Method for controlling a fuel cell device, and fuel cell device

Abstract

The invention relates to a method for controlling a fuel cell device. In at least one step, at least one exhaust gas measurement value (14a; 14b; 14c; 14d; 14e, 22e; 14f) of the composition of an exhaust gas of the fuel cell device is detected, in order to adjust the flow rate of fuel flowing through the fuel cell device on the basis of a controlled variable (16a; 16b; 16c; 16d; 16e; 16f) which is based on the exhaust gas measurement value (14a; 14b; 14c; 14d; 14e, 22e; 14f). According to the invention, in at least one step, the actual value of the controlled variable (16a; 16b; 16c; 16d; 16e; 16f) is ascertained in that an analysis function is analyzed which correlates the exhaust gas measurement value (14a; 14b; 14c; 14d; 14e, 22e; 16f) directly to the controlled variable (16a; 16b; 16c; 16d; 16e; 16f).

Description

[0001] Description

[0002] Method for controlling a fuel cell device and fuel cell device

[0003] State of the art

[0004] In DE 10 2017 222 558 A1, a method for controlling a fuel cell device has already been proposed, in which, in at least one method step, at least one exhaust gas measurement value of a composition of an exhaust gas of the fuel cell device is recorded in order to adjust a fuel flow rate through the fuel cell device and / or an electrical power parameter of the fuel cell device as a function of a control variable dependent on the exhaust gas measurement value.

[0005] Disclosure of the invention

[0006] The invention is based on a method for controlling a fuel cell device, in which in at least one method step at least one exhaust gas measurement value of a composition of an exhaust gas of the fuel cell device is detected in order to adjust a fuel flow rate and / or an electrical power parameter of the fuel cell device by the fuel cell device as a function of a control variable dependent on the exhaust gas measurement value.

[0007] It is proposed that in at least one method step of the method, an actual value of the controlled variable is determined by evaluating an analysis function which directly correlates the exhaust gas measurement value with the controlled variable. The method preferably forms a closed control loop, in particular in contrast to pure feedforward control. Optionally, in addition to the control loop, the method comprises a control of the controlled variable. The fuel cell device preferably comprises a control unit which evaluates the analysis function and, depending on the actual value, controls at least one actuating unit of the fuel cell device in order to adjust the fuel flow rate and / or the electrical power parameter. The actuating unit is preferably a fluid conveying unit, in particular a pump, a compressor, a blower, a fan or the like, or a shut-off device, in particular a valve, a throttle valve or the like.Additionally or alternatively, the control or regulating unit uses an electrical control unit of the fuel cell device, for example, an inverter, an adjustable resistor, or the like, to adjust the electrical power parameter. The electrical power parameter is preferably an electrical current provided by the fuel cell device and / or an electrical voltage associated with the provided electrical current.

[0008] The analysis function is preferably a regression function and / or a correlation function. A regression function as an analysis function is preferably determined in advance of the method, for example on a test bench with additional measuring devices, and stored in a memory of the control unit. A correlation function as an analysis function is preferably determined in advance of the method, for example by means of a simulation and / or theoretical analysis of the fuel cell device, and stored in a memory of the control unit. It is also conceivable for the analysis function to be based partly on empirically collected data and partly on simulated or theoretical data. The analysis function can be stored in the control unit as a closed mathematical expression, as an iterative calculation rule, as a table, or in another form that appears appropriate to the person skilled in the art.The analysis function preferably expresses the controlled variable as a function of the measured exhaust gas value. For example, the analysis function is a first-order, preferably second-order, optionally higher-order polynomial in the measured exhaust gas value. In a specific embodiment of the method, the analysis function is equal to the mathematical identity, and the measured exhaust gas value is used directly as the controlled variable. In at least one embodiment of the method, the measured exhaust gas value is the only variable of the analysis function. In an alternative embodiment of the method, the analysis function comprises at least one further variable. Optionally, the analysis function comprises a function parameter for adjusting the analysis function or for selecting a specific analysis function from a family of analysis functions.An example of a functional parameter is a temperature of a measuring device that records the exhaust gas measurement, a temperature of the exhaust gas, a pressure of the exhaust gas, or the like. The functional parameter can be set to a constant value, in particular a value recorded at the beginning of the method, for the duration of the method, or can be updated during the course of the method. Each variable and / or each functional parameter of the analysis function preferably describes a measurement situation of the exhaust gas measurement, in particular at a measuring point of the exhaust gas measurement. The measurement situation can include, for example, a temperature of the measuring device, a thermodynamic state variable of the exhaust gas at the measuring point of the exhaust gas measurement, an original fuel composition of a fuel supplied to the fuel cell device and converted to form the exhaust gas, or the like.Preferably, the control unit evaluates the analysis function independently, in particular without direct use, of an operating point of the fuel cell device, in particular independently of an electrical power provided by the fuel cell device, independently of fuel utilization of the fuel cell device, independently of the fuel flow rate, independently of an oxygen-carbon ratio of the exhaust gas, or other operating parameters of the fuel cell device. A variable that is indirectly dependent on an operating point of the fuel cell device and is used as a variable or functional parameter of the analysis function, in particular a thermodynamic state variable of the exhaust gas, is preferably detected by means of a sensor, in particular a temperature sensor or pressure sensor, of the fuel cell device.

[0009] A fuel cell unit of the fuel cell device preferably converts the fuel by supplying oxygen to the exhaust gas. The control unit preferably adjusts the fuel flow rate as a function of the measured exhaust gas value such that the exhaust gas contains fuel residues that have not yet been converted. The measured exhaust gas value preferably describes a content or a deficit of at least one constituent of the exhaust gas, in particular relative to a stoichiometric combustion air ratio of the exhaust gas. Particularly preferably, the measured exhaust gas value describes an oxygen content, in particular an atomic oxygen deficit, or an oxygen excess, in particular an atomic oxygen excess, of the exhaust gas.

[0010] The fuel cell unit preferably comprises at least one fuel cell, preferably at least one stack of fuel cells, optionally a combination of several stacks of fuel cells. The at least one fuel cell preferably comprises at least one fuel electrode for direct contact with the fuel and at least one oxygen electrode for direct contact with an oxygen-containing fluid. A measuring point for the exhaust gas measurement value is preferably, in particular immediately, downstream of the fuel electrode. The fact that two objects are "immediately downstream / upstream" of each other with respect to a fluid should preferably be understood to mean that the composition of the fluid at both objects is the same. The objects can be in physical contact with each other or, for example, connected by a fluid line, arranged at a distance from each other.In particular, additional objects can be arranged along the flow direction between the objects, which leave the composition of the fluid unchanged, for example, a temperature sensor or the like. The exhaust gas measurement value can be recorded, for example, in an exhaust line of the fuel cell device from the fuel electrode to an afterburner of the fuel cell device or in a recirculation line of the fuel cell device branching off from the exhaust line. Alternatively, the exhaust gas measurement value is recorded downstream of the afterburner.

[0011] The inventive design allows a sensor system for controlling the fuel cell device to be kept advantageously simple. In particular, the number of sensor elements for controlling the fuel cell device can be advantageously kept low. Furthermore, a variable characterizing fuel gas depletion, for example, fuel utilization of the fuel cell unit, hydrogen content of the exhaust gas, oxygen-carbon ratio of the exhaust gas, or the like, can be determined, so that compliance with a respective limit value can advantageously be achieved even with fluctuating fuel composition and / or changed operating conditions, such as a change in the electrical current supplied by the fuel cell unit. Furthermore, a permissible value range of the variable characterizing fuel gas depletion for operation of the fuel cell device can advantageously be selected to be large.In particular, despite fluctuating fuel composition, a safety buffer separating the limit value from a value of this magnitude that is harmful to the fuel cell device can advantageously be kept small. Furthermore, operational management of the fuel cell device can advantageously be adapted to fluctuations in the fuel composition. In particular, the fuel cell device can advantageously be operated efficiently and / or with low wear. Furthermore, the risk and / or extent of damage to the fuel electrode in the event of a leak in the fuel cell device can advantageously be kept low.

[0012] It is further proposed that a lambda probe be used to detect the exhaust gas measurement value. The lambda probe is preferably a broadband lambda probe. Preferably, an electrical pumping current of a pump cell of the broadband lambda probe is used as the exhaust gas measurement value. Alternatively, the lambda probe is a step-change lambda probe, preferably a step-change lambda probe with internal temperature control. Preferably, an electrical Nernst voltage of a Nernst cell of the step-change lambda probe is used as the exhaust gas measurement value. Optionally, the Nernst voltage or the pumping current is output by the lambda probe as the combustion air ratio or as the atomic oxygen content or the like.As an alternative to a lambda probe, a measuring unit is used which comprises at least one fuel cell operated as a Nernst cell and which optionally comprises at least one further fuel cell upstream of this fuel cell, which is operated as a pump cell and which pumps oxygen into or out of the exhaust gas depending on a measured value from the fuel cell operated as a Nernst cell. The inventive design allows the measured exhaust gas value to be advantageously recorded simply and cost-effectively.

[0013] It is further proposed that the analysis function outputs a hydrogen content of the exhaust gas as a function of the measured exhaust gas value. The hydrogen content is preferably output by the analysis function as a molar fraction of hydrogen molecules in the exhaust gas. Alternatively, the analysis function outputs the hydrogen content as a concentration or ratio and / or in relation to a volume or mass of the exhaust gas. Preferably, the controlled variable is equal to the hydrogen content. Preferably, a setpoint value of the controlled variable is a lower limit value for the hydrogen content in the exhaust gas. Preferably, the control unit sets the fuel flow rate and / or the electrical current provided by the fuel cell unit or the electrical voltage applied to the fuel cell unit such that the hydrogen content of the exhaust gas detected by means of the measured exhaust gas value is equal to or greater than the setpoint value, subject to a tolerance.Thanks to the inventive design, the fuel cell device can advantageously be operated safely and reliably despite dispensing with direct control of fuel usage. This advantageously eliminates the need for cost-intensive sensor elements such as flow meters and / or volumetric flow meters. Furthermore, an advantageously high level of accuracy of the determined controlled variable, i.e., the hydrogen content, can be achieved. Furthermore, the hydrogen content can advantageously be determined independently of other measured values ​​and / or estimates, such as a fuel temperature, a fuel composition, in particular a fuel-exhaust gas mixture composition, a hydrogen-carbon ratio of the fuel, or the like.

