Fuel Cell Data Analysis System and Electronic Device Based on Data Relevance

Abstract

The present application discloses a fuel cell data analysis system and an electronic device based on data correlation. The system includes: a fuel cell big data storage module, an association information storage module, and a data analysis module. The fuel cell big data storage module is used to store component parameters, R & D data, production data, and real-time operation data of the fuel cell. The association information storage module is used to generate and store the association information between multiple levels of target data corresponding to the fuel cell according to the data stored in the fuel cell big data storage module, and store the theoretical model of the fuel cell. The data analysis module is used to construct multiple neural network models according to the association information and the theoretical model, and perform data analysis and fault prediction on the fuel cell through the multiple neural network models. The system defines the association rules between the data at all levels of the fuel cell, facilitating the extraction of corresponding data according to the data analysis requirements, and improving the efficiency and accuracy of data analysis.

Description

Technical Field

[0001] The present application relates to the technical field of fuel cells, and particularly to a fuel cell data analysis system and an electronic device based on data correlation. Background Art

[0002] A proton exchange membrane fuel cell (PEMFC) is a device that directly converts the chemical energy of hydrogen fuel into electrical energy. At present, proton exchange membrane fuel cells have been increasingly widely used in many fields such as transportation, distributed power generation, and industrial production due to their advantages of high efficiency, high power density, environmental friendliness, light weight, and rich resources, and have become one of the most promising new power generation devices. In the field of fuel cell materials, scientific research on predicting material properties through machine learning shows great potential, and in order to improve the competitiveness of fuel cells, the work of using big data analysis for material screening, structural design, and operation management has gradually become the mainstream trend. However, as an emerging industry, some key technologies of fuel cells have not been fully verified, and the data accumulation is insufficient.

[0003] In related technologies, big data analysis is mainly carried out for a certain material screening or a certain structural design and development, lacking a big data analysis platform for a complete cycle of fuel cell development, production, and operation.

[0004] In addition, it is particularly important to identify various faults that may occur in the complex process of fuel cell systems, and to monitor the working efficiency and safety of fuel cells. Analyzing fuel cell performance degradation (or called "life prediction") and health status diagnosis (or called "fault diagnosis") for data provides a method for predicting the remaining life of fuel cells. However, in related technologies, the prediction and diagnosis are mainly carried out for lithium battery vehicles, and the fault analysis and mathematical modeling calculation methods for lithium-ion batteries are not applicable to fuel cells. Therefore, there is an urgent need for a fault prediction and analysis method applicable to fuel cell systems at present.

[0005] Moreover, due to the large number of fuel cell system devices and components, including valves, pipelines, electrical equipment, insulation equipment, etc., the data volume of fuel cell systems is relatively large. When performing big data analysis in related technologies, data screening and noise reduction cannot be accurately and efficiently carried out, resulting in low accuracy of data analysis. Summary of the Invention

[0006] The purpose of the present application is to solve at least one of the above technical problems to some extent.

[0007] To this end, the first object of the present application is to propose a fuel cell data analysis system based on data correlation. The system defines the correlation rules between data at all levels of the fuel cell, facilitating the extraction of corresponding data according to data analysis requirements, improving the efficiency and accuracy of data analysis, and solving the problem of unclear data correlation items in the application of big data analysis in the field of fuel cells.

[0008] The second object of the present application is to propose an electronic device.

[0009] To achieve the above object, an embodiment of the first aspect of the present application proposes a fuel cell data analysis system based on data correlation, which includes:

[0010] A fuel cell big data storage module, a correlation information storage module, and a data analysis module, wherein the fuel cell big data storage module, the correlation information storage module, and the data analysis module are interconnected.

[0011] The fuel cell big data storage module is used to store the component parameters, R & D data, production data, and real-time operation data of the fuel cell.

[0012] The correlation information storage module is used to generate and store the correlation information between the multi-layer target data corresponding to the fuel cell according to the data stored in the fuel cell big data storage module, and store the theoretical model of the fuel cell.

[0013] The data analysis module is used to construct multiple neural network models according to the correlation information and the theoretical model, and perform data analysis and fault prediction on the fuel cell through the multiple neural network models.

