Method for training a soft sensor, fuel cell system, computer program product, computer-readable medium and data carrier signal

Abstract

The present invention relates to a method for training a soft sensor for determining a target measurement variable of a fuel cell system (100), wherein the fuel cell system (100) comprises a first sensor (10) for recording a first measurement variable, a second sensor (20) for recording a second measurement variable, and a control apparatus (50), wherein the control apparatus (50) is designed to determine the target measurement variable using a model stored on the control apparatus (50) at least on the basis of the first measurement variable and the second measurement variable. The fuel cell system also comprises a target variable sensor (40) for determining the target measurement variable. The method comprises, as a step, operating (320) at least one system unit (110) of the fuel cell system (100), which influences the target variable, at at least a first operating point. The method also comprises, as a step, adapting (340) at least one model parameter of the model stored on the control apparatus (50) in such a manner that a discrepancy between the target variable determined by the soft sensor and the target variable determined by the target variable sensor (40) is at least reduced at the first operating point of the system unit (110).

Description

[0001] Method for teaching a soft sensor, fuel cell system, computer program product, computer-readable medium and data carrier signal

[0002] The present invention relates to a method for teaching a soft sensor, a fuel cell system, a computer program product, a computer-readable medium and a data carrier signal.

[0003] A soft sensor, also called a virtual sensor, calculates a target variable based on measured variables recorded by sensors that correlate with the target variable and a correlation model. The sensors used to record a respective measured variable may have a tolerance and / or at least one system unit of a system, e.g., a fluid delivery device of a fuel cell system, may have a tolerance, so the target variable calculated by the soft sensor may be subject to certain inaccuracies.

[0004] The object of the present invention is to at least partially remedy the disadvantages described above. In particular, the object of the present invention is to provide a method for training a soft sensor or a fuel cell system with such a soft sensor, in which a model stored on a control device (of the soft sensor for determining a target measured variable) is particularly advantageously matched, in particular adjusted, to at least respective sensors for acquiring measured variables for determining the target measured variable using the model and / or to a system unit.

[0005] The above object is achieved by a method having the features of claim 1 and a fuel cell system having the features of claim 5, as well as a computer program product, a computer-readable medium, and a data carrier signal having the features of claim 11, 12, and 13, respectively. Further features and details of the invention emerge from the subclaims, the description, and the drawings. Features and details described in connection with the method according to the invention naturally also apply in connection with the fuel cell system according to the invention and / or the computer program product and / or the computer-readable medium and / or the data carrier signal, and vice versa, so that with regard to the disclosure of the individual aspects of the invention, reference is or can always be made reciprocally.According to a first aspect, the present invention shows a method for training a soft sensor for determining a target measured variable of a fuel cell system, wherein the fuel cell system comprises a first sensor for detecting a first measured variable, a second sensor for detecting a second measured variable, and a control device, wherein the control device is designed to determine the target measured variable based on a model stored on the control device, in particular a mathematical model, at least as a function of the first measured variable and the second measured variable. Furthermore, the fuel cell system comprises a target variable sensor for determining, in particular for measuring, the target measured variable. Furthermore, the method comprises, as a step, operating at least one system unit of the fuel cell system that influences the target variable at at least one first operating point.Furthermore, the method comprises as a step an adaptation of at least one model parameter, in particular a value of a model parameter of a characteristic curve, of the model stored on the control device such that a deviation between the target variable determined by the soft sensor and the target variable determined by the target variable sensor is at least reduced, in particular eliminated or substantially eliminated, at least in the first operating point of the system unit.

[0006] The process steps described above and below can, where technically feasible, be carried out individually, together, simply, multiple times, in parallel and / or consecutively in any order.

[0007] In particular, the soft sensor is formed by at least the first sensor for detecting the first measured variable, the second sensor for detecting the second measured variable, and the control device configured to determine the target measured variable based on the model stored on the control device, at least as a function of the first measured variable and the second measured variable. It is conceivable that the soft sensor may have additional sensors for detecting additional measured variables, for example, a third sensor for detecting a third measured variable.

