Proton exchange membrane fuel cell state evaluation method, device, equipment and medium

Abstract

The application relates to a proton exchange membrane fuel cell state evaluation method, device, equipment and medium, which is applied to the technical field of proton exchange membrane fuel cells, and comprises the following steps: collecting core operation data of a proton exchange membrane fuel cell according to a sampling interval, and extracting a core operation data sequence corresponding to a current time according to a preset sliding window. A gated recurrent neural network is used to extract features of the operation data sequence, so that time sequence hidden features are obtained. The time sequence hidden features are processed by KAN, so that a theoretical voltage is obtained. A normalized voltage residual error is calculated according to a measured voltage, the theoretical voltage and a voltage residual error scale corresponding to a current working condition of the proton exchange membrane fuel cell. Physical constraint parameters at the current time are obtained, and a physical constraint penalty term is determined based on the physical constraint parameters. The normalized voltage residual error and the physical constraint penalty term are weighted and averaged to determine a health index, and the health state of the battery is determined based on the health index. The accuracy of battery state evaluation can be improved.

Description

Technical Field

[0001] This application relates to the field of proton exchange membrane fuel cell technology, and in particular to proton exchange membrane fuel cell condition assessment methods, apparatus, equipment and media. Background Technology

[0002] Proton exchange membrane fuel cell systems hold significant promise for vehicles, backup power, distributed energy, and comprehensive hydrogen energy utilization. However, stack lifespan, condition awareness, and early anomaly detection capabilities still limit their large-scale application. As application scenarios shift from laboratory steady-state conditions to on-board dynamic loads and complex gas-water management conditions, health management strategies based solely on voltage thresholds or lifespan extrapolation are insufficient to meet real-time operation and maintenance needs.

[0003] In related technologies, voltage degradation trajectories can be fitted using data-driven methods. However, this method struggles to explain whether the residuals originate from actual degradation, load disturbances, gas path transients, or sensor noise. Therefore, it cannot accurately assess the state of a proton exchange membrane fuel cell. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides a method, apparatus, electronic device, storage medium, and computer program product for assessing the state of a proton exchange membrane fuel cell.

[0005] According to a first aspect of this application, a method for assessing the state of a proton exchange membrane fuel cell is provided, comprising: The core operating data of the proton exchange membrane fuel cell is collected according to the sampling interval, and the core operating data sequence corresponding to the current moment is extracted according to the preset sliding window. The time-series hidden features are obtained by using a gated recurrent neural network to extract features from the running data sequence; The temporal hidden features are processed using KAN (Kolmogorov-Arnold Network) to obtain the theoretical voltage; The normalized voltage residual is calculated based on the measured voltage of the proton exchange membrane fuel cell, the theoretical voltage, and the voltage residual scale corresponding to the current operating condition. Obtain the physical constraint parameters at the current moment, and determine the physical constraint penalty term based on the physical constraint parameters; The normalized voltage residual and the physical constraint penalty term are weighted and averaged to determine the health index, and the health status of the proton exchange membrane fuel cell is determined based on the health index.

[0006] Optionally, the method further includes: When collecting core operating data of proton exchange membrane fuel cells according to sampling intervals, optional operating data of proton exchange membrane fuel cells are also collected according to sampling intervals, and the optional operating data sequence corresponding to the current moment is extracted according to a preset sliding window. The process of using KAN to process the temporal hidden features to obtain the theoretical voltage includes: The theoretical voltage is obtained by processing the time-series hidden features and the optional running data sequence using KAN.

[0007] Optionally, the physical constraint parameters include: theoretical voltage, current, and air outlet pressure; The determination of the physical constraint penalty term based on the physical constraint parameters includes: Based on the theoretical voltage at the current moment, determine the voltage gradient constraint term; Based on the current at the current moment, determine the current gradient constraint term; Based on the current air outlet pressure, determine the air outlet pressure gradient constraint term; The physical constraint penalty term is obtained by weighting the voltage gradient constraint term, the current gradient constraint term, and the air outlet pressure gradient constraint term.

[0008] Optionally, the core operating data includes: voltage, current, air outlet flow rate, and air outlet pressure; the optional operating data includes: temperature, hydrogen pressure, and air inlet pressure, or one or more of these.

