Method for operating at least one electrochemical system, computing unit

Abstract

The invention relates to a method for operating at least one electrochemical system (1), in particular an electrochemical system for producing hydrogen. According to the invention, at least one operating variable is modeled on an analog computer (4), said operating variable being dependent on the operation of the electrochemical plant (1) and being detectable by means of a sensor (6), the analog computer (4) reads out operating data from a control computer (3) for controlling the electrochemical system (1) and uses the operating data to reflect the actual operating state. The invention further relates to a simulation computer (4) for carrying out the steps of the method according to the invention, the simulation computer (4) being designed to carry out the simulation during ongoing operation of the electrochemical plant (1).

Description

Technical Field

[0001] This invention relates to a method for operating at least one electrochemical facility, such as an electrolysis facility for producing hydrogen or other media, a stationary fuel cell facility for generating current and heat, a redox flow system, or a battery-based stationary energy storage device. Furthermore, this invention relates to a computing unit for implementing the steps of the method according to the invention. Background Technology

[0002] Electrolysis facilities for producing hydrogen utilize electrical energy to decompose water into hydrogen and oxygen in an electrochemical process. Since the water used for electrolysis needs to have a specific purity, it is first purified in a purifier. Next, the water passes through a liquid separator, where any remaining gases (such as air or oxygen) are separated. The liquid water is then transported by a device (e.g., a pump) to maintain a predetermined process pressure, and subsequently cooled in a heat transfer unit.

[0003] The core component of an electrolysis facility is a stack of single electrolytic cells or a composite structure of multiple electrolytic cells, within which the electrochemical process takes place. The prepared water or reactants are introduced into the electrochemical process. Electrical energy required for the electrolysis process is supplied to the electrolysis facility via an electrical interface. This interface powers the electrolytic cell stack. Power is always supplied in an overvoltage manner, meaning that the DC voltage needs to be greater than theoretically required to break intramolecular bonds.

[0004] During electrolysis, not all the supplied water is broken down. Therefore, the products oxygen and hydrogen, along with water, are discharged from the electrolytic cell stack. The oxygen, along with excess water, is returned as a gas-liquid mixture from the electrolytic cell stack to a liquid separator, where the water is separated and fed back into the process along with fresh water. From there, the oxygen is either released from the electrolysis facility to the environment or used for other purposes.

[0005] After leaving the electrolytic cell stack, the hydrogen gas, as a mixture of liquid and gaseous water, passes through a second liquid separator, where the liquid water is first separated and, if necessary, returned to the fresh water supply of the electrolysis facility. Depending on the requirements of subsequent processes, the remaining gas phase is post-treated, for example by condensation or purification in a purification facility, and is then discharged from the electrolysis facility as a product.

[0006] Due to the highly dynamic nature of process parameters and state variables within the electrolysis process, there is a significant knowledge gap regarding actual process operation according to current technology, making it difficult to model the entire electrolysis facility as a system. For process monitoring, including reflecting the actual process state and predicting changes based on it, comprehensive measurement of operating and state parameters using suitable sensors is crucial. However, sensors sometimes require significant structural space, are prone to failure, and are maintenance-intensive. Furthermore, a large number of sensors are needed for proper process monitoring within an electrolysis facility. Other electrochemical facilities (e.g., redox flow systems, fuel cell systems, and battery systems) have similar structures, resulting in similar problems there. Summary of the Invention

[0007] Therefore, the present invention addresses the task of eliminating the need for sensors, and especially in the early stages of development, without having to eliminate important operating values. To solve this task, a method for operating at least one electrochemical facility according to the present invention is proposed. Advantageous extensions of the invention are described below. Furthermore, a computational unit for implementing the steps of this method is provided.

[0008] A method for operating at least one electrochemical facility, such as an electrochemical facility for producing hydrogen, is proposed. According to the invention, at least one operating parameter is modeled on a simulation computer. This operating parameter is related to the operation of the electrochemical facility and can be detected by means of sensors. The simulation computer reads operating data from a control computer used to control the electrochemical facility and uses it to reflect the actual operating state. Besides the exemplary electrolysis facility mentioned, the electrochemical facility can also be a stationary fuel cell facility, a redox flow system, or a battery-based stationary energy storage device for generating current and heat. Here, the simulation is performed in parallel with conventional operation. In particular, the simulation does not interfere with or hinder the operating software responsible for controlling the electrochemical facility, which runs on the control computer. To this end, a separate computing unit, i.e., a simulation unit, is provided for the simulation. This simulation unit can be integrated into a separate physical computer or can be configured as a reserved memory within the control computer, in which case write access to the memory configured for controlling the facility is excluded. Preferably, the modeling is hybrid and incorporates both empirical and physical elements. However, the modeling can also be performed purely physically or purely empirically. Here, different operating parameters have different accuracy requirements, and the model created by the simulation needs to meet those accuracy requirements accordingly.

