A method, device, and medium for fuel cell health diagnosis under high-altitude "four low" conditions.

Abstract

This application discloses a method, device, and medium for fuel cell health diagnosis under the "four low" conditions of high-altitude areas, relating to the field of proton exchange membrane fuel cells. The method includes: real-time acquisition of key parameters of the fuel cell, such as voltage, current, ambient temperature, humidity, and oxygen concentration, based on nanosensors; processing various data using BIRCH clustering, multimodal fusion, and GCN model methods to obtain dynamic features, high-altitude environmental features, and static features; extracting the dynamic features into a Hankel matrix field based on the dynamic feature information; extracting the high-altitude environmental features and static features into a co-occurrence matrix field based on the high-altitude environmental features and the co-occurrence matrix field; standardizing and tensorizing the features into a three-dimensional space based on the Hankel matrix field and the co-occurrence matrix field; and fusing a multilayer perceptron using the sample entropy method to obtain the fuel cell diagnosis results for high-altitude areas based on the three-dimensional feature space. This application ensures the safe operation of the fuel cell system.

Description

Technical Field

[0001] This application relates to the field of proton exchange membrane fuel cell technology, and in particular to a method, device and medium for fuel cell health diagnosis under high-altitude "four low" conditions. Background Technology

[0002] The application of fuel cells in high-altitude areas has always been limited by harsh environmental conditions. In high-altitude environments, the atmospheric pressure is much lower than in plains areas, potentially reducing oxygen flow and diffusion rates. This can lead to unstable hydrogen supply within the fuel cell, affecting its normal operation. Simultaneously, the significantly reduced oxygen content can impact the efficiency of redox reactions within the fuel cell, thus affecting its output power and overall efficiency. Secondly, high-altitude environments are generally cold, especially at night when temperatures drop dramatically. The start-up and operational performance of fuel cells at low temperatures is significantly affected, particularly water management, which is prone to condensation and even internal freezing, leading to battery damage. Furthermore, the typically low humidity in high-altitude areas accelerates the drying of the electrolyte membrane, resulting in decreased fuel cell stability and performance. Existing fuel cell monitoring and diagnostic systems struggle to provide timely monitoring and diagnosis of operational conditions under these unique circumstances. In summary, the low air pressure, low oxygen concentration, low temperature, and low humidity in high-altitude areas have many negative impacts on the normal operation of fuel cells, placing higher demands on the health diagnosis system of fuel cells. However, general-purpose fuel cells often struggle to perform high-precision monitoring and timely diagnosis due to the lack of adaptation for high-altitude operating conditions.

[0003] Therefore, proposing a method, device, and medium for fuel cell health diagnosis under the "four low" conditions of high altitude is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0004] The purpose of this application is to provide a method, device, and medium for health diagnosis of fuel cells under the "four low" conditions of high altitude, so as to solve the problem of the difficulty in diagnosing fuel cell systems in high-altitude environments.

[0005] To achieve the above objectives, this application provides the following solution:

[0006] In one aspect, this application provides a method, apparatus, and medium for health diagnosis of fuel cells under high-altitude "four low" conditions, including:

[0007] Acquire key data of different types, such as voltage and current, ambient temperature and humidity, and membrane state, for fuel cells;

[0008] The various types of data were processed using the BIRCH clustering method, multimodal fusion method, and GCN model method to obtain dynamic features, plateau environment features, and static features information;

[0009] Based on the dynamic features, plateau environment features, and static features, features are extracted into the Hankel matrix field and the co-occurrence matrix field, respectively.

[0010] Based on the feature field, the features are standardized and tensorized to three-dimensional space;

[0011] Based on the aforementioned three-dimensional features, a multilayer perceptron is fused using the sample entropy method to obtain the diagnostic results for fuel cells in plateau regions.