[0009] According to a second aspect of this application, a proton exchange membrane fuel cell state assessment device is provided, comprising: The core operating data sequence acquisition module is used to collect core operating data of the proton exchange membrane fuel cell according to the sampling interval, and extract the core operating data sequence corresponding to the current moment according to the preset sliding window; The temporal hidden feature extraction module is used to extract features from the running data sequence using a gated recurrent neural network to obtain temporal hidden features; The theoretical voltage determination module is used to process the time-series hidden features using KAN to obtain the theoretical voltage; The normalized voltage residual determination module is used to calculate the normalized voltage residual based on the measured voltage of the proton exchange membrane fuel cell, the theoretical voltage, and the voltage residual scale corresponding to the current operating condition. The physical constraint penalty term determination module is used to obtain the physical constraint parameters at the current moment and determine the physical constraint penalty term based on the physical constraint parameters; The health status determination module is used to perform a weighted average of the normalized voltage residual and the physical constraint penalty term to determine a health index, and to determine the health status of the proton exchange membrane fuel cell based on the health index.

[0010] Optionally, the device further includes: The optional operating data sequence acquisition module is used to collect optional operating data of the proton exchange membrane fuel cell according to the sampling interval when collecting the core operating data of the proton exchange membrane fuel cell according to the sampling interval, and extract the optional operating data sequence corresponding to the current time according to the preset sliding window. The theoretical voltage determination module is specifically used to process the temporal hidden features and the optional running data sequence using KAN to obtain the theoretical voltage.

[0011] Optionally, the physical constraint penalty term determination module is specifically used to obtain the physical constraint parameters at the current moment, the physical constraint parameters including: theoretical voltage, current and air outlet pressure; determine a voltage gradient constraint term based on the theoretical voltage at the current moment; determine a current gradient constraint term based on the current current moment; determine an air outlet pressure gradient constraint term based on the air outlet pressure at the current moment; and perform a weighted average of the voltage gradient constraint term, the current gradient constraint term and the air outlet pressure gradient constraint term to obtain the physical constraint penalty term.

[0012] Optionally, the core operating data includes: voltage, current, air outlet flow rate, and air outlet pressure; the optional operating data includes: temperature, hydrogen pressure, and air inlet pressure, or one or more of these.

[0013] According to a third aspect of this application, an electronic device is provided, comprising: a processor configured to execute a computer program stored in a memory, wherein the computer program, when executed by the processor, implements the method described in the first aspect.

[0014] According to a fourth aspect of this application, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0015] According to a fifth aspect of this application, a computer program product is provided that, when the computer program product is run on a computer, causes the computer to perform the method described in the first aspect.

[0016] The technical solution provided in this application has the following advantages compared with the prior art: By collecting and extracting the core operating data sequence of a proton exchange membrane fuel cell (PEMFC), a gated recurrent neural network is first used to extract the time-series dependency. Then, a high-order nonlinear mapping is fitted using a KAN spline function to obtain a high-precision theoretical voltage, which is then used to determine the normalized voltage residual. Based on the normalized voltage residual, a physical constraint penalty term unique to the PEMFC field is further introduced. By coupling the two, a health index characterizing the battery state is formed, expanding the output from a simple prediction value to an interpretable health index, thereby improving the accuracy of battery state assessment. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart of a proton exchange membrane fuel cell state assessment method according to an embodiment of this application; Figure 2 This is a schematic diagram of a proton exchange membrane fuel cell state assessment device in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device in an embodiment of this application. Detailed Implementation

[0020] To better understand the above-mentioned objectives, features, and advantages of this application, the solution of this application will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0021] Many specific details are set forth in the following description in order to provide a full understanding of this application, but this application may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only some embodiments of this application, and not all embodiments.

[0022] See Figure 1 , Figure 1 This is a flowchart of a proton exchange membrane fuel cell state assessment method according to an embodiment of this application, including the following steps: Step S102: Collect core operating data of the proton exchange membrane fuel cell according to the sampling interval, and extract the core operating data sequence corresponding to the current moment according to the preset sliding window. Core operating data refers to operating data that is strongly correlated with voltage prediction. Optionally, core operating data includes voltage, current, air outlet flow rate, and air outlet pressure. After collecting core operating data according to the sampling interval, the core operating data can be preprocessed before extracting the core operating data sequence corresponding to the current time according to a preset sliding window. The preprocessing process includes outlier removal, SG filtering (Savitzky-Golay filtering), resampling, and standardization.