[0014] In addition, the distributed air-cooled fuel cell system of the embodiment of the present application also has the following additional technical features:

[0015] Optionally, in some embodiments, the associated information storage module is specifically configured to: establish the association between the material parameters of the fuel cell according to the historical data of the fuel cell and the theoretical model, where the historical data includes the component parameters, the R & D data, and the production data; establish the association between the material parameters of the fuel cell and the components of the fuel cell according to the historical data and the theoretical model; establish the association between the components of the fuel cell and the single-cell structure of the fuel cell according to the test data of the fuel cell and expert knowledge; establish the association between the single-cell structure and the key equipment corresponding to the fuel cell according to the historical data and the theoretical model; establish the association between the key equipment and the power system in which the fuel cell is located, and the association between the external environmental factors and the power system according to the historical data and the attenuation model of the fuel cell, so as to generate the association information between the multi-layer target data in a tree structure.

[0016] Optionally, in some embodiments, the associated information storage module and the data analysis module are connected through a data transmission interface. The associated information storage module further includes: a relevance query sub-module, which is configured to obtain the parameter to be studied; the associated information storage module is further configured to determine the data set corresponding to the parameter to be studied based on the associated information; the data analysis module is further configured to extract the data set from the associated information storage module through the data transmission interface.

[0017] Optionally, in some embodiments, the data analysis module includes: a data preprocessing sub-module, a statistical analysis sub-module, and a machine learning sub-module. The machine learning sub-module is configured to: obtain the theoretical model corresponding to the neural network model to be trained from the associated information storage module; encode the neural network model to be trained according to the corresponding theoretical model.

[0018] Optionally, in some embodiments, the data preprocessing sub-module is configured to: remove the spike data in the data set; fill in the invalid values in the data set; normalize the data through one-hot encoding.

[0019] Optionally, in some embodiments, the fuel cell big data storage module is remotely connected to a remote monitoring system, a component material management system, and a workshop production information management system. The fuel cell big data storage module is specifically configured to: obtain the component parameters of the fuel cell through the component material management system, and obtain the R & D data and the production data through the workshop production information management system.

[0020] Optionally, in some embodiments, the fuel cell big data storage module includes multiple relational databases, a first non-relational database, and a distributed file system. The fuel cell big data storage module is further configured to: receive the real-time operation data of the fuel cell sent by the remote monitoring system; analyze the real-time operation data through an open source stream processing platform or a second non-relational database, and store the analysis results in the relational database, the first non-relational database, and the distributed file system.

[0021] Optionally, in some embodiments, the fuel cell big data storage module is further configured to: detect whether there is duplicate data in the received component parameters, research and development data, production data, and real-time operation data; if there is duplicate data, delete the duplicate data.

[0022] Optionally, in some embodiments, the association information storage module is further configured to: according to the update instruction received by the data transmission interface, perform corresponding deletion, addition, or modification on the association information and the theoretical model.

[0023] The technical solutions provided by the embodiments of the present application at least bring the following beneficial effects: The present application first accesses the data storage systems of the procurement, research and development, production, and operation ends through an integrated open big data storage module, realizing the non-interference and interactivity of data from different sources, and reducing the hardware requirements by deleting duplicate data. This big data storage module stores data based on the fuel cell product structure, improving the comprehensiveness of the stored data and meeting the data analysis requirements in various aspects such as the research and development, production, operation, and after-sales of fuel cells. Moreover, based on technical and production experience and theoretical models, the correlation between fuel cell data is obtained and stored in the associated data storage module in the system. By setting an open interface, it is convenient to add, delete, or modify association information according to the actual operation conditions, and a front-end data correlation query interface is also provided to facilitate accurately and efficiently extracting relevant data when analyzing data later. Additionally, by storing the theoretical model of the associated parameters, it provides retrieval when training the neural network model, facilitating more accurate data analysis. This system also integrates multiple machine learning algorithms for data analysis and prediction, and selects data sets based on data correlation, thereby improving the prediction accuracy of the model, reducing the computational amount of the model, and further improving the accuracy and efficiency of fuel cell data analysis.

[0024] To achieve the above object, a second aspect embodiment of the present invention proposes an electronic device, including the fuel cell data analysis system based on data correlation as described in the above embodiment.

[0025] The additional aspects and advantages of the present application will be partially given in the following description, partially become apparent from the following description, or be understood through the practice of the present application. Brief Description of the Drawings

[0026] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of embodiments in conjunction with the accompanying drawings, where:

[0027] Figure 1 is a schematic structural diagram of a fuel cell data analysis system based on data relevance proposed in an embodiment of the present application;

[0028] Figure 2 is a schematic structural diagram of a specific fuel cell data analysis system based on data relevance proposed in an embodiment of the present application;

[0029] Figure 3 is a schematic diagram of a fuel cell big data storage module docking with an external system proposed in an embodiment of the present application;

[0030] Figure 4 is a schematic flow diagram of an association information storage module generating association information proposed in an embodiment of the present application;

[0031] Figure 5 is a schematic structural diagram of an electronic device proposed in an embodiment of the present application. Detailed Description of the Embodiments

[0032] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain the present application and should not be construed as limiting the present application.

