Method for improved handling of a technical system by means of a technical component

Abstract

The invention comprises a method (100) for improved handling of a technical system (40) by means of a technical component (30), the method having the following steps: - detecting (102) first data (32) of the technical component (30) by a data control unit (50), wherein a characteristic operating behavior of the technical component (30) is determined by means of the first data (32); - transmitting (104) the first data (32) to a first cloud-based data processing unit (70) by means of a first interface (57); - generating (106) second data (33) by pre-processing the first data (32) in the first cloud-based data processing unit (70); - transmitting (108) the generated second data (33) to a second cloud-based data processing unit (80) by means of a second interface (75); - processing (110) the second data (33) by applying a data-based algorithm (84) saved in the second cloud-based data processing unit (80) in order to generate third data (34); - transmitting (112) the generated third data (34) to the data control unit (50) by means of a fourth interface (87); - at least partially replacing (114) preset first data (32) of the technical component (30) with the generated third data (34) in order to improve handling of the technical system (40) by means of the technical component (30) thereby.

Description

Technical Field

[0001] The present invention relates to a method for improving the operation of a technical system by means of technical components. Background Technology

[0002] In a fuel cell, electricity is generated from hydrogen and oxygen. To ensure the safe and robust operation of the fuel cell, multiple sensor values ​​are continuously analyzed and evaluated in the control unit (FCCU). Then, based on these sensor values ​​and appropriately selected and applied algorithms, existing actuators are manipulated, which advantageously influence the physical behavior of the fuel cell.

[0003] Therefore, the hydrogen injection valve (HGI) must be properly energized, for example, to supply the necessary amount of hydrogen to the anode of the fuel cell. This must be done as precisely as possible, because too much hydrogen unnecessarily increases the pressure within the anode, leading to mechanical stress on the membrane. Conversely, too little hydrogen results in low pressure, which also puts mechanical stress on the membrane and may additionally cause hydrogen depletion that damages the cell.

[0004] To optimize the existing regulation and control algorithms in the control device, measurements were taken of the HGI valve, and the data was stored in the control device as a mathematical model. This mathematical model allows for both feedforward control and regulation to be adapted to the specific system, enabling rapid and robust tracking of the hydrogen supply.

[0005] DE 10 2021 203 798 A1 discloses, by way of example, a method for regulating a fuel cell, in which the regulating facility is connected to a network to exchange corresponding control commands for controlling the operating behavior of the fuel cell. Summary of the Invention

[0006] Therefore, the objective of this invention is to provide a solution for operating fuel cell systems efficiently and with improved performance.

[0007] This task is solved by a method with the features of an independent claim for improving the manipulation of a technical system by means of technical components.

[0008] According to a first aspect, the present invention relates to a method for improving the operation of a technical system by means of technical components.

[0009] In the first step, the data control unit detects first data of the technical component, wherein the first data is used to determine the characteristic operating behavior of the technical component.

[0010] In the second step, the first data is transmitted to the first cloud-based data processing unit via the first interface.

[0011] In the third step, the second data is generated by preprocessing the first data in the first cloud-based data processing unit.

[0012] In the fourth step, the generated second data is transmitted to the second cloud-based data processing unit via the second interface.

[0013] In the fifth step, the second data is processed by applying a data-based algorithm stored in the second cloud-based data processing unit to generate the third data.

[0014] In the sixth step, the generated third data is transmitted to the data control unit via the fourth interface.

[0015] In the seventh step, the preset first data of the technical component is at least partially replaced with the generated third data, thereby improving the operation of the technical system via the technical component. This at least partial replacement can be performed, for example, by the data control unit. Therefore, in order to operate the technical component, particularly for example… - Within the scope of the availability of the third data, this third data may be used, and - In the range where the third data is unavailable, the first data from the original detection can be used.

[0016] For example, preprocessed values, variables, and / or parameters can be generated as third data only for a subset of the total values, variables, and / or parameters detected in the form of first data and / or preset values, variables, and / or parameters.

[0017] The basic concept of this invention is that a control unit (FCCU) for operating a fuel cell first detects characterization data used to determine the behavior of the HGI valve. This detected data is then transmitted to a computer network (cloud) via a suitable data connection. There, the data is processed using appropriate algorithms to determine or optimize the parameters of a mathematical model that describe the characterization behavior of the HGI valve. Subsequently, these adapted parameters are sent back to the control unit via another interface, where they are used to regulate and feedforward control the HGI valve, replacing pre-set or detected values.