[0008] The adaptation of the at least one model parameter of the model stored on the control device in such a way that the deviation between the target variable determined by the soft sensor and the target variable determined by the target variable sensor is at least reduced at least in the first operating point of the system unit, is carried out in particular by eliminating the deviation or the error between the target variable determined by the soft sensor and the target variable determined by the target variable sensor by adapting or changing the at least one model parameter of the model, preferably by means of a controller, wherein a controller output of the controller corresponds to the correct or adapted model parameter for the first operating point of the system unit in the event of a (vanishing) deviation between the target variable determined by the soft sensor and the target variable determined by the target variable sensor.The adjusted model parameter can be stored or saved as a result in the control device as a point in an adjusted characteristic curve. Adjusting the model parameter can also be understood as adjusting a (respective) value of the model parameter.

[0009] The term “operating point” of at least one system unit of the fuel cell system can also be understood as an “operating state”, e.g. an operating state of a valve (as a system unit).

[0010] In particular, the target measured variable has a first target measured variable state at the first operating point of the system unit of the fuel cell system. If the system unit has multiple operating points, the target measured variable has a different target measured variable state at each of the multiple operating points of the system unit.

[0011] In particular, the method may comprise, as a further step, that the fuel cell system is transferred into a calibration mode for teaching the soft sensor.

[0012] By adapting the at least one model parameter, the model stored on a control device (of the soft sensor for determining a target measured variable) can be particularly advantageously adapted, in particular adjusted, at least to respective sensors for detecting measured variables for determining the target measured variable by means of the model and / or to a system unit.

[0013] In a method according to the invention, it can be advantageous that at least the at least one system unit of the fuel cell system influencing the target variable is operated at several, in particular many, operating points, wherein the at least one model parameter, in particular a respective value of the at least one model parameter, of the model stored on the control device is adapted such that a respective deviation in a respective operating point of the several, in particular many, operating points of the system unit between the target variable determined by the soft sensor and the target variable determined by the target variable sensor is at least reduced, in particular eliminated or substantially eliminated. By operating the system unit at several, in particular many, operating points, different target measured variable states of the target measured variable can be generated.In other words, the fuel cell system with the at least one system unit can be operated in different states, wherein the target measured variable has different target measured variable states in the different states of the fuel cell system. Thus, a respectively adjusted model parameter can be filed or stored in the control device as a respective point in an adjusted characteristic curve, and the soft sensor can be trained. After adjusting the at least one model parameter, the new or adjusted model parameter can be persisted on the control device, e.g., a control unit, or in the non-volatile memory so that it can be used from the next operation.

[0014] In a method according to the invention, it can be advantageous for the fuel cell system to be operated at least temporarily in a stationary state for training the soft sensor of the fuel cell system. In particular, for training the soft sensor, the fuel cell system is operated in a stationary state at a respective operating point of the plurality, in particular the many, operating points of the at least one system unit, in particular for at least one (respectively) specified period of time. Thus, the training of the soft sensor of the fuel cell system can be carried out particularly advantageously.

[0015] It can be advantageous in a method according to the invention that the soft sensor is taught at intervals of time, wherein for each teaching of the soft sensor the target variable sensor is arranged, in particular mounted, on the fuel cell system to determine the target measured variable, and wherein after each teaching of the soft sensor the target variable sensor is detached, in particular removed, from the fuel cell system. Thus, the target variable sensor can only be mounted on the fuel cell system when required. In particular, the soft sensor of the fuel cell system can be taught, for example, during a workshop visit to a vehicle with such a fuel cell system.Furthermore, to adjust an air mass flow soft sensor, an air mass flow sensor can be arranged, in particular mounted, at a closable opening of an intake pipe of a fluid path of the fuel cell system, or an intake pipe of the fuel cell system can be replaced or exchanged, for example, during a workshop visit, with a modified adjustment intake pipe with an already arranged air mass flow sensor, e.g., a hot-film air mass meter. After the air mass flow sensor has been taught, the air mass flow sensor or the modified adjustment intake pipe with the air mass flow sensor can be removed again, and the fuel cell system can be returned to its original state.

