Method and apparatus for determining the stacking state of an electrolyzer cell stack

The THD analysis with a low-frequency excitation current efficiently detects aging and fault conditions in electrolyzer cell stacks, overcoming the inefficiencies of existing methods by providing rapid and accurate identification of stack health.

DE102025150772A1Pending Publication Date: 2026-06-11AVL LIST GMBH

Patent Information

Authority / Receiving Office
DE · DE
Patent Type
Applications
Current Assignee / Owner
AVL LIST GMBH
Filing Date
2025-12-05
Publication Date
2026-06-11

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

The invention relates to a method (500) for determining the stacking state (151) of an electrolyzer cell stack (200). In this method, the supply current (211) of the electrolyzer cell stack (200) is superimposed with an excitation current (111). The resulting stack voltage (131) across the electrolyzer cell stack (200) is measured. For the measured stack voltage (131), at least one actual THD value (700) related to the excitation current (111) is determined, which indicates the total harmonic distortion of the measured stack voltage (131). Based on the determined actual THD value (700), the stacking state (151) of the electrolyzer cell stack (200) is determined, and finally, the stacking state (151) is output. The invention further relates to a control device (140), a control system (100) and a computer program product in which the method (500) for determining the stack state (151) is implemented.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] The invention relates to a method for determining the stacking state of an electrolyzer cell stack by analyzing the total harmonic distortion. The invention further relates to a control device, a control system, and a computer program product in which this method is implemented.

[0002] The synthetic production of hydrogen and hydrocarbons offers a way to reduce or even replace the use of fossil fuels for energy generation. Corresponding production processes are based on the principle of electrolysis. For example, water vapor can be split into hydrogen and oxygen in an electrolyzer cell by applying an electric current. Industrial electrolysis processes often use high-temperature electrolysis, employing a large number of stacked solid oxide electrolyzer cells (SOECs). These are essentially solid oxide fuel cells (SOFCs) operating in reverse.

[0003] Solid oxide electrolyzer cells operate most efficiently at temperatures between 500°C and 900°C. The electrolyzer cells in a stack are therefore exposed to high thermal, chemical, and mechanical stresses. This can lead to aging and fatigue processes in the cell components and parts used, such as the cell membrane or the electrodes. Furthermore, electrolyzer cells have a complex structure and control system, which can cause problems with mass flows into and out of the cells, making the electrolysis process inefficient.

[0004] Early detection of a problematic operational or health condition of the electrolyzer cell stack is therefore desirable in order to identify and avoid or counteract failures and inefficiencies.

[0005] Electrochemical impedance spectroscopy (EIS) and total harmonic distortion (THD) analysis are known techniques from the state of the art, which can be used to derive information about the health and functionality of fuel cell stacks.

[0006] Both analysis methods are based, among other things, on the fact that electrochemical systems, such as fuel cells, are operated in the linear region of their current / voltage characteristic. When operated in the nonlinear region, an output signal exhibits higher harmonics in response to an input signal. For example, the stack polarization curve already shows small nonlinearities in a fully functional state, but if a stack is degraded, higher harmonics would be generated even though the stack is operating in the linear region.

[0007] The EIS (Energy Information System) uses a measurement of the complex impedance of a fuel cell stack over a wide frequency range (often several decades). Specifically, alternating currents and voltages at various frequencies are generated and measured as input signals. To identify problematic conditions in the fuel cell stack, such as a defective membrane or electrode, the measured impedances are compared with reference curves, each corresponding to a healthy state. The input signals always have a low amplitude to assume quasi-linear system behavior. Therefore, the EIS deliberately ignores non-linear components of the measured output signal, meaning it cannot be directly determined whether a measured system response is linear or not.

[0008] Compared to EIS, THD analysis allows for the identification and analysis of nonlinearities in an electrochemical system. The fundamental principle of THD analysis is to identify deviations of the electrochemical system from expected nonlinear behavior in order to infer problematic conditions. For this purpose, system signal responses to sinusoidal input signals of varying frequencies are measured. Typically, similar to EIS, a frequency band between 0.01 Hz and 10 kHz is traversed. The output signals are examined for the presence of harmonic frequencies of a fundamental frequency, which are always integer multiples of the fundamental frequency. Specifically, in THD analysis, nonlinearity is characterized by evaluating the amplitudes of a number of harmonics relative to the amplitude of the fundamental frequency.For this purpose, the measured signal can be transformed from the time domain to the frequency domain using FFT in order to determine the magnitude of a frequency component of the output signal.