[0009] The operational data used to simulate the actual operating state in the simulation computer is the same as the operational data used by the operating software on the control computer to control the facility. In this respect, the simulation produces a situation picture similar to the situation picture on which the operating software makes control decisions. Preferably, the reading of operational data through simulation is achieved by reading access to the control computer. Therefore, a complete process simulation is not required; rather, the simulation provides the possibility of continuously checking and correcting the simulated operating state on which the model is based against the actual operating state, as necessary. There is no restriction on how or when exactly such a correction is made. At a defined point in time, virtual operating parameters obtain values ​​from the model. Generally, if the value is consistent with the actual value of the real operating parameter within the error tolerance specified in individual cases, the value is considered correct. If all values ​​of the virtual operating parameters are correct throughout the operation, the model should be considered accurate. The operating software can utilize the modeled operating parameters instead of sensor signals, thereby eliminating the need for sensors involved.

[0010] Furthermore, it is proposed that the modeled operating parameters be provided as virtual operating parameters to the operating software, which is implemented on the control computer of the same or another electrochemical facility. This preferred embodiment ensures that the operating software can read and utilize the virtual operating parameters. Preferably, this utilization is integrated into the operating decisions so as to no longer rely on sensor data, for example.

[0011] Therefore, in order to use virtual operating parameters and eliminate the need for corresponding sensors, the modeling of these virtual operating parameters needs to be validated. Ideally, this validation should be achievable without altering the active electrochemical facility in terms of hardware and / or software, and thus without jeopardizing its operational readiness. Therefore, in an extension of the invention, it is proposed to detect the operating parameters using sensors and to check and / or validate the modeling of the operating parameters by comparing the data with the sensor data. Here, this check and / or validation is performed by examining the correctness of the values ​​generated by the modeling. The advantage of this form of check and / or validation is that it closely approximates reality, even for potentially unforeseen events that need to be reproduced as accurately as possible by the modeling. Furthermore, like the simulation process, the validation process for the modeling should also run in the background without interfering with the regular operation of the electrochemical facility, which can be achieved in this way.

[0012] Furthermore, it is proposed that the sensor be installed in the electrochemical facility before verification and / or validation, and then removed again after verification and / or validation. This extended approach has the advantages of eliminating the sensor during normal operation, preserving data, and leveraging the advantage of verifying and / or validating actual measurement data. Therefore, a single sensor can be used, in particular, to verify and / or validate different models on different electrochemical facilities. When using this preferred embodiment in an active, validated facility, care should be taken to avoid short-term changes to the facility hardware and / or control software, while maintaining the achieved validation and / or operational availability. In this embodiment, this is achieved by introducing a sensor, which is used in a time-limited manner, into a support structure on the corresponding facility, so that there is no hardware intervention (Eingriff). The control remains unchanged. In the case of verification, the control is based on modeling. The actual sensor values, which are available in a time-limited manner, are analyzed in a separate computing unit that does not have write access to the active control environment. The output is whether the modeling in the control environment is consistent with the actual sensor values.

[0013] Virtual operating parameters also need to accurately reproduce unexpected states. However, it is typically impossible to verify and / or validate the modeling across all possible physical operating states. Therefore, it is proposed to define at least one reference operating state, based on which the modeling of virtual operating parameters is verified and / or validated. For verification and / or validation, the modeling needs to consistently and accurately represent the operating variables in the reference operating state. Preferably, multiple reference operating states are defined, reflecting not only different operating points during normal operation but also critical operating states and faults. With this preferred implementation, all important operating states can be reliably and modularly incorporated into the verification or validation process. Furthermore, additional reference operating states can be added retroactively if necessary.

[0014] Furthermore, in the case of multiple reference operating states and / or pending verification or validation cycles, the verification and / or validation are parallelized by distributing them across multiple electrochemical facilities. To this end, modeling is run on multiple electrochemical facilities, and reference operating states are assigned to these facilities. On the one hand, this significantly accelerates the verification and / or validation process. On the other hand, some reference operating states may only occur with a relatively low probability. Through the proposed parallelization, the same model can run simultaneously on different facilities during normal operation for verification and / or validation purposes until the background verification and / or validation process is completed. When the verification and / or validation process is complete, the modeling of the virtual operating parameters can be equivalently used as a real sensor.

[0015] In addition, it is suggested that if the operating status is referenced...

