A fuel cell water failure diagnosis method and system based on electrochemical impedance spectroscopy

Abstract

The application discloses a fuel cell water fault diagnosis method and system based on electrochemical impedance spectroscopy, and belongs to the field of fuel cell fault diagnosis. By measuring the electrochemical impedance spectroscopy curve of the fuel cell under different water content states, curve feature points and local change features with high correlation with the water content are extracted, the relationship between the curve features and the water fault states of the fuel cell stack is established through a fuzzy C-means clustering method, and corresponding classification strategies are formulated according to the sensitive differences of the curve features to dry and wet states. The water state can be detected and classified through the local EIS curve features. The method can improve the accuracy and classification efficiency of the fuel cell stack water fault diagnosis, and guarantees the safety and performance stability of the fuel cell.

Description

Technical Field

[0001] This invention belongs to the field of fuel cell fault diagnosis technology, specifically relating to a fuel cell water fault diagnosis method and system based on electrochemical impedance spectroscopy. Background Technology

[0002] Proton exchange membrane fuel cells (PEMFCs) offer advantages such as high energy density, no CO2 or other byproducts, high energy efficiency, and long driving range, making them an ideal solution for hybrid and pure electric drive systems in the new energy field. However, due to the presence of water in the chemical reaction products of PEMFCs and the inherent humidity of the gas fed into the stack, a certain amount of water can accumulate inside the cell. Extremely high or low water content can lead to flooding and membrane drying failures, resulting in decreased stack performance and shortened lifespan. Specifically, excessive water content can clog the proton exchange membrane, affecting the reaction process; while membrane drying reduces the conductivity of the proton exchange membrane and may even cause irreversible membrane damage. Therefore, developing optimal water management strategies requires accurate assessment of whether water failures have occurred and their types to ensure stable fuel cell operation, improve fuel cell output performance, and extend stack lifespan.

[0003] CN 111199122 B provides a multi-physics-based method for diagnosing flooding faults in proton exchange membrane fuel cells (PEMFCs), comprising the following steps: establishing a three-dimensional geometric model of the fuel cell; establishing governing equations and setting physical fields for different regions of the geometric model to obtain a fault diagnosis model; based on finite element analysis of the fault diagnosis model, obtaining the cathode pressure drop curve and cell polarization curve, and determining the occurrence of a flooding fault based on the cell polarization and the rate of change of cathode pressure drop. This method proposes a pressure drop-based approach to determine flooding faults, which can improve the accuracy and reliability of flooding fault diagnosis. However, the modeling part of this method is relatively complex and requires the acquisition of many internal parameters of the fuel cell stack, which limits its widespread application.

[0004] CN 113903954 A discloses an online diagnostic testing device for water faults in hydrogen fuel cells, comprising three sets of pressure detection devices for gas inlets and outlets: an air outlet pressure sensor for detecting outlet air pressure, a hydrogen inlet pressure sensor for detecting inlet hydrogen pressure, a hydrogen outlet pressure sensor for detecting outlet hydrogen pressure, and an air inlet pressure sensor for detecting inlet air pressure. During operation, this device can effectively monitor, warn against, or prevent performance degradation caused by phenomena such as membrane dryness or flooding in the fuel cell. This method is closely integrated with practical engineering and has low cost. However, the detection parameters only include gas pressure, and the accuracy of the detection can be further improved.

[0005] CN 112117475 A discloses a fault detection method for a fuel cell water management subsystem, which realizes fault diagnosis of the hydrogen fuel cell water management subsystem based on local preservative projection and learned vector quantization neural network. This method uses artificial intelligence algorithms to diagnose faults in hydrogen fuel cells, offering advantages such as simple operation, strong scalability, and no need for specialized knowledge. However, directly employing machine learning methods requires a large amount of fault data, and the targeting of data features is relatively weak, resulting in a large workload and computational burden, with room for further efficiency improvement.

