Fuel cell vehicle monitoring method, device, system and equipment and storage medium

Abstract

The invention provides a fuel cell automobile monitoring method, device, system and equipment and a storage medium, and relates to the technical field of fuel cells, and the method comprises the following steps: carrying out data preprocessing on fuel cell data and hydrogen storage data of an automobile, and constructing a multi-dimensional feature vector according to a preprocessing result; the multi-dimensional feature vectors are analyzed, key feature parameters, feature boundaries and time sequence features are obtained, the feature boundaries are used for determining threshold ranges of the key feature parameters under different working conditions, and the time sequence features represent the change trend of the key feature parameters in the time dimension; inputting the key feature parameters, the feature boundary and the time sequence features into a pre-trained fault classification model, and identifying the fault type of the fuel cell; and performing early warning according to the fault type of the fuel cell. In this way, the interaction and coupling mechanism of the multi-parameter coupling relation in the time dimension are considered, deep analysis is carried out on the multi-parameter coupling relation of the fuel cell under the complex working condition, and accurate pre-judgment and diagnosis of potential faults are achieved.

Description

Technical Field

[0001] The present invention relates to the field of fuel cell technology, and in particular to a fuel cell vehicle monitoring method, device, system, equipment and storage medium. Background Art

[0002] Fuel cell vehicles, with their technical characteristics of directly converting hydrogen and oxygen into electrical energy through electrochemical reactions, have significant advantages such as zero emissions, high energy density, and long driving range. They have become a key direction for the future development of the automotive industry. In order to ensure the safety of vehicle operation, the risks faced by fuel cells during actual operation are strictly controlled.

[0003] Currently, fuel cell monitoring typically involves sensors monitoring battery data and uploading it to an onboard terminal system (T-BOX) for real-time collection of fuel cell operating data. This operating data covers key parameters such as fuel cell temperature, hydrogen / air pressure, and system operating status. This data is then transmitted to a remote monitoring center via wireless communication technologies such as cellular networks. The data is then compared with thresholds to generate early warnings, enabling real-time visual monitoring of the vehicle's operating status and providing fault warnings.

[0004] However, existing remote monitoring mainly relies on simple logic such as threshold judgment for fault warning, which makes the judgment of faults rather one-sided and makes it difficult to accurately predict and diagnose potential faults. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a fuel cell vehicle monitoring method, device, system, equipment and storage medium to solve the problem that the current remote monitoring fault warning judgment is relatively one-sided and it is difficult to achieve accurate prediction of potential faults.

[0006] In order to achieve the above-mentioned objectives, in a first aspect, the present invention provides a fuel cell vehicle monitoring method, the method comprising: obtaining fuel cell data and hydrogen storage data of a vehicle; performing data preprocessing on the fuel cell data and hydrogen storage data to obtain a preprocessing result; constructing a multidimensional feature vector based on the preprocessing result, the multidimensional feature vector representing the operating parameters of the fuel cell; analyzing the multidimensional feature vector to obtain key feature parameters, feature boundaries and time series features, the feature boundaries being used to determine the threshold range of the key feature parameters under different operating conditions, and the time series features representing the changing trend of the key feature parameters in the time dimension; inputting the key feature parameters, the feature boundaries and the time series features into a pre-trained fault classification model to identify the fault type of the fuel cell; and issuing an early warning based on the fault type of the fuel cell.

[0007] In this embodiment, after the fuel cell data and hydrogen storage data are preprocessed, a multidimensional feature vector is constructed to comprehensively reflect the current operating parameters of the fuel cell; then, the multidimensional feature vector is analyzed to extract key feature parameters and determine feature boundaries. At the same time, the time series characteristics after dimensionality reduction processing are used to capture the dynamic change trend of key feature parameters in the time dimension, which helps to understand the evolution process of the multi-parameter coupling relationship in the time dimension and reveal the dynamic interaction and coupling mechanism between parameters; finally, the fuel cell fault type is accurately identified through the pre-trained fault classification model, so as to conduct in-depth analysis of the fuel cell multi-parameter coupling relationship under complex working conditions and realize accurate prediction and diagnosis of potential faults.

[0008] In some embodiments, the preprocessing results include time alignment results, spatial alignment results and standardization processing results. The data preprocessing of the fuel cell data and the hydrogen storage data to obtain the preprocessing results includes: unifying the fuel cell data and the hydrogen storage data to the same time base, performing time alignment, and obtaining the time alignment result; unifying the fuel cell data and the hydrogen storage data to the same spatial coordinate system, performing spatial alignment, and obtaining the spatial alignment result; unifying the fuel cell data and the hydrogen storage data to the same scale range, performing standardization processing, and obtaining the standardization processing result.

[0009] With this setup, the fuel cell data and hydrogen storage data from different sensors are unified onto the same time base, ensuring data consistency in the time dimension. All data are then mapped to the same spatial coordinate system for spatial alignment, which can reduce the impact of the different spatial positions of the sensors on the vehicle. Finally, the time-aligned and spatially-aligned fuel cell data and hydrogen storage data are unified in numerical range, making all types of data comparable and providing a reliable data foundation for subsequent feature extraction and fault diagnosis.

[0010] In some embodiments, constructing a multidimensional feature vector according to the preprocessing result includes: arranging and integrating the time alignment result, the space alignment result, and the normalization processing result in chronological order to form a multidimensional feature vector.

[0011] This setting not only preserves the time sequence relationship of each data point, but also ensures that the multidimensional feature vector can fully reflect the operating parameters of the fuel cell, so as to facilitate in-depth analysis of the multi-parameter coupling relationship based on the multidimensional feature vector.

[0012] In some embodiments, the analysis of the multidimensional feature vector to obtain key feature parameters, feature boundaries and time series features includes: screening out key feature parameters from the operating parameters of the fuel cell through principal component analysis and mutual information method; identifying feature boundaries of the operating parameters of the fuel cell in combination with the fuzzy C-means clustering algorithm, so as to determine the threshold ranges of the key feature parameters under normal operating conditions, critical operating conditions and abnormal operating conditions based on the feature boundaries; and extracting time series features through a pre-trained feature extraction model after dimensionality reduction processing of the operating parameters of the fuel cell, so as to determine the changing trend of the key feature parameters in the time dimension based on the time series features.