[0009] Disadvantages of the aforementioned analysis methods include their time-consuming and measurement-intensive nature. This stems primarily from the fact that a system must always be excited with various excitation signals from a wide frequency range, and for a meaningful measurement of the output signal, the system must first reach a steady state. The resulting measurement times, however, are impractical for repeated testing of electrolyzer cell stacks during operation.

[0010] The object of the present invention is to at least partially overcome the disadvantages described above. In particular, the object of the present invention is to detect aging and fault conditions occurring in an electrolyzer cell stack during operation quickly, accurately, robustly, and efficiently.

[0011] The foregoing problem is solved by a method having the features of claim 1, a control device having the features of claim 14, a control system having the features of claim 15 and a computer program product having the features of claim 16.

[0012] Further advantages and features of the invention will become apparent from the dependent claims, 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 control device, the control system, and the computer program product according to the invention, and vice versa, so that the disclosure of the individual aspects of the invention always refers, or can refer, to each other.

[0013] One aspect of the present invention relates to a method for determining the stacking state of an electrolyzer cell stack. The method comprises a step in which the electrolyzer cell stack is supplied with a supply current. The supply current of the electrolyzer cell stack is superimposed with an excitation current. The excitation current has at least one signal component which oscillates at an excitation frequency from a low-frequency range of 0.1 Hz to 10 Hz. The stack voltage established at the electrolyzer cell stack during superposition is measured. Furthermore, a step is performed in which at least one actual THD value, related to the excitation current, is determined for the measured stack voltage. Here, the total harmonic distortion of the measured stack voltage is expressed as a function of the actual THD value.Based on the determined actual THD value, the stacking state of the electrolyzer cell stack is determined and the stacking state thus determined is output.

[0014] In particular, the stack state is only sent as output when a ring buffer is filled with THD values ​​calculated from repeated excitation events.

[0015] Within the scope of the invention, a stack state can be understood to mean, in particular, the state of health, the physical constitution, and / or an indication regarding the proper operation of an electrolyzer cell stack. The determination can thus be made for the entire stack. The electrolyzer cell stack can, in particular, comprise a plurality of electrolyzer cells arranged in a row along a stacking direction. The electrolyzer cell stack can, for example, be a high-temperature electrolyzer cell stack (SOEC). The supply current can preferably be provided to the electrolyzer cell stack as direct current (DC). Preferably, the supply current enables electrolysis operation of the electrolyzer cell stack. The excitation current can preferably be provided to the electrolyzer cell stack as alternating current (AC).The excitation current can thus exhibit a periodic profile, varying between at least two current amplitudes within the period. The superposition of the supply current and the excitation current can be understood, in particular, as the addition of two electrical currents. The stack voltage can be understood, in particular, as the potential difference occurring across the electrodes of the electrolyzer cell stack.

[0016] The actual THD value, or total harmonic distortion, can be expressed in particular as the ratio of the summed powers P harmonisch all harmonics are calculated as the power of the fundamental frequency P0, as expressed by the following formula: THD[%]=PharmonishP0×100

[0017] The actual THD value is therefore calculated in relation to the excitation current.

[0018] In other words, the invention provides a method for the rapid, accurate, and efficient detection of unexpected nonlinearities occurring in an electrolyzer cell stack during operation. The identified nonlinearities allow for the detection of deterioration in the physical condition of the electrolyzer cell stack, such as degradation, as well as operational malfunctions. A normal stack condition can thus be distinguished from a critical stack condition based on the actual THD value. However, unlike in the prior art, detection is not achieved by imprinting excitation currents of a complete frequency sweep on the electrolyzer cell stack, but rather by applying only a single excitation current with a low excitation frequency.This directly leads to a reduction in the required processor power and memory, since a time-frequency transformation of the measured stack voltage for THD analysis only needs to be performed over a significantly smaller frequency range. The inventors also surprisingly discovered that, for electrolyzer cell stacks, the THD values ​​for different stack states show the most pronounced differences for low-frequency excitation currents, while the THD values ​​for excitation currents with excitation frequencies above 10 Hz hardly differ from each other, regardless of the stack state. This is also exemplified in [reference missing]. Fig. Figure 4 shows four THD value curves for four different stacking states as a function of the excitation frequency. Curve (1000) shows a reference value for a non-critical stacking state, while curves (1001 to 1003) illustrate critical stacking states. One reason for this could be that slower excitation, especially in sluggish systems such as chemical reactions at the electrodes, generates good system responses or resonances. Thus, the method according to the invention enables an accurate and meaningful result with significantly reduced measurement and signal processing effort. Accordingly, the throughput time, which is primarily determined by measurement time and post-processing time, can also be significantly reduced for this method.