[0006] CN 116381503 A provides a method for estimating the electrochemical impedance spectroscopy of fuel cells for water management fault diagnosis, including: S1. establishing an equivalent circuit model; S2. identifying the parameters of the equivalent circuit model; S3. single-model impedance estimation; S4. impedance fusion estimation; S5. electrochemical impedance spectroscopy estimation. This invention, through a multi-model fusion impedance estimation method, can combine the advantages of each equivalent circuit model to achieve high accuracy and reliability of impedance estimation across the entire frequency range. However, this method uses an equivalent circuit model (ECM) and requires a complete AC impedance spectrum at both high and low frequencies, making it difficult to achieve online detection for fuel cell water fault diagnosis.

[0007] Existing methods for fuel cell water fault diagnosis mainly fall into three categories: water model-based methods, machine learning-based methods, and experimental result analysis-based methods. Water model-based methods establish a mechanism model of water content within the PEMFC (Potentially Equipped Metal Fuel Cell), representing the generation and transfer of water within the stack through functional relationships. However, this method suffers from difficulties in obtaining parameters, unclear mechanisms, and high computational complexity, making its implementation challenging. Machine learning-based methods treat water management fault diagnosis as a black box problem, collecting operational data such as voltage and current of the stack under fault conditions and using machine learning algorithms to mine fault characteristics. This method requires a large amount of experimental fault data. Currently, experimental result analysis-based methods are mainly based on EIS (Equivalent Information System). EIS provides rich information about the internal structure of the stack, characterizing its internal state. This avoids complex mechanism modeling processes and does not require a large amount of actual operational data. Current methods for analyzing EIS data of the fuel cell stack mainly fall into two categories: methods based on geometric features and methods based on ECM (Equivalent Circuit Model) parameters. The latter establishes an equivalent circuit model of the battery using EIS spectral data, identifies circuit model parameters, and then analyzes the fault state through parameter features. This method requires complete measurement of EIS spectral data across all frequency bands, which is time-consuming and demands high data quality. Furthermore, the fault diagnosis parameters are primarily polarization resistance and ohmic resistance, which presents limitations. Current methods based on EIS geometric features are also limited to selecting features such as high-frequency impedance points. While these methods can meet the requirements for rapid detection, they cannot effectively distinguish between similar water states.

[0008] Therefore, this invention proposes a fuel cell water fault diagnosis method based on electrochemical impedance spectroscopy (EIS). This method can classify fuel cell water by combining the high correlation between EIS curves and fuel cell stack water content, and it does not require all EIS high-frequency and low-frequency region data. Water state detection and classification can be achieved through local curve features. Furthermore, this method has low computational complexity, does not require solving complex models to obtain results, and the required stack parameters are easy to measure. This method can improve the accuracy and efficiency of fuel cell stack water fault diagnosis, which is beneficial to the safe and stable operation of vehicle fuel cell stacks. Summary of the Invention

[0009] In order to solve the technical problems existing in the background art, the present invention aims to provide a method and system for diagnosing water faults in fuel cells based on electrochemical impedance spectroscopy. According to the EIS curves under different water content states of the fuel cell stack, the correlation between the curve feature points and values ​​and the water fault state is established, so as to realize the method of judging the water fault state of the fuel cell stack through the characteristics of the EIS curve; the accuracy of fuel cell stack water fault diagnosis is improved by analyzing multiple sets of features of the electrochemical impedance spectroscopy curve.

[0010] To solve the technical problem, the technical solution of the present invention is as follows:

[0011] A method for diagnosing water faults in fuel cells based on electrochemical impedance spectroscopy, the method comprising:

[0012] S1: Under the actual operating conditions of the fuel cell engine, different water content states of the stack are obtained by controlling different parameters to simulate different water fault conditions.

[0013] S2: Electrochemical impedance spectroscopy (EIS) of the fuel cell stack under different conditions;

[0014] S3: Preprocess the collected EIS data of the fuel cell stack and extract curve features;

[0015] S4: Perform fuzzy C-means clustering on the extracted curve features to obtain curve feature clusters under five states: membrane dryness, slight membrane dryness, normal, slight flooding, and flooding.

[0016] S5: Classify the water content status of fuel cell stacks by combining boundary constraints and the Euclidean distance between different cluster centers of the data to be classified;

[0017] S6: Verify the model's classification performance using experimental data.

[0018] Furthermore, S1 specifically includes the following steps:

[0019] S11: Determine the operating current range of the fuel cell stack to include the operating range from low current to high current.