[0016] According to a second aspect, the present invention shows a fuel cell system, wherein the fuel cell system has a first sensor for detecting a first measured variable, a second sensor for detecting a second measured variable, and a control device, wherein the control device is at least designed to determine the target measured variable based on a model stored on the control device at least as a function of the first measured variable and the second measured variable.The fuel cell system further comprises a target variable sensor for determining the target measured variable and an adaptation unit for adapting at least one model parameter of the model stored on the control device such that at least one deviation in a first operating point of the system unit between the target variable determined by the soft sensor and the target variable determined by the target variable sensor is at least reduced, in particular eliminated or substantially eliminated. In particular, the fuel cell system is designed to carry out a method according to the invention.

[0017] In particular, the adaptation unit can form part of the control device. In particular, the control device can be configured to adapt the at least one model parameter of the model stored on the control device such that the deviation between the target variable determined by the soft sensor and the target variable determined by the target variable sensor is at least reduced at least at the first operating point or at several operating points of the system unit (respectively).

[0018] In particular, the fuel cell system can be used or deployed in a vehicle, preferably a motor vehicle, e.g., a passenger car, a truck, or a motorcycle. Preferably, during normal operation, the fuel cell system is free of the target variable sensor for determining the target measured variable. In other words, during normal operation, the fuel cell system or a vehicle with such a fuel cell system does not have the target variable sensor, e.g., an air mass flow sensor. Normal operation of a fuel cell system of a vehicle is, for example, operating the fuel cell system while the vehicle is traveling on a road.In a fuel cell system according to the invention, it can be advantageous for the fuel cell system to have a fluid delivery device as the system unit for delivering the cathode gas, a pressure sensor as the first sensor for detecting a pressure of a cathode gas of a fuel cell stack of the fuel cell system as the first measured variable, a temperature sensor as the second sensor for detecting a temperature of the cathode gas as the second measured variable, a rotational speed sensor as a third sensor of the fuel cell system for detecting a rotational speed of the fluid delivery device, an air mass flow sensor as a target variable sensor for determining an air mass flow as the target measured variable, and an air mass flow model as a model stored on the control device. In other words, the fuel cell system can have an air mass flow soft sensor, wherein the air mass flow soft sensor orat least one model parameter of the air mass flow model is adapted, preferably adjusted, by means of an air mass flow sensor, for example by means of a hot-film air mass meter. The fluid delivery device as a system unit of the fuel cell system is in particular a compressor for compressing the cathode gas, for example air, of a cathode gas path of the fuel cell system for supplying the fuel cell stack with the cathode gas. For example, in order to teach the air mass flow soft sensor, a speed can be passed through as a step profile in a speed range, preferably across the entire speed range, of the fluid delivery device. At each speed step and thus at a different pressure ratio from the fluid delivery device, a respective deviation or error between the air mass flow sensor and the target variable sensor can be compensated for, in particular in a stationary manner, by a respective adaptation orChanges in at least one model parameter, in particular a (respective) value of the model parameter, for example by means of a controller, can be reduced, in particular eliminated. Furthermore, a sensor value of a hot-film air mass meter serving as an air mass flow sensor can be smoothed using a low-pass filter in order to filter out superimposed noise. The noise can be generated by geometrically determined deviations from the optimal installation position of the hot-film air mass meter. In particular, after the step profile has been completely traversed, the characteristic curve is filled with the respective (if necessary) adjusted model parameters. Thus, the validity range of the air mass flow soft sensor can be particularly advantageous over a complete speed range of the fluid conveying device, even at low speeds. Furthermore, intermediate values ​​can be determined or calculated by interpolation, in particular linear interpolation.can be determined, so that the soft sensor can be trained particularly quickly. In a fuel cell system according to the invention, it can be advantageous for the target variable sensor for determining the target measurement variable to be mounted on the fuel cell system in a detachable, particularly non-destructive, manner. Thus, the target variable sensor can be mounted (non-destructively) or dismounted (non-destructively) as needed.