[0012] Optionally, the processing of the various types of data using BIRCH clustering, multimodal fusion, and GCN models to obtain dynamic features, plateau environment features, and static features specifically includes:

[0013] The data is subjected to BIRCH cluster analysis to capture the relationships between different variables;

[0014] Based on the relationships between the different variables, dynamic feature information is determined;

[0015] The data is standardized for each modality to obtain the transformed data;

[0016] The weights of each modality are dynamically adjusted based on the data transformed by the introduced attention fusion mechanism to determine the characteristic information of the plateau environment;

[0017] The feature information of the data node is aggregated with the information of the neighboring nodes to obtain the transformed data.

[0018] The co-occurrence field is calculated based on the transformed data to determine static feature information;

[0019] Optionally, based on the dynamic features, features are extracted into the Hankel matrix field, specifically including:

[0020] Based on the voltage subsequence and current subsequence, construct the voltage Hankel matrix and the current Hankel matrix, and map them to the two-dimensional matrix field Hankel matrix field;

[0021] Optionally, based on the plateau environment characteristics, features are extracted to the symbiotic field, specifically including:

[0022] Based on plateau environmental characteristic data of temperature, air pressure, oxygen concentration, and humidity at N time steps, construct a co-occurrence matrix of a single pair of features and a set of co-occurrence matrices of all feature pairs.

[0023] Optionally, based on the static features, features are extracted to the co-occurrence field, specifically including:

[0024] Based on the type of static feature information, discrete static features are extracted into the co-occurrence matrix field using the One-Hot encoding method, and numerical static features are extracted into the co-occurrence matrix field using the standardization method.

[0025] This invention also provides a fuel cell health diagnosis system under high-altitude "four low" conditions, comprising:

[0026] The acquisition module is used to acquire data such as voltage and current, membrane state, etc. of the fuel cell system, as well as environmental data such as air pressure, oxygen concentration, humidity, and temperature in the "four low" environment of the plateau.

[0027] The conversion module is used to extract features from the data to obtain dynamic features, plateau environment features, and static features, and convert them to Hankel matrix field and co-occurrence matrix field, and then standardize and tensorize them to three-dimensional space.

[0028] Optionally, the conversion module specifically includes:

[0029] A normalization processing unit is used to normalize the data to obtain the transformed data;

[0030] The calculation unit is used to calculate the Hankel matrix field and the co-occurrence matrix field based on the transformed data;

[0031] The dynamic feature information determination unit is used to determine dynamic feature information based on the Hankel matrix field.

[0032] The plateau environment characteristic information determination unit is used to determine plateau environment characteristic information based on the co-occurrence matrix field.

[0033] The static feature information determination unit is used to determine static feature information based on the co-occurrence matrix field.

[0034] The first identification module is used to identify fuel cells using the sample entropy method based on feature information, and obtain the sample entropy value of the fuel cell.

[0035] The second identification module is used to identify problems in the fuel cell using the multilayer perceptron method based on feature information.

[0036] The fusion module is used to fuse the sample entropy value output and the problem score value output to obtain the diagnostic results of the fuel cell system.

[0037] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the fuel cell health diagnosis method, device, and medium under high-altitude "four low" conditions as described in any of the above claims.

[0038] Thirdly, this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the fuel cell health diagnosis method, apparatus, and medium under high-altitude "four low" conditions as described in any of the above claims.

[0039] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0040] This application discloses a method, device, and medium for fuel cell health diagnosis under the "four low" conditions of high altitude. The method uses nanosensors to collect key data such as voltage, current, temperature, humidity, and oxygen concentration of the fuel cell in real time. Feature data is extracted from this data using methods such as BIRCH clustering, multimodal fusion, and graph convolutional networks (GCN) to obtain dynamic feature information, high-altitude environmental feature information, and static feature information. A Hankel matrix field is constructed based on the dynamic feature information, and co-occurrence matrix fields are constructed based on the high-altitude environmental feature information and the static feature information, respectively. Based on the Hankel matrix field and the co-occurrence matrix field, the features are standardized and tensorized to a three-dimensional space. Based on the feature three-dimensional space, a multilayer perceptron is fused using the sample entropy method to obtain the fuel cell diagnosis results for high-altitude areas. This invention comprehensively considers dynamic, high-altitude environmental, and static features, improving the accuracy of fuel cell health diagnosis under harsh high-altitude conditions. It features real-time, high efficiency, and intelligence, providing comprehensive condition monitoring and fault diagnosis capabilities for fuel cell systems. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is an overall diagram of a fuel cell health diagnosis method, device, and medium under the "four low" conditions of high altitude, according to the present invention.