[0018] The following advantages are obtained by the method according to the present invention: The system identifies changes to the HGI valve's behavior and automatically tracks those changes in the control unit. This ensures consistent regulation quality throughout operation.

[0019] Compared to algorithms in control devices, processing detected data in the cloud offers significantly more memory and computing power. This also enables the use of sophisticated optimization algorithms. Furthermore, parameters can be manually approved by experienced engineers early in the development process, eliminating the need for costly monitoring algorithms in the early stages of development.

[0020] By employing the method according to the invention or the corresponding system architecture for implementing the method, the commissioning of new systems can be significantly accelerated. Furthermore, component drift behavior can be identified and visualized during the testing phase because the detected data is highly compressed through optimization of the mathematical behavioral model. Partial drift can be identified and compensated for in the system after production has commenced, thereby improving product lifespan. Even when drift is out of specification, faults or deviations from pre-set normal states can be quickly identified, and corresponding responses or measures to eliminate the identified deviations can be initiated.

[0021] Furthermore, the method according to the invention, or through a corresponding architecture, can also be applied to large fleets of vehicles, since the required computing power can be dynamically scaled as needed by a virtual data processing unit.

[0022] One possible configuration of the method is that the first data is a measurable parameter that is influenced by and / or affects the technical component, such that the first data describes the characteristic operational behavior of the technical component. This achieves the advantage of mapping the operational behavior of the technical component in an application-oriented manner and according to the corresponding application domain.

[0023] One possible configuration of the method is that the first data displays parameters of a mathematical model of the technical component, which describes the characteristic operational behavior of the technical component. This achieves the following advantages: mapping the operational behavior of the technical component in an application-oriented manner and according to the corresponding application domain.

[0024] One possible configuration of the method involves preprocessing the first data including at least one of the following measures: error and integrity checking, interference signal cleanup, and data format conversion. This achieves the advantages of generating the most accurate possible mapping of the actual operational behavior of the technical component and enabling effective manipulation of it.

[0025] One possible configuration of the method involves processing the second data to the third data by providing adapted or optimized parameters, in the form of the third data, for a mathematical model of the technical component based on the second data. This achieves the advantage of producing a mapping as accurate as possible to the actual operational behavior of the technical component.

[0026] One possible configuration of the method involves processing the second data into the third data using a virtual data processing unit installed within the second cloud-based data processing unit. This virtual data processing unit receives the second data from the first cloud-based data processing unit and accesses the data-based algorithm to process the second data. This achieves the advantage of enabling efficient and secure data processing. Furthermore, this architecture allows the method to be applied to large fleets, as the required computing power can be dynamically scaled as needed by the virtual data processing unit.

[0027] In one possible configuration of the method, the virtual data processing unit can be connected to the output unit via a third interface to output the generated third data. This achieves the advantage of providing the generated data for verification in a simple manner.

[0028] One possible configuration of the method is that the technical system is constructed as a fuel cell or fuel cell stack, which is controlled by a hydrogen injection valve (HGI) constructed as a technical component. This achieves the advantage of consistently providing optimal and application-appropriate control over the fuel cell or fuel cell stack.

[0029] According to a second aspect, the present invention relates to a computer program comprising machine-readable instructions, which, when executed on one or more computers and / or computer instances, cause the one or more computers or computer instances to perform the method according to the invention.

[0030] According to a third aspect, the present invention relates to a machine-readable data carrier and / or download product having the aforementioned computer program.

[0031] According to a fourth aspect, the present invention relates to one or more computers and / or computer instances having the computer program, and / or having the machine-readable data carrier and / or downloadable product.

[0032] Next, measures to improve the present invention will be further shown in conjunction with the description of preferred embodiments of the present invention and with reference to the accompanying drawings. Attached Figure Description

[0033] The attached diagram shows: Figure 1 A schematic flowchart of a method 100 for improving control of a technical system 40 by means of a technical component 30 according to an embodiment of the present invention; and Figure 2 : An illustrative architecture for implementing method 100 according to one embodiment of the present invention. Detailed Implementation

[0034] Figure 1 A schematic flowchart is shown for a method 100 for improving the control of a technical system 40 by means of a technical component 30.

[0035] Here, the technical system 40 can preferably and exemplary be constructed as a fuel cell or fuel cell stack, which is controlled by a hydrogen injection valve (HGI) constructed as technical component 30. However, the present invention can be applied to different application areas and various technical applications.