[0023] The core runtime data sequence corresponding to the current time t refers to the sequence consisting of the runtime data from the current time t and multiple times preceding the current time t. It can be represented as follows: ; L represents the preset sliding window, which can be 6, and the sliding step size can be 1. express The voltage from time t to the current time t, express The current from time t to the current time t, express Air outlet flow rate from time t to the current time t. express The air outlet pressure from time t to the current time t.

[0024] Step S104: Use a gated recurrent neural network to extract features from the running data sequence to obtain temporal hidden features.

[0025] The degradation process of proton exchange membrane fuel cells is influenced by the coupling of multiple physical fields, including electrochemistry, thermodynamics, and fluid dynamics, resulting in significant nonlinear and non-stationary characteristics in their time-series data. In this embodiment, a gated recurrent neural network (GRN) is first used to time-encode the operating data sequence, obtaining a time-series hidden feature. This feature condenses the dynamic relationship between voltage, current, air outlet flow rate, and air outlet pressure within a preset sliding window. The GRN can be, for example, an LSTM (Long Short-Term Memory) network or a GRU (Gated Recurrent Unit). When the data volume is large, an LSTM can be selected to capture more subtle long-term dependencies. When the data volume is small, a GRU can be selected to improve computational efficiency. When using an LSTM, two layers can be used to capture more abstract time dependencies.

[0026] Both gated recurrent neural networks (GRNNs) and KANs are pre-trained networks. The training process for GRNNs and KANs is as follows: Data preparation: The raw experimental data were cleaned, SG filtered, resampled, feature selected and standardized, and a sliding time window sample was constructed. Model initialization: Set the hyperparameters such as the number of hidden units and layers of the gated recurrent neural network, the number of layers, nodes, and spline grids of the KAN, and randomly initialize the network parameters. Forward propagation: The input sequence of the training samples is fed into the model, which then passes through a gated recurrent neural network and a KAN in sequence to obtain the predicted voltage value. Loss Calculation and Backpropagation: The mean square error (MSE) between the predicted and actual voltages is calculated as the loss function. The gradient of the loss with respect to the network parameters is calculated using the backpropagation algorithm. Parameter update: Update the model network parameters using the Adam optimizer based on the gradient obtained in the previous step. Validation and Early Stopping: After each training epoch, evaluate the model performance on the validation set. If the validation loss does not improve for several consecutive epochs, stop training and save the model parameters with the minimum validation loss. Testing and Prediction: Load the optimal model parameters and perform prediction evaluation on the test set. For future predictions, a recursive approach is used, taking the current prediction value as part of the input for the next time step to progressively generate a voltage prediction sequence for multiple future steps.

[0027] Step S106: Use KAN to process the time-series hidden features to obtain the theoretical voltage.

[0028] Next, a KAN (Kind-Analog Recurrent Neural Network) is used to learn a complex nonlinear mapping function from the time-series feature space to the output voltage space. The side functions in the KAN can be in B-spline form, and the order of the B-splines can be 2. The complex nonlinear mapping from the time-series features to the output variable (i.e., the theoretical voltage) can be fitted using B-splines. Optionally, the KAN can also be two-layered to improve the fitting ability of the function. Combining a gated recurrent neural network with KAN can accurately predict the theoretical voltage of a fuel cell and has good generalization ability.

[0029] Optionally, while collecting core operating data of the proton exchange membrane fuel cell according to sampling intervals, optional operating data of the proton exchange membrane fuel cell are also collected according to sampling intervals, and the optional operating data sequence corresponding to the current moment is extracted according to a preset sliding window. Optional operating data includes one or more of temperature, hydrogen pressure, and air inlet pressure. Accordingly, the temporal hidden features and optional operating data sequence are processed using KAN to obtain the theoretical voltage. By selecting and supplementing other optional operating data in KAN to enhance constraints, the overall prediction performance can be improved.

