parameters and fundamental principles are explained below: The fundamental influencing parameters are described by the general gas equation n = (p * V) / (R * T). Temperature, in particular, has an influence and can be taken into account by only evaluating measurements at the same temperature. Furthermore, the incoming amount of hydrogen also has an influence. If the flow through the supply is lower, pressure buildup occurs more slowly. The evaluation is part of the software of the control unit 240. In particular, pressure sensors 230 on the anode side 103 and / or the cathode side 105 are used to measure the pressure during the start-up process.

[0058] The exemplary embodiments described and shown in the figures are selected only as examples. Different exemplary embodiments can be combined with each other in their entirety or with regard to individual features. Furthermore, one exemplary embodiment can be supplemented by features of another exemplary embodiment.

[0059] Furthermore, the process steps presented here can be repeated and carried out in a different order than that described.

[0060] If an embodiment comprises an “and / or” link between a first feature and a second feature, this is to be read as meaning that the embodiment according to one embodiment has both the first feature and the second feature and according to another embodiment has either only the first feature or only the second feature.

Claims

Claims 1. A method (540) for operating a fuel cell system (200), wherein the fuel cell system (200) comprises at least one fuel cell module (100) designed as a proton exchange membrane fuel cell, the method (540) comprising the following steps: Determining (546) at least one characteristic variable (247) of the fuel cell system (200) during a start-up process of the fuel cell system (200) using at least one sensor signal (235) representing a pressure on an anode side (103) of the at least one fuel cell module (100) and / or a pressure on a cathode side (105) of the at least one fuel cell module (100), wherein during the start-up process, first a hydrogen supply and then an air supply to the at least one fuel cell module (100) is switched on; and Generating (548) an operating parameter (249) for operating the fuel cell system (200) by evaluating the at least one determined characteristic variable (247) using an evaluation rule, wherein the operating parameter (249) represents an internal tightness state of the fuel cell system (200).

2. Method (540) according to claim 1, wherein in the step (546) of determining a pressure rise rate in response to the switching on of the hydrogen supply is determined as a characteristic variable (247), wherein the evaluation rule used in the step (548) of generating comprises a comparison of the pressure rise rate with a predefined reference value, wherein the operating parameter (249) generated in the step (548) of generating represents an internal leakage if the pressure rise rate falls below the reference value.

3. Method (540) according to one of the preceding claims, wherein in the step (546) of determining, a pressure profile on the cathode side (105) after switching on the hydrogen supply and before switching on the air supply is determined as a characteristic variable (247), wherein the evaluation rule used in the step (548) of generating comprises a comparison of the pressure profile with a predefined threshold value, wherein the operating parameter (249) generated in the step (548) of generating represents an internal leakage if the pressure profile exceeds the threshold value.

4. Method (540) according to one of the preceding claims, wherein in the step (546) of determining, a time period from the switching on of the air supply until reaching a target pressure on the cathode side (105) is determined as a characteristic (247), wherein the evaluation rule used in the step (548) of generating comprises a comparison of the time period with a predefined reference time period, wherein the operating parameter (249) generated in the step (548) of generating represents an internal leakage if the time period falls below the reference time period 5. The method (540) according to one of the preceding claims, wherein in the step (546) of determining, pressure profiles and / or temperature profiles on the anode side (103), on the cathode side (105), in the hydrogen supply and / or in a cooling water supply during the start-up process are determined as characteristic variables (247), wherein the evaluation rule used in the step (548) of generating comprises a classification of the characteristic variables (247) into different classes by means of machine learning, wherein the operating parameter (249) generated in the step (548) of generating represents an internal leak if at least one of the characteristic variables (247) falls into a class representing a leak.

6. Method (540) according to one of the preceding claims, wherein in step (546) of determining pressure profiles and / or temperature profiles on the anode side (103), on the cathode side (105), in the hydrogen supply and / or in a cooling water supply during the starting process are determined as parameters (247), wherein the evaluation rule used in the generating step (548) comprises monitoring the parameters (247) for conspicuous behavior patterns by means of machine learning, wherein the operating parameter (249) generated in the generating step (548) represents an internal leakage if at least one of the parameters (247) shows a conspicuous behavior pattern.

7. Method (540) according to one of the preceding claims, comprising a step (544) of reading in the at least one sensor signal (235) via an interface (241) from at least one pressure sensor (230) of the fuel cell system (200).

8. Control device (240) which is designed to carry out the steps of a method (540) according to one of the preceding claims in corresponding devices (244, 246, 248).

9. A fuel cell system (200) comprising: at least one fuel cell module (100) configured as a proton exchange membrane fuel cell; and a control unit (240) according to claim 8.

10. Computer program product with program code for carrying out the method (540) according to one of claims 1 to 7, when the program product is executed on a device.

Images