[0019] In a fuel cell system according to the invention, it can be advantageous for the target variable sensor to be connected to an energy source that is at least independent of the fuel cell system for operating the target variable sensor. For example, the target variable sensor can have its own energy source for operating the target variable sensor. Thus, for example, an air mass flow sensor can be arranged particularly easily or with particularly little effort at a closable opening in an intake pipe of a cathode gas path of the fuel cell system for arranging the air mass flow sensor for teaching the soft sensor.

[0020] In a fuel cell system according to the invention, it can be advantageous for the target variable sensor to be configured to wirelessly transmit data on the determined target variable of the target variable sensor, for example, to the adaptation unit or the control device of the fuel cell system. This allows the soft sensor to be trained particularly easily and with particularly little effort.

[0021] The fuel cell system according to the second aspect of the invention thus has the same advantages as those already described for the method according to the first aspect of the invention.

[0022] According to further aspects, the present invention shows a computer program product, a computer-readable medium and a data carrier signal, wherein the computer program product, the computer-readable medium and the data carrier signal have the same advantages as have already been described for the method according to the first aspect of the invention or the fuel cell system according to the second aspect of the invention or the computer program product or the computer-readable medium or the data carrier signal according to the further aspects of the invention.

[0023] The computer program product comprises commands, in particular instructions, wherein the commands, in particular instructions, cause a fuel cell system according to the invention to execute a method according to the invention. The computer program product can be implemented as computer-readable instruction code in any suitable programming language, such as JAVA, C++, etc., wherein, in particular, the instruction code of the computer program product can program a computer or other programmable devices, such as a control device of a fuel cell system according to the invention, such that the desired functions are executed.

[0024] The computer program product can be implemented either by means of an instruction code, i.e. a software (computer program), or by means of one or more special electronic circuits, i.e. in hardware, or in any hybrid form, i.e. by means of software components and hardware components.

[0025] The computer program product may be stored on a computer-readable storage medium, e.g. a volatile or non-volatile memory and / or a processor of the control device.

[0026] Furthermore, the computer program product can be provided in a network such as the Internet, from which it can be downloaded by a user as needed, wherein, in particular, when downloading the computer program product, the data carrier signal transmits the computer program product according to the invention.

[0027] Further measures improving the invention will become apparent from the following description of some exemplary embodiments of the invention, which are schematically illustrated in the figures. All features and / or advantages apparent from the claims, the description, or the drawings, including design details, spatial arrangements, and method steps, may be essential to the invention both individually and in various combinations. It should be noted that the figures are merely descriptive and are not intended to limit the invention in any way.

[0028] They show schematically:

[0029] Fig. 1 a method, and

[0030] Fig. 2 a fuel cell system.

[0031] In the following figures, identical reference numerals are used for the same technical features, even in different exemplary embodiments. Fig. 1 schematically discloses a method for teaching a soft sensor for determining a target measured variable of a fuel cell system (100), as described, for example, in relation to Fig. 2. In particular, the fuel cell system (100) is operated at least temporarily in a stationary state for teaching the soft sensor of the fuel cell system (100). The fuel cell system (100) comprises a first sensor (10) for detecting a first measured variable, a second sensor (20) for detecting a second measured variable, and a control device (50), wherein the control device (50) is designed to determine the target measured variable based on a model stored on the control device (50), at least as a function of the first measured variable and the second measured variable.The fuel cell system (100) further comprises a target variable sensor (40) for (directly) determining or (directly) measuring the target measured variable. Furthermore, the method comprises, as a first step, operating (320) at least one system unit (110) of the fuel cell system (100) that influences the target variable at at least one first operating point. Furthermore, the fuel cell system (100) comprises, as a step, adapting (340) at least one model parameter or a value of a model parameter of a characteristic curve of the model stored on the control device (50) such that a deviation between the target variable determined by the soft sensor and the target variable determined by the target variable sensor (40) is at least reduced, preferably eliminated, at least at the first operating point of the system unit (110).