[0030] Step S108: Calculate the normalized voltage residual based on the measured voltage, theoretical voltage, and voltage residual scale corresponding to the current operating condition of the proton exchange membrane fuel cell.

[0031] The voltage residual scale differs depending on the operating condition (e.g., steady-state, quasi-dynamic, dynamic). Therefore, the corresponding voltage residual scales for different operating conditions can be statistically analyzed in the healthy sample. The current operating condition can be determined based on the core operating data of the proton exchange membrane fuel cell, and the normalized voltage residual can be calculated using the voltage residual scale corresponding to the current operating condition. This can be expressed by the following formula: ; This represents the normalized voltage residual at the current time t. This represents the measured voltage at the current time t. This represents the theoretical voltage at the current time t. This indicates the voltage residual scale corresponding to the current operating condition. This indicates a pre-set, smaller value used to prevent... When the divisor is 0, the divisor is 0.

[0032] Step S110: Obtain the physical constraint parameters at the current moment, and determine the physical constraint penalty term based on the physical constraint parameters.

[0033] In the field of proton exchange membrane fuel cells, several unique physical constraint parameters exist. Therefore, a physical constraint penalty term can be constructed based on these parameters. Coupled with the normalized voltage residual, the accuracy of the cell health status assessment can be improved. Optionally, the physical constraint parameters include: theoretical voltage, current, and air outlet pressure.

[0034] Based on the theoretical voltage at the current moment, determine the voltage gradient constraint term; based on the current current at the current moment, determine the current gradient constraint term; based on the air outlet pressure at the current moment, determine the air outlet pressure gradient constraint term. This is expressed as follows: ; ; ; in, This represents the voltage gradient constraint term. This represents the current gradient constraint term. This represents the air outlet pressure gradient constraint term.

[0035] Then, a weighted average is performed on the voltage gradient constraint term, current gradient constraint term, and air outlet pressure gradient constraint term to obtain the physical constraint penalty term. This is expressed as follows: ; in, This represents the physical constraint penalty term. , and Indicates the weighting coefficient. .

[0036] Step S112: Perform a weighted average of the normalized voltage residual and the physical constraint penalty term to determine the health index, and determine the health status of the proton exchange membrane fuel cell based on the health index.

[0037] According to the formula: Health index can be determined ; The range is 0 to 1. The smaller the value, the healthier the battery; 0 represents an ideal healthy state, and 1 represents a completely degraded state. and Indicates the weighting coefficient. . and This can be determined by a validation set or by experts. It can also generate a results interpretation table, listing the normalized voltage residuals and physical constraint penalty terms. The contribution. And the division of 0 to 1 into three levels: normal, attention, and alarm, corresponding to the respective numerical ranges, based on... The size outputs the corresponding level.

[0038] The proton exchange membrane fuel cell (PEMFC) state assessment method of this application collects and extracts the core operating data sequence of the PEMFC. First, it uses a gated recurrent neural network to extract the time-series dependency, and then uses a KAN spline function to fit a high-order nonlinear mapping to obtain a high-precision theoretical voltage, thereby determining the normalized voltage residual. Based on the normalized voltage residual, a physical constraint penalty term unique to the PEMFC field is further introduced. By coupling the two, a health index characterizing the battery state is formed, expanding the output result from a simple prediction value to an interpretable health index, thereby improving the accuracy of the battery state assessment.

[0039] It should be noted that although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0040] Corresponding to the above method embodiments, this application also provides a proton exchange membrane fuel cell state assessment device, see [link to relevant documentation]. Figure 2 The proton exchange membrane fuel cell condition assessment device 200 includes: The core operating data sequence acquisition module 202 is used to collect core operating data of the proton exchange membrane fuel cell according to the sampling interval, and extract the core operating data sequence corresponding to the current time according to the preset sliding window; The temporal hidden feature extraction module 204 is used to extract features from the running data sequence using a gated recurrent neural network to obtain temporal hidden features; The theoretical voltage determination module 206 is used to process the timing hidden features using KAN to obtain the theoretical voltage; The normalized voltage residual determination module 208 is used to calculate the normalized voltage residual based on the measured voltage, theoretical voltage and voltage residual scale corresponding to the current operating condition of the proton exchange membrane fuel cell. The physical constraint penalty term determination module 210 is used to obtain the physical constraint parameters at the current moment and determine the physical constraint penalty term based on the physical constraint parameters; The health status determination module 212 is used to perform a weighted average of the normalized voltage residual and the physical constraint penalty term to determine the health index, and to determine the health status of the proton exchange membrane fuel cell based on the health index.