[0016] - It is harmful to the material.

[0017] - Significantly hinders ongoing operations, and / or

[0018] - This occurs only very rarely during the normal operation of the electrochemical facility.

[0019] Then, the checks and / or verifications based on the reference operating conditions are performed on the test bench.

[0020] With this preferred embodiment, reference operating states that might interfere with or jeopardize the active operation of the electrochemical facility can also be used for verification and / or validation. Furthermore, loading rare reference operating states onto the test bench can achieve significant time savings, as the duration of verification and / or validation is no longer related to statistically improbable events.

[0021] In an extended embodiment of the invention, modeling is proposed for verifying sensors. Preferably, the sensor is a robustly mounted sensor. Sensors are prone to failure and aging. This preferred embodiment allows for the detection of such failures and aging in existing sensors. To this end, a validated model is run in parallel with the sensor within the simulation. Sensor data is used by a control computer for routine operation. Simultaneously, the sensor data is read from the simulation and compared with the modeled data. If the data is inconsistent within an error tolerance that can be specified in individual cases, the sensor is considered faulty.

[0022] Furthermore, it is proposed to predict sensor drift using modeling, and sensor verification includes checking for drift. Preferably, the sensor is a rigidly mounted sensor. Drift is an age-related change in sensor sensitivity. Taking this drift into account when verifying the sensor prevents sensors from being declared faulty: sensors that only age within a predicted and generally accepted framework, but remain reliable and usable. This thus prevents excessive scrap.

[0023] In an extended embodiment of the invention, it is proposed that at least one electrochemical facility within a facility group comprising multiple electrochemical facilities uses modeled operating parameters, rather than operating parameters measured by sensing, as virtual operating parameters. With this preferred embodiment, physical sensors can be omitted within the facility group, and preferably, for safety reasons, the model is verified on a small number of facilities using real, robust, and / or temporarily installed sensors. If an electrochemical facility within the facility group is at least temporarily equipped with a sensor, that sensor is generally also available for use by other facilities to implement the steps of the method according to the invention. For example, a sensor on one facility can be used to verify a model on another facility, or a fault in a sensor on one facility can be verified using a model running on another facility. It is noteworthy that a facility group is by no means limited to a large number of facilities with a central control unit. The term remains limited to multiple facilities of similar type or inspection of the same facility by the same operator. More precisely, a facility group can be understood as any number of electrochemical facilities that can exchange data, for example, via the cloud, and where uniform modeling and verification are applied.

[0024] Furthermore, a simulation computer is proposed for implementing the steps of the method according to the invention, wherein the simulation computer is configured to perform simulations during the ongoing operation of an electrochemical facility. Here, the simulation computer may be a separate physical computing unit, or it may be reserved memory within a control computer, in which case write access to the memory configured for controlling the facility by software using that reserved memory is excluded. In this way, it is ensured that the method according to the invention does not hinder the normal operation of the electrochemical facility, maintains the integrity of the facility's control, and, in particular, the issued operation activation / review certificate remains retained.

[0025] Furthermore, the simulation computer has read access to the memory in the control computer, which implements the operating software of the electrochemical facility. This ensures that actual operating data is available for use by the method according to the invention. Here, direct read access to the memory avoids transmission delays, ensuring that the simulation computer is always provided with the same current operating data as the operating software. Attached Figure Description

[0026] Exemplary embodiments of the method according to the present invention are described below with reference to the accompanying drawings.

[0027] Figure 1 A schematic diagram showing the preferred information flow in an electrochemical facility is shown.

[0028] Figure 2 A schematic diagram illustrating another preferred information flow for an electrochemical facility. Detailed Implementation

[0029] Figure 1 The information flow in an electrochemical facility 1 according to the present invention is illustrated. During operation, operational data is sometimes collected within the electrochemical facility 1 by multiple sensors 2. This operational data is transferred to a control computer 3 (on which operating software runs) and reflects the actual operating state, based on which the operating software controls the ongoing operation of the electrochemical facility 1. A computing unit 4 is connected to the control computer 3. The computing unit 4 has read access to the control computer 3 and reads the operational data located there. Therefore, the computing unit obtains the same actual operating state as the control computer 3 prior to it. Based on this actual operating state, a simulation of the process or corresponding subprocess carried out in the respective electrochemical facility 1 is created on the computing unit 4.

[0030] Next, within this simulation, sensor 6, which can be omitted, is represented as model 6' within the modeling framework. The goal of modeling is for computing unit 4 to accurately calculate the data of sensor 6, which can be omitted, based on the actual operating state of the simulation. This is ensured consistently through model validation. To this end, a reference operating state is first identified, which is applied to the scale of validation. In this computing unit, model 6' is decoupled from the changes induced by the simulation and is now used again as model 6' to be tested during the validation process.

