Rapid online prediction method for state of health and electrochemical performance of proton exchange membrane fuel cell

Abstract

The invention relates to a proton exchange membrane fuel cell health state and electrochemical performance rapid online prediction method, and belongs to the technical field of cell life prediction, and the method comprises the following steps: S1, employing a DRT analysis method to analyze a Nyquist diagram of a galvanic pile, and calculating the impedance of each reaction process in the galvanic pile; s2, according to correlation analysis of the resistance and the voltage, establishing a relationship between ECM recession model parameters and the voltage, and estimating the health state of the fuel cell; s3, obtaining each reaction condition in the galvanic pile by using data from continuous online monitoring; the voltage of the galvanic pile is predicted through a hybrid neural network model, all reaction resistance values are obtained through extrapolation, and the electrochemical reaction related process in the galvanic pile is known in advance; and carrying out on-line estimation on the health state of the fuel cell.

Description

Technical Field

[0001] The invention belongs to the technical field of battery life prediction and relates to a method for rapid online prediction of the health status and electrochemical performance of a proton exchange membrane fuel cell. Background Art

[0002] Proton exchange membrane fuel cells have become an important pillar in the future energy field with their unique advantages, such as high energy conversion efficiency and zero pollutant emissions. In the development of fuel cell technology, life prediction has become a crucial research direction. Advanced life prediction models and technologies can not only more accurately assess the remaining service life and health status of fuel cells, but also provide early warnings of potential failures and performance degradation. This means that researchers or operators can perform necessary maintenance and replacement in advance based on the prediction results, thereby avoiding the impact of unexpected downtime or performance degradation on system stability and reliability.

[0003] From the perspective of the parameters obtained from the stack, the evaluation methods can be divided into two types: voltage-based and impedance-based. There are many studies that use voltage as a measurable quantity and use it as a health indicator for the stack. Most studies on voltage-based fuel cell life prediction and health status evaluation are based on the fact that voltage data is easy to obtain online and simple to operate. However, voltage analysis can only analyze the overall performance of the stack, and cannot identify the contribution and degradation degree of each electrochemical reaction inside the stack. Therefore, the impedance of the electrochemical characterization parameter inside the stack is introduced as a new evaluation indicator. Usually, the impedance spectrum data obtained from the electrochemical test of the fuel cell is presented in the form of a Nyquist diagram. This diagram can intuitively show the impedance changes of the system at different frequencies, help analyze the resistance and capacitance characteristics of the electrochemical process, and the charge transfer process and polarization process inside the battery. However, there are certain challenges in parsing the Nyquist diagram. Since the data points on the diagram are relatively scattered, it is difficult to extract key information. Therefore, other statistical or visualization tools are needed for data analysis and processing. A common visualization tool is an equivalent circuit model, which uses a combination of one or more resistors and capacitors to represent the response of the system at different frequencies to simplify the analysis and understanding of complex electrochemical processes. However, the components and circuit structures used in ECM require researchers to select and arrange them based on experience, and the parameters in the model (such as resistance, capacitance, etc.) usually need to be determined through experiments and complex parameter identification methods. EIS measurements usually require steady-state operating conditions and a certain test time, so they cannot be used online, and real-time stack impedance data cannot be obtained. Summary of the invention

[0004] In view of this, an object of the present invention is to provide a method for rapid online prediction of the health status and electrochemical performance of a proton exchange membrane fuel cell.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] A method for rapid online prediction of the health status and electrochemical performance of a proton exchange membrane fuel cell comprises the following steps:

[0007] S1: Use the DRT analysis method to analyze the Nyquist diagram of the fuel cell stack and calculate the impedance of each reaction process inside the fuel cell stack;

[0008] S2: Based on the correlation analysis between resistance and voltage, the relationship between ECM degradation model parameters and voltage is established to estimate the health status of the fuel cell;

[0009] S3: Use data from continuous online monitoring to obtain the reaction conditions inside the stack; predict the stack voltage through a hybrid neural network model, extrapolate the resistance values ​​of each reaction, and know the relevant processes of the electrochemical reactions in the stack in advance; and make online estimates of the health status of the fuel cell.

[0010] Further, step S1 specifically includes the following steps:

[0011] S11: Verify the Kramers-Kronig relationship of the experimentally obtained EIS data.