[0032] In the method shown in Fig. 1, it can be advantageous that at least the at least one system unit (110) of the fuel cell system (100) influencing the target variable is operated at a plurality of operating points (321), wherein the at least one model parameter, in particular a respective value of the at least one model parameter, of the model stored on the control device (50) is adapted (341) in such a way that a respective deviation in a respective operating point of the plurality of operating points of the system unit (110) between the target variable determined by the soft sensor and the target variable determined by the target variable sensor (40) is at least reduced.

[0033] It can also be advantageous in the method shown in Fig. 1 that the soft sensor is taught at intervals of time, wherein for each teaching of the soft sensor, the target variable sensor (40) is arranged on the fuel cell system (100) to determine the target measured variable (300), and wherein after each teaching of the soft sensor, the target variable sensor (40) is detached, in particular removed, from the fuel cell system (100) (360). Fig. 2 schematically discloses a fuel cell system (100) with a fuel cell stack (101), wherein in particular the fuel cell system (100) is designed to carry out or execute a method as described, for example, in relation to Fig. 1. Furthermore, Fig. 2 schematically shows at least in part a cathode gas path for supplying the fuel cell stack (101) with a cathode gas.The fuel cell system (100) comprises a fluid delivery device arranged in the cathode gas path as a system unit (110) for delivering the cathode gas. Furthermore, the fuel cell system (100) comprises a pressure sensor as a first sensor (10) for detecting a pressure of the cathode gas of the fuel cell stack (101) of the fuel cell system (100) as a first measured variable. Furthermore, the fuel cell system (100) comprises a temperature sensor as a second sensor (20) for detecting a temperature of the cathode gas as a second measured variable. Furthermore, the fuel cell system (100) comprises a rotational speed sensor as a third sensor (30) of the fuel cell system (100) for detecting a rotational speed of the fluid delivery device as a third measured variable. Furthermore, the fuel cell system (100) comprises an air mass flow sensor as a target variable sensor (40) for determining an air mass flow as the target measured variable.The fuel cell system (100) further comprises a control device (50), wherein the control device (50) is at least designed to determine the air mass flow as the target measured variable based on an air mass flow model stored on the control device (50) at least as a function of the first measured variable, the second measured variable and the third measured variable, and wherein the control device (50) is designed to adapt at least one model parameter of the air mass flow model stored on the control device (50) such that at least one deviation in a first operating point of the system unit (110) between the target variable determined by the soft sensor and the target variable determined by the target variable sensor (40) is at least reduced.

[0034] In the fuel cell system (100) shown in Fig. 2, it may be advantageous for the target variable sensor (40) for determining the target measured variable to be detachably, in particular non-destructively, attached to the fuel cell system (100). In the fuel cell system (100) shown in Fig. 2, it may further be advantageous for the target variable sensor (40) to be connected to an energy source that is at least independent of the fuel cell system (100) for the operation of the target variable sensor (40). In the fuel cell system (100) shown in Fig. 2, it may further be advantageous for the target variable sensor (40) to be designed to transmit data relating to the determined target measured variable of the target variable sensor (40) by radio, for example to the control device (50).

[0035] List of reference symbols

[0036] 10 first sensor

[0037] 20 second sensor

[0038] 30 third sensor

[0039] 40 Target sensor

[0040] 50 Control device

[0041] 60 adjustment unit

[0042] 100 fuel cell system

[0043] 101 fuel cell stacks

[0044] 110 System unit

[0045] 300 Arranging target variable sensor on fuel cell system 320, 321 Operating at least one system unit influencing the target variable in a first operating point or several operating points

[0046] 340, 341 Adjusting at least one model parameter of the model stored on the control device

[0047] 360 Solve target sensor of fuel cell system

Claims

Claims 1. A method for teaching a soft sensor for determining a target measurement variable of a fuel cell system (100), wherein the fuel cell system (100) comprises: - a first sensor (10) for detecting a first measured variable, - a second sensor (20) for detecting a second measured variable, - a control device (50), wherein the control device (50) is designed to determine the target measurement value based on a model stored on the control device (50) at least as a function of the first measurement value and the second measurement value, - a target variable sensor (40) for determining the target measured variable, and wherein the method comprises: - operating (320) at least one system unit (110) of the fuel cell system (100) influencing the target variable at at least one first operating point, - adapting (340) at least one model parameter of the model stored on the control device (50) such that a deviation between the target variable determined by the soft sensor and the target variable determined by the target variable sensor (40) is at least reduced at least in the first operating point of the system unit (110).