[0013] With this setting, key characteristic parameters are screened out through principal component analysis and mutual information method, so as to screen out the most representative data of the fuel cell; and the threshold range of the key characteristic parameters under normal operating conditions, critical operating conditions and abnormal operating conditions is determined under the fuzzy C-means clustering algorithm to provide an accurate basis for fault judgment; the fuel cell operating condition parameters are then subjected to dimensionality reduction processing to reduce redundant data while retaining important information, so as to extract time series features and capture potential abnormal signals or performance degradation signs of the fuel cell, so as to achieve in-depth analysis of the fuel cell operating status, which is helpful for early prediction of fuel cell failures.

[0014] In some embodiments, the inputting of the key characteristic parameters, the characteristic boundaries and the time series characteristics into a pre-trained fault classification model to identify the state and fault type of the fuel cell includes: analyzing the key characteristic parameters, the characteristic boundaries and the time series characteristics according to the fault classification model to determine whether the fault type of the fuel cell is no fault, a minor fault or a major fault; wherein, the no fault characterization that the key characteristic parameters are within the threshold range of the normal operating condition, the minor fault characterization that the key characteristic parameters are within the threshold range of the critical operating condition, and the major fault characterization that the key characteristic parameters are within the threshold range of the abnormal operating condition.

[0015] With this setting, key feature parameters, feature boundaries, and time series features are used as inputs and input into a pre-trained fault classification model for comprehensive analysis to determine the current state of the fuel cell and classify it as one of three types: no fault, minor fault, or major fault. This allows the current fault type of the fuel cell to be determined based on the threshold range of the key feature parameters, helping to identify fuel cell faults and issue early warnings.

[0016] In some embodiments, the early warning according to the fault type of the fuel cell includes: determining the early warning level corresponding to the fault type of the fuel cell according to a pre-established correspondence between the fault type and the early warning level; and triggering early warning measures based on the early warning level.

[0017] This setting allows for timely notification of fuel cell failures and the specific type of failure through triggered early warning measures, thereby facilitating timely handling of faulty fuel cell vehicles and significantly improving risk management efficiency.

[0018] In a second aspect, the present invention provides a fuel cell vehicle monitoring device, comprising an acquisition module, a processing module and an early warning module, wherein the acquisition module is used to acquire fuel cell data and hydrogen storage data; the processing module is used to perform data preprocessing on the fuel cell data and hydrogen storage data to obtain a preprocessing result; a multidimensional feature vector is constructed based on the preprocessing result, and the multidimensional feature vector represents the operating parameters of the fuel cell; the device is used to analyze the multidimensional feature vector to obtain key feature parameters, feature boundaries and time series characteristics, the feature boundaries are used to determine the threshold range of the key feature parameters under different operating conditions, and the time series characteristics represent the changing trend of the key feature parameters in the time dimension; the device is also used to input the key feature parameters, the feature boundaries and the time series characteristics into a pre-trained fault classification model to identify the fault type of the fuel cell; the early warning module is used to issue an early warning based on the fuel cell fault type.

[0019] In a third aspect, the present invention provides a fuel cell vehicle monitoring system, comprising a data acquisition device, a data transmission device and a monitoring device; the data acquisition device is used to collect fuel cell data and hydrogen storage data of the vehicle; the data transmission device is used to transmit the fuel cell data and hydrogen storage data collected by the data acquisition device to the monitoring device, and the monitoring device is used to preprocess the fuel cell data and hydrogen storage data to obtain a preprocessing result; a multidimensional feature vector is constructed based on the preprocessing result, and the multidimensional feature vector represents the operating parameters of the fuel cell; the multidimensional feature vector is analyzed to obtain key feature parameters, feature boundaries and time series features, the feature boundaries are used to determine the threshold range of the key feature parameters under different operating conditions, and the time series features represent the changing trend of the key feature parameters in the time dimension; the key feature parameters, the feature boundaries and the time series features are input into a pre-trained fault classification model to identify the fault type of the fuel cell; and an early warning is issued based on the fault type of the fuel cell.

[0020] In a fourth aspect, the present invention provides an electronic device comprising a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and the processor can execute the computer program to implement the fuel cell vehicle monitoring method described in the first aspect.

[0021] In a fifth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the fuel cell vehicle monitoring method as described in the first aspect is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 A flow chart of a fuel cell vehicle monitoring method provided by an embodiment of the present invention;

[0024] Figure 2 A flow chart of a method for constructing a multidimensional feature vector provided by an embodiment of the present invention;

[0025] Figure 3 A flow chart of a method for analyzing multidimensional feature vectors provided by an embodiment of the present invention;

[0026] Figure 4 A flowchart of a method for providing early warning according to fault type provided by an embodiment of the present invention;

[0027] Figure 5 A schematic diagram of the functional modules of a fuel cell vehicle monitoring device provided by an embodiment of the present invention;

[0028] Figure 6 A structural block diagram of a fuel cell vehicle monitoring system provided by an embodiment of the present invention;

[0029] Figure 7 This is a flow chart of information control of a fuel cell vehicle monitoring system provided by an embodiment of the present invention;

[0030] Figure 8 A schematic diagram of the control principle of hydrogen leakage provided by an embodiment of the present invention;

[0031] Figure 9 A block diagram of an electronic device provided by an embodiment of the present invention.

[0032] Icon: 1000 - monitoring device; 1100 - acquisition module; 1200 - processing module; 1300 - early warning module; 2000 - electronic device; 2100 - processor; 2200 - memory; 2300 - bus; 2400 - communication interface. DETAILED DESCRIPTION

[0033] The following will be combined with the accompanying drawings to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0034] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but is merely intended to represent selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative work are within the scope of protection of the present invention.

[0035] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

[0036] The following is a brief introduction to some concepts that may be involved in the embodiments of the present invention.