[0014] It is further proposed that the analysis function outputs an electron gas coefficient of the exhaust gas as a function of the measured exhaust gas value. The electron gas coefficient of a fluid preferably indicates the number of electrons per mole of the fluid that can be bound by oxygen. Preferably, the controlled variable is equal to the electron gas coefficient. Preferably, a setpoint value of the controlled variable is a lower limit value for the electron gas coefficient in the exhaust gas. Preferably, the control unit sets the fuel flow rate and / or the electrical current provided by the fuel cell unit or the electrical voltage applied to the fuel cell unit such that the electron gas coefficient of the exhaust gas detected by means of the measured exhaust gas value is equal to or greater than the setpoint value, up to a tolerance. In an alternative embodiment, the analysis function is equal to the mathematical identity, so that the controlled variable, i.e.The actual value of the controlled variable is identical to the measured exhaust gas value. The inventive design allows the fuel cell device to be operated safely and reliably despite dispensing with direct control of fuel usage. This advantageously eliminates the need for cost-intensive sensor elements such as flow meters and / or volumetric flow meters. Furthermore, many components of the fuel or exhaust gas can be advantageously taken into account when determining the actual value.

[0015] It is further proposed that the exhaust gas measurement value and / or a result of the analysis function be corrected using a machine learning process. The machine learning process is used, for example, to correct an error E between a real value h ziei and a value h recorded by the lambda sensor or determined by the analysis function s, ie the exhaust gas measured value and / or the actual value of the controlled variable, and thus to improve the latter values, in particular in the form h ziei = h s+ E. The machine learning process is preferably trained in advance of the method with training data for estimating the error E at various operating points of the fuel cell device. The machine learning process can be set up as a function of at least one measured variable of the lambda probe, such as the pumping current, an electrical pumping voltage of the pumping cell, an exhaust gas temperature of the exhaust gas and / or the Nernst voltage, and / or as a function of at least one operating parameter of the fuel cell device, such as a component temperature, a temperature, a pressure, a flow rate of the exhaust gas, the fuel, or the like. The error E preferably represents an output variable to which the machine learning process is trained. The machine learning process is designed, for example, as a multivariate linear regression, as a neural network, and / or as a Gaussian process.Due to the design according to the invention, the actual value of the controlled variable and optionally the exhaust gas measured value can be determined advantageously precisely and advantageously reliably.

[0016] It is further proposed that the analysis function or a further analysis function output a system-wide fuel utilization of the fuel cell device as a function of the exhaust gas measurement value. The analysis function and the further analysis function can be evaluated using the same exhaust gas measurement value or using exhaust gas values ​​from different measuring devices, in particular different lambda sensors. During operation of the fuel cell device, the fuel cell unit has a local fuel utilization which in particular compares a composition of the exhaust gas with a composition of the fuel entering the fuel cell unit, in particular a fuel-exhaust gas mixture. The system-wide fuel utilization preferably compares a composition of fresh fuel, upstream of a refeed of exhaust gas, with a composition of the exhaust gas.The system-wide fuel utilization is in particular greater than or, in particular if no exhaust gas is fed back, equal to the local fuel utilization. Preferably, a setpoint of the controlled variable is an upper limit for the system-wide fuel utilization. Preferably, the control unit sets the fuel flow rate and / or the electrical current provided by the fuel cell unit or the electrical voltage applied to the fuel cell unit such that the system-wide fuel utilization of the exhaust gas detected by means of the measured exhaust gas value is equal to or less than the setpoint, up to a tolerance. The configuration according to the invention allows for advantageously simple implementation of the control.

[0017] It is further proposed that, in at least one method step of the method, a dependency of at least one control variable on the controlled variable is adapted, in particular iteratively. The control unit preferably uses the control variable to adjust the fuel flow rate, for example by means of a fluid delivery unit, in particular a pump, a compressor, a blower, a fan, or the like, or by means of a shut-off device, in particular by means of a valve, a throttle valve, or the like. The control variable can, in particular, be a flow rate of the fuel and / or the exhaust gas, a rotational speed of the fluid delivery unit, a position of the shut-off device, or the like.Preferably, at least one calculation instruction is stored in a memory of the control unit, by means of which calculation instruction the control unit determines the manipulated variable as a function of the controlled variable, in particular as a function of the actual value and / or the setpoint value of the controlled variable. Preferably, in at least one method step of the method, the control unit modifies the calculation instruction by updating a parameter of the calculation instruction in order to change the dependence of the manipulated variable on the controlled variable. For example, the control unit determines at least one fuel parameter of the fuel and / or the exhaust gas, which characterizes a quality of the fuel converted into the exhaust gas, in order to modify the calculation instruction. Preferably, the control unit determines the fuel parameter as a function of the actual value and / or the setpoint value of the controlled variable.Due to the design according to the invention, the control can be carried out with advantageously few preliminary estimates and / or preliminary measurements of the fuel with respect to an advantageously wide range of fuel compositions, advantageously precisely and in particular with advantageously small safety buffers.

[0018] Furthermore, it is proposed that the analysis function comprise at least one fuel parameter, in particular the one already mentioned, which characterizes a quality of a fuel converted to the exhaust gas. The fuel parameter is preferably designed as a hydrogen-carbon ratio of the, in particular fresh, fuel. Alternatively, the fuel parameter is designed as a calorific value, a gross calorific value, a Wobbe index, a hydrogen content, a hydrocarbon content, or the like of the, in particular fresh, fuel. The fuel parameter is preferably a functional parameter of the analysis function. The analysis function can comprise the fuel parameter as a factor of the measured exhaust gas value and / or as a summand independent of the measured exhaust gas value.Preferably, at least one form of the analysis function is stored in the memory of the control unit, which the control unit uses when the fuel parameter is set to zero, for example when the fuel parameter has not been determined, in particular has not yet been determined. Preferably, at least one further form of the analysis function is stored in the memory of the control unit, which the control unit uses when the fuel parameter is taken into account when determining the actual value of the controlled variable, for example when the fuel parameter was determined in the course of the control, when the fuel parameter is specified by an operator or when a standard parameter of the fuel parameter is stored in the memory of the control unit. Optionally, the control unit determines at least one further fuel parameter, in particular for adapting the dependence of the at least one manipulated variable of the control on the controlled variable.The additional fuel parameter is, for example, a thermal capacity, a molar mass, a carbon content, a hydrogen content, an oxygen content, an electron gas coefficient of the fuel, or the like. The inventive design allows the analysis function to be specifically adapted to the determined fuel parameter. In particular, an advantageously high accuracy of the analysis function can be achieved.

[0019] Furthermore, it is proposed that at least one, in particular the one already mentioned, fuel parameter of a fuel converted into the exhaust gas is determined iteratively during the control process. The fuel parameter is preferably determined several times during the process, particularly preferably during each run of the control loop. The fuel parameter is preferably determined in order to more precisely determine the actual value of the controlled variable and / or the manipulated variable. Particularly preferably, the fuel parameter is determined as a function of the most recently determined actual value of the controlled variable in order to more precisely determine the next actual value of the controlled variable and / or the manipulated variable dependent on the most recently determined actual value.The design according to the invention provides additional information about the fuel, which can be used to improve the control and / or to adjust other components of the fuel cell device.

[0020] It is further proposed that the analysis function be corrected as a function of an exhaust gas temperature of the exhaust gas. Preferably, a temperature sensor of the fuel cell device detects the exhaust gas temperature. Preferably, the analysis function has the exhaust gas temperature as a functional parameter. The exhaust gas temperature can be included in the analysis function as a factor of the exhaust gas measured value and / or as an additional summand independent of the exhaust gas measured value. Preferably, the exhaust gas temperature is detected at least once at the start of the method and / or after completion of an operating point change in order to adjust the analysis function or to select one of several analysis functions. The exhaust gas temperature can be updated with each run of the control loop, after a defined number of run-throughs of the control loop, at a fixed time interval and / or triggered by a temperature change in the detected exhaust gas temperature.Preferably, the exhaust gas temperature is measured on or in the measuring device, in particular the lambda sensor, of the exhaust gas measured value. Alternatively, a temperature measured at the fuel cell unit is used as the exhaust gas temperature. The inventive design allows for an advantageously high degree of accuracy of the actual value of the controlled variable. In particular, a state of the exhaust gas can be taken into account when determining the exhaust gas measured value during an evaluation of the exhaust gas measured value.

[0021] It is further proposed that a sensor unit, in particular the aforementioned measuring device, of the fuel cell device is temperature-controlled for detecting the exhaust gas measurement value. Preferably, the measuring device for the exhaust gas measurement value, in particular the lambda probe, comprises an internal electrical heating element for temperature-controlling the exhaust gas. Alternatively, the exhaust gas is temperature-controlled by an electrical heating element of the fuel cell device that is formed separately from the measuring device for the exhaust gas measurement value, in particular the lambda probe, in particular upstream of the measuring device with respect to the exhaust gas. Preferably, the exhaust gas temperature is controlled to a constant value by the measuring device or the control unit. The control unit can use a temperature setpoint or an actual temperature value of the exhaust gas temperature to evaluate the analysis function. Due to the embodiment according to the invention, the exhaust gas measurement value can advantageously be detected independently of temperature.

[0022] Furthermore, it is proposed that at least one further exhaust gas measurement value is recorded in order to determine a value to be set for the fuel flow rate. The further exhaust gas measurement value is preferably recorded independently of the exhaust gas measurement value. The further exhaust gas measurement value is preferably recorded by a further measuring device, in particular by a further lambda probe. Preferably, an electrical pump current of a pump cell of the further lambda probe, an electrical Nernst voltage of a Nernst cell of the further lambda probe, a combustion air ratio or an atomic oxygen content or the like is output by the further measuring device as an exhaust gas measurement value. Preferably, the further exhaust gas measurement value is recorded at a measuring point at which a composition of the exhaust gas is different from the measuring point of the exhaust gas measurement value.The control unit preferably uses the further exhaust gas measurement value to determine a fuel parameter, in particular the one already mentioned, the further one already mentioned, or an additional fuel parameter. The control unit preferably uses the further exhaust gas measurement value to determine the electron gas coefficient and / or an electron flow linked to the electron gas coefficient of electrons available for oxidation of the fuel, in particular fresh fuel. The control unit preferably compares the exhaust gas measurement value and the further exhaust gas measurement value to determine the fuel parameter. The embodiment according to the invention advantageously allows different variables to be recorded independently of one another for control purposes. In particular, consistency of the exhaust gas measurement values ​​can advantageously be checked.

[0023] It is further proposed that the further exhaust gas measurement value be recorded downstream of an afterburner of the fuel cell device. The afterburner converts the fuel residues contained in the exhaust gas, preferably with the addition of oxygen, at least substantially completely, in particular to more than 90%, preferably to more than 95%, particularly preferably to more than 99%. The afterburner particularly preferably uses a low-oxygen fluid as the oxygen source, which fluid exits at the oxygen electrode. The low-oxygen fluid is formed from the oxygen-containing fluid by transferring oxygen to the fuel within the fuel cell unit. The term "low-oxygen" is preferably an abbreviation for "low in oxygen relative to the oxygen-containing fluid" or "having less oxygen than the oxygen-containing fluid." The afterburner preferably converts the exhaust gas and the low-oxygen fluid to form an afterburner exhaust gas.The additional exhaust gas measurement value is preferably recorded in the afterburner exhaust gas. Particularly preferably, the additional exhaust gas measurement value is recorded downstream of at least one heat exchanger of the fuel cell device, relative to the afterburner exhaust gas. The inventive design advantageously allows a simple recording of a molar flow, in particular an electron flow, of the fuel. Furthermore, advantageously high dynamic control response, in particular a control response time of less than one minute, can be achieved. Furthermore, the additional exhaust gas measurement value can be recorded outside a high-temperature zone of the fuel cell device, in particular under advantageously stable ambient conditions.