[0033] It should be noted that since there are many components and parameters in a fuel cell and many mechanisms of the battery are not yet fully understood, data screening and noise reduction are crucial in the big data analysis of fuel cells. The applicant has found that in order to perform reliable big data analysis on fuel cells, it is necessary to determine the relevance of fuel cell data and continuously update the relevance with the deepening understanding in actual applications. For this reason, the present application proposes a fuel cell data analysis system based on data relevance, which solves the problem of unclear data association items in the application of big data analysis in the field of fuel cells, defines association rules, facilitates data extraction, and improves the efficiency and accuracy of data analysis.

[0034] The fuel cell data analysis system based on data relevance in the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0035] Figure 1 is a schematic structural diagram of a fuel cell data analysis system based on data relevance proposed in an embodiment of the present application, as Figure 1As shown in the figure, the system includes: a fuel cell big data storage module 10, an associated information storage module 20, and a data analysis module 30. Among them, the fuel cell big data storage module 10, the associated information storage module 20, and the data analysis module 30 are interconnected to achieve data transmission.

[0036] Among them, the fuel cell big data storage module 10 is used to store the component parameters, R & D data, production data, and real-time operation data of the fuel cell.

[0037] Among them, the real-time operation data includes the monitoring data of the fuel cell in actual applications. For example, the single-cell voltage and stack voltage of the battery collected by sensors. And the real-time operation data also includes the real-time operation data, status data of the equipment powered by the fuel cell (such as new energy vehicles or drones), and environmental data of the environment where it is located. For example, the real-time operation data also includes the driving mileage, vehicle condition, warning information of the vehicle powered by the fuel cell, and road conditions and weather in the current environment.

[0038] The component parameters include the physical and chemical parameters of each component of the fuel cell, such as the proton exchange membrane, catalyst, bipolar plate, carbon paper, and gasket of the fuel cell. For example, the size of the gasket and the activity of the catalyst. The R & D data and production data include the data generated during the design of the fuel cell and the production of the battery in the workshop, such as design drawings, processing technologies, and processing materials, etc.

[0039] In an embodiment of the present application, when specifically obtaining the above-mentioned component parameters, R & D data, production data, and real-time operation data, the fuel cell big data storage module 10 can be remotely connected to a remote monitoring system, a component material management system ERP, and a workshop production information management system MES. The fuel cell big data storage module 10 can obtain the component parameters of the fuel cell through the component material management system and obtain the R & D data and production data through the workshop production information management system.

[0040] Moreover, the remote monitoring system platform can first collect the operating data of the whole vehicle and the operating data of the fuel cell in real time through devices such as in-vehicle terminals, and after backing up, send them to the fuel cell big data storage module 10. Then, the fuel cell big data storage module 10 analyzes the real-time operating data through the open-source stream processing platform Kafka or the non-relational database Redis, and further obtains the analysis results of the real-time operating data through real-time stream analysis. Among them, the analysis results include the judgment results and statistical results of the real-time operating data. For example, the fault detection results analyzed according to the warning information, or the statistical vehicle emission data and driving trajectory data, etc. Then, the fuel cell big data storage module 10 stores the analysis results in the relational database MySQL, non-relational database HBase, and distributed file system HDFS preset in this module. Among them, in the embodiment of the present application, in order to distinguish the non-relational data for real-time stream analysis and data storage, Hbase is used as the first non-relational database, and Redis is used as the second non-relational database.

[0041] It should be noted that in actual applications, since the fuel cell big data storage module 10 can receive data in different ways, there may be duplicates in the received data. Therefore, in an embodiment of the present application, the fuel cell big data storage module 10 is further configured to detect whether there are duplicate data in the received component parameters, R & D data, production data, and real-time operating data. If it is determined that there are duplicate data, the duplicate data is deleted to reduce bandwidth consumption and lower the hardware requirements for the system.

[0042] The association information storage module 20 is configured to generate and store the association information between the multi-layer target data corresponding to the fuel cell based on the data stored in the fuel cell big data storage module, and store the theoretical model of the fuel cell.