[0019] According to a preferred embodiment, the excitation current can have only one signal component. Efficiency and time advantages can thus be further increased, since with this design the measurement and THD analysis only need to be performed for a single excitation frequency.

[0020] Furthermore, it can be advantageous if the excitation frequency is within a low-frequency range of 0.1 Hz to 5 Hz, 0.1 Hz to 1 Hz, or 0.25 Hz to 0.75 Hz. Alternatively, the excitation frequency can be 0.5 Hz. Experiments have surprisingly shown that meaningful and discriminatory actual THD values ​​for different stacking states of the electrolyzer cell stack are available for these excitation frequencies.

[0021] According to a further preferred embodiment, the supply current can be superimposed with the excitation current for at least a defined minimum excitation duration. This minimum excitation duration can be correlated with the period of at least one of the signal components of the excitation current. For example, the minimum excitation duration can be at least five times the period occurring at the excitation frequency. This ensures that the stack voltage is reliably measured in a steady state of the electrolyzer cell stack and after the elapsed stack-inherent latencies.

[0022] It is further advantageous if, according to a preferred embodiment, determining the actual THD value involves transforming the measured stack voltage into the time-frequency domain. The measured stack voltage can be transformed into the frequency domain by a Fourier transform. Furthermore, to determine the actual THD value, at least a defined number of the largest harmonic overtones can be identified, and the actual THD value can be calculated for at least the identified harmonics.

[0023] Alternatively or additionally, it is also conceivable to calculate the actual THD value using the Goertzel algorithm in relation to the excitation current.

[0024] The Goertzel algorithm is known in digital signal processing as a special form of the discrete Fourier transform. Unlike, for example, the fast Fourier transform, the Goertzel algorithm only calculates individual, discrete spectral components. This allows the Goertzel algorithm to calculate only the magnitude of the spectral components present at the excitation frequency. This enables fast, efficient, and resource-saving computation. Furthermore, the Goertzel algorithm can deliver reliable results even with fewer measurement repetitions during an excitation.

[0025] Preferably, the determination of the actual THD value using the Goertzel algorithm can be performed in parallel with a determination using the aforementioned time-frequency domain transformation. Furthermore, preferably, the lower of the two actual THD values ​​determined in this way can be selected as the determined actual THD value. This allows the calculated results of the actual THD value to be compared against each other, and corrective action can be taken in the event of significant deviations. Even if the measurement of the stacking stress was insufficient or could be performed very quickly, reliable results for the actual THD value can still be obtained.

[0026] According to a preferred embodiment, the excitation current can be an alternating current. For example, the excitation current can be a sinusoidal signal or at least exhibit a sinusoidal signal. This is advantageous because with sinusoidal excitation, the spectral components at the harmonic frequencies can be more easily separated from noise components.

[0027] According to a further preferred embodiment, the excitation current can have an excitation amplitude in the range between 1% and 10% of the amplitude of the supply current, or an excitation amplitude of at least 5% of the amplitude of the supply current. This allows, similar to EIS, excitation of the electrolyzer cell stack in the quasi-linear range.

[0028] According to a preferred embodiment, the stacking state of the electrolyzer cell stack can be determined by identifying a stacking state as critical if at least one fault condition is met with the determined actual THD value. Preferably, the fault condition can include at least one THD reference value.

[0029] Within the scope of the invention, a critical stack condition can be understood as a constitution of the electrolyzer cell stack that deviates from normal operation. For example, a critical stack condition could be the occurrence of an insufficient supply of oxygen or water vapor to the anode or cathode, or the presence of moisture in the oxygen electrode. Alternatively or additionally, degradation of the separating membrane of the electrolyzer cells or of the electrodes of the electrolyzer cell stack could constitute a critical stack condition. In this way, an efficient differentiation between various critical stack conditions can be implemented.