[0024] It is further proposed that the additional exhaust gas measurement be recorded by means of a lambda probe, in particular the one already mentioned. The additional lambda probe can be designed as a broadband lambda probe or as a step-type lambda probe. Alternatively, the additional exhaust gas measurement is recorded by one or more fuel cells, which are / are operated like a lambda probe. The inventive design allows the additional exhaust gas measurement to be recorded advantageously in a simple and cost-effective manner.

[0025] Furthermore, a fuel cell device is proposed with at least one control unit for carrying out a method according to the invention and with at least one sensor unit for detecting the exhaust gas measurement value. The sensor unit preferably comprises the measuring device for the exhaust gas measurement value, in particular the lambda probe. The sensor unit preferably comprises the measuring device for the further exhaust gas measurement value, in particular the further lambda probe. The sensor unit optionally comprises at least one temperature sensor for detecting the exhaust gas temperature. The sensor unit preferably comprises at least one ammeter, a voltmeter or the like for detecting an electrical current provided by the fuel cell unit, or a variable associated therewith, for example an electrical voltage, an electrical power or the like.The sensor unit optionally comprises at least one flow meter, for example for detecting the inflow rate of fresh fuel to the fuel cell unit or for detecting an inflow rate of oxygen-containing fluid to the fuel cell unit. A “control unit” is to be understood in particular as a unit with at least one control electronics unit. A “control electronics unit” is to be understood in particular as a unit with a processor unit and with a memory as well as with an operating program stored in the memory. Optionally, the control unit comprises at least one control element, in particular a proportional element, a differential element and / or an integral element for adjusting the manipulated variable. Alternatively, the control unit determines a setpoint value of the manipulated variable and transfers this to an internal control system of the control unit.

[0026] The fuel cell unit comprises at least one fuel cell, preferably a stack of, in particular identical, fuel cells, or a combination of several stacks of, in particular identical, fuel cells. The at least one fuel cell is preferably designed as a high-temperature fuel cell, in particular a solid oxide fuel cell or a molten carbonate fuel cell. Alternatively, the fuel cell is a phosphoric acid fuel cell, a direct methanol fuel cell, or a polymer electrolyte membrane fuel cell. The fuel cell device preferably comprises the recirculation line for feeding the exhaust gas back into the fuel. The fuel cell device preferably comprises, in particular as an actuating unit, at least one fuel feed unit or a fuel shut-off device for adjusting the inflow rate of fresh fuel.The fuel cell device preferably comprises, in particular as an adjusting unit or as a further adjusting unit, at least one recirculation conveying unit arranged in the recirculation line for adjusting the recirculation rate of the exhaust gas. The fuel cell device preferably comprises at least one oxygen conveying unit or an oxygen shut-off device for adjusting the inflow rate of oxygen-containing fluid. The fuel cell device preferably comprises the afterburner, which is arranged downstream of the fuel cell unit, in particular connected to an outlet of the fuel electrode and to an outlet of the oxygen electrode of the fuel cell unit, and is provided for converting fuel residues in the exhaust gas.Optionally, particularly when using a hydrocarbon-containing fuel, the fuel cell device comprises a desulfurizer and / or a reformer for processing the fuel before feeding the fuel into the fuel cell unit. Preferably, the fuel cell device comprises at least one heat exchanger for transferring heat from the afterburner exhaust gas to the oxygen-containing fluid and / or to the fuel. The inventive design makes it possible to provide a fuel cell device that can be controlled with advantageously little effort and / or advantageously precisely using an advantageously simple and cost-effective sensor unit.

[0027] The method according to the invention and / or the fuel cell device according to the invention are not intended to be limited to the application and embodiment described above. In particular, the method according to the invention and / or the fuel cell device according to the invention may comprise a number of individual elements, components, units, and method steps that differs from the number stated herein to fulfill a functionality described herein. Furthermore, in the value ranges specified in this disclosure, values ​​within the stated limits are also to be considered disclosed and can be used arbitrarily.

[0028] Drawings

[0029] Further advantages will become apparent from the following description of the drawings. The drawings illustrate six exemplary embodiments of the invention. The drawings, the description, and the claims contain numerous features in combination. Those skilled in the art will also expediently consider the features individually and combine them into useful further combinations.

[0030] They show:

[0031] Fig. 1 is a schematic representation of a fuel cell device according to the invention,

[0032] Fig. 2 is a schematic dot diagram illustrating a correlation between a hydrogen content of an exhaust gas of the fuel cell device according to the invention as a controlled variable and a pumping current of a lambda probe of the fuel cell device according to the invention as an exhaust gas measurement value,

[0033] Fig. 3 is a schematic point diagram in which function values ​​of an analysis function of the hydrogen content are compared with reference values ​​of the hydrogen content which were measured with a reference measuring device,

[0034] Fig. 4 is a schematic flow diagram of a method according to the invention, which in particular uses the hydrogen content or an electron gas coefficient as a control variable,

[0035] Fig. 5 is a schematic point diagram illustrating a correlation between the electron gas coefficient of the exhaust gas of the fuel cell device according to the invention as a controlled variable and the pumping current of the lambda probe as an exhaust gas measured value, Fig. 6 is a schematic point diagram in which function values ​​of an analysis function of the electron gas coefficient are compared with reference values ​​of the electron gas coefficient, which were recorded with a reference measuring device,

[0036] Fig. 7 is a schematic point diagram illustrating a correlation between a system-wide fuel utilization of the fuel cell device according to the invention as a controlled variable and the pumping current of the lambda probe as an exhaust gas measurement value,

[0037] Fig. 8 is a schematic scatter diagram in which function values ​​of an analysis function of the system-wide fuel utilization are compared with reference values ​​of the system-wide fuel utilization, which were recorded with a reference measuring device,

[0038] Fig. 9 is a schematic flow diagram of an alternative embodiment of a method according to the invention, which in particular uses the system-wide fuel utilization of the fuel cell device according to the invention as a control variable,

[0039] Fig. 10 is a schematic flow diagram showing an extension of the method shown in Figure 9 by a pre-control value of the system-wide fuel utilization,

[0040] Fig. 11 is a schematic flow diagram of a further alternative embodiment of a method according to the invention, which in particular uses the system-wide fuel utilization of the fuel cell device according to the invention as a control variable and iteratively determines at least one fuel parameter of the fuel,

[0041] Fig. 12 is a schematic point diagram in which function values ​​of the analysis function of the system-wide fuel utilization, which was refined by the fuel parameter determined by the method shown in Fig. 9, are compared with the reference values ​​of the system-wide fuel utilization, Fig. 13 is a schematic representation of a further embodiment of a fuel cell device according to the invention with a further lambda probe,

[0042] Fig. 14 is a schematic flow diagram of a further alternative embodiment of a method according to the invention, which uses the further lambda probe of the fuel cell device from Fig. 13,

[0043] Fig. 15 is a schematic flow diagram of an additional alternative embodiment of a method according to the invention with two situation-dependent control loops,

[0044] Fig. 16 is a schematic flow diagram of a hazard control in the course of the method according to the invention from Fig. 15 and

[0045] Fig. 17 is a schematic flow diagram of an operating point control in the course of the inventive method from Fig. 15.

[0046] Description of the embodiments

[0047] Figure 1 shows a fuel cell device 12a. The fuel cell device 12a preferably comprises at least one fuel cell unit 30a. The fuel cell unit 30a comprises at least one fuel cell, preferably a plurality of fuel cells, which are particularly preferably arranged in at least one stack. For the sake of clarity, the fuel cell unit 30a is functionally illustrated here as a single fuel cell. The fuel cell unit 30a, in particular each fuel cell of the fuel cell unit 30a, comprises at least one oxygen electrode 32a and at least one fuel electrode 34a. Particularly preferably, the at least one fuel cell, in particular all fuel cells of the fuel cell unit 30a, is designed as a solid oxide fuel cell (SOFC).

[0048] The fuel cell device 12a preferably comprises an oxygen delivery unit 36a, in particular a fan, a blower, or a compressor, for delivering an oxygen-containing fluid to the oxygen electrode 32a. The oxygen-containing fluid is particularly preferably ambient air, which is drawn in by the oxygen delivery unit 36a. Alternatively, the oxygen-containing fluid is an industrial gas with a defined oxygen content. The oxygen delivery unit 36a is arranged upstream of the oxygen electrode 32a with respect to the oxygen-containing fluid.

[0049] The fuel cell device 12a preferably comprises a fuel delivery unit 38a, in particular a fan, a blower, or a compressor, for delivering a fuel to the fuel electrode 34a. The fuel is preferably hydrogen and / or natural gas, alternatively at least one hydrocarbon as a pure substance or as a mixture and / or ammonia. The fuel delivery unit 38a is arranged upstream of the fuel electrode 34a with respect to the fuel. The fuel cell unit 30a is preferably provided for providing electrical energy by converting the fuel into an exhaust gas by supplying oxygen from the oxygen-containing fluid.

[0050] Optionally, the fuel cell device 12a comprises a desulfurizer 40a. The desulfurizer 40a is preferably arranged downstream of the fuel feed unit 38a and upstream of the fuel cell unit 30a with respect to the fuel. Optionally, the fuel cell device 12a comprises a reformer 42a for reforming the fuel. The reformer 42a is preferably arranged downstream of the fuel feed unit 38a with respect to the fuel, in particular downstream of the desulfurizer 40a, and upstream of the fuel cell unit 30a.

[0051] The fuel cell device 12a preferably comprises a recirculation line 44a and a recirculation conveying unit 46a, in particular a fan, a blower, or a compressor, arranged in or on the recirculation line 44a, for recirculating the exhaust gas exiting the fuel electrode 34a into the fuel upstream of the fuel electrode 34a. An inlet of the recirculation line 44a is preferably upstream of the reformer 42a and downstream of the desulfurizer 40a with respect to the fuel.