[0037] On-board hydrogen system controller (Hydrogen Management System, HMS): A control unit that mainly monitors and manages various parameters and status of the on-board hydrogen system, such as monitoring the storage pressure of hydrogen to ensure that hydrogen is stored within a safe pressure range.

[0038] Battery Management System (BMS): Mainly used to monitor and manage the status of batteries, such as monitoring key parameters such as voltage, current and temperature of batteries (usually battery packs used to store electrical energy).

[0039] Vehicle-mounted terminal system (T-BOX): Continuously collects core vehicle data such as speed, mileage, fuel level, etc., connects the vehicle with external networks, and realizes data interaction and intelligent services.

[0040] The Fuel Cell Functional Control Unit (FCU) monitors the fuel cell's output voltage, current, and key parameters such as internal hydrogen and air pressure and temperature in real time. Based on the vehicle's power requirements and the fuel cell's actual operating status, it precisely adjusts the hydrogen and air supply to ensure efficient and stable fuel cell power generation.

[0041] Vehicle Control Unit (VCU): Responsible for coordinating and managing the entire power system, it implements integrated control of multiple vehicle functions, such as vehicle start-stop control, energy recovery control, and coordinated control of the air conditioning system and power system.

[0042] Currently, fuel cell monitoring systems typically consist of two core components: the fuel cell controller (FCU) and the data transmission module, the vehicle-mounted terminal system (T-BOX). The FCU collects real-time fuel cell system operating data, including key parameters such as fuel cell temperature, hydrogen / air pressure, and system operating status. The T-BOX transmits this data to a remote monitoring center via wireless communication technologies such as cellular networks. The center then compares the data with thresholds to generate early warnings, enabling real-time visual monitoring of the vehicle's operating status and providing fault warnings.

[0043] However, when monitoring fuel cells, fault warnings are only based on simple logic such as threshold judgment. There is a lack of in-depth analysis of the multi-parameter coupling relationship under complex operating conditions of the fuel cell system, making it difficult to accurately predict and diagnose potential faults.

[0044] To this end, an embodiment of the present invention provides a fuel cell vehicle monitoring method, such as Figure 1 As shown, Figure 1 This is a flowchart of a fuel cell vehicle monitoring method provided by an embodiment of the present invention. By extracting key characteristic parameters, determining characteristic boundaries and time series characteristics, understanding the evolution of multi-parameter coupling relationships in the time dimension, and revealing the dynamic interactions and coupling mechanisms between parameters, this method deeply analyzes the multi-parameter coupling relationships of fuel cells under complex operating conditions, and achieves accurate prediction and diagnosis of potential faults. The method includes steps S100 to S500:

[0045] S100: Obtain fuel cell data and hydrogen storage data of the vehicle.

[0046] In this embodiment, fuel cell data and hydrogen storage data can be obtained through the BMS and HMS, transmitted to the vehicle gateway via the Controller Area Network (CAN), and then aggregated and sent to the T-BOX after protocol conversion. Fuel cell data includes the voltage, current, and temperature of the fuel cell. The BMS can collect data collected by the voltage sensor, current sensor, and temperature sensor at the fuel cell. Hydrogen storage data includes the temperature, pressure, and concentration of hydrogen storage. The HMS can collect data collected by the temperature sensor, pressure sensor, and concentration sensor in the high-pressure hydrogen pipeline and hydrogen cylinder.

[0047] S200 , preprocessing the fuel cell data and the hydrogen storage data to obtain a preprocessing result; constructing a multidimensional feature vector based on the preprocessing result, wherein the multidimensional feature vector represents the operating parameters of the fuel cell.

[0048] In this embodiment, preprocessing involves temporal and spatial alignment and standardization of real-time data (including hydrogen concentration, pressure, temperature, battery voltage, current, and temperature) collected by controllers such as the FCU, HMS, BMS, and VCU. The preprocessing results include temporal alignment, spatial alignment, and standardization.

[0049] For example, see Figure 2 , Figure 2 The flowchart of the method for constructing a multi-dimensional feature vector provided by an embodiment of the present invention, S200 may specifically include the following steps S201 to S204:

[0050] S201: Unify the fuel cell data and the hydrogen storage data to the same time base, perform time alignment, and obtain a time alignment result.

[0051] In this embodiment, time alignment synchronizes fuel cell data and hydrogen storage data collected from different sensors or data acquisition devices. Because each sensor may have different sampling frequencies or clock offsets, time synchronization using an interpolation algorithm aligns all data to the same time base. The input data for time alignment is the fuel cell data and hydrogen storage data, and the output is a time-aligned data sequence in which each data point has a uniform timestamp.

[0052] S202: Unify the fuel cell data and the hydrogen storage data into the same spatial coordinate system, perform spatial alignment, and obtain a spatial alignment result.

[0053] In this embodiment, spatial alignment further processes the time-aligned data to eliminate spatial deviations caused by differences in sensor installation positions or inconsistent coordinate systems. Specifically, a geometric transformation matrix (such as a rotation matrix and translation vector) is applied to each sensor data so that all data can be mapped to a common reference coordinate system. The input of spatial alignment is the time-aligned data sequence, and the output is a data set expressed in a unified reference coordinate system.

[0054] S203: Unify the fuel cell data and the hydrogen storage data into the same scale range, perform standardization processing, and obtain a standardization processing result.

[0055] In this embodiment, the normalization process uses a normalization or standardization algorithm to reduce dimensional differences and numerical range differences between data. For example, Z-score normalization or Min-Max normalization is used to convert each type of data into a numerical range with the same distribution characteristics, thereby unifying the data into the same scale range, thereby preventing certain features from dominating the entire feature space due to excessively large values. The input of the normalization process is the spatially aligned data, and the output is a set of normalized feature values.

[0056] S204: Arrange the time alignment result, the space alignment result, and the normalization processing result in chronological order and integrate them to form a multi-dimensional feature vector.