[0043] Figure 2 This is a flowchart illustrating the diagnostic process identification and judgment of the present invention;

[0044] Figure 3 A schematic diagram of the Hankel matrix field for fuel cell current and voltage;

[0045] Figure 4 A symbiotic matrix field diagram illustrating the characteristics of the plateau environment for fuel cells;

[0046] Figure 5 A flowchart of the fuel cell health diagnosis method, apparatus, and medium provided by the present invention;

[0047] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0048] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0049] The purpose of this application is to provide a method, device and medium for health diagnosis of fuel cells under the "four low" conditions of high altitude, which aims to enhance the system's ability to identify abnormal states under the harsh high-altitude environment of low air pressure, low oxygen concentration, low temperature and low humidity, and ensure the accuracy and timeliness of diagnosis.

[0050] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0051] like Figure 5 As shown, the present invention discloses a method, apparatus, and medium for fuel cell health diagnosis under high-altitude "four low" conditions, comprising:

[0052] Step 101: Acquire data such as voltage and current, membrane state, etc. of the fuel cell system, as well as real-time data of ambient air pressure, oxygen concentration, humidity, and temperature in the high-altitude environment through nanosensors.

[0053] Step 102: Based on different types of data such as battery system voltage and current, ambient temperature and humidity, and membrane state, dynamic features, plateau environment features, and static features are obtained by using BIRCH cluster analysis, multimodal fusion, and GCN model respectively.

[0054] Step 103: Extract features to the Hankel matrix field based on the dynamic features, and extract features to the symbiotic matrix field based on the plateau environment features and static features.

[0055] Step 104: Based on the different feature fields, transform them into a unified tensor structure, and integrate the tensors into three-dimensional space to form three-dimensional feature data.

[0056] Step 105: Based on the features in the three-dimensional space, the multilayer perceptron is fused using the sample entropy method to obtain the diagnostic results of the fuel cell in the plateau region.

[0057] Step 105 specifically includes:

[0058] Based on the three-dimensional feature data, the sample entropy method and the multilayer perceptron method are used to obtain a combination of sample entropy value output and problem score value output; based on the output combination data, the fuel cell diagnostic rules are used to determine the current state of the fuel cell and obtain the fuel cell diagnostic result.

[0059] Part 1: Transforming Numerical Information into Feature Information

[0060] The first step is to use the BIRCH algorithm to represent the summary information of each cluster using cluster feature triples. The cluster feature triples are defined as follows:

[0061] CF = (N,LS,SS)

[0062] Where N is the number of data points in the current cluster (fuel cell operating state cluster), representing the number of sampling points in this operating state;

[0063] A multimodal fusion method is employed to extract plateau environmental features to address the complexity and multidimensional influencing factors of the plateau environment. To facilitate fusion, features from different modes need to be standardized. Assume x... (i) The mean and standard deviation are μ (i) and σ (i) Then the features of each modality can be standardized:

[0064]

[0065] An attention fusion mechanism is introduced to dynamically adjust the weights of each modality, automatically assigning weights to highlight key modal features. A learnable weight parameter w is set. (i) Weight each modal feature:

[0066]

[0067] The GCN model is based on graph convolution operations, aggregating the feature information of a node with the information of its neighboring nodes. The forward propagation process of each GCN layer is as follows:

[0068]

[0069] Among them, H (l) Let H be the node feature matrix of the l-th layer, and H be the initial layer feature matrix. (0) This is the input representation of static features. To add the adjacency matrix with self-joins, I is the identity matrix. for The degree matrix is ​​defined as W (l) Let be the weight matrix of the l-th layer, which needs to be learned through training, and σ be the ReLU nonlinear activation function.