[0036] In step 102, the data control unit 50 detects the first data 32 of the technical component 30, wherein the first data 32 is used to determine the characterization operation behavior of the technical component 30.

[0037] Optionally, measurable parameters are transmitted here, which are affected by HGI and / or influence HGI.

[0038] Optionally, the first data 32 is a measurable parameter that is affected by and / or influences the technical component 30, such that the first data 32 describes the characteristic operational behavior of the technical component 30.

[0039] In step 104, the first data 32 is transmitted to the first cloud-based data processing unit 70 via the first interface 57 (see...). Figure 2 ).

[0040] In step 106, the second data 33 is generated by preprocessing the first data 32 in the first cloud-based data processing unit 70.

[0041] Optionally, the preprocessing of the first data 32 may include at least one of the following measures: error and integrity checking, interference signal cleanup, and data format conversion.

[0042] In step 108, the generated second data 33 is transmitted to the second cloud-based data processing unit 80 through the second interface 75.

[0043] In step 110, the second data 33 is processed by applying the data-based algorithm 84 stored in the second cloud-based data processing unit 80 to generate the third data 34.

[0044] Optionally, the processing of the second data 33 to the third data 34 may involve providing, based on the second data 33, adapted or optimized parameters of the mathematical model 31 for the technical component 30, in the form of the third data 34.

[0045] Furthermore, the processing of the second data 33 to the third data 34 can be performed by a virtual data processing unit 82 installed in the second cloud-based data processing unit 80. This virtual data processing unit 82 (e.g., a cloud application, which may be configured with corresponding computing or data processing capabilities depending on the application location) receives the second data 33 from the first cloud-based data processing unit 80 and accesses a data-based algorithm 84 in order to process the second data 33.

[0046] The virtual data processing unit 82 can also be connected to the output unit 86 via the third interface 85 so as to output the generated third data 34.

[0047] In step 112, the generated third data 34 is transmitted to the data control unit 50 through the fourth interface 87.

[0048] In step 114, the generated third data 34 at least partially replaces the preset first data 32 of the technical component 30 in order to improve the control of the technical system 40 by the technical component 30.

[0049] Figure 2 (combined) Figure 1 The diagram illustrates a schematic architecture for implementing method 100 according to one embodiment of the present invention.

[0050] The method 100 according to the invention can be used to improve the control of a technical system 40, such as a fuel cell or fuel cell stack, by means of a technical component 30, such as an HGI valve.

[0051] For this purpose, the technical system 40 is connected to a data control unit 50, which can be configured as a control device. The data control unit 50 detects first data 32, which determines the characteristic operational behavior of the technical component 30. This first data 32 may, for example, be: -Pre-HGI pressure - The current on the reactor, which is converted into the flow rate at the HGI, and - The current used to control the HGI.

[0052] The first data 32 is transmitted to the first cloud-based data processing unit 70 via the first interface 57, optionally via a data logger 55 acting as a data providing unit, which may be configured as a computer network.

[0053] The first cloud-based data processing unit 70 is configured as a first area, also known as the Data Landing Zone. This first area has lower security requirements than the second cloud-based data processing unit 80, which can be regarded as a second area, also known as the Data Processing Zone.

[0054] In this example, the first cloud-based data processing unit 70 consists of a data storage unit 72 and a data preprocessing unit 74.

[0055] The first data 32 is stored in the data memory 72.

[0056] Preprocessing of the first data 32 to the second data 33 is performed in the data preprocessing unit 74: here, the first data 32 can be checked in terms of its data quality, the first data 32 can be cleaned, and the first data 32 can be converted into a data format that is favorable for further processing.

[0057] Then, the second data 33 generated in the first region is transmitted to the second cloud-based data processing unit 80 through the second interface 75.

[0058] exist Figure 2 In the example, the second cloud-based data processing unit 80 includes a virtual data processing unit 82 and a data-based algorithm 84.

[0059] To prevent unauthorized access to algorithm 84, the two zones 70 and 80 are separated from each other, for example, by technical security structures such as firewalls. Therefore, the algorithm is typically not directly accessible from the outside, which enhances system security.

[0060] Algorithm 84 in the second cloud-based data processing unit 80 is used to process and interpret the received second data 33 in order to generate third data 34. Here, the third data 34 shows the updated parameters of the mathematical model 31 of the technical component 30, which describes the characterization of the operational behavior of the technical component 30.