[0030] According to a further preferred embodiment, the method can be repeated in iterations with a defined verification cycle. The excitation current can preferably be configured identically in each iteration. The verification cycle can, for example, have a frequency of 2 minutes, 5 minutes, 10 minutes, or 15 minutes. This allows the stack state to be checked regularly and enables a rapid response to the occurrence of critical stack states. Furthermore, it is advantageous that comparable measurement results can be generated with an identically configured excitation current, so that measurement outliers or special processes have less of an impact, making the method more robust and reliable.

[0031] Preferably, the stacking state of the electrolyzer cell stack can be determined by calculating a moving average actual THD from a sequence of successive actual THD values. The number of actual THD values ​​used for averaging can be predetermined by a specified number of averaging operations. The actual THD values ​​used for averaging are preferably determined in successive iterations of the procedure. A critical stacking state can be identified if at least one error condition is met by the calculated actual THD average. This reduces the influence of measurement outliers or other exceptional values ​​of the stack voltage, making the procedure more robust and reliable.

[0032] Preferably, the calculation of the moving THD average can include a storage step. During storage, the current THD value of an iteration can be saved in a ring buffer. This ring buffer preferably has a number of indexed memory locations, preferably corresponding to the number of averaging iterations. The memory locations can preferably be filled with the currently calculated current THD value in sequential order of their indices. This allows for an efficient and resource-saving implementation for calculating the current THD average, ensuring the data remains up-to-date using simple means. The processing time of the procedure can also be further reduced. In particular, the batch state is only sent as output when the first ring buffer is filled with THD values.

[0033] Furthermore, a fault condition can preferably be met if the actual THD mean value determined in the current iteration is at least twice as large as an initial THD value. The initial THD value can preferably be the actual THD mean value resulting from averaging the actual THD values ​​determined in the first iterations until the averaging number is reached. In this way, a reference value can be automatically generated at the beginning of the process. This utilizes the assumption that the electrolyzer cell stack has a high probability of functioning correctly at the beginning of the process. Furthermore, any deterioration of the stack's condition relative to its initial state can be determined.

[0034] Alternatively or additionally, an error condition can be met if the actual THD average is greater than a defined THD threshold. For example, the THD threshold could be 10%. This allows for a simple and efficient check of the stack's condition.

[0035] Alternatively or additionally, an error condition can be met if a gradient that develops over a period of time between the actual THD mean value determined in the current iteration and at least one of a defined number of actual THD mean values ​​determined in previous iterations exceeds a particularly positive THD gradient limit. In this way, significant variations in the actual THD mean values ​​occurring over a period of time can be detected.

[0036] Preferably, it may be necessary for the aforementioned failure conditions to be met individually or together in order for a stack state to be identified or classified as critical.

[0037] According to a further preferred embodiment, the supply current can be superimposed on the excitation current during steady-state operation of the electrolyzer cell stack. In this way, the method can be carried out after the electrolyzer cell stack has settled. This reduces the amount of interference signals that would occur when measuring the stack voltage under transient conditions.

[0038] According to a preferred embodiment, in addition to the determined stack condition, a potential fault location, a warning, and / or an adjusted control parameter for controlling the electrolyzer cell stack can also be output. In this way, after the detection of unexpected nonlinearities, further deterioration or contamination of the electrolyzer cell stack can be prevented by isolating faults and intervening in the operation of the electrolyzer cell stack through adjustment of control parameters.

[0039] According to a further preferred embodiment, the excitation current can be measured in parallel and simultaneously with the stack voltage at a preferably identical measurement frequency. The measurement frequency can be at least 1000 Hz. In this way, in addition to the frequency spectra, phase shifts between the excitation current and the AC component of the stack current can also be determined, enabling further and more detailed statements about the stack condition.

[0040] Another aspect of the invention relates to a control device for determining the stacking state of an electrolyzer cell stack. The electrolyzer cell stack is supplied with a supply current, and a stack voltage can be tapped at its electrodes. The control device includes a processor adapted to regulate an excitation current via an excitation current output. The excitation current regulation is adapted such that the excitation current has at least one signal component oscillating at an excitation frequency from a low-frequency range of 0.1 Hz to 10 Hz. The processor is further adapted to measure at least the stack voltage via a measurement input during regulation. The processor is further adapted to determine at least one actual THD value for the measured stack voltage, referenced to the excitation current, wherein the actual THD value indicates the total harmonic distortion of the measured stack voltage.The processor is also adapted to determine the stack state of the electrolyzer cell stack based on the measured actual THD value and to output this determined stack state via an output. Preferably, the processor can further be adapted to adjust a control parameter for controlling the electrolyzer cell stack and to output this parameter via a control output or the output output.