[0057] In this embodiment, time order refers to the arrangement of the timestamps corresponding to the data points in chronological order. For example, assuming that at a certain time t1 there are three eigenvalues ​​x1, y1, and z1, and at the next time t2 there are corresponding eigenvalues ​​x2, y2, and z2, the resulting multidimensional feature vector is [x1, y1, z1, x2, y2, z2]. If data from more time points is available, the dimensions of the feature vector are further expanded according to the same rules. This shows that the multidimensional feature vector not only preserves the time series characteristics of the data, but also effectively reflects the relative relationships between the features, providing high-quality input data for subsequent analysis or modeling tasks.

[0058] S300. Analyze the multidimensional feature vector to obtain key feature parameters, feature boundaries, and time series features. The feature boundaries are used to determine the threshold range of the key feature parameters under different working conditions. The time series features represent the changing trend of the key feature parameters in the time dimension.

[0059] In this embodiment, key characteristic parameters include, but are not limited to, the fuel cell's voltage fluctuation rate, current density distribution, temperature gradient, and the rate of change of hydrogen pressure and concentration. Feature boundaries include threshold ranges for key characteristic parameters under normal, critical, and abnormal operating conditions. Time series features include a set of parameters that describe the changing trends of key characteristic parameters over time.

[0060] For example, see Figure 3 , Figure 3 The flowchart of the method for analyzing a multidimensional feature vector provided by an embodiment of the present invention, S300 specifically includes the following steps S301 to S303:

[0061] S301. Filter out key characteristic parameters of the fuel cell operating parameters through principal component analysis and mutual information method.

[0062] In this embodiment, principal component analysis (PCA) extracts the principal components that best reflect the variance of the original data by performing a linear transformation on each dimension of the multidimensional feature vector. These principal components represent the primary direction of variation in the data, thereby preliminarily identifying potential key features that significantly impact fuel cell performance. Mutual information analysis is then used to further evaluate the correlation between each of these potential key features and the target output variable, eliminating redundant or fuel cell-irrelevant features. The input to this step is the multidimensional feature vector, and the output is a set of filtered key feature parameters.

[0063] S302 , identifying characteristic boundaries of the fuel cell operating parameters in combination with the fuzzy C-means clustering algorithm, and determining threshold ranges of key characteristic parameters under normal operating conditions, critical operating conditions, and abnormal operating conditions based on the characteristic boundaries.

[0064] In this embodiment, the operating parameters (key characteristic parameters under different operating conditions) are grouped and classified by the fuzzy C-means clustering algorithm to identify the characteristic boundaries corresponding to normal operating conditions, critical operating conditions and abnormal operating conditions. The fuzzy C-means clustering algorithm models the spatial distribution of key characteristic parameters, calculates the membership of each data point to different categories (such as normal, critical or abnormal), and divides the boundary conditions of each category accordingly, thus obtaining the characteristic boundaries. For example, for a certain key characteristic parameter, if its value falls within the normal operating condition range with a higher degree of membership, it is considered that the parameter is in a normal state; conversely, if its value is close to the boundary of the critical or abnormal operating condition, it indicates that there may be a risk of performance degradation or failure. The input of this step is a set of key characteristic parameters, and the output is a clear characteristic boundary condition and its corresponding threshold range.

[0065] S303 , after performing dimensionality reduction processing on the operating parameters of the fuel cell, extracting time series features through a pre-trained feature extraction model, so as to determine the change trend of the key feature parameters in the time dimension according to the time series features.

[0066] In this embodiment, the extraction of time series features is a process of in-depth exploration of the time dimension characteristics of the multidimensional feature vector. Since the multidimensional feature vector itself is arranged and integrated in chronological order, the changing trend or periodic law of the key feature parameters can be analyzed from the time dimension. Therefore, before the analysis, in order to reduce the dimensional complexity of the multidimensional feature vector, the operating parameters are first subjected to dimensionality reduction processing. The technical means used for dimensionality reduction processing can adopt the Linear Discriminant Analysis (LDA) method to retain key information while reducing the data dimension. The data after dimensionality reduction is then input into a pre-trained feature extraction model for analysis, wherein the model can be a time series analysis model based on deep learning, such as a long short-term memory network (LSTM) or a convolutional neural network (CNN), which can capture the dynamic change law of data in the time dimension. The above model is then used to calculate the difference or slope between eigenvalues ​​at adjacent moments, capturing the dynamic characteristics of the eigenvalues ​​over time. Alternatively, frequency domain analysis methods such as Fourier transforms can be used to extract the periodic components in the key characteristic parameters, thereby determining their temporal trends. The input for this step is a multidimensional feature vector, and the output is a set of parameters describing the characteristics of the time series, such as the rate of change of hydrogen concentration / pressure fluctuation amplitude, and the location of sudden changes in battery voltage / current / temperature.

[0067] S400 , inputting key feature parameters, feature boundaries, and time series features into a pre-trained fault classification model to identify the fault type of the fuel cell.

[0068] In this embodiment, the fault classification model analyzes key characteristic parameters, characteristic boundaries, and time series characteristics to determine the fuel cell fault type as no fault, minor fault, or major fault. Key characteristic parameters indicating no fault are within a threshold range for normal operating conditions, key characteristic parameters indicating minor faults are within a threshold range for critical operating conditions, and key characteristic parameters indicating major faults are within a threshold range for abnormal operating conditions.

[0069] Specifically, the fault classification model is pre-trained using historical fault data to form a feature knowledge base of fault patterns. This historical fault data is derived from actual fuel cell systems in operation and includes key characteristic parameters, feature boundaries, and time series features under normal, critical, and abnormal operating conditions. The fault classification model supports correlation analysis of multi-parameter coupled anomalies, such as fuel cell voltage, current, and temperature anomalies, hydrogen storage pressure fluctuations, and rapid changes in hydrogen concentration, enabling prediction of potential faults as early as 100 milliseconds.