[0052] The fuel cell device 12a preferably comprises an afterburner 24a for converting fuel residues contained in the exhaust gas. The afterburner 24a is preferably arranged downstream of the fuel electrode 34a with respect to the fuel or the exhaust gas, and in particular downstream of a branch into the recirculation line 44a. The afterburner 24a is preferably arranged downstream of the oxygen electrode 32a with respect to the oxygen-containing fluid. The fuel cell device 12a preferably comprises an exhaust-to-oxygen heat exchanger 48a for transferring heat from an afterburner exhaust gas of the afterburner 24a exiting the afterburner 24a to the oxygen-containing fluid upstream of the fuel cell unit 30a. The exhaust gas oxygen heat exchanger 48a is preferably arranged downstream of the oxygen delivery unit 36a and upstream of the fuel cell unit 30a with respect to the oxygen-containing fluid.The fuel cell device 12a preferably comprises an exhaust-fuel heat exchanger 50a for transferring heat from the afterburner exhaust gas of the afterburner 24a exiting the afterburner 24a to the fuel upstream of the fuel cell unit 30a. The exhaust-fuel heat exchanger 50a is preferably arranged downstream of the desulfurizer 40a and upstream of the feed opening of the recirculation line 44a with respect to the fuel. The exhaust-fuel heat exchanger 50a is arranged here, for example, downstream of the exhaust-oxygen heat exchanger 48a with respect to the afterburner exhaust gas. Alternatively, the exhaust-oxygen heat exchanger 48a is arranged downstream of the exhaust-fuel heat exchanger 50a with respect to the afterburner exhaust gas.

[0053] The fuel cell device 12a comprises at least one control unit 28a. The control unit 28a is preferably provided for controlling the fuel feed unit 38a and / or the recirculation feed unit 46a in order to adjust a fuel flow rate of the fuel through the fuel electrode 34a, in particular at a predetermined value of an electrical current provided by the fuel cell unit 30a. Alternatively or additionally, the control unit 28a is provided for controlling the electrical current provided by the fuel cell unit 30a, hereinafter referred to as stack current l stk, to be set. The control unit 28a is provided for carrying out a method 10a, which is explained in more detail in the following Figures 2 to 6. The fuel cell device 12a comprises at least one sensor unit for detecting an exhaust gas measured value 14a (cf. Figure 4). The sensor unit preferably comprises at least one lambda probe 18a, preferably a broadband lambda probe. The lambda probe 18a preferably comprises a Nernst cell and a pump cell and a measuring chamber arranged between them. A measuring point of the lambda probe 18a is preferably arranged in the exhaust gas exiting the fuel electrode 34a in order to detect the exhaust gas measured value 14a. The measuring chamber of the lambda probe 18a preferably has a diffusion barrier in order to set an exchange rate of exhaust gas in the measuring chamber with exhaust gas outside the measuring chamber.The Nernst cell is preferably provided to detect a Nernst voltage that is established due to a difference in the oxygen content of the exhaust gas in the measuring chamber compared to the oxygen content of a reference gas, for example, ambient air and / or an industrial gas with a defined oxygen content. The pump cell is preferably provided to pump oxygen into or out of the measuring chamber in order to generate a stoichiometric combustion air ratio in the measuring chamber. The pump cell is preferably controlled as a function of the detected Nernst voltage in order to achieve the stoichiometric combustion air ratio. The lambda probe 18a is preferably provided to generate an electrical pump current I. Pumpof the pump cell, which is adjusted to achieve the stoichiometric combustion air ratio, as the exhaust gas measured value 14a. Alternatively, particularly when the lambda probe 18a is designed as a step-down lambda probe, i.e., without a pump cell, the lambda probe 18a is provided to output the Nernst voltage of the Nernst cell. The measuring point of the lambda probe 18a can be arranged between the fuel electrode 34a and the branch into the recirculation line 44a, between the branch into the recirculation line 44a and the afterburner 24a, or within the recirculation line 44a.

[0054] Figure 2 shows a schematic dot diagram illustrating a correlation between reference values ​​52a of a hydrogen content x H2 of the exhaust gas of the fuel cell device 12a and the electrical pumping current Ipump of the lambda probe 18a as exhaust gas measurement value 14a. The reference values ​​52a of the hydrogen content x H2can be measured, for example, on a test bench using a mass spectrometer or the like. The method 10a is preferably based on an analysis function of a controlled variable 16a, in particular the hydrogen content x H2 , depending on the exhaust gas measured value 14a, in particular the pump current I Pum p-

[0055] Figure 3 shows a schematic scatter diagram in which function values ​​of the analysis function of the hydrogen content x H2iregr the reference values ​​52a of the hydrogen content x H2 The analysis function is preferably used for pairs of values ​​of the hydrogen content x H2 and the pump current Ip U mp at different exhaust gas temperatures T typical for the fuel cell device 12a during operation ab of the exhaust gas. The value pairs are preferably determined at exhaust gas temperatures T abin a range of 400° to 1000°C, preferably from 500° to 800°C, particularly preferably from 590°C to 630°C. The analysis function is preferably used for pairs of values ​​of the hydrogen content x H2 and the pump current I Pump determined with different fuel compositions. The value pairs are preferably recorded at a hydrogen-carbon ratio H / C of the fuel in a range of 3 to 5, preferably from 3.25 to 4.5, particularly preferably from 3.53 to 4.35. In its simplest form, the analysis function is a first-order polynomial in the exhaust gas measured value 14a, here, for example, the pump current I Pump- Optionally, the analysis function comprises second-order, third-order, or higher-order terms in the exhaust gas measured value 14a. A distribution of the value pairs around the analysis function preferably has a fluctuation range of less than 3%, preferably less than 1%, particularly preferably 0.5% or less. In particular, the coefficient of determination (R 2 ) of the correlation function is greater than 0.99, preferably greater than 0.999, particularly preferably greater than 0.9995, most preferably at least 0.9997. Optionally, the analysis function comprises at least one correction term in the exhaust gas temperature T ab and / or in a fuel parameter that characterizes the quality of a fuel converted into the exhaust gas, wherein the above-specified coefficient of determination is preferably achieved without such correction terms. The fuel parameter is, for example, the hydrogen-carbon ratio H / C or a Wobbe index of the fuel.

[0056] Figure 4 shows a flowchart of the method 10a for controlling the fuel cell device 12a. In at least one method step of the method 10a, a fuel flow rate of the fuel through the fuel cell device 12a, in particular through the fuel cell unit 30a, is set as a function of the controlled variable 16a dependent on the exhaust gas measured value 14a.

[0057] After a start 54a of the method 10a, in at least one method step of the method 10a, the at least one exhaust gas measured value 14a of a composition of the exhaust gas of the fuel cell device 12a is recorded. The lambda probe 18a is used to record the exhaust gas measured value 14a. Preferably, the pump current I Pumpas exhaust gas measurement value 14a from the lambda probe 18a to the control unit 28a. The lambda probe 18a is temperature-controlled to detect the exhaust gas measurement value 14a. Preferably, the lambda probe 18a comprises an electrical heating element for temperature-controlling the exhaust gas in the measuring chamber. Alternatively, the fuel cell device 12a comprises a heating element formed separately from the lambda probe 18a for temperature-controlling the exhaust gas. Preferably, the exhaust gas temperature T ab , especially in a tempered state of the exhaust gas, is detected by means of a temperature sensor. Alternatively, a temperature target value to be achieved by tempering is defined as the exhaust gas temperature T ab Alternatively, a stack temperature of the fuel cell unit 30a is used as the exhaust gas temperature T ab used.

[0058] The method 10a preferably comprises an actual value determination step 56a. In the actual value determination step 56a, an actual value of the controlled variable 16a is determined. The actual value is determined by evaluating the analysis function. The analysis function directly correlates the exhaust gas measured value 14a with the controlled variable 16a. The analysis function outputs a hydrogen content x H2 of the exhaust gas depending on the exhaust gas measured value 14a. The analysis function is dependent on an exhaust gas temperature T ab of the exhaust gas. The exhaust gas measurement value 14a and / or a result of the analysis function is optionally corrected using a machine learning process.

[0059] The control unit 28a preferably determines a control difference 20a from a predetermined setpoint 58a and the actual value determined by the analysis function. The setpoint 58a is preferably a lower limit for the hydrogen content x H2. The regulation is intended in particular to limit the hydrogen content x H2 of the exhaust gas at or above the limit value. Optionally, the control unit 28a selects the setpoint 58a depending on an operating point of the fuel cell device 12a and / or modifies the setpoint 58a by an operating point-dependent risk factor.

[0060] The method 10a preferably comprises an adjustment step 60a. The adjustment step 60a is preferably carried out when the control difference 20a is greater than a predetermined tolerance. The tolerance is preferably predetermined as a function of a measurement accuracy of the lambda probe 18a and / or an accuracy of the analysis function. In the adjustment step 60a, the control unit 28a preferably adjusts the fuel flow rate of the fuel. In the adjustment step 60a, the control unit 28a preferably adjusts an inflow rate of fresh fuel by means of the fuel feed unit 38a or a valve of the fuel cell device 12a arranged downstream of the fuel feed unit 38a. Optionally, the control unit 28a uses a pilot control value to adjust the fuel flow rate, to which the change in the fuel flow rate is added. For example, the control unit 28a determines the pilot control value as a function of the stack current l stk, the stack temperature and / or a moisture content of the fuel, in particular according to a method as known from DE 10 2011 088120 A1. The control optionally additionally provides a recirculation rate of the fuel cell device 12a by means of the recirculation conveying unit 46a and / or the stack flow l provided by the fuel cell unit 30a. stk a.

[0061] If the control difference 20a is smaller than the specified tolerance, the control unit 28a preferably concludes that a system-wide fuel use FU systhe fuel cell device 12a has assumed a fuel usage target value. If the control difference 20a is smaller than the specified tolerance, the control unit 28a preferably concludes that the fuel cell device 12a is in steady-state operation 62a. In particular, an end 64a of a run of the method 10a is reached when the fuel cell device 12a is in steady-state operation 62a.

[0062] Figures 5 and 6 illustrate that the method 10a can be modified by selecting an alternative controlled variable 16a', in particular without any other changes. Figure 5 shows a schematic dot diagram illustrating a correlation between reference values ​​52a' of an electron gas coefficient K e - the exhaust gas of the fuel cell device 12a and the electrical pumping current I Pumpof the lambda probe 18a as exhaust gas measured value 14a. The reference values ​​52a' of the electron gas coefficient K e - can be recorded, for example, on a test bench or the like. Method 10a is alternatively or additionally based on an analysis function of the electron gas coefficient K e - as an alternative control variable 16a' depending on the exhaust gas measured value 14a, in particular on the pump current I Pum p- The electron gas coefficient K e - is in particular a number of electrons per mole of fuel which are available for binding with oxygen during a conversion in the fuel cell unit 30a.

[0063] Figure 6 shows a schematic dot diagram in which function values ​​of the analysis function of the electron gas coefficient K e - the reference values ​​52a' of the electron gas coefficient K e- be compared. A distribution of the reference values ​​52a' around the alternative analysis function preferably has a fluctuation range of less than 5%, preferably less than 4%, particularly preferably of at most 3.5%. A setpoint for forming the control difference 20a preferably specifies a lower limit of the electron gas coefficient K e - before.