[0070] Part Two: Feature Information Extraction to Feature Field

[0071] The dynamic characteristic data of fuel cells (such as voltage and current) are divided into multiple clusters after BIRCH clustering, with each cluster representing a dynamic behavior pattern. Voltage time series: {V1, V2, ..., V N}, Current time series: {I1,I2,…,I N}. Using BIRCH clustering, these time series were divided into L clusters, each containing one or more subsequences. For example, the dynamic characteristics of the l-th cluster are: voltage series: {V l,1 V l,2 ,…,V l,Nl}, Current sequence: {I l,1 ,I l,2 ,…,I l,Nl Extract the voltage and current subsequences and construct the Hankel matrix H. Vl and H Il :

[0072]

[0073] like Figure 3 As shown, the voltage and current Hankel matrices of each cluster are mapped onto a two-dimensional matrix field to form the Hankel matrix field H:

[0074] H = {(H} V1 H I1 ),(H V2 H I2 ),…,(H VL H IL )}

[0075] By leveraging dimensionality reduction and pattern analysis of the co-occurrence matrix, the system can process multidimensional feature data more efficiently. For feature pairs (X... i ,X j ), whose co-occurrence matrix has elements C p This indicates that the two features fall within the interval R. p and R q Co-occurrence frequency:

[0076]

[0077] Among them, R p and R q It is feature X i and X j The range of values ​​for δ(x). k ∈R p ,y k ∈R q ) is an indicator function:

[0078]

[0079] There is plateau environmental characteristic data at N time steps, including: temperature T={T1,T2,…,T N}, air pressure P = {P1, P2, ..., P N Oxygen concentration O={O1,O2,…,O N Humidity H = {H1, H2, ..., H} N}. For two features (such as temperature and air pressure), a co-occurrence matrix C is constructed. ij Discretize the two features to intervals R respectively. p and R q Iterate through each time step k = 1, 2, ..., N, and count the number of times the two features co-occur in the corresponding interval. For example, ... Figure 4 For a feature pair (temperature T and air pressure P), the co-occurrence matrix elements are:

[0080]

[0081] For four features T, P, O, H, construct the set of co-occurrence matrices for all feature pairs:

[0082] C = {C TP C TO C TH C PO C PH C OH}

[0083] For discrete static features (such as material type), One-Hot encoding is used to transform categorical features into continuous vector forms for extraction into the co-occurrence matrix field; for numerical static features, standardization is used to ensure that different features are in the same numerical range for extraction into the co-occurrence matrix field.

[0084] Part 3: Tensorizing Feature Information to Three-Dimensional Space

[0085] In the three-dimensional spatial model, each coordinate axis represents a different feature type: the X-axis represents dynamic features, the Y-axis represents plateau environment features, and the Z-axis represents static features. This three-dimensional space will provide a reference for the comprehensive analysis of the fuel cell's state across multiple dimensions, including time, environment, and design parameters. When constructing the three-dimensional space, the different feature data are first transformed into a unified tensor structure. Then, the dynamic features, plateau environment features, and static features are concatenated according to their feature dimensions to form a comprehensive tensor.

[0086]

[0087] Where N is the number of samples, T is the time step, and F is the feature dimension.

[0088] Part 4: Fuel Cell Health Diagnosis Based on Sample Entropy Method Fueled by Multilayer Perceptron Method

[0089] Sample entropy is used to measure the complexity of the dynamic characteristics of a system and to assess the stability of fuel cell operation.

[0090] (1) Set the embedding dimension m and the tolerance r:

[0091] m: Used to define the size of the mode within the time window.

[0092] r: Sets the allowable error range, controlling the strictness of pattern matching, typically set to 0.1-0.2 times the data standard deviation.

[0093] r = k × σ(X)

[0094] Where σ(X) is the standard deviation of the time series, and k is a scaling factor (e.g., 0.1 or 0.2).