[0070] For example, on the one hand, the fault classification model can simultaneously receive multiple key characteristic parameters at the input end and comprehensively consider the interaction between them. For example, the set of key characteristic parameters at a certain moment may include voltage fluctuation rate, current density distribution, temperature gradient, hydrogen concentration change rate, etc. The fault classification model captures coupled abnormal patterns that cannot be reflected by a single parameter through joint analysis of these parameters. On the other hand, the fault classification model combines characteristic boundary conditions to further enhance its sensitivity to normal operating conditions, critical operating conditions, and abnormal operating conditions. When certain key characteristic parameters approach or exceed the characteristic boundaries, the fault classification model can quickly identify potential fault risks. On the other hand, the fault classification model also uses time series features to model dynamic change trends, thereby making up for the shortcomings of static parameter analysis. For example, if a key characteristic parameter fluctuates violently or oscillates periodically in a short period of time, the fault classification model can regard it as abnormal behavior and identify the fault type in a timely manner.

[0071] It should be noted that when identifying the type of fuel cell fault, since temperature and pressure are proportional, the hydrogen storage bottle pressure and hydrogen concentration in the key characteristic parameters are used as the basis for fault judgment, because the hydrogen storage bottle pressure and hydrogen concentration can best reflect whether the fuel cell has the risk of explosion. Among them, the preset threshold range of the hydrogen storage bottle pressure is 0-70Mpa (excluding 70Mpa, corresponding to normal operating conditions), 70-90Mpa (excluding 90Mpa, corresponding to critical operating conditions) and above 90Mpa (including 90Mpa, corresponding to abnormal operating conditions). The preset threshold range of hydrogen concentration is 0-0.5ppm (excluding 0.5ppm, corresponding to normal operating conditions), 0.5-4ppm (excluding 4ppm, corresponding to critical operating conditions) and above 4ppm (including 4ppm, corresponding to abnormal operating conditions).

[0072] S500: Issue a warning based on the type of fuel cell failure.

[0073] See Figure 4 , Figure 4 The flowchart of the method for providing early warning according to the fault type provided in the embodiment of the present invention may specifically include the following steps S501 to S502:

[0074] S501 : Determine the warning level corresponding to the fault type of the fuel cell according to a pre-established correspondence between the fault type and the warning level.

[0075] In this embodiment, a hierarchical early warning decision matrix is ​​pre-established to match fault types with warning levels. Specifically, the following can be used: Level 1 warning (yellow): the hydrogen storage tank pressure reaches 70 MPa and the hydrogen concentration is within the range of 0.5-1 ppm; Level 2 warning (orange): the hydrogen storage tank pressure reaches between 70-90 MPa and the hydrogen concentration is within the range of 1-4 ppm. Parameter coupling abnormality is indicated only when both the hydrogen storage tank pressure and hydrogen concentration are within the range; Level 3 warning (red): the hydrogen storage tank pressure reaches above 90 MPa and the hydrogen concentration reaches above 4 ppm, corresponding to high-risk signals such as hydrogen leakage.

[0076] It should be noted that when the hydrogen storage tank pressure value and hydrogen concentration value are compared with the preset threshold range, if the hydrogen storage tank pressure value reaches the second-level warning, if the hydrogen concentration value is only the first-level warning or no warning, then the warning will be issued according to the second-level warning level reached by the hydrogen storage tank pressure value, and the hydrogen storage tank pressure value and hydrogen concentration value data will be uploaded to the cloud monitoring platform to provide fault clues to assist in troubleshooting. If the hydrogen storage tank pressure value is only the first-level warning or no warning, then the focus is on the warning level reached by the hydrogen concentration value, and a warning will be issued according to the warning level reached by the hydrogen concentration value.

[0077] S502: Trigger warning measures based on the warning level.

[0078] In this embodiment, the corresponding warning action for the first level warning (yellow) is to trigger an alarm on the vehicle dashboard. The corresponding warning action for the second level warning (orange) is to activate local sound and light alarms and create a fault work order in the cloud. The corresponding warning action for the third level warning (red) is to immediately close the solenoid valve to cut off the hydrogen supply.

[0079] Specifically, when the FCU, HMS, VCU and FCU detect that parameters deviate from the preset thresholds (such as abnormal hydrogen storage pressure or excessive hydrogen concentration) through the sensor network, a multi-level early warning mechanism is immediately triggered: first, the fault code is displayed visually through the on-board instrument panel, and the abnormal data is packaged into CAN messages that comply with the SAE J1939 protocol and transmitted to the cloud monitoring platform via T_BOX, achieving dual early warning coverage of local warnings and remote operation and maintenance.

[0080] Based on the above method, the embodiment of the present invention also provides a device corresponding to the above method, such as Figure 5 As shown, Figure 5This is a schematic diagram of the functional modules of a fuel cell vehicle monitoring device 1000 provided in an embodiment of the present invention. It should be noted that the basic principles and technical effects of the fuel cell vehicle monitoring device 1000 provided in this embodiment are the same as those of the aforementioned method embodiment. For the sake of brevity, reference should be made to the corresponding content in the method embodiment for any sections not mentioned in this embodiment. The fuel cell vehicle monitoring device 1000 includes an acquisition module 1100, a processing module 1200, and an early warning module 1300.

[0081] The acquisition module 1100 is used to acquire fuel cell data and hydrogen storage data. It can be understood that the acquisition module 1100 executes the above step S100.

[0082] The processing module 1200 is used to perform data preprocessing on the fuel cell data and hydrogen storage data to obtain preprocessing results; construct a multidimensional feature vector based on the preprocessing results, and the multidimensional feature vector represents the operating parameters of the fuel cell; analyze the multidimensional feature vector to obtain key feature parameters, feature boundaries and time series features, the feature boundaries are used to determine the threshold range of the key feature parameters under different operating conditions, and the time series features represent the changing trend of the key feature parameters in the time dimension; and input the key feature parameters, the feature boundaries and the time series features into a pre-trained fault classification model to identify the fault type of the fuel cell. It can be understood that the processing module 1200 is used to execute the above steps S200 to S400,

[0083] The early warning module 1300 is used to issue an early warning according to the fuel cell fault type. It can be understood that the early warning module 1300 is used to execute the above step S500.