[0061] Then, the generated third data 34 is transmitted to the data control unit 50 via the fourth interface 87, and optionally to the data logger 55.

[0062] Subsequently, the generated third data 34 at least partially replaces the preset first data 32 of the technical component 30 in order to improve the control of the technical system 40 (fuel cell) via the technical component 30 (HGI valve).

[0063] As in Figure 2As further illustrated, the virtual data processing unit 82 may optionally be connected to the output unit 86 via a third interface 85, such as a display, tablet, or smart device, to output the generated third data 33 for evaluation. This is significant in terms of security, because in this invention, the generated third data 34 is directly fed into the control device 50, and thus directly intervenes in or influences the operation of the technical system 40 through the control of the component 30.

[0064] Optionally, the second cloud-based data processing unit 80 may include a so-called monitor, i.e., a data arrival register. This monitor identifies when new data arrives in the second area 80, i.e., the so-called data processing area, and triggers processing. For this purpose, a container may be started, for example, in an environment containing an Octave kernel.

[0065] A suitable environment is a Databricks cluster, batch processing pool, Kubernetes / Docker server, or simply a Linux virtual machine. The container loads both newly arriving data and an algorithm, for example, in the form of an Octave file. This algorithm describes how to determine the parameters of the HGI valve model based on measurements. The data is then heavily compressed using this algorithm. Therefore, in the HGI valve example, only four parameters are determined based on a measurement file containing several kilobytes of data.

[0066] These parameters, along with some of the measured characterization parameters, can be visualized in the cloud via dashboards (based on Tableau, Plotly, Dash, or Seaborn). This allows for the rapid identification of anomalies, drift, and faulty functions in the algorithm.

[0067] After checking the results, which can be completed manually in the early stages and by appropriate algorithms in the later stages, the data is transmitted back to the control device (FCCU) 50 via an appropriate data connection and the data logger is sent back.

[0068] In control device 50, data is either stored in RAM (working page) during the early development phase or transferred to Flash (reference page) at a later stage (e.g., after SOP). Changed parameters affect the model-based regulator in the control device because both the feedforward control behavior and the closed-loop regulation behavior depend on the mathematical model of HGI valve 30. Thus, by adapting the regulation behavior to the changed behavior of HGI valve 30, the quality of maintaining the invariant system behavior 40 can be achieved.

[0069] In control device 50, the following equation is analyzed and evaluated: This equation calculates the adjustment parameter (current = x) for controlling the HGI valve based on the desired flow rate y and a given pressure condition p (e.g., the pressure before the HGI valve), i.e., x = f(y, p). Here, the function f can correlate the parameter y with p in a linear or nonlinear manner to obtain the result x.

[0070] To estimate the function f(y,p), data is transmitted from control device 50 to the cloud, for example, via mobile communication (4G or 5G). Therefore, in this case, multiple individual measurements / samples of current, flow rate, and pressure are transmitted.

[0071] Importantly, the triplets of the three parameters are transmitted separately, thus allowing for the allocation in the cloud of which values ​​were acquired at the same time point / sampling step. Furthermore, status information can also be transmitted, indicating that these values ​​were acquired under relevant operating conditions, such as the absence of specific manual application intervention at the time, or the fuel cell or fuel cell stack being in a particular operating state (e.g., purge valves and / or exhaust valves were closed).

[0072] Therefore, valid values ​​for x, y, and p are then provided in the cloud for multiple measurements / samples. For example, these parameters can be associated (y=k(p,x)) as follows: Based on this, for example, four parameters mm, mc, cm, and cc can be estimated in the cloud using a least-squares estimator with the aid of multiple measurements / sampling. The advantage of this approach / configuration is that a linearity problem exists, which is subsequently easier to solve. Based on this form, the form x = f(y, p) related to the control device is then derived through transformation, where the estimated parameters mm, mc, cm, and cc are then associated with the inputs p and y in the function f.

[0073] In this specific case, the function is nonlinear. The estimated parameters mm, mc, cm, and cc are then transmitted back to the FCCU via mobile communication, for example, through the MQTT protocol, so that x=f(y,p) can be re-analyzed and evaluated based on the valve's current behavior.

[0074] Alternatively, the function f can be represented in different ways, such as by a Gaussian process or neural network that has inputs y and p and provides a result x.

[0075] Alternatively, a characteristic field can be used, in which the current x is labeled according to p and y. Then, interpolation is performed between the characteristic points in a known manner.