[0041] Optionally, the proton exchange membrane fuel cell condition assessment device 200 further includes: The optional operating data sequence acquisition module is used to collect optional operating data of the proton exchange membrane fuel cell according to the sampling interval when collecting the core operating data of the proton exchange membrane fuel cell according to the sampling interval, and extract the optional operating data sequence corresponding to the current time according to the preset sliding window. The theoretical voltage determination module 206 is specifically used to process the time-series hidden features and optional running data sequences using KAN to obtain the theoretical voltage.

[0042] Optionally, the physical constraint penalty term determination module 210 is specifically used to obtain the physical constraint parameters at the current moment, including: theoretical voltage, current and air outlet pressure; determine the voltage gradient constraint term based on the theoretical voltage at the current moment; determine the current gradient constraint term based on the current current moment; determine the air outlet pressure gradient constraint term based on the air outlet pressure at the current moment; and perform a weighted average of the voltage gradient constraint term, current gradient constraint term and air outlet pressure gradient constraint term to obtain the physical constraint penalty term.

[0043] Optionally, core operating data include: voltage, current, air outlet flow rate, and air outlet pressure; optional operating data include: temperature, hydrogen pressure, and air inlet pressure, or one or more of these.

[0044] The specific details of each module or unit in the above-mentioned device have been described in detail in the corresponding methods, so they will not be repeated here.

[0045] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0046] This application also provides an electronic device, including: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute the above-described proton exchange membrane fuel cell state assessment method.

[0047] Reference Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device in an embodiment of this application. The specific embodiments of this application do not limit the specific implementation of the electronic device.

[0048] like Figure 3 As shown, the electronic device may include: a processor 302, a communication interface 304, a memory 306, and a communication bus 308.

[0049] The processor 302, communication interface 304, and memory 306 communicate with each other via communication bus 308.

[0050] Communication interface 304 is used to communicate with other electronic devices or servers.

[0051] The processor 302 is used to execute program 310, specifically the relevant steps in the above method embodiments.

[0052] Specifically, program 310 may include program code that includes computer operation instructions.

[0053] Processor 302 may be a central processing unit, a specific integrated circuit, or one or more integrated circuits configured to implement the embodiments of this application. The one or more processors included in the smart device may be processors of the same type, such as one or more CPUs; or they may be processors of different types, such as one or more CPUs and one or more ASICs.

[0054] Memory 306 is used to store program 310. Memory 306 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0055] Specifically, program 310 can be used to cause processor 302 to execute the steps in the above-described proton exchange membrane fuel cell state assessment method embodiment.

[0056] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the devices and modules described above can be referred to the corresponding process descriptions in the foregoing method embodiments, and will not be repeated here.

[0057] In this embodiment of the application, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the above-described proton exchange membrane fuel cell state assessment method.

[0058] It should be noted that the computer-readable storage medium shown in this application can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable storage medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, radio frequency, etc., or any suitable combination thereof.

[0059] In this embodiment of the application, a computer program product is also provided, which, when run on a computer, causes the computer to execute the above-described proton exchange membrane fuel cell state assessment method.

[0060] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0061] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for assessing the state of a proton exchange membrane fuel cell, characterized in that, include: The core operating data of the proton exchange membrane fuel cell is collected according to the sampling interval, and the core operating data sequence corresponding to the current moment is extracted according to the preset sliding window. The time-series hidden features are obtained by using a gated recurrent neural network to extract features from the running data sequence; The theoretical voltage is obtained by processing the temporal hidden features using a Kolmogorov-Arnold network (KAN). The normalized voltage residual is calculated based on the measured voltage of the proton exchange membrane fuel cell, the theoretical voltage, and the voltage residual scale corresponding to the current operating condition. Obtain the physical constraint parameters at the current moment, and determine the physical constraint penalty term based on the physical constraint parameters; The normalized voltage residual and the physical constraint penalty term are weighted and averaged to determine the health index, and the health status of the proton exchange membrane fuel cell is determined based on the health index.