[0084] Exemplarily, the processing module 1200 is specifically used to unify the fuel cell data and the hydrogen storage data to the same time base, perform time alignment, and obtain a time alignment result; unify the fuel cell data and the hydrogen storage data to the same spatial coordinate system, perform spatial alignment, and obtain a spatial alignment result; unify the fuel cell data and the hydrogen storage data to the same scale range, perform standardization processing, and obtain a standardized processing result; and arrange the time alignment results, spatial alignment results, and standardized processing results in chronological order to integrate them to form a multi-dimensional feature vector.

[0085] It can be understood that the processing module 1200 can execute the above steps S201 to S204.

[0086] Exemplarily, the processing module 1200 is used to screen out key characteristic parameters of the fuel cell operating parameters through principal component analysis and mutual information method; to identify characteristic boundaries of the fuel cell operating parameters in combination with the fuzzy C-means clustering algorithm, so as to determine the threshold ranges of the key characteristic parameters under normal operating conditions, critical operating conditions and abnormal operating conditions based on the characteristic boundaries; and to extract time series features through a pre-trained feature extraction model after performing dimensionality reduction processing on the fuel cell operating parameters, so as to determine the changing trend of the key characteristic parameters in the time dimension based on the time series features.

[0087] It can be understood that the processing module 1200 can execute the above steps S301 to S303.

[0088] Exemplarily, the processing module 1200 is used to analyze key characteristic parameters, characteristic boundaries and time series characteristics according to the fault classification model to determine whether the fault type of the fuel cell is no fault, a minor fault or a major fault; wherein, the key characteristic parameters representing no fault are within the threshold range of normal operating conditions, the key characteristic parameters representing a minor fault are within the threshold range of critical operating conditions, and the key characteristic parameters representing a major fault are within the threshold range of abnormal operating conditions.

[0089] Illustratively, the warning module 1300 is used to determine the warning level corresponding to the fault type of the fuel cell according to a pre-established correspondence between the fault type and the warning level; and is also used to trigger warning measures based on the warning level.

[0090] It can be understood that the early warning module 1300 can execute the above steps S501 to S502.

[0091] Based on the above method, see Figure 6 , Figure 6 This is a block diagram of the structure of a fuel cell vehicle monitoring system provided by an embodiment of the present invention. The embodiment of the present invention also provides a fuel cell vehicle monitoring system corresponding to the above-mentioned method, comprising a data acquisition device, a data transmission device, and a monitoring device. The data acquisition device is used to collect fuel cell data and hydrogen storage data from the vehicle. The data transmission device is used to transmit the fuel cell data and hydrogen storage data collected by the data acquisition device to the monitoring device. The monitoring device is used to preprocess the fuel cell data and hydrogen storage data to obtain a preprocessing result; construct a multidimensional feature vector based on the preprocessing result, and the multidimensional feature vector represents the operating parameters of the fuel cell; analyze the multidimensional feature vector to obtain key feature parameters, feature boundaries, and time series features; the feature boundaries are used to determine the threshold range of the key feature parameters under different operating conditions; the time series features represent the changing trend of the key feature parameters over time; the key feature parameters, feature boundaries, and time series features are input into a pre-trained fault classification model to identify the type of fuel cell fault; and an early warning is issued based on the type of fuel cell fault.

[0092] In this embodiment, the data acquisition device includes an HMS, a BMS, a VCU, an FCU, and a sensor array to collect fuel cell data and hydrogen storage data. The data transmission device includes a T-BOX, and the monitoring device includes a monitoring platform. The T-BOX is responsible for hierarchical caching and priority scheduling of the collected fuel cell data and hydrogen storage data to the monitoring device. It can be understood that the data acquisition device transmits the accurately collected key parameters to the monitoring platform through the T-BOX, and the monitoring platform is connected to the host computer (such as a mobile phone app).

[0093] For example, when T-BOX transmits data, it can implement a 100Hz high-frequency sampling strategy for key parameters such as hydrogen storage temperature / pressure / concentration and fuel cell voltage / current / temperature by building a hierarchical cache and priority scheduling mechanism, and adopt variable-cycle sampling (50-100Hz dynamically adjustable) for parameters such as vehicle insulation resistance, power battery system voltage, and temperature. As a data interaction hub, T-BOX follows vehicle-mounted communication protocols such as SAE J1939, and performs standardized packaging processing on signals from HMS, BMS, and other subsystems to form data frames that comply with cloud interaction specifications, providing a reliable data foundation for subsequent remote monitoring and fault diagnosis.

[0094] Specifically, the hierarchical cache design adopts a three-level cache architecture (cache / dynamic cache / persistent cache), in which the cache (Static Random Access Memory, SRAM) stores key parameters such as hydrogen concentration and hydrogen storage pressure (sampling frequency 100Hz), the dynamic cache (Double Data Rate4, DDR4) temporarily stores medium-frequency parameters such as temperature and voltage (dynamic adjustment of 50-100Hz), and the persistent cache (NAND Flash) regularly stores historical data (1Hz). Direct Memory Access (DMA) is used between the three levels of cache to achieve zero CPU overhead data migration, ensuring millisecond-level response for critical data. Priority scheduling is based on standard-defined data priority tags (Class of Service, CoS), hydrogen leakage signals (QoS level 7), and fuel cell voltage mutations (QoS level 6) to give priority to occupying CAN bus bandwidth. QoS stands for Quality of Service.

[0095] It can be understood that the sensor array transmits real-time data to the FCU through the analog-to-digital conversion module (ADC). The FCU preprocesses the data and sends key parameters to the T-BOX through the vehicle CAN bus. The T-BOX parses the data and packages it according to the preset protocol, and transmits it to the monitoring platform through the APN4 channel using the HTTPS protocol.