[0041] Within the scope of the invention, rules can be understood in particular as controlling, controlling and adjusting physical quantities or parameters.

[0042] A further aspect of the invention relates to a control system for determining the stacking state of an electrolyzer cell stack. The control system includes an excitation current source for generating an excitation current. Furthermore, the control system includes a superposition device for superimposing the supply current of the electrolyzer cell stack with the excitation current. The control system also includes a measuring device for measuring the stack voltage of the electrolyzer cell stack, which is established during superposition. Additionally, the control system includes the control device described above.

[0043] Another aspect of the invention relates to an electrolyzer cell system comprising an electrolyzer cell stack and the aforementioned control system.

[0044] Another aspect of the invention relates to a computer program product which includes instructions which, when the program is executed by a computer, cause it to perform the method described above.

[0045] The computer program product can be implemented as computer-readable instruction code in any suitable programming language, such as Java or C++. The computer program product can be stored on a computer-readable storage medium, volatile or non-volatile memory, or on built-in memory / processor. The instruction code can program a computer or other programmable devices, such as control devices, to execute the desired functions. Furthermore, the computer program product can be made available on a network, such as the internet, from which it can be downloaded by a user as needed. The computer program product can be implemented using a computer program (i.e., software), one or more specific electronic circuits (i.e., hardware), or in any hybrid form.by means of software components and hardware components.

[0046] Another aspect of the invention relates to a computer-readable storage medium which includes instructions which, when executed by a computer, cause it to perform the method described above.

[0047] The aforementioned devices and systems can produce the same effects and advantages as those already described at the outset for the method according to the invention.

[0048] Further advantages, features, and details of the invention will become apparent from the following description, in which exemplary embodiments of the invention are described with reference to the drawings. The drawings schematically show: Fig. 1 an embodiment of a control device and a control system according to the invention, Fig. 2 an embodiment of a method according to the invention, Fig. 3A an initial memory allocation with repeatedly determined actual THD values ​​for calculating a THD initial value in an embodiment of the method according to the invention, Fig. 3B the memory out Fig. 3A in its first complete occupancy, Fig. 3C the memory allocation from Fig. 3B after filling the memory after another iteration of the procedure, and Fig. 4 Example curves of THD indices for critical stacking states and a non-critical stacking state as a function of the excitation frequency.

[0049] The figures show different views of different aspects and embodiments of the invention.

[0050] Fig. Figure 1 shows an example of a control system 100 according to the invention for determining a stack state of an electrolyzer cell stack 200.

[0051] The device to be monitored can comprise a plurality of stacked electrolyzer cells 201. During operation, the electrolyzer cell stack 200 is supplied with a supply current 211 from a power supply source 210 via its electrodes 221, 222 to carry out an electrolysis process. The power supply source 210 can, for example, be a DC power source which can be connected to the electrodes 221, 222 via power lines. Preferably, the power supply source 210 can have a control input via which the supply current 211 can be adjusted.

[0052] The control system 100 further comprises an excitation current source 110 for generating an excitation current 111. The excitation current source 110 can, for example, be a sine wave generator or an AC power source. The excitation current source 110 can be arranged separately from or integrated with the supply current source 210. Preferably, the excitation current source 110 can have a control input via which the excitation current 111 can be adjusted.

[0053] The control system 100 also includes a superposition device 120 to superimpose the supply current 211 with the excitation current 111. For example, the superposition device 120 can be a node or a signal adder. Fig. Figure 1 shows the superposition device 120 as an exemplary component of the excitation current source 110. However, the superposition device 120 can also be arranged separately from the excitation current source 110.

[0054] The control system 100 further comprises a measuring device 130 for measuring a stack voltage 131 of the electrolyzer cell stack 200. Fig. 1. The stack voltage 131 is measured as the potential difference between the tap points 133 and 134. The measuring device 130 can, for example, be a potentiometer with a digital or analog measuring output. Additionally, the control system 100 can have a current measuring device 135, from which the supply current 211 superimposed on the excitation current 111 can be measured. For example, the current measuring device 135 can be an ammeter with a digital or analog measuring output.