[0095] (2) Construct the embedding vector:

[0096] Construct an embedding vector sequence of length m from the time series X:

[0097] X i =[x i ,x i+1 ,…,x i+m-1 for 1≤i≤N-m+1

[0098] (3) Calculate the sample entropy SampEn:

[0099]

[0100] Where A is the number of matched pattern pairs of length m+1, and B is the number of matched pattern pairs of length m. At low sample entropy, the system operates stably with high pattern repeatability; at high sample entropy, the system exhibits more fluctuations, and its state is more complex or abnormal. Preliminary judgments of abnormal states can be made based on the sample entropy calculation results.

[0101] The multilayer perceptron method extracts complex nonlinear relationships between dynamic and static features and environmental features based on multiple layers of neurons.

[0102] The network structure consists of an input layer, hidden layers, and an output layer. The input layer receives a three-dimensional feature tensor T. combined After flattening The input tensor after flattening, if it is [N,T,F], will be [N,T×F]. The ReLU function is commonly used as the activation function in hidden layers.

[0103] a (l) =ReLU(z) (l) )

[0104] The hidden layer structure is designed as follows:

[0105] Hidden layer 1: 64 neurons, using the ReLU activation function.

[0106] Hidden layer 2: 32 neurons, using the ReLU activation function.

[0107] Hidden layer 3: 16 neurons, using the ReLU activation function.

[0108] During the forward propagation process, data passes through layers of the network until it reaches the output layer:

[0109] (1) Output of neurons in layer l:

[0110] z (l) =W (l) a (l-1) +b (l)

[0111] Among them, W (l) Let b be the weight matrix of the l-th layer. (l) The bias of the l-th layer, a (l-1) This is the activation value (output) of the previous layer.

[0112] (2) Assume the output layer of the multilayer sensor model is a vector Z = [z1, z2, ..., z n ], where n is the number of categories (e.g., "normal", "minor anomaly", "serious anomaly"), and the output layer uses the Softmax activation function:

[0113]

[0114] Among them, P i z is the probability score for the i-th category. i It is the i-th logit (inactive score) in the output layer of the multilayer perceptron model, where n is the total number of categories.

[0115] (3) Output layer results:

[0116]

[0117] Output the probability of each category to complete the classification task, and output the problem score to complete the regression task.

[0118] In backpropagation, the loss function is selected based on the task type:

[0119] (1) Classification task: Cross-entropy loss function:

[0120]

[0121] (2) Regression task: Mean Squared Error (MSE):

[0122]

[0123] The gradient of the loss with respect to the weights is calculated through backpropagation. The Adam optimizer is used to optimize the algorithm, and the learning rate is dynamically adjusted to improve the convergence speed. Multimodal data and sample entropy values ​​are used as input features for forward and backpropagation training. The error is calculated based on the loss function, and the model parameters and weights are updated using gradient descent.

[0124]

[0125] Where η is the learning rate.

[0126] The sample entropy result is combined with other features as an input vector, which is then fed into a multilayer perceptron model. After inference by the multilayer perceptron model, it can determine whether the output state is abnormal and provide a problem score. If the system state is abnormal, or the problem score exceeds a set threshold, a warning is triggered. Figure 2 First, feature data is processed using the sample entropy method and then input into a multilayer perceptron model for state inference. Based on different ranges of the sample entropy value (SampEn) and the problem score (P), namely SampEn≥1.0 or P≥0.7, 0.5≤SampEn<1.0 or 0.3≤P<0.7, and SampEn<0.5 and P<0.3, the system triggers a severe anomaly warning, a minor anomaly warning, or determines no anomaly, respectively. In the case of a severe anomaly, the system sends an emergency notification, automatically generates a fault report, and reminds the user to investigate immediately. In the case of a minor anomaly, the system sends an information notification to remind the user to monitor the abnormal state. If no anomaly is found, the system state is recorded and continuously monitored. Furthermore, the process includes data correction and verification mechanisms to ensure data accuracy and allows for continuous model optimization through feedback learning, improving diagnostic accuracy and system responsiveness.