[0096] Illustratively, the sensor array of HMS includes pressure sensors, temperature sensors and hydrogen concentration sensors located in high-pressure hydrogen pipelines and hydrogen cylinders. The pressure sensor is used to collect hydrogen storage pressure in the range of 0-90MPa in real time with an accuracy of ±0.1%, and is used to monitor pressure changes in the fuel cell stack and hydrogen tank; the temperature sensor is used to collect medium temperature in the range of -50°C to 150°C in real time, and is used to monitor the temperature of the fuel cell stack, cooling system and key components, providing basic data for hydrogen storage safety assessment; to address the risk of hydrogen leakage, the hydrogen concentration sensor achieves monitoring in the concentration range of 0-4ppm in leakage-prone areas such as hydrogen storage bottle mouths, pipeline connection points and fuel cell compartments, which is used to monitor hydrogen leakage and ensure vehicle safety.

[0097] For example, the sensor array of the BMS includes a current sensor, a voltage sensor, and a temperature sensor located on the fuel cell stack. The current sensor uses closed-loop fluxgate technology to support high-precision bidirectional current detection (accuracy ±0.1%) for monitoring the output current of the fuel cell. The voltage sensor uses a differential amplifier circuit to achieve synchronous monitoring of the total voltage of the fuel cell pack (accuracy ±0.05%) and the voltage of the single cell (resolution 1mV). The temperature sensor array covers the battery probes and cooling pipes (water outlet, water inlet), providing real-time temperature data for the thermal management strategy.

[0098] It should be noted that the HMS, BMS, FCU, and VCU transmit collected signals via the CAN bus to the vehicle gateway, which then aggregates and sends them to the T-BOX after protocol conversion. The T-BOX encrypts the signals and sends them to the monitoring platform, which then downloads and parses the files to obtain fuel cell and hydrogen storage data.

[0099] In some embodiments, see Figure 7 , Figure 7This is the information control flow chart of the fuel cell vehicle monitoring system provided by an embodiment of the present invention. Based on the T-BOX hardware platform, a dedicated APN (APN4) network configuration is added to the monitoring platform for the directional transmission of fuel cell-related data. Specifically, the T-BOX obtains fuel cell data (such as fuel cell voltage, current, temperature, etc.) and hydrogen storage data (such as hydrogen pressure, hydrogen concentration, and hydrogen temperature) via the CAN bus, encapsulates the data frames according to the SAE J1939 protocol, and transmits them to the monitoring platform through the APN4 channel. This APN supports 4G communication and HTTPS protocol transparent transmission. The data transmission cycle is set to 1 second. If a vehicle anomaly occurs, the data transmission cycle within 1 hour before and after the event will be adjusted to 0.1 seconds to ensure the real-time requirements of data collection and transmission to the monitoring platform when the vehicle anomaly occurs. The hardware retains the original T-BOX's CAN bus interface, Ethernet interface, and SIM card module to maintain compatibility with vehicle controllers, power batteries, instruments, and other equipment.

[0100] In some embodiments, during fuel cell vehicle production, the production system can establish a correlation between the vehicle VIN code, fuel cell number, T-BOX number, vehicle model code, and timestamp through an offline code scanning process, forming a data traceability link. The T-BOX sends the vehicle VIN code, fuel cell number, T-BOX number, vehicle model code, and timestamp to the monitoring platform, establishing a unique mapping relationship between vehicle, fuel cell and other key component information, and T-BOX. This correlation data serves as the basis for subsequent fault tracing and component replacement management.

[0101] For example, when key components such as fuel cells or T-BOXs are replaced, maintenance personnel re-scan the code and enter the new part number through the production system, triggering the data update process: the production system sends a component change instruction to the monitoring platform, and the platform verifies and updates the association relationship in the database to ensure that subsequent data accurately corresponds to the latest hardware configuration.

[0102] The following is an example of a high-risk working condition such as hydrogen leakage. Figure 8 , Figure 8 This is a control principle diagram of the monitoring platform provided by an embodiment of the present invention. Assuming that a fuel cell vehicle has a dangerous hydrogen leak, once the hydrogen concentration sensor captures the leakage signal, the HMS immediately initiates the response program.

[0103] First, the HMS sends a diagnostic message containing the leak location code, real-time concentration value, and leak rate to the vehicle control unit (VCU) via the CAN bus. The leak rate can be calculated using the differential method of dual hydrogen concentration sensors: dC / dt = (C(t) - C(t)) / Δx·v, where C / C is the reading of two sensors arranged along the airflow direction, Δx is the distance between them, and v is the gas flow rate (calibrated through CFD simulation). This helps distinguish between instantaneous leaks (>2000ppm / s) and slow permeation (<500ppm / s), triggering different levels of response.

[0104] Secondly, the HMS closes the solenoid valve of the hydrogen storage bottle group (response time ≤ 0.5s);

[0105] After receiving the leak diagnostic message, the VCU simultaneously sends a STOP control command (compliant with the ISO 22840 fuel cell system communication protocol) to the fuel cell controller (FCU). The FCU closes the fuel cell's anode hydrogen inlet valve and air inlet valve, cutting off the cathode hydrogen and air supply, and halts the boost DC / DC converter, achieving zero fuel cell power output safety protection within 200 milliseconds.

[0106] Based on the same inventive concept as disclosed above, correspondingly, the embodiment of the present invention further provides a block diagram of an electronic device 2000 for executing the above method, please refer to Figure 9 , Figure 9 This is a block diagram of an electronic device 2000 provided in an embodiment of the present invention. The electronic device 2000 includes a processor 2100, a memory 2200, a bus 2300, and a communication interface 2400. The processor 2100 and the memory 2200 are connected via the bus 2300, and the processor 2100 communicates with external devices via the communication interface 2400.

[0107] The processor 2100 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits in the processor 2100 or by software instructions. The above processor 2100 can be a general-purpose processor 2100, including a central processing unit 2100 (CPU), a network processor 2100 (NP), etc.; it can also be a digital signal processor 2100 (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0108] The memory 2200 is used to store computer programs, for example, the fuel cell vehicle monitoring device in the embodiment of the present invention. The fuel cell vehicle monitoring device includes at least one software functional module that can be stored in the memory 2200 in the form of software or firmware. After receiving the execution instruction, the processor 2100 executes the program to implement the fuel cell vehicle monitoring method in the embodiment of the present invention.