[0012] S12: The complex impedance Z(ω) of the electrochemical process is converted into a distribution function g(τ) about the relaxation time, and the size, position and related impedance characteristics of each electrochemical process are analyzed; the relationship between the distribution function g(τ) about the relaxation time and the complex impedance Z(ω) is expressed as:

[0013]

[0014] Where R0 represents the ohmic resistance of the fuel cell, R pol represents the total polarization impedance of the fuel cell, τ = RC, is the time constant of a single R / / C, and f is the test frequency; g(τ) in equation (1) is written as γ(lnτ), as shown in equation (2):

[0015]

[0016] S13: The PEMFC stack is equivalent to a circuit consisting of three R / / CPE circuits and one ohmic resistor in series; the multi-peak curve obtained by DRT analysis is integrated; the area under each peak is calculated in turn to obtain the impedance values ​​of the anode activation reaction, the cathode activation reaction and the mass transfer reaction.

[0017] Further, the step S2 of establishing the relationship between the ECM decay model parameters and the voltage specifically includes the following steps:

[0018] S21: The ECM constructed based on the DRT spectrum includes four resistors Rohm, R1, R2, R3 and three constant phase elements CPE1, CPE2, CPE3, where Rohm represents the ohmic resistance of the fuel cell, the parallel circuit composed of R1 and CPE1 represents the anode activation loss part in the electrochemical reaction process, and R2 and R3 are connected in parallel with CPE2 and CPE3 respectively to represent the activation loss and mass transfer loss related to the cathode in the fuel cell reaction process;

[0019] S22: Perform a linear fit between the resistance obtained from the DRT analysis and the actual voltage of the stack.

[0020] Furthermore, the step of predicting the stack voltage in step S3 includes:

[0021] S31: Determine the appropriate time step: set the time window to 5-20, and predict the next moment data of the time window value;

[0022] S32: Divide the measured voltage data into groups of 10 hours each as input, and output the 11th hour of each group;

[0023] S33: Predict in a cyclic manner, using the known data from n to n+9 hours to predict the voltage at the n+10th hour, using the known data from n+1 to n+9 hours and the predicted voltage value at the n+10th hour to predict the voltage at the n+11th hour, using the known data from n+2 to n+9 hours and the predicted voltage values ​​at the n+10th and n+11th hours to predict the voltage at the n+12th hour, and so on, to obtain voltage prediction data from n+10 hours to n+19 hours;

[0024] S34: If the predicted voltage value does not appear to be lower than the voltage threshold, it is considered that the fuel cell stack can continue to be tested within the future time step setting range; the actually measured voltage values ​​from n+10 hours to n+19 hours are used as known data, and the method of step S33 is used to predict from n+20 hours to n+29 hours until the predicted voltage appears to be lower than the voltage threshold.

[0025] The beneficial effect of the present invention is that the present invention can accurately distinguish the coupled kinetic reactions in the PEMFC reaction process and calculate their impedance values ​​without the need for a priori models, while saving a lot of parameter adjustment time. By using data from continuous online monitoring, the impedance spectrum of the stack at future moments is predicted, the electrochemical reaction of the stack is analyzed, and the root causes that affect the service life or health of the stack are explored, which is expected to intervene in advance. It has broad application prospects in the fields of proton exchange membrane fuel cell life prediction and health status estimation.

[0026] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0028] Figure 1 The overall flow chart of the online prediction method for the health status of proton exchange membrane fuel cells;

[0029] Figure 2 The following is a graph of EIS data analysis results based on the DRT method; (a) is the value of the regularization parameter λ in the DRT analysis (taking 515 hours as an example), and (b) is the DRT analysis result of the FC1 stack;

[0030] Figure 3 Further analysis of the EIS data based on the DRT method;

[0031] Figure 4 is the equivalent circuit diagram of PEMFC;

[0032] Figure 5 is the Nyquist diagram after DRT analysis;

[0033] Figure 6 is a relationship diagram between resistance parameters and voltage; (a) is the prediction and fitting result of R2, (b) is the prediction and fitting result of R3, (c) is the prediction and fitting result of total polarization impedance Rpol, and (d) is the prediction and fitting result of ohmic impedance Rohm;

[0034] Figure 7 Nyquist diagram based on voltage linear fitting, Nyquist diagram obtained by DRT analysis and original Nyquist diagram;

[0035] Figure 8 Predicting structures for hybrid neural networks;

[0036] Fig. 9 Nyquist plot of the original test EIS data based on the experimental test voltage, the predicted voltage linear fit. DETAILED DESCRIPTION

[0037] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0038] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0039] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0040] The present invention provides an online prediction method for the health status of a proton exchange membrane fuel cell, comprising the following steps:

[0041] 1) This embodiment uses the data provided by the 2014 IEEE PHM Data Challenge to conduct experiments.