[0127] This invention also provides a fuel cell health diagnosis system under high-altitude "four low" conditions, comprising:

[0128] The acquisition module is used to acquire data such as voltage and current, membrane state, etc. of the fuel cell system, as well as environmental data such as air pressure, oxygen concentration, humidity, and temperature in the "four low" environment of the plateau.

[0129] The conversion module is used to extract features from the data to obtain dynamic features, plateau environment features, and static features, and convert them to Hankel matrix field and co-occurrence matrix field, and then standardize and tensorize them to three-dimensional space.

[0130] The first identification module is used to identify fuel cells using the sample entropy method based on feature information, and obtain the sample entropy value of the fuel cell.

[0131] The second identification module is used to identify problems in the fuel cell using the multilayer perceptron method based on feature information.

[0132] The fusion module is used to fuse the sample entropy value output and the problem score value output to obtain the diagnostic results of the fuel cell system.

[0133] In one exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a method, apparatus, and medium for fuel cell health diagnosis under high-altitude "four low" conditions.

[0134] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements a method, apparatus, and medium for health diagnosis of fuel cells under high-altitude "four low" conditions.

[0135] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements a method, apparatus, and medium for health diagnosis of fuel cells under high-altitude "four low" conditions.

[0136] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 6 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a method, device, and medium for fuel cell health diagnosis under high-altitude "four low" conditions.

[0137] Those skilled in the art will understand that Figure 6The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0138] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0139] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0140] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0141] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0142] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for health diagnosis of fuel cells under high-altitude "four low" conditions, characterized in that, include: Acquire key data on different types of fuel cell conditions, including voltage and current, ambient temperature and humidity, and membrane state. The various types of data were processed using the BIRCH clustering method, multimodal fusion method, and GCN model method to obtain dynamic features, plateau environment features, and static features information; Based on the dynamic features, plateau environment features, and static features, features are extracted into the Hankel matrix field and the co-occurrence matrix field, respectively. Based on the feature field, the features are standardized and tensorized to three-dimensional space; Based on the aforementioned three-dimensional features, a multilayer perceptron is fused using the sample entropy method to obtain the diagnostic results for fuel cells in plateau regions. Based on the information of each feature in three-dimensional space, the diagnostic information is obtained by combining the multilayer perceptron method with the sample entropy method. The diagnostic results are shown by the sample entropy value (SampEn) and the problem score (P). SampEn is the complexity evaluation of the fuel cell system output by the sample entropy method, and P is the score result of whether the fuel cell has an anomaly output by the multilayer perceptron method. The first step is to calculate the sample entropy of the fuel cell system: The second step is to calculate the problem score for the fuel cell system: Wherein, the ReLU function is used as the activation function, W (l) Let b be the weight matrix of the l-th layer. (l) The bias of the l-th layer, a (l−1) P is the activation value (output) of the previous layer. i z is the probability score for the i-th category. i It is the i-th logit in the output layer of the multilayer perceptron model, representing the score of non-activated data, and n is the total number of categories; This allows us to obtain fuel cell diagnostic results based on the sample entropy method and the fusion of multilayer perceptrons: Fuel cell diagnostic results = We completed fuel cell diagnosis based on the sample entropy method and fusion of multilayer sensors.

2. The method for fuel cell health diagnosis under high-altitude "four low" conditions according to claim 1, characterized in that, The data of each type are processed using BIRCH clustering, multimodal fusion, and GCN model methods to obtain dynamic features, plateau environment features, and static features, specifically including: The data is subjected to BIRCH cluster analysis to capture the relationships between different variables; Based on the relationships between the different variables, dynamic feature information is determined; The data is standardized for each modality to obtain the transformed data; An attention fusion mechanism is introduced to dynamically adjust the weight of each modality based on the transformed data, thereby determining the characteristic information of the plateau environment. The feature information of the data node is aggregated with the information of the neighboring nodes to obtain the transformed data.