2. Method according to claim 1, characterized in that at least the at least one system unit (110) of the fuel cell system (100) influencing the target variable is operated at a plurality of operating points (321), wherein the at least one model parameter of the model stored on the control device (50) is adapted (341) in such a way that a respective deviation in a respective operating point of the plurality of operating points of the system unit (110) between the target variable determined by the soft sensor and the target variable determined by the target variable sensor (40) is at least reduced.

3. Method according to one of the preceding claims, characterized in that that the fuel cell system (100) is operated at least temporarily in a stationary state for teaching the soft sensor of the fuel cell system (100).

4. Method according to one of the preceding claims, characterized in that the training of the soft sensor is carried out at intervals of time, wherein for each training of the soft sensor the target variable sensor (40) is arranged on the fuel cell system (100) for determining the target measurement variable (300), and wherein after each training of the soft sensor the target variable sensor (40) is detached, in particular removed, from the fuel cell system (100) (360).

5. Fuel cell system (100), wherein the fuel cell system (100) comprises: - a first sensor (10) for detecting a first measured variable, - a second sensor (20) for detecting a second measured variable, - a control device (50), wherein the control device (50) is designed to determine the target measurement value based on a model stored on the control device (50) at least as a function of the first measurement value and the second measurement value, - a target variable sensor (40) for determining the target measured variable, - an adaptation unit (60) for adapting at least one model parameter of the model stored on the control device (50) such that at least one deviation in a first operating point of the system unit (110) between the target variable determined by the soft sensor and the target variable determined by the target variable sensor (40) is at least reduced.

6. Fuel cell system (100) according to claim 5, characterized in that the fuel cell system (100) comprises: - a fluid conveying device as the system unit (110) for conveying the cathode gas, - a pressure sensor as the first sensor (10) for detecting a pressure of a cathode gas of a fuel cell stack (101) of the fuel cell system (100) as the first measured variable, - a temperature sensor as the second sensor (20) for detecting a temperature of the cathode gas as the second measured variable, - a rotational speed sensor as a third sensor (30) of the fuel cell system (100) for detecting a rotational speed of the fluid conveying device as a third measured variable, - an air mass flow sensor as a target variable sensor (40) for determining an air mass flow as the target measured variable, - an air mass flow model stored on the control device (50).

7. Fuel cell system (100) according to claim 5 or 6, characterized in that the target variable sensor (40) for determining the target measured variable is arranged detachably, in particular non-destructively detachably, on the fuel cell system (100).

8. Fuel cell system (100) according to one of claims 5 to 7, characterized in that the target variable sensor (40) is connected in terms of energy technology to an energy source that is at least independent of the fuel cell system (100) for the operation of the target variable sensor (40).

9. Fuel cell system (100) according to one of claims 5 to 8, characterized in that the target variable sensor (40) is designed to transmit data on the determined target measurement variable of the target variable sensor (40) by radio.

10. Fuel cell system (100) according to one of claims 5 to 9, characterized in that the fuel cell system (100) is designed to carry out a method according to one of claims 1 to 4.

11. A computer program product, the computer program product comprising instructions, the instructions causing a fuel cell system (100) according to any one of claims 5 to 10 to execute a method according to any one of claims 1 to 4.

12. A computer-readable medium, wherein the computer-readable medium stores the computer program product according to claim 11.

13. A data carrier signal, wherein the data carrier signal transmits the computer program product according to claim 11.

Images