[0055] The control system 100 further comprises a control device 140 with a processor 147. The control device 140 can be, for example, a computing unit, a control unit, a microcontroller, or a computer.

[0056] The control device 140 has signal inputs and signal outputs to control the components of the control system 100.

[0057] The control device 140 has an excitation current output 141, which is preferably connected to the excitation current source 110 via a signal line. Furthermore, the control device 140 has a measurement input 143, which is preferably connected to the measurement device 130 via a signal line.

[0058] The processor 147 is adapted to execute the inventive methods for determining a stack state. This is exemplified in Fig. 2 is represented as a sequence of process steps 501 to 506.

[0059] In a first process step 501, the electrolyzer cell stack 200 is supplied with the supply current 211.

[0060] In a parallel or subsequent process step 502, the supply current 211 is superimposed with the excitation current 111. Preferably, this process continues until the electrolyzer cell stack 200 reaches steady state. Processor 147 is configured to control the excitation current 111 via the excitation current output 141. Specifically, processor 147 controls the excitation current 111 such that it exhibits a sinusoidally oscillating signal component with an excitation frequency within a frequency range of 0.1 Hz to 10 Hz. The resulting excitation of the electrolyzer cell stack 200 can last for a defined minimum duration before being terminated, thus generating a time-limited excitation event. Preferably, the processor 147 can set an excitation amplitude of at least 5% of the amplitude of the supply current 211.

[0061] In a parallel or subsequent process step 503, the stack voltage 131, which develops during the superposition (of the excitation event), is measured. The processor 147 is adapted to measure the stack voltage 131 via the measurement input 143 while controlling the excitation current 111. Preferably, the control device 140 can also have a stack input current measurement input 144, via which the current measured by the current measuring device 135 can be acquired and evaluated by the processor 147. Preferably, the processor 147 can acquire the current and the stack voltage 131 in parallel at a measurement rate of 1 kHz.

[0062] In a subsequent process step 504, an actual THD value 700, referenced to the excitation current 111, can be determined for the measured stack voltage 131, which indicates the total harmonic distortion of the measured stack voltage 131. For this purpose, the processor 147 can, for example, apply a Fourier transform and / or the Goertzel algorithm to the stack voltage 131. The spectral power corresponding to the excitation frequency f_0 can be compared to the power of the harmonics of the excitation frequency f_0.

[0063] In process step 505, a stack state 151 of the electrolyzer cell stack 200 is determined based on the calculated actual THD value of 700. For this purpose, the processor 147 can check whether a user-defined fault condition exists for the calculated actual THD value of 700. In the simplest case, for example, the calculated actual THD value of 700 can be compared with a THD reference value. If one or more fault conditions are met, the processor 147 can determine the presence of a critical stack state 151.

[0064] The resulting stack state 151 is then output via an output 145. For example, in Fig. Figure 1 shows an output monitor 150, on which a warning is displayed when a critical stack state 151 is detected. Alternatively or additionally, the stack state 151 can also be output to a control section 149 of the processor 147, from which, for example, a control parameter SP for the operation of the electrolyzer cell stack 200 is adjusted. The control parameter SP can then be output via a control output 142. This is shown by way of example in process step 506. In the example of the Fig. Figure 1 further demonstrates by way of example that the supply current 211 can be adjusted based on the stack state 151. In this way, for example, the electrolyzer cell stack 200 can be switched off if, under the prevailing operating conditions, excessive degradation of the electrolyzer cell stack 200 is imminent.

[0065] Preferably, the procedure 500 can be repeated in iterations with a verification interval of 2 min to 5 min, each time using an identically configured excitation current 111.