[0109] The memory 2200 may include a high-speed random access memory 2200 (RAM) or a non-volatile memory 2200. Alternatively, the memory 2200 may be a storage device built into the processor 2100 or a storage device independent of the processor 2100.

[0110] The bus 2300 may be an ISA bus 2300, a PCI bus 2300, an EISA bus 2300, or the like. Figure 9 The use of only one bidirectional arrow does not mean that there is only one bus 2300 or only one type of bus 2300 .

[0111] The electronic device 2000 may be a computer device such as a mobile phone, a tablet computer, a laptop computer, or a desktop computer.

[0112] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the fuel cell vehicle monitoring method described above. The computer-readable storage medium may include a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, among other media capable of storing program code.

[0113] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A fuel cell vehicle monitoring method, characterized in that: The method comprises: Obtain the vehicle's fuel cell data and hydrogen storage data; Performing data preprocessing on the fuel cell data and the hydrogen storage data to obtain a preprocessing result; constructing a multidimensional feature vector according to the preprocessing result, wherein the multidimensional feature vector represents the operating parameters of the fuel cell; Analyzing the multidimensional feature vector to obtain key feature parameters, feature boundaries, and time series features, wherein the feature boundaries are used to determine the threshold range of the key feature parameters under different operating conditions, and the time series features represent the changing trend of the key feature parameters in the time dimension; Inputting the key feature parameters, the feature boundaries, and the time series features into a pre-trained fault classification model to identify the fault type of the fuel cell; An early warning is issued according to the failure type of the fuel cell.

2. The method according to claim 1, characterized in that The preprocessing results include a time alignment result, a space alignment result, and a normalization processing result. The preprocessing of the fuel cell data and the hydrogen storage data to obtain the preprocessing results includes: Unifying the fuel cell data and the hydrogen storage data to the same time base, performing time alignment, and obtaining the time alignment result; Unifying the fuel cell data and the hydrogen storage data into the same spatial coordinate system, performing spatial alignment, and obtaining the spatial alignment result; The fuel cell data and the hydrogen storage data are unified into the same scale range and standardized to obtain the standardized processing result.

3. The method according to claim 2, characterized in that The constructing of a multidimensional feature vector according to the preprocessing result comprises: The time alignment result, the space alignment result and the normalization processing result are arranged and integrated in time sequence to form a multi-dimensional feature vector.

4. The method according to claim 1, wherein The analysis of the multi-dimensional feature vector to obtain key feature parameters, feature boundaries and time series features includes: The operating parameters of the fuel cell are screened out by principal component analysis and mutual information method to obtain key characteristic parameters; identifying characteristic boundaries of the fuel cell operating parameters in combination with a fuzzy C-means clustering algorithm, and determining threshold ranges of the key characteristic parameters under normal operating conditions, critical operating conditions, and abnormal operating conditions according to the characteristic boundaries; After the operating parameters of the fuel cell are subjected to dimensionality reduction processing, time series features are extracted through a pre-trained feature extraction model, so as to determine the change trend of the key characteristic parameters in the time dimension according to the time series features.

5. The method according to claim 4, characterized in that Inputting the key characteristic parameters, the characteristic boundaries, and the time series characteristics into a pre-trained fault classification model to identify the fault type of the fuel cell includes: Analyzing the key characteristic parameters, the characteristic boundaries, and the time series characteristics according to the fault classification model to determine whether the fault type of the fuel cell is no fault, a minor fault, or a major fault; Among them, the fault-free characterization indicates that the key characteristic parameters are within the threshold range of the normal operating condition, the minor fault characterization indicates that the key characteristic parameters are within the threshold range of the critical operating condition, and the serious fault characterization indicates that the key characteristic parameters are within the threshold range of the abnormal operating condition.

6. The method according to claim 5, characterized in that The early warning according to the fault type of the fuel cell includes: Determining the warning level corresponding to the fault type of the fuel cell according to a pre-established correspondence between the fault type and the warning level; Early warning measures are triggered based on the early warning level.

7. A fuel cell vehicle monitoring device, characterized in that: include: An acquisition module is used to acquire fuel cell data and hydrogen storage data; A processing module, configured to perform data preprocessing on the fuel cell data and the hydrogen storage data to obtain a preprocessing result; constructing a multidimensional feature vector based on the preprocessing result, wherein the multidimensional feature vector represents the operating parameters of the fuel cell; analyzing the multidimensional feature vector to obtain key feature parameters, feature boundaries, and time series features, wherein the feature boundaries are used to determine the threshold range of the key feature parameters under different operating conditions, and the time series features represent the changing trend of the key feature parameters in the time dimension; and inputting the key feature parameters, the feature boundaries, and the time series features into a pre-trained fault classification model to identify the fault type of the fuel cell; The early warning module is used to issue an early warning according to the fault type of the fuel cell.

8. A fuel cell vehicle monitoring system, characterized in that: Including data acquisition device, data transmission device and monitoring device; The data acquisition device is used to collect fuel cell data and hydrogen storage data of the vehicle; The data transmission device is used to transmit the fuel cell data and hydrogen storage data collected by the data acquisition device to the monitoring device; The monitoring device is used to perform data preprocessing on the fuel cell data and hydrogen storage data to obtain a preprocessing result; A multidimensional feature vector is constructed based on the preprocessing result, and the multidimensional feature vector represents the operating parameters of the fuel cell; the multidimensional feature vector is analyzed to obtain key feature parameters, feature boundaries and time series characteristics, the feature boundaries are used to determine the threshold range of the key feature parameters under different operating conditions, and the time series characteristics represent the changing trend of the key feature parameters in the time dimension; the key feature parameters, the feature boundaries and the time series characteristics are input into a pre-trained fault classification model to identify the fault type of the fuel cell; and an early warning is issued according to the fault type of the fuel cell.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and the processor can execute the computer program to implement the fuel cell vehicle monitoring method according to any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the fuel cell vehicle monitoring method according to any one of claims 1 to 6 is implemented.

Images