[0042] The experimental test stack used includes 5 single cells, model BZ-100, and the membrane electrode assembly of the cell is composed of Gore Primea5761 membrane and SGL 10BB graphite paper as a gas diffusion layer. The active area of ​​a single cell is 100 cm^2. The nominal current density of the battery is 0.7A / cm^2, and the maximum current density is 1A / cm^2.

[0043] The battery pack was subjected to aging experiments. The battery stack operating at a nominal current of 0.7A / cm^2 was called FC1, and the battery stack with a current density of 0.7A / cm^2 and 10% oscillation at a frequency of 5kHz was called FC2.

[0044] The other experimental operating conditions were controlled to remain unchanged, and the experimental data were collected for analysis. For the collection of EIS data, FC1 and FC2 were both performed at 10mHz to 50kHz at a current density of 0.7A / cm^2.

[0045] 2) The DRT analysis method is used to analyze the Nyquist diagram of the fuel cell stack and calculate the impedance of each reaction process inside the fuel cell stack.

[0046] The steps to calculate the impedance of each reaction process inside the battery stack include:

[0047] 2.1) The Kramers-Kronig relationship of the EIS data obtained from the FC1 experiment was verified to ensure the physical credibility of the EIS data.

[0048] 2.2) The complex impedance Z(ω) of the electrochemical process is converted into a distribution function g(τ) about the relaxation time, and the size, location and related impedance characteristics of each electrochemical process are analyzed. The relationship between the distribution function g(τ) about the relaxation time and the complex impedance Z(ω) is shown in formula (1):

[0049]

[0050] Where R0 represents the ohmic resistance of the fuel cell, R pol represents the total polarization impedance of the fuel cell, τ = RC, is the time constant of a single R / / C, and f is the test frequency. g(τ) in the above formula can also be written as γ(lnτ), as shown in formula (2) and formula (3):

[0051]

[0052] γ(τ)=τg(τ) (3)

[0053] The DRT method is used to analyze EIS data. The analysis results under different regularization parameter values ​​are as follows: Figure 2 As shown in (a)-(b) in the figure, λ is finally determined to be 10 -3 Further in-depth analysis revealed Figure 3 The analysis results shown in the figure show that the Nyquist diagram obtained from the experimental data can be approximately equivalent to three series R / / CPE circuits. The green part in the figure represents the anode activation loss of the fuel cell, and the blue and red semicircles represent the cathode activation loss and mass transfer loss of the fuel cell respectively. The PEMFC stack equivalent circuit model (ECM) is as follows Figure 4 shown.

[0054] 2.3) The multi-peak curve after DRT analysis is integrated and the area under each peak is calculated to be the impedance value of each reaction.

[0055] Table 1 includes some results of DRT analysis, including the intersection of high-frequency data and the real axis Rohm, and the intersection of low-frequency data and the real axis Rpol extracted from the original experimental test EIS data. ΔRpol and ΔRohm are the percentage deviations between the true values ​​of Rpol and Rohm and the analyzed values. The fitting accuracy of the total polarization impedance is higher than 98%, and the fitting accuracy of the ohmic impedance is 98.5%, indicating that the Nyquist diagram obtained by DRT analysis can accurately fit the key parameters of the stack reaction.

[0056] Table 1

[0057]

[0058]

[0059] By combining the obtained component data with the established ECM model, the Nyquist diagram after DRT analysis can be obtained, as shown in Figure 5 From the red part, we can see that DRT analysis can accurately describe the electrochemical behavior inside the battery stack by connecting multiple R / / CPE circuits in series without a priori model.

[0060] 3) Based on the correlation analysis between resistance and voltage, the relationship between ECM degradation model parameters and voltage is established to estimate the health status of the fuel cell.

[0061] 3.1) By fitting the relationship between the first five groups of voltages and the key resistance parameters of the ECM, a linear equation was obtained, and the accuracy of the linear fitting was greater than 85%, such as Figure 6 As shown in (a)-(d) in .

[0062] 3.2) According to the fitted linear equation and the voltage data of FC1 at 648h, 823h and 991h, the resistance values ​​of each reaction part at 648h, 823h and 991h can be obtained. Figure 7 The red scatter points are shown in the figure, and the resistance values ​​of each part fitted by the DRT method are shown in the green scatter plot.