[0066] Referring back to the previously described error conditions using a THD reference value, it has proven advantageous in tests to preferably calculate an initial THD value of 890 as the THD reference value and to use actual THD mean values ​​of 800 instead of the actual THD values ​​of 700. Efficient calculation was achieved by using a ring buffer of 600. This is exemplified in the Fig. Figures 3A to 3C are shown. The ring buffer 600 shown there has a number of indexed memory locations. In each iteration of the procedure 500, these memory locations are filled sequentially (see Figure 3). Fig. 3B and Fig. 3C for actual THD value 716). The initial THD value of 890 can preferably be calculated as the mean of the initially determined actual THD values ​​701 to 710. In Fig. Figure 3B shows the subsequent (eleventh) iteration of method 500, for which a first actual THD mean value 801 is calculated as a moving average of the ten actual THD values ​​702 to 722. The invention is not limited to the number of memory locations and averaging shown, but is only intended to be illustrated in the figures. In further iterations of method 500, additional actual THD mean values ​​can be calculated, such as those shown in Fig. 3B shows actual THD mean values ​​802 and other actual THD mean values ​​indicated by dots. Fig. Figure 3C shows the calculation of the actual THD mean values ​​801 to 806 after the ring buffer 600 (with the fifteenth iteration) was completely filled for the first time.

[0067] Based on Fig. Section 4 makes it clear that the selection of the excitation frequency f_0 from a low-frequency range is not arbitrary. The in Fig. The four depicted THD value curves 1000 to 1003 show clear differences in the frequency range from 0.1 Hz to 10 Hz, making it possible to distinguish between non-critical and critical stack states 151. Furthermore, different critical stack states 151 can be differentiated from one another, thus enabling the localization of any fault occurring in the electrolyzer cell stack 200. Fig. Figure 4 shows this using a reference curve of a non-critical stack state 1000 as well as three different THD curves of critical stack states 1001 to 1003.

[0068] The preceding explanation of the embodiments describes the present invention solely by way of examples. Naturally, individual features of the embodiments can be freely combined with one another, provided this is technically feasible, without departing from the scope of the present invention. Reference symbol list 100 Control system 110 Excitation current source 111 Excitation current 120 superposition device 130 measuring device 131 Stack tension 133, 134 Tapping points 135 Current measuring device 140 Control device 141 Excitation current output 142 Control output 143 Measurement input 144 stack input current measurement input 145 Output 147 processor Section 149 150 Monitor 151 Stack state 200 electrolyzer cell stacks 201 Electrolyzer cell 210 Power supply 211 Supply power 221, 222 electrodes 500 procedures 501 to 506 Procedural steps 600 ring buffers 700 Actual THD Value 701 to 716 elements of an actual THD value sequence 800 Actual THD Average 801, 802, 803, 806 Elements of an actual THD mean sequence 890 THD initial value 1000 Reference history of a non-critical stack state 1001 to 1003 THD curves of critical stack states f_0 Excitation frequency SP control parameters