[0076] The implementation shown here with four parameters is merely an example. Depending on the chosen modeling approach, more or fewer parameters may be required.

[0077] In another configuration, it can be determined in advance in the cloud: for example, whether there is sufficient excitation in the signal, i.e., whether the entire operating range can be estimated based on existing data. If not, i.e., if, for example, the pressure p is constant, then separate countermeasures must be taken.

[0078] If, for example, there is insufficient excitation in the signal and one or more of the four parameters mm, mc, cm, and cc cannot be estimated, then the parameters that can be determined by the excitation are estimated first. The remaining parameters that have not yet been estimated are estimated when the excitation is permissible in the signal.

[0079] If no estimate exists for the corresponding parameters, the following initial values ​​are used: these initial values ​​can, for example, be derived from HGI measurements under the new conditions. Therefore, values ​​always exist for the four parameters mm, mc, cm, and cc, making it always possible to control the HGI valve, or to calculate the control current (current = x).

[0080] Therefore, in summary, by using the method described at the beginning or the corresponding architecture used to implement the method, the goal is to determine the parameters of the HGI valve based on existing measurement parameters during the operation of the fuel cell system and thus enable the mathematical model to continuously track the real behavior.

[0081] Therefore, the controller is automatically adapted to the valves used in the new system, and potential drift in valve behavior is identified and compensated for during operation. By processing existing measurement parameters in the control equipment in the cloud, complex and computationally expensive algorithms are used to derive the mathematical model of the HGI valve. However, in addition to HGI valves, the method described herein can also be used to identify model parameters of other components of the control equipment. Examples include: throttle valves in the cathode, anode, and thermal systems; pumps / compressors in the cathode, anode, and thermal systems; volumes in the anode and cathode; and reactor characteristics such as polarization characteristic lines or water diffusion.

Claims

1. A method (100) for improving the control of a technical system (40) by means of a technical component (30), the method comprising the following steps: - The first data (32) of the technical component (30) is detected (102) by the data control unit (50), wherein, The characterization and operational behavior of the technical component (30) are determined using the first data (32); - The first data (32) is transmitted (104) to the first cloud-based data processing unit (70) via the first interface (57); - The second data (33) is generated (106) by preprocessing the first data (32) in the first cloud-based data processing unit (70); - The generated second data (33) is transmitted (108) to the second cloud-based data processing unit (80) via the second interface (75); - The second data (33) is processed (110) by applying the data-based algorithm (84) stored in the second cloud-based data processing unit (80) to generate the third data (34). - The generated third data (34) is transmitted (112) to the data control unit (50) via the fourth interface (87). - Replace (114) the preset first data (32) of the technical component (30) at least partially with the generated third data (34) in order to improve the operation of the technical system (40) by the technical component (30).

2. The method (100) according to claim 1, wherein, The first data (32) is a measurable parameter that is affected by and / or influences the technical component (30) such that the first data (32) describes the characterization of the technical component (30)'s operational behavior.

3. The method (100) according to claim 2, wherein, The processing (110) from the second data (33) to the third data (34) involves providing, based on the second data (33), parameters of an adapted or optimized mathematical model (31) for the technical component (30) in the form of the third data (34).

4. The method (100) according to any one of the preceding claims, wherein, The preprocessing of the first data (32) includes at least one of the following measures: error and integrity checks, interference signal cleanup, and data format conversion.

5. The method (100) according to any one of the preceding claims, wherein, The processing (110) of the second data (33) to the third data (33) is performed by a virtual data processing unit (82) installed in the second cloud-based data processing unit (80), wherein the virtual data processing unit receives the second data (33) from the first cloud-based data processing unit (80) and accesses the data-based algorithm (84) in order to process the second data (33).

6. The method (100) according to claim 5, wherein, The virtual data processing unit (82) can be connected to the output unit (86) through the third interface (85) so as to output the generated third data (33).

7. The method (100) according to any one of the preceding claims, wherein, The technical system (40) is configured as a fuel cell or fuel cell stack, which is controlled by a hydrogen injection valve (HGI) configured as a technical component (30).

8. A computer program comprising machine-readable instructions that, when executed on one or more computers and / or computer instances, cause the computer or computer instance to perform the method according to any one of claims 1 to 7.

9. A machine-readable data carrier and / or download product having the computer program according to claim 8.

10. One or more computers and / or computer instances having a computer program according to claim 8, and / or having a machine-readable data carrier and / or downloadable product according to claim 9.

Images