3. The method for fuel cell health diagnosis under high-altitude "four low" conditions according to claim 1, characterized in that, Based on the dynamic characteristics, features are extracted into the Hankel matrix field, specifically including: According to the voltage subsequence {V l,1 V l,2 ,…,V l,Nl } and current subsequence {I l,1 ,I l,2 ,…,I l,Nl Construct the Hankel matrix H Vl and H Il : Based on the Hankel matrix H of voltage and current Vl and H Il Mapped to the two-dimensional matrix field Hankel matrix field H: Based on the Hankel matrix field, for each Hankel matrix H l Singular value decomposition (SVD) is performed to extract principal components and analyze the temporal patterns of dynamic features, including: Where, Σ Vl and Σ Il It is a singular value matrix, which contains the main change patterns of the matrix. The largest singular value can reflect the main dynamic behavior of the voltage or current characteristics within the cluster.

4. The method for fuel cell health diagnosis under high-altitude "four low" conditions according to claim 1, characterized in that, Based on the aforementioned plateau environmental characteristics, features are extracted into the symbiotic field, specifically including: There are N time steps with temperature T = {T1, T2, ..., T...} N }, air pressure P={P1,P2,…,P N Oxygen concentration O = {O1, O2, ..., O N Humidity H = {H1, H2, ..., H} N The data consists of plateau environmental characteristics, and a co-occurrence matrix for each feature pair and a set of co-occurrence matrices for all feature pairs are constructed. in, Let be the co-occurrence matrix of temperature and air pressure, and C be the co-occurrence matrix of four characteristic temperatures T, pressure P, oxygen concentration O, and humidity H.

5. The method for fuel cell health diagnosis under high-altitude "four low" conditions according to claim 1, characterized in that, Based on the static features, features are extracted to the co-occurrence array, specifically including: Based on the type of static feature information, discrete static features are extracted into the co-occurrence matrix field using the One-Hot encoding method, and numerical static features are extracted into the co-occurrence matrix field using the standardization method.

6. The method for health diagnosis of fuel cells under high-altitude "four low" conditions according to claim 1, characterized in that, Based on the aforementioned feature field, the features are standardized and tensorized to three-dimensional space: Where, x (i) The mean and standard deviation are respectively μ (i) and σ (i) , T combined It is a three-dimensional tensor. T dynamic , T environment , T static These are identified as dynamic characteristics, plateau environment characteristics, and static characteristics.

7. The method for fuel cell health diagnosis under high-altitude "four low" conditions according to claim 1, characterized in that, Based on the task type, select the loss function in backpropagation and update the model parameters based on gradient descent, including: Where L1 is the cross-entropy loss function used for classification tasks, L2 is the mean squared error function used for regression tasks, and W... (l) Let be the weight matrix of the l-th layer of the neural network, η be the learning rate, and L be the loss function. This represents the gradient of the loss function with respect to the weights.

8. A fuel cell health diagnosis system under high-altitude "four low" conditions, characterized in that, include: The acquisition module is used to acquire voltage and current, membrane state data, and environmental data of the "four lows" (low air pressure, low oxygen concentration, low humidity, and low temperature) of the fuel cell system. The conversion module is used to extract features from the data to obtain dynamic features, plateau environment features, and static features, and convert them to Hankel matrix field and co-occurrence matrix field, and then standardize and tensorize them to three-dimensional space. The first identification module is used to identify fuel cells using the sample entropy method based on feature information, and obtain the sample entropy value of the fuel cell. The second identification module is used to identify problems in the fuel cell using the multilayer perceptron method based on feature information. The fusion module is used to fuse the sample entropy value output and the problem score value output to obtain the diagnostic results of the fuel cell system.

9. A computer device, comprising: The memory, the processor, and the computer program stored in the memory and executable on the processor are characterized in that the processor executes the computer program to implement the fuel cell health diagnosis method under high-altitude "four low" conditions as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements a method for health diagnosis of fuel cells under high-altitude "four low" conditions as described in any one of claims 1-7.

Images