2. The method according to claim 1, characterized in that, The method further includes: When collecting core operating data of proton exchange membrane fuel cells according to sampling intervals, optional operating data of proton exchange membrane fuel cells are also collected according to sampling intervals, and the optional operating data sequence corresponding to the current moment is extracted according to a preset sliding window. The process of using a Kolmogorov-Arnold network (KAN) to process the temporal hidden features to obtain the theoretical voltage includes: The theoretical voltage is obtained by processing the temporal hidden features and the optional running data sequence using the Kolmogorov-Arnold network (KAN).

3. The method according to claim 1, characterized in that, The physical constraint parameters include: theoretical voltage, current, and air outlet pressure; The determination of the physical constraint penalty term based on the physical constraint parameters includes: Based on the theoretical voltage at the current moment, determine the voltage gradient constraint term; Based on the current at the current moment, determine the current gradient constraint term; Based on the current air outlet pressure, determine the air outlet pressure gradient constraint term; The physical constraint penalty term is obtained by weighting the voltage gradient constraint term, the current gradient constraint term, and the air outlet pressure gradient constraint term.

4. The method according to claim 2, characterized in that, The core operating data includes: voltage, current, air outlet flow rate, and air outlet pressure; the optional operating data includes: temperature, hydrogen pressure, and air inlet pressure, or one or more of these.

5. A proton exchange membrane fuel cell condition assessment device, characterized in that, The device includes: The core operating data sequence acquisition module is used to collect core operating data of the proton exchange membrane fuel cell according to the sampling interval, and extract the core operating data sequence corresponding to the current moment according to the preset sliding window; The temporal hidden feature extraction module is used to extract features from the running data sequence using a gated recurrent neural network to obtain temporal hidden features; The theoretical voltage determination module is used to process the temporal hidden features using the Kolmogorov-Arnold network (KAN) to obtain the theoretical voltage. The normalized voltage residual determination module is used to calculate the normalized voltage residual based on the measured voltage of the proton exchange membrane fuel cell, the theoretical voltage, and the voltage residual scale corresponding to the current operating condition. The physical constraint penalty term determination module is used to obtain the physical constraint parameters at the current moment and determine the physical constraint penalty term based on the physical constraint parameters; The health status determination module is used to perform a weighted average of the normalized voltage residual and the physical constraint penalty term to determine a health index, and to determine the health status of the proton exchange membrane fuel cell based on the health index.

6. The apparatus according to claim 5, characterized in that, The device further includes: The optional operating data sequence acquisition module is used to collect optional operating data of the proton exchange membrane fuel cell according to the sampling interval when collecting the core operating data of the proton exchange membrane fuel cell according to the sampling interval, and extract the optional operating data sequence corresponding to the current time according to the preset sliding window. The theoretical voltage determination module is specifically used to process the temporal hidden features and the optional running data sequence using the Kolmogorov-Arnold network (KAN) to obtain the theoretical voltage.

7. The apparatus according to claim 5, characterized in that, The physical constraint penalty term determination module is specifically used to obtain the physical constraint parameters at the current moment, which include: theoretical voltage, current, and air outlet pressure; determine a voltage gradient constraint term based on the theoretical voltage at the current moment; determine a current gradient constraint term based on the current current moment; determine an air outlet pressure gradient constraint term based on the air outlet pressure at the current moment; and perform a weighted average of the voltage gradient constraint term, the current gradient constraint term, and the air outlet pressure gradient constraint term to obtain the physical constraint penalty term.

8. The apparatus according to claim 6, characterized in that, The core operating data includes: voltage, current, air outlet flow rate, and air outlet pressure; the optional operating data includes: temperature, hydrogen pressure, and air inlet pressure, or one or more of these.

9. An electronic device, characterized in that, include: A processor for executing a computer program stored in a memory, wherein the computer program, when executed by the processor, implements the method of any one of claims 1-4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1-4.

Images