[0063] 3.3) If Figure 7 As shown in the figure, the blue line is the EIS data obtained from the experiment, the green scattered points are the Nyquist plot of ECM obtained according to DRT analysis, and the red scattered point Nyquist plot is fitted by the resistance value obtained by linear calculation of the voltage value. The specific fitting data is shown in Table 2. Figure 6From Table 2, we can see that the resistance value calculated by the linear model is fitted to the Nyquist plot, and the fitting effect is significant at high frequencies, but there is a slight deviation at low frequencies. The calculation proves that the deviation between the ohmic resistance Rohm of the battery stack obtained by fitting and the experimental data Rohm does not exceed 2%, and the deviations of the total polarization impedance Rpol are 0.0193%, 0.9093% and 2.0201%, respectively, all below 3%.

[0064] Table 2

[0065]

[0066] 4) Use the hybrid neural network model to predict the voltage at the future moment. The steps of predicting the voltage at the future moment include:

[0067] 4.1) Determine the appropriate time step: Set the time window to 10, and predict the data of the next moment of the time window value, and use known data for verification.

[0068] 4.2) Divide the known 550 hours of voltage data into groups of 10 hours each as input, and output the 11th hour of each group, for a total of 540 groups of input and output.

[0069] 4.3) Use a cyclic method to predict, using the known data from 541 to 550 hours to predict the voltage at the 551st hour, using the known data from 542 to 550 hours and the predicted voltage value at the 551st hour to predict the voltage at the 552th hour, using the known data from 543 to 550 hours and the predicted voltage values ​​at the 551st and 552nd hours to predict the voltage at the 553rd hour, and so on, to obtain voltage prediction data from 551 to 560 hours. If the predicted voltage value does not fall below the voltage threshold, it is considered that the stack can continue to be tested within the set range of future time steps. Use the actual measured voltage values ​​from 551 to 560 hours as known data, and use the above method to predict from 561 to 570 hours until the predicted voltage falls below the voltage threshold. The prediction process is as follows: Figure 8 As shown in (a) in the figure, the prediction results are as follows Figure 8 As shown in (b) in .

[0070] 4.4) Predict the impedance spectrum at a specified time, such as 666h, 830h and 1016h. Extract the voltage data at these three times in the FC2 aging experiment for linear calculation, and the obtained resistance value can be used to obtain the corresponding Nyquist diagram. Similarly, the predicted voltage can also be used to obtain the corresponding Nyquist diagram after linear calculation, such as Fig. 9 shown.

[0071] 4.5) The parameters of each part obtained by linear fitting based on the predicted voltage are shown in Table 3. It can be seen that in the low frequency band, the maximum total polarization impedance difference of the fitting curve is 6.8044%, while the maximum ohmic impedance fitting difference in the high frequency band is 1.1471%.

[0072] Table 3

[0073]

[0074] 4.6) According to the error calculation formula (4-6) provided by the 2014 IEEE PHM Challenge, the Euclidean distance between the experimental impedance and the predicted impedance at the specified frequency is first calculated, and then the calculated Euclidean distance is weighted with 550 hours to calculate the error Err at each specific time. t Finally, the final score Score is given by the weighted average of the errors SOH .

[0075]

[0076] Specific parameters include:

[0077] Fr∈{0.05Hz,0.789Hz,5.18Hz,505Hz}: specifies the frequency characteristics to be estimated;

[0078] t∈{666h,830h,1016h}: specifies a specific estimated time point;

[0079] ActSoH Fr / t ∈{ReZ Fr / t ,Im Fr / t}: Real and imaginary parts of complex impedance extracted from EIS data;

[0080] Predicted value of complex impedance

[0081] The final score calculated by this method is Score SOH =6.76948×10 -6 . The parameter fitting errors of the method proposed in this study are kept within a reasonable range, which fully verifies its effectiveness and reliability. This method allows the establishment of a regression model through historical data under the premise of an unknown precise circuit model of the fuel cell, and then uses the voltage data of continuous online monitoring to deeply analyze the dynamic process of each electrochemical reaction inside the fuel cell. This capability is of great significance for accurately diagnosing the potential degradation mechanism of the fuel cell, predicting its performance change trend, and taking immediate intervention measures, providing strong technical support for improving the maintenance efficiency of the fuel cell and extending its service life.

[0082] In the above embodiments, the description's reference to "this embodiment" indicates that a particular feature, structure, or characteristic described in conjunction with the embodiment is included in at least some embodiments, but not necessarily all embodiments. Multiple occurrences of "this embodiment" do not necessarily all refer to the same embodiment.