Claims

Method (500) for determining a stack state (151) of an electrolyzer cell stack (200), characterized by supplying the electrolyzer cell stack (200) with a supply current (211), superimposing the supply current (211) of the electrolyzer cell stack (200) with an excitation current (111) which has at least one excitation frequency (f_0) from a low frequency range of 0.exhibits a signal component oscillating from 1 Hz to 10 Hz, measuring the stack voltage (131) of the electrolyzer cell stack (200) which is established at the electrolyzer cell stack (200) when superimposed, determining at least one actual THD value (700) related to the excitation current (111) for the measured stack voltage (131), wherein the actual THD value (700) indicates the total harmonic distortion of the measured stack voltage (131), determining the stack state (151) of the electrolyzer cell stack (200) on the basis of the determined actual THD value (700), and outputting the determined stack state (151). Method (500) according to claim 1, wherein the excitation current (111) has only one signal component, and / or wherein the excitation frequency (f_0) has a frequency from a low frequency range of 0.1 Hz to 5 Hz, from 0.1 Hz to 1 Hz, or from 0.25 Hz to 0.75 Hz, or has a frequency of 0.5 Hz. Method (500) according to claim 1 or claim 2, wherein the supply current (211) is superimposed with the excitation current (111) for at least a defined minimum excitation duration, wherein preferably the minimum excitation duration correlates with a period of at least one of the signal components of the excitation current (111). Method (500) according to one of the preceding claims, wherein the determination of the actual THD value (700) comprises: • transforming the measured stack voltage (131) into the time-frequency domain, preferably by a Fourier transform, furthermore determining at least a defined number of the largest harmonic overtones and calculating the actual THD value (700) for at least the determined overtones, and / or • calculating the actual THD value (700) using the Goertzel algorithm in relation to the excitation current (111), which is preferably carried out in parallel with the determination by means of the time-frequency domain transformation. Method (500) according to one of the preceding claims, wherein the excitation current (111) is an alternating current and / or has a sinusoidal signal, wherein the excitation current (111) preferably has an excitation amplitude in the range between 1% and 10% of the amplitude of the supply current (211) or has an excitation amplitude of at least 5% of the amplitude of the supply current (211). Method (500) according to one of the preceding claims, wherein determining the stack state (151) of the electrolyzer cell stack (200) comprises: • Identifying a stack state (151) as a critical stack state (151) when at least one fault condition is met with the actual THD value (700), wherein the fault condition has at least one THD reference value. Method (500) according to one of the preceding claims, wherein the method (500) is preferably repeated in iterations with a defined verification clocking, preferably with an identically configured excitation current (111), wherein the verification clocking preferably has a clock frequency of 2 min, 5 min, 10 min or 15 min. Method (500) according to claim 6 and claim 7, wherein determining the stack state (151) of the electrolyzer cell stack (200) comprises: • calculating a moving actual THD mean (800) from a number of actual THD values ​​(700) defined by an averaging number, which were determined in successive iterations of the method (500), and • identifying as a critical stack state (151) if the at least one failure condition is satisfied by the calculated actual THD mean (800). Method (500) according to claim 8, wherein the calculation of the moving THD average includes a storage in which the actual THD value (700) of an iteration is stored in a ring buffer (600), wherein the ring buffer (600) has a number of indexed storage locations which preferably corresponds to the number of averaging operations. Method (500) according to claim 8 or claim 9, wherein the error condition is satisfied if: • the actual THD mean (800) is at least twice as large as an initial THD value (890), wherein the initial THD value (890) is preferably the actual THD mean (800) resulting from averaging the actual THD values ​​(700) determined in the first iterations up to reaching the iteration in the averaging number; • the actual THD mean (800) is greater than a defined THD limit, which is preferably a THD value of 10%; and / or • a gradient that develops over a period of time between the actual THD mean (800) determined in the current iteration and at least one of a defined number of actual THD mean values ​​(800) determined in previous iterations is particularly positive THD gradient limit exceeded. Method (500) according to one of the preceding claims, wherein the supply current (211) is superimposed with the excitation current (111) in a steady-state operation of the electrolyzer cell stack (200). Method (500) according to one of the preceding claims, wherein the determined stack state (151), a potential fault location, a warning and / or an adapted control parameter (SP) for controlling the electrolyzer cell stack (200) is output. Method (500) according to one of the preceding claims, wherein the excitation current (111) is preferably measured simultaneously with an identical measuring frequency in parallel with the stack voltage (131), and / or wherein the measurement is carried out with a measuring frequency of at least 1000 Hz. Control device (140) for determining a stack state (151) of an electrolyzer cell stack (200) using a method according to one of claims 1 to 13, which is supplied with a supply current (211) and from whose electrodes (221, 222) a stack voltage (131) can be tapped, characterized by a processor (147) which is adapted to control an excitation current (111) via an excitation current output (141), wherein the control of the excitation current (111) is adapted such that the excitation current (111) has at least one excitation frequency (f_0) from a low frequency range of 0.exhibits a signal component oscillating from 1 Hz to 10 Hz, - to measure at least the stack voltage (131) during the regulation of the excitation current (111) via a measurement input (143), - to determine at least one actual THD value (700) for the measured stack voltage (131) related to the excitation current (111), wherein the actual THD value (700) indicates the total harmonic distortion of the measured stack voltage (131), - to determine a stack state (151) of the electrolyzer cell stack (200) based on the determined actual THD value (700), - to output the determined stack state (151) via an output output (145), and - preferably to adjust a control parameter (SP) for controlling the electrolyzer cell stack (200) and to output it via a control output (142) or the output output (145). to spend. Control system (100) for determining a stack state (151) of an electrolyzer cell stack (200), characterized by an excitation current source (110) for generating an excitation current (111), a superposition device (120) for superimposing a supply current (211) of the electrolyzer cell stack (200) with the excitation current (111), a measuring device (130) for measuring the stack voltage (131) of the electrolyzer cell stack (200), which is established at the electrolyzer cell stack (200) during superposition, and a control device (140) according to claim 14. Computer program product comprising instructions which, when the program is executed by a computer, cause it to execute the method (500) according to any one of the preceding claims 1 to 13.