[0043] Among them, the multi-layer target data corresponding to the fuel cell includes the data involved in different structural levels that make up the fuel cell. Each layer includes the material layer, component layer, and system layer of the fuel cell, etc. There is an association between the levels of the multi-layer target data.

[0044] Among them, the theoretical model of the fuel cell is the fuel cell physical model involved in the multi-layer target data of the fuel cell, including the Nernst - Planck equation for ion transport in the proton exchange membrane, the Darcy's law and Navier - Stokes equations related to two-phase flow in the gas diffusion layer, etc.

[0045] In an embodiment of the present application, when the association information storage module 20 establishes association information, it specifically is used for: first, based on the historical data and theoretical model of the fuel cell, establishing the association between the material parameters of the fuel cell. The historical data includes component parameters, R & D data, and production data. Then, based on the historical data and theoretical model, establishing the association between the material parameters of the fuel cell and the components of the fuel cell. Next, based on the test data and expert knowledge of the fuel cell, establishing the association between the components of the fuel cell and the single-cell structure of the fuel cell. Then, based on the historical data and theoretical model, establishing the association between the single-cell structure and the key equipment corresponding to the fuel cell. Finally, based on the historical data and the attenuation model of the fuel cell, establishing the association between the key equipment and the power system in which the fuel cell is located, as well as the association between the external environmental factors and the power system, so as to generate the association information between the multi-layer target data in a tree structure.

[0046] Specifically, the historical data may further include the experience of technicians and literature records during the design and production of the fuel cell. The key equipment corresponding to the fuel cell refers to the equipment that has a greater impact on the performance of the fuel cell, including a controller, an air compressor, a hydrogen injector, etc.

[0047] Furthermore, the association information storage module 20 may also provide the function of querying association information. In an embodiment of the present application, the association information storage module 20 and the data analysis module 30 are connected through a data transmission interface. The association information storage module 20 further includes an association query sub-module 21. The association query sub-module 21 is used to obtain the parameter to be studied. The association information storage module 20 is further used to determine the data set corresponding to the parameter to be studied based on the association information, and send the data set to the data analysis module 30. In this example, the association query sub-module 21 includes a data association query interface to determine the parameter to be studied currently. As a possible implementation manner, the association query sub-module 21 may establish a connection with the client to obtain the data queried from the front end. As another possible implementation manner, the association query sub-module 21 may also be set in the client to obtain the data input by the user through the human-computer interaction interface.

[0048] Thus, the association information storage module 20 can extract various parameters or frequent items related to the currently analyzed data from the fuel cell big data storage module 10 according to the association rules, generate a data set corresponding to the data analysis requirement, and the data analysis module 30 then extracts the data set from the association information storage module 20 through the data transmission interface, eliminating the complex process of data set selection.

[0049] In one embodiment of the present application, the associated information storage module 20 is further configured to delete, add, or modify the associated information and the theoretical model according to the update instruction received by the data transmission interface. Specifically, by receiving the update instruction of the associated information sent by the client or the data analysis module 30 through an open interface, the associated information can be updated according to the new association rules that are more accurate and applicable determined through verification in actual applications or analyzed by the data analysis module, so as to improve the accuracy and timeliness of the associated information established in the present application.

[0050] The data analysis module 30 is configured to construct a plurality of neural network models according to the associated information and the theoretical model, and perform data analysis and fault prediction on the fuel cell through the plurality of neural network models.

[0051] In one embodiment of the present application, the data analysis module 30 includes: a data preprocessing sub-module 31, a statistical analysis sub-module 32, and a machine learning sub-module 33. Among them, the machine learning sub-module 33 is configured to obtain the theoretical model corresponding to the neural network model to be trained from the associated information storage module 20, and encode the neural network model to be trained according to the corresponding theoretical model. Thus, a high-precision network model can be established by combining the relevant theoretical knowledge stored in the system.

[0052] In one embodiment of the present application, the data preprocessing sub-module 31 is specifically configured to: first remove the spike data in the data set, then fill in the invalid values in the data set, and then standardize the data through one-hot encoding.

[0053] In the embodiment of the present application, the data analysis module 30 first preprocesses the extracted data set, including data filtering and encoding, etc., performs data analysis through a variety of statistical analysis algorithms according to the processed data, and establishes a neural network model through machine learning for analysis and prediction such as fuel cell performance degradation analysis and remaining service life prediction, realizing functions such as fuel cell data analysis and fault prediction.