[0064] Figures 7 to 17 show further embodiments of the method according to the invention. The following descriptions and the drawings are essentially limited to the differences between the embodiments, whereby with regard to components, method steps or quantities with the same designation, in particular with regard to components, method steps or quantities with the same reference symbols, reference can in principle also be made to the drawings and / or the description of the other embodiments, in particular Figures 1 to 6. To distinguish the embodiments, the letter a is placed after the reference symbols of the embodiment in Figures 2 to 6. In the embodiments in Figures 7 to 17, the letter a is replaced by the letters b to f.The methods illustrated in Figures 7 to 12 and Figures 15 to 17 can be carried out with a fuel cell device as shown in Figure 1 without further modifications. In Figure 1, the reference symbols are therefore followed by the letters a to d and f.

[0065] Figure 7 shows a schematic scatter diagram illustrating a correlation between a system-wide fuel utilization FU sys a fuel cell device 12b as a controlled variable 16b of a method 10b and a pump current I pump a lambda probe 18b of a fuel cell device 12b as exhaust gas measurement value 14b of the method 10b. For details of the fuel cell device 12b, reference is made to Figure 1. The method 10b is preferably based on an analysis function of the controlled variable 16b, ie the system-wide fuel utilization FU sys , depending on the exhaust gas measured value 14b, in particular the pump current I Pump- The analysis function of the system-wide fuel utilization FU sys Depending on the exhaust gas measured value 14b, it is referred to as the FU analysis function for differentiation.

[0066] Figure 8 shows a schematic dot diagram in which function values ​​FU sys regr the FU analysis function of the system-wide fuel utilization FU sys Reference values ​​52b of system-wide fuel utilization FU sys The FU analysis function is preferably a first-order polynomial, particularly preferably second-order, in the exhaust gas measured value 14b, here for example the pump current I Pum p- Optionally, the FU analysis function comprises third-order or higher-order terms in the exhaust gas measured value 14b. Preferably, the FU analysis function comprises at least one, in particular linear, correction term in an exhaust gas temperature T aban exhaust gas of the fuel cell device 12b. A distribution of the value pairs around the FU analysis function preferably has a fluctuation range of less than 3%, preferably less than 2%, particularly preferably 1% or less. Optionally, the FU analysis function comprises at least one fuel parameter that characterizes a quality of a fuel converted into the exhaust gas. The fuel parameter is, for example, a hydrogen-carbon

[0067] H / C ratio or a Wobbe index of the fuel. For example, a control unit 28b of the fuel cell device 12b uses as FU-

[0068] Analysis function following calculation rule: where c1 to a8 are analysis coefficients to be determined, for example by regression, which are stored in a memory of the control unit 28b. If the fuel parameter of the fuel is not used, the analysis coefficients a4, a6, and a7 are set equal to zero before determining the remaining analysis coefficients. The fuel parameter or a characteristic of the fuel, via which the control unit 28b can retrieve the fuel parameter from an internal or external database, is preferably specified by a user before executing the method 10b. Alternatively, the fuel parameter is already integrated into the analysis coefficients a4, a6, and a7, or after combining the terms with the same dependencies into the analysis coefficients a4, a2, and a5.

[0069] Figure 9 shows a flowchart of the method 10b for controlling the fuel cell device 12b. In at least one method step of the method 10b, at least one exhaust gas measured value 14b, in particular the pump flow lpump> of a composition of the exhaust gas of the fuel cell device 12b, is recorded. Preferably, the exhaust gas temperature T ab In a FU setting step 74b of the method 10b, a fuel flow rate through the fuel cell device 12b is set as a function of the controlled variable 16b dependent on the exhaust gas measured value 14b. In the method 10b, the system-wide fuel utilization FU sys used as controlled variable 16b. In a FU actual value determination step 66b of the method 10b, an actual value of the controlled variable 16b is determined by evaluating the FU analysis function, which directly compares the exhaust gas measured value 14b with the controlled variable 16b, ie the system-wide fuel utilization FU sys , correlated.

[0070] The control unit 28b preferably forms a VFD control difference 68b from a VFD setpoint 70b and the actual value of the controlled variable 16b determined by means of the VFD analysis function. The VFD setpoint 70b is preferably an upper limit for the controlled variable 16b. The VFD setpoint 70b is preferably stored in a memory of the control unit 28b. If the VFD control difference 68b is greater than a predetermined tolerance, the VFD control step 74b is preferably executed. To execute the VFD control step 74b, the method 10b preferably comprises a control variable determination step 72b. In the control variable determination step 72b, the control unit 28b preferably determines a control variable as a function of at least one electrical performance parameter 76b of the fuel cell device 12b. Preferably, the electrical power parameter 76b is equal to an electrical stack current I provided by a fuel cell unit 30b of the fuel cell device 12b stkIn an advantageously simple embodiment, the control unit 28b uses exclusively the electrical power parameter 76b and the controlled variable 16b as variables to determine the manipulated variable. For example, the control unit 28b determines an electron flow of the fuel as a manipulated variable according to the following calculation rule: where N Ze u en is the number of fuel cells in the fuel cell unit 30b, and F is the Faraday constant. Particularly preferably, in the manipulated variable determination step 72b, the control unit 28b determines a change in the manipulated variable to be set as a function of the FI setpoint 70b and the actual value of the controlled variable 16b determined using the FI analysis function. For example, the control unit 28b uses the following expression to determine the change in the manipulated variable to be set: where FU sys ist the actual value of the controlled variable 16b and FU syssou denotes the FU setpoint 70b. In the FU adjustment step 74b, the control unit 28b preferably controls a fuel feed unit 38b of the fuel cell device 12b to adjust the fuel flow rate. Preferably, the control unit 28b transmits the change in the manipulated variable to be adjusted to the fuel feed unit 38b.

[0071] If the FI control difference 68b is smaller than the tolerance, the control unit 28b preferably sets a system flag 78b to true. If the FI control difference 68b is greater than the tolerance, the control unit 28b preferably sets the system flag 78b to false. Preferably, the control unit 28b performs a logical AND operation 82b of the system flag 78b with a recirculation flag 80b. In particular, the control unit 28b ends a run of the method 10b when the system flag 78a and the recirculation flag 80b are true. For example, the control unit 28b sets the recirculation flag 80b to true if a control difference 20b between another controlled variable and a setpoint 58b is smaller than a tolerance. The other controlled variable is, for example, a hydrogen content x H2 or an electron gas coefficient K e- of the exhaust gas. An actual value of the further controlled variable is preferably determined in an actual value determination step 56b of the method 10b, for example as described in the description of Figures 2 to 6. In particular, the controlled variable 16b and the further controlled variable can be determined as a function of the same exhaust gas measured value or as a function of exhaust gas measured values ​​from different measuring devices, in particular lambda sensors. If the control difference 20b is greater than a tolerance value, the control unit 28b preferably adjusts a recirculation rate of the fuel cell device 12b in a setting step 60b of the method 10b. The control of the controlled variable 16b and the further controlled variable can be carried out in parallel, with an overlap in time, alternately, or sequentially. Preferably, the recirculation flag 80b is automatically kept at false as long as the system flag 78b is set to false.In particular, the control unit 28b executes at least one control loop of the further controlled variable after the system flag 78b is set to true.

[0072] With regard to further features of the method 10b and the fuel cell device 12b, reference is made to Figures 1 to 6 and their description.

[0073] Figure 10 shows a control loop of a method 10c for controlling a fuel cell device 12c. For details of the fuel cell device 12c, reference is made to Figure 1. In at least one method step of the method 10c, at least one exhaust gas measured value 14c of a composition of an exhaust gas of the fuel cell device 12c is recorded. In at least one method step of the method 10c, a fuel flow rate through the fuel cell device 12c is adjusted as a function of a controlled variable 16c dependent on the exhaust gas measured value 14c. The controlled variable 16c is preferably a system-wide fuel utilization FUsysof the fuel cell device 12c. In an actual value determination step 66c of the method 10c, an actual value of the controlled variable 16c is determined by evaluating an analysis function that directly correlates the exhaust gas measured value 14c with the controlled variable 16c. Preferably, the method 10c comprises a pilot control 86c that specifies a pilot control value of the fuel flow rate as a function of an actual value 70c of the controlled variable 16c. Preferably, the method 10c comprises a control phase 84c that outputs a correction of the fuel flow rate dependent on an actual value 68c of the actual control difference 68c. A control unit 28c of the fuel cell device 12c comprises, for example, a control proportional element, a control differential element, and / or a control integral element to execute the control phase 84c.Preferably, the fuel cell device 12c, in particular a fuel feed unit 38c of the fuel cell device 12c, is adjusted as a function of a sum of the pilot control value and the correction.

[0074] For further features of the method 10c and the fuel cell device 12c, reference is made to Figures 1 to 9 and their descriptions. In particular, the method 10c represents an extension of the method 10b described in Figure 9.

[0075] Figure 11 shows a flowchart of a method 10d for controlling a fuel cell device 12d. For details of the fuel cell device 12d, reference is made to Figure 1. In at least one method step of the method 10d, at least one exhaust gas measured value 14d, preferably a pump current I Pumpa lambda probe 18d of the fuel cell device 12d, a composition of an exhaust gas of the fuel cell device 12d is detected. In at least one setting step 60d, 74d of the method 10d, a fuel flow rate through the fuel cell device 12d is set as a function of a controlled variable 16d dependent on the exhaust gas measured value 14d. In the method 10d, a system-wide fuel utilization FU is preferably used as the controlled variable 16d. sys used. In at least one FU actual value determination step 66d of method 10d, an actual value of the controlled variable 16d is determined by evaluating an analysis function that directly correlates the exhaust gas measured value 14d with the controlled variable 16d. Preferably, the analysis function is independent of a fuel parameter of the fuel. Preferably, a control unit 28d of the fuel cell device 12b evaluates the following expression as an analysis function in the FU actual value determination step 66d: where a1, a2, a3, a5, a8 are analysis coefficients and T ab an exhaust gas temperature of the exhaust gas. Preferably, in a further FU actual value determination step 98d of method 10d, a further actual value of the controlled variable 16d is determined by evaluating a further analysis function that directly correlates the exhaust gas measured value 14d with the controlled variable 16d. Preferably, the further analysis function is dependent on the fuel parameter of the fuel. Preferably, the control unit 28d of the fuel cell device 12b evaluates the following expression as a further analysis function in the further FU actual value determination step 98d:

[0076] Where b to b8 are analysis coefficients and - denotes the fuel parameter as the hydrogen-carbon ratio of the fuel. Preferably, the analysis coefficients a lta2, a3, a5, a8 of the analysis function and the analysis coefficients b to b8 of the further analysis function are stored in a memory of the control unit 28d. The control unit 28d preferably selects the analysis function to determine the actual value of the controlled variable 16d if the fuel parameter is not available. The control unit 28d preferably selects the further analysis function to determine the actual value of the controlled variable 16d if the fuel parameter is available. Optionally, the control unit 28d queries the fuel parameter from an operator of the fuel cell device 12d or an external gas analyzer.