[0083] In the above-described embodiments, although the invention has been described in conjunction with specific embodiments of the invention, many substitutions, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. For example, other storage structures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed. Embodiments of the invention are intended to encompass all such substitutions, modifications, and variations that fall within the broad scope of the appended claims.

[0084] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, any one of the methods in this embodiment is implemented.

[0085] This embodiment also provides an electronic terminal, including: a processor and a memory;

[0086] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.

[0087] The computer-readable storage medium in this embodiment can be understood by ordinary technicians in this field: all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

[0088] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes each step of the above method.

[0089] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0090] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0091] The present invention can be used in many general or special computing system environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like.

[0092] The present invention may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A method for rapid online prediction of the health status and electrochemical performance of a proton exchange membrane fuel cell, characterized in that: The following steps are involved: S1: Use the DRT analysis method to analyze the Nyquist diagram of the fuel cell stack and calculate the impedance of each reaction process inside the fuel cell stack; S2: Based on the correlation analysis between resistance and voltage, the relationship between ECM degradation model parameters and voltage is established to estimate the health status of the fuel cell; S3: Use data from continuous online monitoring to obtain the reaction conditions inside the stack; predict the stack voltage through a hybrid neural network model, extrapolate the resistance values ​​of each reaction, and know the relevant processes of the electrochemical reactions in the stack in advance; and make online estimates of the health status of the fuel cell.

2. The method for rapid online prediction of the health status and electrochemical performance of a proton exchange membrane fuel cell according to claim 1, characterized in that: Step S1 specifically includes the following steps: S11: Verify the Kramers-Kronig relationship of the experimentally obtained EIS data. S12: The complex impedance z(ω) of the electrochemical process is converted into a distribution function g(τ) about the relaxation time, and the size, position and related impedance characteristics of each electrochemical process are analyzed; the relationship between the distribution function g(τ) about the relaxation time and the complex impedance Z(ω) is expressed as: Where R0 represents the ohmic resistance of the fuel cell, R pol represents the total polarization impedance of the fuel cell, τ = RC, is the time constant of a single R / / C, and f is the test frequency; g(τ) in equation (1) is written as γ(lnτ), as shown in equation (2): S13: The PEMFC stack is equivalent to a circuit consisting of three R / / CPE circuits and one ohmic resistor in series; the multi-peak curve obtained by DRT analysis is integrated; the area under each peak is calculated in turn to obtain the impedance values ​​of the anode activation reaction, the cathode activation reaction and the mass transfer reaction.

3. The method for rapid online prediction of the health status and electrochemical performance of a proton exchange membrane fuel cell according to claim 1, characterized in that: Step S2 of establishing the relationship between the ECM decay model parameters and the voltage specifically includes the following steps: S21: The ECM constructed based on the DRT spectrum includes four resistors Rohm, R1, R2, R3 and three constant phase elements CPE1, CPE2, CPE3, where Rohm represents the ohmic resistance of the fuel cell, the parallel circuit composed of R1 and CPE1 represents the anode activation loss part in the electrochemical reaction process, and R2 and R3 are connected in parallel with CPE2 and CPE3 respectively to represent the activation loss and mass transfer loss related to the cathode in the fuel cell reaction process; S22: Perform a linear fit between the resistance obtained from the DRT analysis and the actual voltage of the stack.

4. The method for rapid online prediction of the health status and electrochemical performance of a proton exchange membrane fuel cell according to claim 1, characterized in that: The step of predicting the stack voltage in step S3 includes: S31: Determine the appropriate time step: set the time window to 5-20, and predict the next moment data of the time window value; S32: Divide the measured voltage data into groups of 10 hours each as input, and output the 11th hour of each group; S33: Predict in a cyclic manner, using the known data from n to n+9 hours to predict the voltage at the n+10th hour, using the known data from n+1 to n+9 hours and the predicted voltage value at the n+10th hour to predict the voltage at the n+11th hour, using the known data from n+2 to n+9 hours and the predicted voltage values ​​at the n+10th and n+11th hours to predict the voltage at the n+12th hour, and so on, to obtain voltage prediction data from n+10 hours to n+19 hours; S34: If the predicted voltage value does not appear to be lower than the voltage threshold, it is considered that the fuel cell stack can continue to be tested within the future time step setting range; the actually measured voltage values ​​from n+10 hours to n+19 hours are used as known data, and the method of step S33 is used to predict from n+20 hours to n+29 hours until the predicted voltage appears to be lower than the voltage threshold.

Images