[0054] In order to more clearly describe the specific process of realizing data analysis by the fuel cell data analysis system based on data correlation in the present application, the following takes a specific fuel cell data analysis system based on data correlation as an example for illustration. Figure 2 It is a schematic structural diagram of a specific fuel cell data analysis system based on data correlation proposed in the embodiment of the present application. It should be noted that, for the convenience of describing the functions of each module in the system, the data processed or stored by each module, as well as the data processing process, are also described in this schematic diagram, which is not intended to limit the system structure.

[0055] As Figure 2As shown in the figure, the fuel cell big data storage module 10 can be an open data storage platform. The fuel cell big data storage module 10 includes call interfaces for the ERP database and the MES database. By calling ERP data and MES data, component parameters, R & D data, and production data are extracted and stored. The fuel cell big data storage module 10 stores common physical models of various components in the fuel cell, including the gas diffusion layer, proton exchange membrane, catalyst, bipolar plate, etc., as well as parameters of each physical model. For example, it includes material parameters of components such as carbon paper, proton exchange membrane, and bipolar plate, as well as performance parameters of single cells and stacks, component parameters of the stack system, etc. Among them, the interface of the fuel cell big data storage module 10 is extensible to support the interaction of multi-file system clients. Files can be read and written to the fuel cell big data storage module 10 to continuously improve the data stored in this module in practical applications.

[0056] Furthermore, when the fuel cell big data storage module 10 obtains real-time operation data, as Figure 3 shown, as a possible implementation method, the fuel cell big data storage module 10 can be remotely docked with the remote monitoring platform, ERP database, and MES database. Among them, the remote monitoring system platform collects the overall vehicle operation data of the vehicle powered by the fuel cell in real time through the in-vehicle terminal, and obtains the operation information of single cells and stacks collected by the remote information processing module T-BOX through the inspection system.

[0057] In specific implementation, the working data of the fuel cell vehicle monitoring unit can be transmitted through a CAN interface transmission device pre-set in the fuel cell vehicle. After being packaged according to the communication protocol, the data is sent to the remote monitoring center in real time through GPRS wireless communication technology. Then, at the remote monitoring center, the received information is parsed to obtain the monitoring data collected by the T-BOX, and the monitoring data is classified. The sensitive data is locally stored and backed up, and the non-sensitive data is stored in the cloud storage platform for standby. Among them, the sensitive data may include the real-time operation data of the whole vehicle and the fuel cell required for analyzing the fuel cell. Furthermore, the remote monitoring center sends the collected real-time operation data to the fuel cell big data storage module 10 for storage. If the remote monitoring center fails, the received fuel cell product operation data can be sequentially stored in the in-vehicle terminal in the order of time to avoid data loss. After the remote monitoring center recovers from the failure, the cached data is uploaded by the in-vehicle terminal. Moreover, the remote monitoring center can perform real-time data analysis and real-time storage of the uploaded fuel cell products through the display of the monitoring page, and perform fault diagnosis by analyzing the existing data, thereby realizing remote monitoring of the fuel cell products. The remote monitoring center can also send the generated diagnostic data and monitoring data to the fuel cell big data storage module 10 to assist in the data analysis of the fuel cell. The fuel cell big data storage module 10 can also be connected to the user's mobile terminal to facilitate the user to query and call the required data in real time.

[0058] Furthermore, the fuel cell big data storage module 10 performs real-time stream analysis on the received real-time operation data. Specifically, the message queue can be input into Kafka or Redis for real-time stream analysis, and the obtained analysis results, such as real-time trajectory monitoring, fault detection results, emission data monitoring, as well as the obtained warning information and fuel cell real-time statistical data, etc., are persistently stored. Specifically, it can be stored offline through a non-relational database preset by the storage platform, such as Hbase and Redis, and a relational database, such as MySQL, etc.

[0059] Among them, the association information storage module 20 classifies and stores the data designed for the fuel cell according to the structural level. Starting from the material layer, the association relationship between the data of each layer is established from bottom to top until the power system, and the correlation of key parameters is determined. Thus, the association information storage module 20 determines that the association level is clear, and the association information can be represented by an association tree diagram.

[0060] To more clearly illustrate the specific implementation process of the association information storage module 20 generating the association information between the multi-layer target data corresponding to the fuel cell, the following is an example of a specific method for the association information storage module to generate association information, as Figure 4As shown in the figure, the associated information storage module 20 can execute the following steps in this method:

[0061] Step S1: Combine the experience of laboratory technicians, literature records, and theoretical models to store the correlations between material parameters.