[0031] During validation, the results of model 6" are compared with the results of sensor 6 used for inspection. Here, sensor 6 can be securely mounted or temporarily installed in the electrolysis facility 1 for this purpose only. If model 6" produces results consistent with the data from sensor 6 used for inspection in all previously identified reference operating states, the model is considered validated. From this point onward, the validated model 6" can be provided as a virtual sensor, and its results can be provided as virtual operating parameters, rather than the corresponding sensor data from the control unit 3 during operation. In this case, model 6" is loaded onto a single control unit 3, which thereafter uses the virtual operating parameters provided by model 6" instead of the data from sensor 6, which can be omitted. The sensor, which can be omitted, is no longer used and can be omitted. In another preferred embodiment, model 6" is used to monitor sensor 6, for example, to monitor the drift of sensor 6. In this case, sensor 6" remains mounted, and the control unit 3 still uses the data from that sensor.

[0032] Instead of the same electrochemical facility 1, the described processing method can also be divided into multiple different electrochemical facilities 1.1, 1.2, 1.3 within a facility group (see reference numerals in parentheses). When implemented within a facility group, model 6' is loaded via network 5 after creation onto the corresponding computational unit 4 of at least one electrolysis facility 1.2 for verification as model 6' to be verified, and after verification, it is instantiated and used as verified model 6'' in the computational unit 4 of at least one electrolysis facility 1.3.

[0033] Figure 2 Another exemplary implementation is shown. In this additional exemplary implementation, the already verified model 6"', which is running in the control unit 3, is temporarily checked. For this purpose, a sensor 6 is temporarily installed on the electrochemical facility 1 for checking. Simultaneously, a model copy 6"' of model 6"' is applied to the computing unit 4. While the control unit 3 of the electrochemical facility 1 is still running with the data from model 6"', the computing unit 4 reads the data from the sensor 6 installed for checking purposes and compares the obtained data with the data from the model copy 6"'.

Claims

1. A method for operating at least one electrochemical facility (1), particularly an electrochemical facility for producing hydrogen, characterized in that, At least one operating parameter is modeled on a simulation computer (4), the operating parameter being related to the operation of the electrochemical facility (1) and detectable by means of a sensor (6), wherein the simulation computer (4) reads operating data from a control computer (3) for controlling the electrochemical facility (1) and uses the operating data to reflect the actual operating state.

2. The method according to any one of the preceding claims, characterized in that, The modeled operating parameters are provided as virtual operating parameters to the operating software, which is implemented on the control computer (3) of the same or another electrochemical facility (1).

3. The method according to any one of the preceding claims, characterized in that, The operating parameters are detected by the sensor (6), and the modeling of the operating parameters is checked and / or verified by comparing the data with that of the sensor (6).

4. The method according to claim 3, characterized in that, The sensor (6) is installed in the electrochemical facility (1) before the verification and / or validation, and is removed again after the verification and / or validation.

5. The method according to any one of the preceding claims, characterized in that, At least one reference operating state is specified, and the modeling of the virtual operating parameters is checked and / or verified based on the reference operating state.

6. The method according to claim 5, characterized in that, In the case of multiple reference operating states and / or pending verification or validation cycles, the verification and / or validation are parallelized by distributing the verification and / or validation across multiple electrochemical facilities (1).

7. The method according to claim 5 or 6, characterized in that, If the reference operating state - It is harmful to the material. - Significantly hinders ongoing operations, and / or - This only occurs rarely during the normal operation of the electrochemical facility. Then, the checks and / or verifications based on the reference operating conditions are performed on the test bench.

8. The method according to any one of the preceding claims, characterized in that, The modeling is used to verify the sensor (6).

9. The method according to claim 8, characterized in that, The modeling is used to predict the drift of the sensor (6), and the verification of the sensor (6) includes checking the compliance of the drift.

10. The method according to any one of the preceding claims, characterized in that, At least one electrochemical facility (1) within a facility group comprising multiple electrochemical facilities (1) uses modeled operating parameters, rather than operating parameters measured by sensing, as virtual operating parameters.

11. A simulation computer (4) for implementing the steps of the method according to the invention, wherein, The simulation computer (4) is configured to perform the simulation during the ongoing operation of the electrochemical facility (1).

12. The analog computer (4) according to claim 11, characterized in that... Read access to the memory in the control computer (3), the memory implementing the operating software of the electrochemical facility (1).

Images