[0077] In at least one operating mode of the control unit 28d, the control unit 28d iteratively determines the fuel parameter of the fuel converted into the exhaust gas during the control process. A control loop of the method 10d preferably comprises one of the actual fuel converter value determination steps 66d, 98d, a determination of an actual fuel converter control difference 68d, and at least one of the setting steps 60d, 74d. For example, in a first run of the control loop of the method 10d, the control unit 28d uses the analysis function or the further analysis function and a standard value of the fuel parameter stored in a memory of the control unit 28d. The control unit 28d preferably determines an iteration value of the fuel parameter to determine a control variable of the control.Preferably, in a further run of the control loop of method 10d, in the further fuel controller actual value determination step 98d, the control unit 28d inserts the iteration value of the fuel parameter into the further analysis function in order to determine the actual value of the controlled variable 16d. In particular, by this and / or by determining at least one further fuel parameter, for example, a heat capacity of the fuel, during the control process, a dependency of the at least one control variable on the controlled variable 16d is adjusted, in particular iteratively.

[0078] If the FI control difference 68d from the determined actual value of the controlled variable 16d and a FI setpoint 70d is greater than a tolerance, the control unit 28d preferably carries out a series of manipulated variable determination steps described below in order to determine the manipulated variable.

[0079] Preferably, the control unit 28d uses an electrical power parameter 76d provided by a fuel cell unit 30d of the fuel cell device 12d as an input variable for determining the manipulated variable. The electrical power parameter 76d is preferably expressed as a stack current l stk , for example, from an inverter or an ammeter of the fuel cell device 12d. Preferably, the control unit 28d uses a fuel flow parameter 88d of the fuel as an input variable for determining the manipulated variable. The fuel flow parameter 88d is preferably expressed as a fuel volume flow V BS , for example, by a flow meter of the fuel cell device 12d.

[0080] Preferably, a local FU setpoint 92d is stored in a memory of the control unit 28d, which is a setpoint of a local fuel use FU stkof the fuel cell unit 30d. The FU setpoint 70d of the system-wide fuel utilization FU sysand / or the local FU setpoint 92d can be stored as a value or as a calculation rule, for example to determine the FU setpoint 70d and / or the local FU setpoint 92d depending on an operating point of the fuel cell device 12d. For example, the FU setpoint 70d is a value between 75% and 95%, preferably between 80% and 90%. The local FU setpoint 92d is preferably smaller than the FU setpoint 70d. The local FU setpoint 92d is preferably a value between 50% and 80%, preferably between 60% and 70%. A recirculation setpoint 94d is preferably stored in a memory of the control unit 28d, which prescribes a setpoint for a recirculation rate of the exhaust gas through a recirculation line 44d of the fuel cell device 12d.The recirculation setpoint 94d can be stored as a value or as a calculation rule, for example, to determine the recirculation setpoint 94d as a function of an operating point of the fuel cell device 12d, in particular as a function of the local FI setpoint 92d and / or the FI setpoint 70d. For example, the control unit 28d sets the recirculation setpoint 94d based on the following calculation rule: U. s tk,SOll U s yS'SOll r S0U = FU s ys,S0ll ■ (FU stk.soll ~ 1) where r so u the recirculation setpoint 94d, FU stk so u the local FI setpoint 92d and FI sys sou is the FU setpoint 70d. Preferably, a standard value of a fuel composition 90d is stored in a memory of the control unit 28d. The fuel composition 90d is, for example, equal to the fuel parameter; alternatively or additionally, the fuel composition 90d comprises a plurality of values ​​for describing the fuel, in particular a respective molar fraction of the components of the fuel. Preferably, at least one characteristic curve of a fuel feed unit 38d of the fuel cell device 12d and / or at least one characteristic curve 96d of a recirculation feed unit 46d of the fuel cell device 12d is stored in a memory of the control unit 28d. The characteristic curve 96d describes, in particular, a pressure-volume flow dependency of the recirculation feed unit 46d.

[0081] Preferably, the method lOd comprises an electron current determination step lOOd as a manipulated variable determination step, in which the control unit 28d determines an electron current n Ke _ of the fuel. Preferably, the control unit 28d determines the electron flow n Ke _ depending on the actual value of the controlled variable 16d and the power parameter 76d. For example, the control unit 28d determines the electron current n Ke _ based on the calculation rule: where N Ze u en is the number of fuel cells in the fuel cell unit 30d and F is the Faraday constant.

[0082] Preferably, the method 10d comprises an electron gas coefficient determination step 102d as a manipulated variable determination step, in which the control unit 28d determines the electron gas coefficient K e - of the fuel. For example, the control unit 28d determines the electron gas coefficient Ke - using the calculation rule: where n BS a fuel flow of the fuel. Optionally, the method 10d comprises a conversion step 104d in which the control unit 28d converts the fuel flow parameter 88d into the fuel flow n BS For example, the fuel flow parameter 88d is recorded as a volume flow, a mass flow or a particle number flow and is converted by the control unit 28d in the conversion step 104d into the fuel flow n by means of a molar volume of the fuel, a molar mass of the fuel or the Avogadro constant. BS converted. The control unit 28d determines the fuel flow n BS particularly depending on the fuel composition 90d. Alternatively, the fuel flow parameter 88d is directly defined as fuel flow n BS recorded.

[0083] Preferably, the method 10d comprises, as a manipulated variable determination step, a change coefficient determination step 106d, in which the control unit 28d determines a change coefficient K & of the molar mass flow as a function of the electron gas coefficient K e - of the fuel. The coefficient of change K & preferably describes a change in the molar flow of the fuel, which is caused by a reformer 42d of the fuel cell device 12d and by the fuel cell unit 30d. Preferably, a change analysis function, in particular a change regression function, is stored in the control unit 28d, by means of which the control unit 28d determines the change coefficient K & The change coefficient K &can be expressed as the sum of the products of the molar fractions and the change in the material flow of the fuel components, which can be recorded, for example, on a test bench to determine the change analysis function. The change analysis function is preferably expressed as a function of the electron gas coefficient K e - stored in the control unit 28d. The change analysis function is, for example, a first-order, second-order, or higher-order polynomial in the electron gas coefficient K e - . Alternatively, the change analysis function is a power function in the electron gas coefficient K e - , where an exponent of the electron gas coefficient K e - is determined by regression. Preferably, the coefficient of determination (R 2 ) of the change analysis function is greater than 0.9, preferably greater than 0.99, particularly preferably greater than 0.993, optionally greater than 0.995.

[0084] Preferably, the method 10d comprises, as a manipulated variable determination step, an oxygen-carbon hybrid coefficient determination step 108d, in which the control unit 28d determines a hybrid coefficient K <p des Brennstoffs ermittelt. Der Sauerstoff-Kohlenstoff-Hybridkoeffizient K<p fasst vorzugsweise einen Kohlenstoffanteil, einen Sauerstoffanteil und / oder ein Kohlenstoff-Sauerstoff-Verhältnis des Brennstoffs zusammen. Vorzugsweise ermittelt die Regeleinheit 28d den Sauerstoff-Kohlenstoff-Hybridkoeffizienten K<p anhand einer der folgenden Rechenvorschriften:

[0085] Preferably, the method lOd comprises an oxygen-carbon ratio determination step llOd as a manipulated variable determination step, in which the control unit 28d determines an oxygen-carbon ratio = K <p + K0) / K c of the fuel, where K o a molar fraction of oxygen atoms in the fuel and K ca molar proportion of carbon atoms in the fuel. Preferably, the control unit 28d determines the oxygen-carbon ratio using the calculation rule: where c to c4 are regression factors of a hybrid analysis function R^regr, which are stored in a memory of the control unit 28d. The hybrid analysis function K (f>i reg r preferably gives the oxygen-carbon hybrid coefficient as a, in particular linear, function of the oxygen-carbon ratio and the electron gas coefficient K e - Preferably, the hybrid analysis function K (f>i r egr has the following form:

[0086] R <t>,regr = C 1 + c 2 ' K e ~ + C3■ <5 + C4■ <5 ■ K e -

[0087] Preferably, the method 10d comprises, as a control variable determination step, a carbon atom content determination step 112d, in which the control unit 28d determines the molar fraction K c of carbon atoms in the fuel. Preferably, the control unit 28d determines the molar fraction K c of carbon atoms depending on the oxygen-carbon ratio and the change coefficient K & . In particular, the control unit 28d determines the molar fraction K c of carbon atoms according to the calculation rule:

[0088] KC ~ ^2~

[0089] Preferably, the method 10d comprises, as a manipulated variable determination step, an oxygen atom fraction determination step 114d, in which the control unit 28d determines the molar fraction K o of oxygen atoms in the fuel. Preferably, the control unit 28d determines the molar fraction K o of oxygen atoms depending on the change coefficient K & . In particular, if the change coefficient K & is determined at least by means of a quadratic polynomial or a power function as a coefficient analysis function, the control unit 28d determines the molar fraction K o of oxygen atoms according to the calculation rule:

[0090] K o = 2 ■ K c - K &

[0091] Alternatively, the control unit 28d determines the molar fraction K o of oxygen atoms depending on the oxygen-carbon ratio <5, for example according to:

[0092] Preferably, the method 10d comprises, as a manipulated variable determination step, a hydrogen atom fraction determination step 116d, in which the control unit 28d determines a molar fraction K H of hydrogen atoms in the fuel. Preferably, the control unit 28d determines the molar fraction K H of hydrogen atoms depending on the molar proportions K C ,K O of carbon atoms and oxygen atoms and the electron gas coefficient K e -, in particular in accordance with:

[0093] K H = K e - - 4K C + K o

[0094] Preferably, the method 10d comprises, as a control variable determination step, a fuel parameter determination step 118d, in which the control unit 28d determines a hydrogen-carbon ratio H / C as a fuel parameter. Preferably, the control unit 28d determines the fuel parameter by dividing the molar fraction K H of hydrogen by the molar fraction K c of carbon. Preferably, the method 10d comprises an inert gas determination step 120d as a manipulated variable determination step, in which the control unit 28d determines a molar inert gas content K N , in particular a molar nitrogen content, of the fuel. Preferably, the control unit 28d determines the inert gas content K N by means of a sum of the molar fractions of all components of the fuel, for example based on:

[0095] Preferably, the method 10d comprises a fuel mass determination step 122d as a manipulated variable determination step, in which the control unit 28d determines a molar mass M BS of the fuel. Preferably, the control unit 28d determines the molar mass M BS of the fuel by summing the products of molar mass and molar fraction of atomic species in the fuel. The molar mass of the atomic species contained in the fuel is preferably stored in the memory of the control unit 28d. For example, the control unit 28d determines the molar mass M BS of the fuel using the following calculation rule:

[0096] M BS = M H ■ K H + M o ■ K o + M c ■ K c + M N ■ K N where M H the molar mass of hydrogen, M o the molar mass of oxygen, M c the molar mass of carbon and M N is the molar mass of nitrogen. Preferably, the control unit 28d updates a value of the molar mass M BS of the fuel contained in the fuel composition 90d and / or on which a variable of the fuel composition 90d depends, with the value of the molar mass M determined in the fuel mass determination step 122d BS of the fuel. Optionally, the control unit 28d determines at least one further fuel parameter of the fuel, in particular a molar heat capacity, preferably an isobaric molar heat capacity, of the fuel. The control unit 28d determines the further fuel parameter, for example, by means of correlation as a function of the molar mass M BS and / or the molar fractions of atomic species of the fuel. Preferably, the control unit 28d updates a value of the further fuel parameter contained in the fuel composition 90d and / or on which a variable of the fuel composition 90d depends, with the value of the further fuel parameter determined in the fuel mass determination step 122d.