[0062] Specifically, in this step, store the relevant parameter categories that affect the performance of the fuel cell. As an implementation method, only record the parameter names and types without storing specific values to reduce the occupied storage space.

[0063] During specific implementation, the historical records of proton exchange membrane materials, catalyst materials, bipolar plate metal / graphite materials, carbon paper materials, and sealant materials, etc., can be obtained from the membrane electrode laboratory, catalyst laboratory, bipolar plate laboratory, and open material database in sequence to determine the correlations between material parameters. For example, it can include the molecular structure of the proton membrane polymer and its conductivity, the specific surface area of the catalyst material, etc.

[0064] Step S2: Combine the experience of laboratory technicians, literature records, and theoretical models to establish the correlations between the lower-layer key material parameters and the parameters of the components at this layer.

[0065] Among them, the components can include components that make up the fuel cell, such as the membrane electrode, bipolar plate, and current collector plate of the fuel cell.

[0066] Specifically, in this step, the lower-layer structural material parameters affect the key parameters of the components at this layer, such as the ion exchange rate, catalyst activity, etc., and the components at this layer affect the relevant parameters of the fuel cell, such as the structural strength, fluid rate, etc.

[0067] Step S3: Combine the single-cell performance obtained from product testing and the experience of relevant experts to establish the correlations between the lower-layer component parameters and the single-cell structure parameters at this layer.

[0068] Among them, the single-cell performance includes parameters such as the voltage and current of the single cell.

[0069] Specifically, in this step, the lower-layer structural parameters affect the key parameters of the components at this layer, such as the catalyst loading and the durability of the membrane electrode, etc., while the components at this layer affect the relevant parameters of the fuel cell's single-cell current, voltage, and decay rate, etc.

[0070] Step S4: Combine the experience of production technicians, literature records, and theoretical models to establish the correlations between the lower-layer single-cell structure and the key equipment at this layer.

[0071] Among them, the key equipment includes the stack, controller, air compressor, hydrogen injector, and hydrogen cylinder, etc.

[0072] Specifically, in this step, the structure of the lower layer affects the battery performance of the components in this layer, and the components in this layer affect relevant parameters such as the hydrogen storage capacity and stack power of the fuel cell.

[0073] Step S5: Based on the experience of production and after-sales personnel, literature records, and attenuation models, establish the correlation between the key equipment parameters of the lower layer and the power system parameters of this layer, as well as the correlation between external factors and power system parameters.

[0074] Specifically, in this step, the structure of the lower layer affects parameters such as the stack power, water and heat management, and air flow rate of the power system in this layer, while the components in this layer affect relevant parameters such as the durability and safety of the fuel cell.

[0075] Thus, the correlation between the data of each level of the fuel cell is obtained based on technical and production experience and theoretical models, which is conducive to realizing reliable big data analysis.

[0076] Moreover, the association information storage module 20 also includes an open interface, which can delete, add, and modify parameter types and correlations according to the received instructions, making the association storage database 20 scalable and easy to update. In practical applications, as the understanding of fuel cell technology deepens, data extraction can be completed by updating the association information without changing the data tags of the fuel cell data storage platform, further improving the convenience of data extraction.

[0077] In addition, as Figure 2 shown, the association information storage module 20 also stores the theoretical models involved in the target data of each level of the fuel cell system, including the Nernst - Planck equation for ion transport in the proton exchange membrane, the Darcy's law and Navier - Stokes equations for two - phase flow in the gas diffusion layer, etc. Among them, the association information storage module 20 can store the corresponding theoretical models classified by the levels of the target data. For example, the Fick's law, gas leakage model, and voltage decay model corresponding to the power system, or the density function corresponding to the density material, or the fracture criterion corresponding to the structural design, etc., so as to improve the speed of obtaining the corresponding theoretical models.

[0078] The association information storage module 20 also includes a data extraction and output interface, which can extract data from the fuel cell big data storage module 10 and output the determined data set to the data analysis module 30.

[0079] Among them, the associated information storage module 20 can also mine the fuel cell data required for the machine learning of the data analysis module 30, and extract the corresponding data set and send it to the data analysis module 30. In specific implementation, the associated information storage module 20 can, based on the association information between the multi-layer target data generated by analysis, query the parameters to be studied from the client through the association storage system interface, obtain the corresponding parameters or frequent items according to the association rules, and then extract the data set from the fuel cell big data storage module 10 according to the determined parameters or frequent items. Thus, the complexity of selecting the data set is reduced, and the associated data can be quickly retrieved through the data association storage platform, which is beneficial for users who are not fuel cell professional technicians to perform data analysis.