[0020] S12: Determine the appropriate operating parameters and boundary values ​​of the fuel cell stack under different operating currents. The operating parameters include inlet air humidity, coolant inlet temperature, and excess air ratio.

[0021] S13: Adjust the stack operating parameters to achieve a dry film state at low current and an electrode flooding state at high current.

[0022] S14: Obtain real-time open-circuit voltage data of the fuel cell stack, and verify the presence of a water fault state by combining the open-circuit voltage drop.

[0023] Furthermore, in step S2, the EIS data acquisition method includes:

[0024] The measuring electrodes were connected to the first five consecutive single cells and the last five consecutive single cells of the stack, and measurements were performed separately. The EIS measurement parameters used were 0.2Hz-10kHz, Galvanostatic mode, with a 5% perturbation current below 120A and a 2% perturbation current above 120A.

[0025] Furthermore, in step S3, the curve features specifically include: high-frequency impedance value and corresponding frequency, the rate of change of high-frequency impedance value of the 5 batteries, and the phase angle value and corresponding frequency at the maximum value of the charge transfer semicircular arc.

[0026] 5. The fuel cell water fault diagnosis method based on electrochemical impedance spectroscopy according to claim 1, characterized in that, in step S4, the fuzzy C-means clustering process specifically includes:

[0027] S41: Normalize the curve feature data and fuzzify the input parameters using a Gaussian membership function;

[0028] S42: The subtractive clustering algorithm is used to determine the location of the center of each cluster;

[0029] S43: Use Chebyshev distance to establish a similarity matrix:

[0030]

[0031] Furthermore, in step S42, the subtractive clustering algorithm specifically includes:

[0032] S421: For n multidimensional spatial data points (x1, x2, ..., xn) after preprocessing and dimensionality reduction... n ), for candidate cluster centers x i The density index at this location is defined as:

[0033]

[0034] Where, r a For x iThe radius of the circular region that has a significant impact on the density index at the center of the point;

[0035] S422: The density index of each sample point is calculated according to equation (1), and the point with the largest density index is defined as the cluster center c. k The density index is D ck Update the density index at each data point as follows:

[0036]

[0037] S423: Find the maximum density index at each point and select the next cluster center c. k+1 Repeat step S422;

[0038] S424: Stop iterating when the maximum value of the density index obtained after iterative calculation satisfies equation (4);

[0039]

[0040] Furthermore, step S5 specifically includes:

[0041] S51: Determine the extreme fault classification of water conditions by using the characteristic point values ​​of high-frequency impedance points and the frequency variation range;

[0042] S52: Calculate the characteristic values ​​of the corresponding curves based on membrane dryness and flooding conditions;

[0043] S53: Classify the water content status of fuel cell stacks by combining regional boundary constraints and Euclidean distances from different cluster centers.

[0044] Furthermore, in step S6, the verification experimental data used is a set of EIS measurement result curves under different currents, with only the air excess ratio and coolant temperature being changed. The current range is from 0.5 current density to 0.9 current density, with each increase being 0.1 current density.

[0045] A fuel cell water fault diagnosis system based on electrochemical impedance spectroscopy, the system comprising:

[0046] Experimental simulation module: Under the actual operating conditions of the fuel cell engine, different parameters are controlled to obtain different water content states of the fuel cell stack, simulating different water fault conditions;

[0047] Measurement module: Measures electrochemical impedance spectroscopy (EIS) of the fuel cell stack under different conditions;

[0048] Feature extraction module: preprocesses the collected EIS data of the fuel cell stack and extracts curve features;

[0049] Clustering module: Performs fuzzy C-means clustering on the extracted curve features to obtain curve feature clusters under five states: membrane dryness, slight membrane dryness, normal, slight flooding, and flooding;

[0050] Classification module: Based on boundary constraints and the Euclidean distance between different cluster centers, the water content status of the fuel cell stack is classified.

[0051] Validation module: Verify the model's classification performance using experimental data.

[0052] A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method described in any one of the preceding descriptions.