[0097] Preferably, the method 10d comprises a fuel-exhaust gas mass determination step 124d as a manipulated variable determination step, in which the control unit 28d determines a molar mass M BAM of a fuel-exhaust gas mixture consisting of fresh fuel and the exhaust gas fed back via the recirculation line 44d. Preferably, the control unit 28d determines the molar mass M BAM of the fuel-exhaust gas mixture depending on the recirculation setpoint 94d, for example using the following calculation rule: where MQ2 is the molar mass of oxygen molecules and n 02itr is a material flow of oxygen molecules which pass within the fuel cell unit 30d from an oxygen electrode 32d to a fuel electrode 34d of the fuel cell unit 30d. The material flow n 02itr of oxygen molecules is determined by the control unit 28d preferably in an oxygen transfer determination step 126d of the method 10d as a function of the electrical power parameter 76d, for example using the calculation rule:

[0098] Optionally, the control unit 28d determines a molar heat capacity, preferably an isobaric molar heat capacity, of the exhaust gas, in particular immediately upstream of the recirculation conveying unit 46d. The control unit 28d preferably determines a molar mass of the exhaust gas in order to determine the molar heat capacity of the exhaust gas, for example by means of correlation as a function of the molar mass of the exhaust gas.

[0099] Preferably, the control unit 28d updates a value of the molar mass M B AM of the fuel-exhaust gas mixture, a value of the molar heat capacity of the exhaust gas and / or a value of the molar mass of the exhaust gas, on which the characteristic curve 96d of the recirculation conveying unit 46d depends, with the value of the molar mass M determined in the fuel-exhaust gas mass determination step 124d BAM of the fuel-exhaust gas mixture, the value of the molar heat capacity of the exhaust gas determined in the fuel-exhaust gas mass determination step 124d or the value of the molar mass of the exhaust gas determined in the fuel-exhaust gas mass determination step 124d.

[0100] In the FU setting step 74d, the control unit 28d preferably determines a setpoint n B s,soii the fuel flow rate, in particular in the form of a material flow. The control unit 28d preferably adjusts the fuel feed unit 38d as a function of the setpoint n BS , S oii of the fuel flow rate. The control unit 28d determines the setpoint of the fuel flow rate preferably as a function of the FU setpoint 70d, for example using the calculation rule:

[0101] In the setting step 60d, the control unit 28d preferably determines a setpoint n BAMiS oii a mixture flow rate, in particular in the form of a material flow, of the fuel-exhaust gas mixture through the fuel cell unit 30d, in particular when the fuel-exhaust gas mixture enters a reformer 42d of the fuel cell device 12d. The control unit 28d preferably sets the recirculation conveying unit 46d as a function of the setpoint n BAMiSO ii the mixture flow rate of the fuel-exhaust gas mixture. The control unit 28d determines the setpoint of the exhaust gas flow rate preferably as a function of the recirculation setpoint 94d (r soii ) and the setpoint n BS , S oii the fuel flow rate, for example using the calculation rule:

[0102] Optionally, the control unit 28d determines a recirculation flow rate, in particular in the form of a material flow, of the exhaust gas through the recirculation line 44d by means of a correlation to the fuel composition 90d and a composition of the fuel-exhaust gas mixture based on the characteristic curve 96d of the recirculation conveying unit 46d in order to determine the recirculation rate and to control the recirculation conveying unit 46d as a function thereof.

[0103] The control loop is preferably repeated until the FU control difference 68d is smaller than the tolerance. Figure 12 shows a schematic dot diagram in which function values ​​FU sys regr the further FU analysis function of the system-wide fuel usage FU, which depends on the fuel parameter sys Reference values ​​52d of the system-wide fuel utilization FU sys For a representation of the FU analysis function independent of the fuel parameter, reference is made to Figures 7 and 8. A distribution of value pairs around the further FU analysis function preferably has a fluctuation range of less than 2%, preferably less than 1%, particularly preferably 0.6% or less.

[0104] For further features of the fuel cell device 12d and the method 10d, reference is made to Figures 1 to 10 and their description. In particular, the method 10d is not limited to regulating system-wide fuel utilization. For example, instead of the system-wide fuel utilization as described above, the hydrogen content can be used as a controlled variable, in particular to adjust the recirculation rate, particularly advantageously quickly and directly.

[0105] Figure 13 shows a fuel cell device 12e. The fuel cell device 12e comprises at least one control unit 28e. The control unit 28e is provided for carrying out a method 10e, which is explained in more detail in the following Figure 14. The fuel cell device 12e comprises at least one sensor unit for detecting an exhaust gas measurement value 14e (see Figure 14). The sensor unit preferably comprises a further lambda probe 26e. A measuring point of the further lambda probe 26e is preferably arranged downstream of the afterburner 24e with respect to an afterburner exhaust gas of an afterburner 24e of the fuel cell device 12e. Preferably, the measuring point of the further lambda probe 26e is arranged downstream of an exhaust gas-fuel heat exchanger 50e and / or an exhaust gas-oxygen heat exchanger 48e of the fuel cell device 12e with respect to the afterburner exhaust gas of the afterburner 24e.

[0106] Figure 14 shows a flow chart of the method 10e for controlling a fuel cell device 12e. For details of the fuel cell device 12e, reference is made to Figure 1. In at least one method step of the method 10e, at least one exhaust gas measured value 14e, preferably a pump current I Pump a lambda probe 18e of the fuel cell device 12e, a composition of an exhaust gas of the fuel cell device 12e is detected. In at least one setting step 60e, 74e of the method 10e, a fuel flow rate through the fuel cell device 12e is set as a function of a controlled variable 16e dependent on the exhaust gas measured value 14e. In the method 10e, a system-wide fuel utilization FU is preferably used as the controlled variable 16e. sys used. In at least one FU actual value determination step 66e of method 10e, an actual value of the controlled variable 16e is determined by evaluating an analysis function which directly correlates the exhaust gas measured value 14e with the controlled variable 16e.

[0107] At least one further exhaust gas measured value 22e is recorded in order to determine a value to be set for the fuel flow rate. The further exhaust gas measured value 22e is recorded by means of the further lambda probe 26e. The further exhaust gas measured value 22e is recorded downstream of the afterburner 24e of the fuel cell device 12e. The further exhaust gas measured value 22e is preferably a pump current of a pump cell of the further lambda probe 26e. The method 10e preferably comprises an oxygen determination step 128e, in which the control unit 28e determines an oxygen content x 02 ab of molecular oxygen in the afterburner exhaust gas is determined depending on the further exhaust gas measured value 22e. Alternatively, the lambda sensor 26e outputs the oxygen content x 02 , ab immediately.

[0108] Preferably, the control unit 28e determines a setpoint n BSiSO ii the fuel flow rate and or the fuel flow rate, in particular as fuel flow n BS , depending on the oxygen content x 02 ab of the afterburner exhaust gas. Preferably, the control unit 28e sets the setpoint value ^■BS,SOII of the fuel flow rate in the setting step 74e according to the following calculation rule: where n SF an input stream of an oxygen-containing fluid to an oxygen electrode 32e of a fuel cell unit 30e of the fuel cell device 12e and x O2>ein is an input content of molecular oxygen in the oxygen-containing fluid and a molar change of the afterburner exhaust gas relative to a molar amount of the fuel and the oxygen-containing fluid. The input material flow n SF of the oxygen-containing fluid is preferably detected by a flow meter of the fuel cell device 12e. The input content %o2,em of molecular oxygen in the oxygen-containing fluid is preferably stored in a memory of the control unit 28e. Alternatively, the fuel cell device 12e comprises at least one oxygen sensor for detecting the input content x O2>ein . The change in mass can be expressed as an afterburner exhaust gas mass flow divided by the sum of the fuel flow rate and the input mass flow n SF of the oxygen-containing fluid. The change in the molar mass of the afterburner exhaust gas depends on a fuel composition 90e. Preferably, a starting value for the change in the molar mass of the afterburner exhaust gas, for example, = 1, is stored in a memory of the control unit 28e, in particular together with the fuel composition 90e. Preferably, the control unit 28e updates the change in the molar mass during the control process, in particular depending on a determined molar mass of the fuel.

[0109] Preferably, the control unit 28e determines the fuel flow n in a conversion step 104e BS depending on the oxygen content x 02 , ab of the afterburner exhaust gas, for example using the following calculation rule:

[0110] Alternatively or additionally, the fuel stream n BS , in particular in the form of a volume flow, is detected by a flow meter of the fuel cell device 12e. The exhaust gas measured value 14e, the further exhaust gas measured value 22e and / or a result of the analysis function is optionally corrected by means of a machine learning process, in particular as already described above.

[0111] For further features of the fuel cell device 12e, reference is made to Figure 1 and its description. For further features of the method 10e, reference is made to Figures 2 to 12 and their description.

[0112] Figure 15 shows a flow diagram of a method 10f for controlling a fuel cell device 12f. For details of the fuel cell device 12f, reference is made to Figure 1. In at least one method step of the method 10f, at least one exhaust gas measurement value 14f of a composition of an exhaust gas from the fuel cell device 12f is recorded, preferably in the form of a pumping current Ipump of a lambda probe 18f of the fuel cell device 12f. In at least one setting step 60f of the method 10f, a fuel flow rate through the fuel cell device 12f is set as a function of a controlled variable 16f dependent on the exhaust gas measurement value 14f. In at least one power setting step 160f of the method 10f, an electrical power parameter 76f of a fuel cell unit 30f of the fuel cell device 12f is preferably set as a function of the controlled variable 16f dependent on the exhaust gas measurement value 14f.Preferably, the electrical power parameter 76f is equal to an electrical stack current I provided by a fuel cell unit 30f of the fuel cell device 12f. stk . The power parameter 76f is set, for example, by an inverter or a controllable resistor of the fuel cell device 12f. As the controlled variable 16f in the method 1Of, a hydrogen content x H2 used. In at least one actual value determination step 56f of the method 1Of, an actual value of the controlled variable 16f is determined by evaluating an analysis function which directly correlates the exhaust gas measured value 14f with the controlled variable 16f.