[0080] Continue to refer to Figure 2 As shown, the data analysis module 30 includes a data preprocessing module 31, a statistical analysis module 32, a machine learning module 33, and an association rule update module 34. Among them, the data preprocessing module 31 preprocesses the extracted data set, including data filtering and encoding. In specific implementation, the burr data can be removed through a preset artificial experience threshold and statistical analysis, the null values can be filled by interpolation or average method, and the data can be standardized through one-hot encoding. The statistical analysis module 32 can perform accident statistics and warning statistics through various statistical analysis algorithms such as the Weibull distribution, and analyze the data of the fuel cell. The machine learning module 33 integrates various algorithms such as CNN, RNN, ANN, and SVM, and provides parameter adjustment annotations at the activation function, loss function, and learning rate. Through the associated information storage module 20, the relevant theoretical models of the prediction and analysis content can be visualized on the client. The neural network model can be developed in combination with the physical model, and the visualized relevant theoretical model can be encoded into the neural network model to develop a high-precision network model. The association rule update module 34 can update the newly determined association rules through the analysis of the neural network model to the associated information storage module 20 to improve the accuracy of the stored association information. The data analysis module 30 can also store the trained machine learning network model available for prediction to the fuel cell big data storage module to reduce the storage capacity of the data analysis module and ensure the operation performance of the data analysis module 10.

[0081] Thus, this fuel cell data analysis system based on data association realizes the retrieval of data related to research questions in the fuel cell big data analysis platform through the interface of the associated information storage module, and can update the machine learning neural network according to the associated theoretical model, improving the model prediction accuracy and facilitating data analysis.

[0082] In summary, the fuel cell data analysis system based on data correlation in the embodiments of the present application first accesses the data storage systems of procurement, R & D, production, and operation through an integrated open big data storage module, realizing the non-interference and interactivity of data from different sources, and reducing the hardware requirements by deleting duplicate data. This big data storage module stores data based on the fuel cell product structure, improving the comprehensiveness of the stored data and meeting the data analysis requirements in various aspects such as R & D, production, operation, and after-sales of fuel cells. Moreover, based on technical and production experience and theoretical models, the correlation between fuel cell data is obtained and stored in the associated data storage module in the system. By setting open interfaces, it is convenient to add, delete, or modify associated information according to the actual operating conditions. A front-end data correlation query interface is also provided to facilitate the accurate and efficient extraction of relevant data when analyzing data later. Additionally, by storing the theoretical models of associated parameters, retrieval is provided during the training of the neural network model, facilitating more accurate data analysis. This system also integrates multiple machine learning algorithms for data analysis and prediction, and selects data sets based on data correlation, thereby improving the prediction accuracy of the model, reducing the computational amount of the model, and further enhancing the accuracy and efficiency of fuel cell data analysis.

[0083] Based on the above embodiments, in order to more clearly illustrate the setting method and data analysis process of the fuel cell data analysis system based on data correlation of the present application in an actual application scenario, a specific embodiment is described below.

[0084] In this embodiment, the T-BOX can be first installed in the fuel cell power system, and an IoT card can be installed in the system to receive signals. Then, a remote monitoring system is arranged to collect real-time vehicle information, process the real-time data through the queue information Kafka, and display it in real time on the remote monitoring platform interface. An offline data storage is carried out by building a cluster environment in the fuel cell data analysis platform, and the ERP system and MES system in the data platform are connected to access component numbers, R & D, and production test data. Then, according to actual needs and in combination with theoretical models, the relationships between fuel cell parameters are analyzed. For example, there is a correlation between the relative humidity, temperature, flow rate, current, and pressure of the anode and cathode in a single cell. The data correlation and related theoretical models are stored in the database, and can be extracted through the interface during subsequent data analysis. The extracted correlation data is preprocessed and then further data mining and prediction are carried out through the neural network model tool.

[0085] To implement the above embodiments, an electronic device is also proposed in the embodiments of the present invention. Figure 5 The structural schematic diagram of an electronic device proposed in the embodiments of the present application.

[0086] As Figure 5As shown, the electronic device 1000 may include the fuel cell data analysis system 2000 based on data correlation as described in the above embodiments. The electronic device in the embodiments of the present application may be an application server or the like. By running the fuel cell data analysis system based on data correlation therein, the accuracy and efficiency of the electronic device in analyzing fuel cell data are improved.