[0053] Compared with the prior art, the advantages of the present invention are as follows:

[0054] This invention provides a method and system for diagnosing fuel cell water faults based on electrochemical impedance spectroscopy (EIS). This method can classify fuel cell water conditions by combining the high correlation between EIS curves and fuel cell stack water content, without requiring all EIS high-frequency and low-frequency data. Water state detection and classification can be achieved through local curve features. Furthermore, this method has low computational complexity, does not require solving complex models, and the required stack parameters are easy to measure. This approach improves the accuracy and efficiency of fuel cell stack water fault diagnosis, contributing to the safe and stable operation of automotive fuel cell stacks. Attached Figure Description

[0055] Figure 1 A flowchart of the method of the present invention;

[0056] Figure 2 The experimental flowchart for obtaining water fault data using the method of this invention;

[0057] Figure 3 The flowchart of the method of the present invention for obtaining curve feature clusters through fuzzy C-means clustering;

[0058] Figure 4 A flowchart of fuel cell water fault analysis according to an embodiment of the present invention;

[0059] Figure 5 A diagram showing the clustering analysis results of fuel cell water faults in an embodiment of the invention;

[0060] Figure 6 A diagram of the fuel cell water fault clustering quality index (profile coefficient) of this invention embodiment;

[0061] Figure 7 Classification results of fuel cell water fault verification set in embodiments of the present invention. Detailed Implementation

[0062] The specific implementation of the present invention is described below with reference to embodiments:

[0063] It should be noted that the structures, proportions, sizes, etc. shown in this specification are only used to complement the content disclosed in the specification for those skilled in the art to understand and read, and are not intended to limit the conditions under which the present invention can be implemented. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0064] Furthermore, the terms such as "upper," "lower," "left," "right," "middle," and "one" used in this specification are merely for clarity of description and are not intended to limit the scope of the invention. Any changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention.

[0065] Example 1:

[0066] See Figure 1 This invention provides a method for diagnosing water faults in fuel cells based on electrochemical impedance spectroscopy, comprising the following steps:

[0067] S1. Under the actual operating conditions of the fuel cell engine, different parameters are controlled to obtain different water content states of the fuel cell stack, simulating different water fault conditions;

[0068] join Figure 2 The specific steps included in the fuel cell stack fault simulation experiment are as follows:

[0069] S1.1. Determine the operating current range of the fuel cell stack to include the operating range from small current to large current;

[0070] S1.2. Determine the appropriate operating parameters and boundary values ​​of the fuel cell stack under different operating currents. The operating parameters include inlet air humidity, coolant inlet temperature, and excess air ratio.

[0071] S1.3. Adjust the stack operating parameters so that the film is dry at low current and the electrodes are flooded at high current;

[0072] S1.4. Observe the real-time open-circuit voltage data of the fuel cell stack and verify the presence of a water fault state by combining the open-circuit voltage drop.

[0073] S2. Measure the electrochemical AC impedance spectrum of the fuel cell stack under different conditions;

[0074] S3. Process the collected EIS (Electrochemical Impedance Spectrum) data of the fuel cell stack and extract the curve feature points and feature values;

[0075] S4. Fuzzy C-means clustering is performed on the extracted curve features to obtain curve feature clusters under five states: membrane dryness, slight membrane dryness, normal, slight flooding, and flooding.

[0076] See Figure 3 The fuzzy C-means clustering process specifically includes:

[0077] S4.1. Normalize the curve feature data and fuzzify the input parameters using a Gaussian membership function;

[0078] S4.2. The subtractive clustering algorithm is used to determine the location of the center of each cluster;

[0079] The specific implementation process of subtractive clustering includes:

[0080] S4.2.1. For n multidimensional spatial data points (x1, x2, ..., xn) after preprocessing and dimensionality reduction. n ), for candidate cluster centers x i The density index at this location is defined as:

[0081]

[0082] Where, r a For x i The radius of the circular region that has a significant impact on the density index at the center of the point;

[0083] S4.2.2. Calculate the density index of each sample point according to equation (2), and define the point with the largest density index as the cluster center c. k The density index is D ck Update the density index at each data point;

[0084]

[0085] S4.2.3. Find the maximum value of the density index at each point and select the next cluster center c. k+1 Repeat S4.2.2;

[0086] S4.2.4. Stop iterating when the maximum value of the density index obtained after iterative calculation satisfies (4);

[0087]

[0088] S4.3. Use Chebyshev distance to establish a similarity matrix;

[0089]

[0090] S4.4. Optimize fuzzy inference rules based on clustering results.