[0113] The method 10f preferably comprises an operating point control 130f. The operating point control 130f is provided to adjust the actual value of the controlled variable 16f to a setpoint 58f of the controlled variable 16f. The operating point control 130f preferably sets the fuel flow rate, in particular an inflow rate of fresh fuel, by means of a fuel feed unit 38f of the fuel cell device 12f. Preferably, a control unit 28f of the fuel cell device 12f (cf. Fig. 1) determines a value to be set for the fuel flow rate by adding an operating point pre-control value 164f and an operating point adjustment value 166f, wherein the operating point adjustment value 166f is preferably smaller than the operating point pre-control value 164f, in particular less than 20%, preferably less than 10%, optionally less than 5% of the operating point pre-control value 164f. The operating point control 130f preferably determines the operating point adjustment.The method 1Of preferably comprises an operating point pre-control 132f for determining the operating point pre-control value 164f. The operating point pre-control value 164f is determined, for example, as described in the description of Figure 11, in particular by iterative approximation.In particular, the operating point feedforward control 132f comprises an electron gas coefficient determination step, a conversion step, a change coefficient determination step, a hybrid coefficient determination step, an oxygen-carbon ratio determination step, a carbon atom fraction determination step, an oxygen atom fraction determination step, a hydrogen atom fraction determination step, a fuel parameter determination step, an inert gas determination step, a fuel mass determination step, a fuel-exhaust gas mass determination step and / or an oxygen transfer determination step, as described in the description of Figure 11 and are not shown again here for the sake of clarity.Preferably, the control unit 28f determines an electron flow n in an electron flow determination step 100f of the method 1Of as a function of a system-wide fuel utilization of the fuel cell device 12f. Ke _ of the fuel, as explained, for example, in the description of Figure 11. The system-wide fuel utilization is determined by the control unit 28f, for example, by evaluating the exhaust gas measured value 14f with an additional analysis function, which correlates the exhaust gas measured value 14f with the system-wide fuel utilization, as explained, for example, in the description of Figure 11. The electron stream n Ke _ the control unit 28f preferably passes to the operating point pre-control 132f.

[0114] The method 10f comprises a hazard control 134f. The hazard control 134f is provided to counteract the reaching of a harmful operating condition of the fuel cell device 12f. Preferably, the hazard control 134f is provided to keep the actual value of the controlled variable 16f above a hydrogen limit value 136f, in particular independently of the setpoint value 58f of the operating point control 130f. The hazard control 134f preferably has a shorter response time than the operating point control 130f. The hazard control 134f sets the performance parameter 76f of the fuel cell device 12f.Preferably, a control unit 28f of the fuel cell device 12f determines a value to be set for the power parameter 76f by adding a power pre-control value 142f and a hazard adjustment value 162f, wherein the hazard adjustment value 162f is preferably smaller than the power pre-control value 142f, in particular less than 20%, preferably less than 10%, optionally less than 5% of the power pre-control value 142f. The method 10f preferably includes a power pre-control 168f for determining the pre-control value. The hazard control 134f preferably determines the hazard adjustment value 162f.

[0115] Figure 16 shows a flow chart of the hazard control 134f. The hazard control 134f preferably has the actual value of the controlled variable 16f, the hydrogen limit value 136f, and optionally the power pre-control value 142f as input values. The hazard control 134f preferably has the hazard adjustment value 162f as output value. The hazard control 134f preferably forms a hazard control difference 138f from the hydrogen limit value 136f and the actual value of the controlled variable 16f. The hazard control 134f preferably includes a control-related proportionality element 140f. The proportionality element 140f preferably applies a positive proportionality factor K to the hazard control difference 138f. P , which expresses a rate of change of the power parameter 76f as a function of the controlled variable 16f. The proportionality factor K P can be determined, for example, as the difference quotient or as the differential of the performance parameter 76f with respect to the controlled variable 16f. For example, the hazard control 134f uses the following proportionality factor: where K App u is an application-dependent factor that can be set, for example, by an operator of the fuel cell device 12f. In the simplest case, K App u = 1. The hazard control 134f preferably comprises a limiting filter 148f, in particular a dynamic saturation element, which with the proportionality factor K P hazard control difference 138f is filtered and / or limited before being output as hazard adjustment value 162f. Preferably, hazard control 134f uses a value of zero as the upper filter limit 146f of the limit filtering 148f, so that hazard adjustment value 162f can only be zero or negative. This ensures that hazard control 134f can only reduce or leave unchanged an oxygen transfer within fuel cell unit 30f, so that the hydrogen content x H2 of the exhaust gas is increased and, in particular, kept above the hydrogen limit value 136f. The power pre-control value 142f, which is subjected to a factor of -1 and / or an adaptation pre-control value ratio 144f of the hazard control 134f, is preferably used as the lower filter limit of the limiting filter 148f. The adaptation pre-control value ratio 144f specifies, in particular, the above-specified size ratio of the hazard adaptation value 162f to the power pre-control value 142f. The adaptation pre-control value ratio 144f is preferably stored in a memory of the control unit 28f by an operator, manufacturer, or installer of the fuel cell device 12a, in particular depending on the application. The hazard control 134f preferably has a control time of less than 15 seconds, particularly preferably less than 10 seconds, particularly preferably less than 5 seconds.

[0116] Figure 17 shows a flow chart of the operating point control 130f. The operating point control 130f preferably has the actual value of the controlled variable 16f and the setpoint 58f of the controlled variable 16f as input values. The operating point control 130f preferably has the operating point pre-control value 164f as input values. The operating point control 130f preferably has an operating point reference value 158f as input values. The operating point reference value 158f is preferably an initial value of the operating point adjustment value 166f and is determined, for example, in a previous run of the operating point control 130f or is stored as a predetermined standard value in a memory of the control or regulating device. The operating point reference value 158f is preferably equal to zero in a first run of the operating point control 130f. The operating point control 130f preferably has the operating point adjustment value 166f as output value.Optionally, the operating point adjustment value 166f is saved as the new operating point reference value 158f.

[0117] The operating point controller 130f preferably forms a control difference 20f from the setpoint 58f and the actual value of the controlled variable 16f. The operating point controller 130f preferably includes a control-related integral element 154f. Before being transferred to the integral element 154f, the operating point controller 130f preferably applies a negative integration coefficient 150f to the control difference 20f. The integral element 154f is preferably designed as a discrete integrator. In addition to the control difference 20f, the integral element 154f preferably processes the operating point reference value 158f. The operating point controller 130f preferably limits a maximum output value of the integral element 154f with the operating point pre-control value 164f, which is subjected to a further adjustment pre-control value ratio 152f.The further adjustment pre-control value ratio 152f specifically determines the above-specified ratio of the operating point adjustment value 166f to the operating point pre-control value 164f. Preferably, the operating point controller 130f limits a minimum output value of the integral element 154f with the negative of the operating point pre-control value 164f, which is subjected to the further adjustment pre-control value ratio 152f or an additional value different from the further adjustment pre-control value ratio 152f.

[0118] The operating point control 130f preferably comprises an operating point limitation filter 156f, in particular a dynamic saturation element, which filters and / or limits the output value of the integral element 154f before outputting it as a hazard adjustment value 162f. The operating point control 130f preferably uses the operating point pre-control value 164f as the upper filter limit of the operating point limitation filter 156f, to which the further adjustment pre-control value ratio 152f is applied. The negative of the operating point pre-control value 164f is preferably used as the lower filter limit of the operating point limitation filter 156f, to which the further adjustment pre-control value ratio 152f is applied or an additional value different from the further adjustment pre-control value ratio 152f.

[0119] For further features of the fuel cell device 12f, reference is made to Figure 1 and its description. For further features of the method 10f, reference is made to Figures 2 to 14 and their description.< / t>

Claims

Claims 1. A method for controlling a fuel cell device, wherein in at least one method step at least one exhaust gas measurement value (14a; 14b; 14c; 14d; 14e, 22e; 14f) of a composition of an exhaust gas of the fuel cell device is recorded in order to set a fuel flow rate through the fuel cell device and / or an electrical power parameter of the fuel cell device as a function of a controlled variable (16a; 16b; 16c; 16d; 16e; 16f) dependent on the exhaust gas measurement value (14a; 14b; 14c; 14d; 14e, 22e; 14f), characterized in that in at least one method step an actual value of the controlled variable (16a; 16b; 16c; 16d; 16e) is determined by evaluating an analysis function which evaluates the exhaust gas measurement value (14a; 14b; 14c; 14d; 14e, 22e; 14f) directly correlates with the controlled variable (16a; 16b; 16c; 16d; 16e; 16f).

2. Method according to claim 1, characterized in that a lambda probe (18a; 18b; 18c; 18d; 18e; 18f) is used to detect the exhaust gas measured value (14a; 14b; 14c; 14d; 14e, 22e; 14f).

3. Method according to claim 1 or 2, characterized in that the analysis function determines a hydrogen content of the exhaust gas as a function of the exhaust gas measured value (14a; 14b; 14c; 14f).

4. Method according to claim 1 or 2, characterized in that the analysis function comprises an electron gas coefficient (K e ~) of the exhaust gas as a function of the exhaust gas measured value (14a; 14b; 14c; 14f).

5. Method according to one of the preceding claims, characterized in that the exhaust gas measured value (14a; 14b; 14c; 14d; 14e, 22e; 14f) and / or a result of the analysis function is corrected by means of a machine learning process.

6. Method according to one of the preceding claims, characterized in that the analysis function or a further analysis function comprises a system-wide fuel utilization (FU sys ) of the fuel cell device as a function of the exhaust gas measured value (14b; 14c; 14d; 14e, 22e).

7. Method according to one of the preceding claims, characterized in that in at least one method step a dependency of at least one manipulated variable of the control on the controlled variable (16d; 16e; 16f) is adapted, in particular iteratively.

8. Method according to one of the preceding claims, characterized in that the analysis function comprises at least one fuel parameter which characterizes a quality of a fuel converted into the exhaust gas.

9. Method according to one of the preceding claims, characterized in that at least one fuel parameter of a fuel converted into the exhaust gas is determined iteratively in the course of the control.

10. Method according to one of the preceding claims, characterized in that the analysis function is dependent on an exhaust gas temperature (T ab ) of the exhaust gas is corrected.

11. Method according to one of the preceding claims, characterized in that a sensor unit of the fuel cell device is tempered to detect the exhaust gas measured value (14a; 14b; 14c; 14d; 14e, 22e; 14f).

12. Method according to one of the preceding claims, characterized in that at least one further exhaust gas measured value (22e) is recorded in order to determine a value of the fuel flow rate to be set.

13. The method according to claim 12, characterized in that the further exhaust gas measured value (22e) is detected downstream of an afterburner (24e) of the fuel cell device. Method according to claim 12 or 13, characterized in that the further exhaust gas measured value (22e) is detected by means of a lambda probe (26e). A fuel cell device with at least one control unit (28a; 28b; 28c; 28d; 28e; 28f) for carrying out a method according to one of the preceding claims and with at least one sensor unit for detecting the exhaust gas measured value (14a; 14b; 14c; 14d; 14e, 22e; 14f).

Images