[0087] It should be noted that in the description of the present application, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0088] Any process or method description shown in the flowchart or described in other ways herein can be understood to represent a module, segment, or part of code including one or more executable instructions for implementing a specific logical function or process. The scope of the preferred embodiments of the present application includes additional implementations, where the functions may be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed. This should be understood by those skilled in the technical field to which the embodiments of the present application belong.

[0089] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0090] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A fuel cell data analysis system based on data correlation, characterized in that Including: A fuel cell big data storage module, an associated information storage module, and a data analysis module, wherein the fuel cell big data storage module, the associated information storage module, and the data analysis module are interconnected. The fuel cell big data storage module is used to store the component parameters, R & D data, production data, and real-time operation data of the fuel cell. Among them, the real-time operation data includes the monitoring data of the fuel cell in actual application, and also includes the real-time operation data, status data, and environmental data of the environment where the equipment powered by the fuel cell is located. The associated information storage module is used to generate and store the association information between the multi-layer target data corresponding to the fuel cell according to the data stored in the fuel cell big data storage module, and store the theoretical model of the fuel cell. The multi-layer target data corresponding to the fuel cell includes the data involved in different structural levels that make up the fuel cell. Each layer includes the material layer, component layer, and system layer of the fuel cell. There is an association between the levels of the multi-layer target data. Among them, the theoretical model of the fuel cell includes the Nernst-Planck equation for ion transport in the proton exchange membrane, the Darcy's law and Navier-Stokes equation for two-phase flow in the gas diffusion layer. The data analysis module is used to construct multiple neural network models based on the association information and the theoretical model, and perform data analysis and fault prediction on the fuel cell through the multiple neural network models. The associated information storage module is specifically used for: Establish the association between the material parameters of the fuel cell according to the historical data of the fuel cell and the theoretical model. The historical data includes the component parameters, R & D data, and production data. Establish the association between the material parameters of the fuel cell and the components of the fuel cell according to the historical data and the theoretical model. Establish the association between the components of the fuel cell and the single cell structure of the fuel cell according to the test data of the fuel cell and expert knowledge. Establish the association between the single cell structure and the key equipment corresponding to the fuel cell according to the historical data and the theoretical model. Establish the association between the key equipment and the power system where the fuel cell is located, and the association between the external environmental factors and the power system according to the historical data and the attenuation model of the fuel cell, so as to generate the association information between the multi-layer target data in a tree structure.

2. The system according to claim 1, wherein The associated information storage module and the data analysis module are connected through a data transmission interface. The associated information storage module further includes: a relevance query sub-module. The relevance query sub-module is used to obtain the parameters to be studied. The associated information storage module is further used to determine the data set corresponding to the parameter to be studied based on the association information. The data analysis module is further used to extract the data set from the associated information storage module through the data transmission interface.

3. The system according to claim 2, wherein The data analysis module includes: a data preprocessing sub-module, a statistical analysis sub-module, and a machine learning sub-module. The machine learning sub-module is used for: Obtain the theoretical model corresponding to the neural network model to be trained from the associated information storage module; Encode the neural network model to be trained according to the corresponding theoretical model.

4. The system according to claim 3, wherein The data preprocessing sub-module is used for: Remove the spurious data in the dataset; Fill in the invalid values in the dataset; Normalize the data through one-hot encoding.

5. The system according to claim 1, wherein The fuel cell big data storage module is remotely connected to the remote monitoring system, the component material management system and the workshop production information management system. Specifically, the fuel cell big data storage module is used for: obtaining the component parameters of the fuel cell through the component material management system, and obtaining the R & D data and the production data through the workshop production information management system.

6. The system according to claim 5, wherein The fuel cell big data storage module includes multiple relational databases, a first non-relational database and a distributed file system. The fuel cell big data storage module is further used for: Receive the real-time operation data of the fuel cell sent by the remote monitoring system; Analyze the real-time operation data through an open-source stream processing platform or a second non-relational database, and store the analysis results in the relational database, the first non-relational database and the distributed file system.

7. The system according to claim 5, wherein The fuel cell big data storage module is further used for: Detect whether there are duplicate data in the received component parameters, R & D data, production data and real-time operation data; If there are duplicate data, delete the duplicate data.

8. The system according to claim 1, wherein The associated information storage module is further used for: Delete, add or modify the associated information and the theoretical model correspondingly according to the update instruction received by the data transmission interface.

9. An electronic device, characterized in that, Including the fuel cell data analysis system based on data relevance according to any one of claims 1-8.

Images