[0091] S5. Classify the water content status of fuel cell stacks by combining boundary constraints and data to be classified using Euclidean distances between different cluster centers;

[0092] The water fault classification method specifically includes:

[0093] S5.1. Determine the extreme fault classification of water conditions by using characteristic point values ​​such as high-frequency impedance points and frequency variation range;

[0094] S5.2. Calculate the characteristic values ​​of the corresponding curves based on the membrane drying and flooding conditions;

[0095] S5.3. Classify the state characteristics based on the Euclidean distance of different cluster centers according to the classification data, and determine the water fault type.

[0096] S6. Verify the model's classification effect using experimental data.

[0097] Example 2:

[0098] See Figure 4 A detailed step-by-step diagram illustrating the proposed method for fault diagnosis of a set of fuel cell EIS data is provided, including:

[0099] S1: Design five simulation experiments of fuel cell stack water content states by controlling humidity, temperature, excess air ratio and operating current;

[0100] S2: During the experiment, the EIS data of the fuel cell stack under different water content conditions were measured, and the open circuit voltage drop data was used to verify whether a water fault condition occurred.

[0101] S3: Process the EIS experimental data and extract the required curve features, perform principal component analysis on the features, and determine the degree of correlation with the water content state;

[0102] S4: Divide the experimental data into training set and validation set, and obtain the feature clusters of different water content states in the training set data by C-means clustering;

[0103] S5: First, perform preliminary dry / wet state classification based on feature points, and then perform state classification based on the local feature weights of the curve;

[0104] S6: Verify the classification model's effectiveness by testing the classification results on the validation set.

[0105] See Figure 5 The image shows the clustering results of water fault diagnosis based on a set of fuel cell EIS data using the proposed method:

[0106] The established model can accurately cluster the training set data into curve feature clusters under five states: membrane dryness, slight membrane dryness, normal, slight flooding, and flooding. The feature clusters are far apart and have high compactness within each cluster, proving that the proposed method has good clustering quality and high classification efficiency.

[0107] See Figure 6 The figure shows the silhouette coefficient evaluation results of the clustering effect of the proposed method for water fault diagnosis based on a set of fuel cell EIS data:

[0108] The proposed method has an average cluster profile coefficient of 0.9692, which proves that the clustering effect is excellent.

[0109] See Figure 7 The experimental results of the proposed method for water fault diagnosis based on a set of fuel cell EIS data are shown in the figure. Water fault diagnosis was performed on EIS data of the fuel cell under different operating currents. There were 18 sets of data, including 17 normal sets and 1 abnormal set. The results show that the water content status of all 17 sets of normal data was successfully identified and was consistent with the experimental simulation state. The membership degree of 16 sets of data was higher than 0.75. The water fault identification rate of the proposed method was 100%, and the water status identification accuracy was 94.4%, which proves that the proposed method has extremely high accuracy.

[0110] Example 3:

[0111] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0112] To this end, embodiments of the present invention provide a storage medium storing multiple instructions that can be loaded by a processor to execute steps in a fuel cell water fault diagnosis method based on electrochemical impedance spectroscopy provided in embodiments of the present invention.

[0113] For example, this instruction can perform the following steps:

[0114] S1: Under the actual operating conditions of the fuel cell engine, different water content states of the stack are obtained by controlling different parameters to simulate different water fault conditions.

[0115] S2: Electrochemical impedance spectroscopy (EIS) of the fuel cell stack under different conditions;

[0116] S3: Preprocess the collected EIS data of the fuel cell stack and extract curve features;

[0117] S4: Perform fuzzy C-means clustering on the extracted curve features to obtain curve feature clusters under five states: membrane dryness, slight membrane dryness, normal, slight flooding, and flooding.

[0118] S5: Classify the water content status of fuel cell stacks by combining boundary constraints and the Euclidean distance between different cluster centers of the data to be classified;

[0119] S6: Verify the model's classification performance using experimental data.

[0120] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0121] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0122] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0123] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

[0124] Many other changes and modifications can be made without departing from the concept and scope of this invention. It should be understood that this invention is not limited to the specific embodiments, and the scope of this invention is defined by the appended claims.

Claims

1. A method for diagnosing water faults in fuel cells based on electrochemical impedance spectroscopy, characterized in that, The method includes: S1: Under the actual operating conditions of the fuel cell engine, different water content states of the stack are obtained by controlling different parameters to simulate different water fault conditions. S2: Electrochemical impedance spectroscopy (EIS) of the fuel cell stack under different conditions; S3: Preprocess the collected EIS data of the fuel cell stack and extract curve features; S4: Perform fuzzy C-means clustering on the extracted curve features to obtain curve feature clusters under five states: membrane dryness, slight membrane dryness, normal, slight flooding, and flooding. S5: Classify the water content status of fuel cell stacks by combining boundary constraints and the Euclidean distance between different cluster centers of the data to be classified; S6: Verify the model's classification performance using experimental data; S1 specifically includes the following steps: S11: Determine the operating current range of the fuel cell stack to include the operating range from low current to high current. S12: Determine the appropriate operating parameters and boundary values ​​of the fuel cell stack under different operating currents. The operating parameters include inlet air humidity, coolant inlet temperature, and excess air ratio. S13: Adjust the stack operating parameters to achieve a dry film state at low current and an electrode flooding state at high current. S14: Obtain real-time open-circuit voltage data of the fuel cell stack and verify the presence of a water fault state by combining the open-circuit voltage drop. In S2, the EIS data acquisition method includes: The measuring electrodes were connected to the first five consecutive single cells and the last five consecutive single cells of the stack, and measurements were performed separately. The EIS measurement parameters used were 0.2Hz-10kHz, Galvanostatic mode, with a 5% perturbation current when the current was below 120A and a 2% perturbation current when the current was above 120A. In S3, the curve features specifically include: high-frequency impedance value and corresponding frequency, the rate of change of high-frequency impedance value of 5 batteries, and the phase angle value and corresponding frequency at the maximum value of the charge transfer semicircular arc.

2. The method for diagnosing water faults in fuel cells based on electrochemical impedance spectroscopy according to claim 1, characterized in that, In S4, the fuzzy C-means clustering process specifically includes: S41: Normalize the curve feature data and fuzzify the input parameters using a Gaussian membership function; S42: The subtractive clustering algorithm is used to determine the location of the center of each cluster; S43: Use Chebyshev distance to establish a similarity matrix.

3. The method for diagnosing water faults in fuel cells based on electrochemical impedance spectroscopy according to claim 1, characterized in that, S5 specifically includes: S51: Determine the extreme fault classification of water conditions by using the characteristic point values ​​of high-frequency impedance points and the frequency variation range; S52: Calculate the characteristic values ​​of the corresponding curves based on membrane dryness and flooding conditions; S53: Classify the water content status of fuel cell stacks by combining regional boundary constraints and Euclidean distances from different cluster centers.

4. The method for diagnosing water faults in fuel cells based on electrochemical impedance spectroscopy according to claim 1, characterized in that, In S6, the verification experimental data used is a set of EIS measurement result curves with different currents, only changing the excess air ratio and coolant temperature.

5. A fuel cell water fault diagnosis system based on electrochemical impedance spectroscopy, characterized in that, The system is used to perform the method according to any one of claims 1-4, the system comprising: Experimental simulation module: Under the actual operating conditions of the fuel cell engine, different parameters are controlled to obtain different water content states of the fuel cell stack, simulating different water fault conditions; Measurement module: Measures electrochemical impedance spectroscopy (EIS) of the fuel cell stack under different conditions; Feature extraction module: preprocesses the collected EIS data of the fuel cell stack and extracts curve features; Clustering module: Performs fuzzy C-means clustering on the extracted curve features to obtain curve feature clusters under five states: membrane dryness, slight membrane dryness, normal, slight flooding, and flooding; Classification module: Based on boundary constraints and the Euclidean distance between different cluster centers, the water content status of the fuel cell stack is classified. Validation module: Verify the model's classification performance using experimental data.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method of any one of claims 1-4.

Images