Battery internal state identification method and device, computer equipment and storage medium

Abstract

The invention relates to a battery internal state identification method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring first measurement data of a first battery; inputting the first measurement data into a state reconstruction model, and obtaining internal state data, reconstructed by the state reconstruction model, of the first battery; wherein the state reconstruction model is obtained by training the second measurement data and the internal state data corresponding to the second measurement data as a training set, and the second measurement data and the internal state data corresponding to the second measurement data are obtained based on a simulation model of the second battery; the internal state data comprises at least one of the oxygen concentration of a cathode catalyst layer, the water content of a proton exchange membrane, the water vapor concentration of a cathode flow channel and the liquid water saturation of the cathode catalyst layer, and the algorithm architecture of the state reconstruction model corresponds to the internal state data. And the accuracy of identifying the internal state data is improved.

Description

Technical Field

[0001] The present application relates to the technical field of battery state recognition, and particularly to a method, device, computer device, and storage medium for recognizing the internal state of a battery. Background Art

[0002] In future long-distance heavy-load, ship, aircraft, power generation, and other operating scenarios, the requirements for fuel cells are high power, high efficiency, and long life. Since the fuel cell stack is developing in the direction of large area, the phenomenon of uneven internal gas-water-thermal state is aggravated. The uneven distribution of gas, water, electricity, etc. inside the large-area fuel cell may lead to phenomena such as local membrane drying, flooding, and gas shortage occurring simultaneously. How to quickly reconstruct the internal state of the large-area fuel cell and identify possible local faults is an urgent problem to be solved currently.

[0003] However, there is a lack of a recognition method in traditional technologies that can accurately identify the internal state of a battery. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide a recognition method, device, computer device, and storage medium that can accurately identify the internal state of a battery.

[0005] In a first aspect, the present application provides a method for recognizing the internal state of a battery, the method comprising:

[0006] Obtaining first measurement data of a first battery;

[0007] Inputting the first measurement data into a state reconstruction model to obtain internal state data of the first battery reconstructed by the state reconstruction model;

[0008] Wherein, the state reconstruction model is trained by using second measurement data and internal state data corresponding to the second measurement data as a training set, the second measurement data and the internal state data corresponding to the second measurement data are obtained based on a simulation model of a second battery, the internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content in the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer, and the algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0009] In one of the embodiments, the state reconstruction model includes at least one of the following sub-models:

[0010] A first support vector machine for reconstructing the oxygen concentration in the cathode catalyst layer of the first battery from the first measurement data;

[0011] A first random forest model for reconstructing the first measurement data to obtain the proton exchange membrane water content of the first battery;

[0012] A second support vector machine for reconstructing the first measurement data to obtain the water vapor concentration in the cathode flow channel of the first battery;

[0013] A second random forest model for reconstructing the first measurement data to obtain the liquid water saturation in the cathode catalyst layer of the first battery.

[0014] In one embodiment, the training process of the state reconstruction model includes:

[0015] Obtaining second measurement data according to a simulation model; the second measurement data is the measurement data of multiple battery partitions of the simulation model, and the battery partitions are obtained by partitioning the second battery based on the active area of the battery;

[0016] Inputting the second measurement data into the initial state reconstruction model to obtain the initial internal state data of each of the battery partitions inside the battery;

[0017] Optimizing the initial state reconstruction model according to the internal state data corresponding to the second measurement data and the initial internal state data to obtain the state reconstruction model.

[0018] In one embodiment, the obtaining of the second measurement data according to the simulation model includes:

[0019] Constructing a simulation model corresponding to the internal state of the second battery;

[0020] Obtaining a simulated current density value through the simulation model;

[0021] Determining a current density error based on the simulated current density value and the experimental current density value corresponding to the second battery;

[0022] When the current density error is less than an error threshold, obtaining the second measurement data of the second battery based on the simulation model.

[0023] In one embodiment, the simulated current density value includes an average simulated current density value and a partitioned simulated current density value, the experimental current density value includes an average experimental current density value and a partitioned experimental current density value, and the determining of the current density error based on the simulated current density value and the experimental current density value corresponding to the second battery includes:

[0024] Determining a first current density error according to the average simulated current density value and the average experimental current density value;

[0025] Determine a second current density error according to the simulated value of the partition current density and the experimental value of the partition current density.

[0026] In one embodiment, the inputting the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model includes:

[0027] Input the first measurement data and the position information of the position to be predicted into the state reconstruction model to obtain the internal state data of the position to be predicted; the position to be predicted is determined according to the order of the risk of fault occurrence.

[0028] In one embodiment, after the inputting the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model, the method further includes:

[0029] Determine the fault threshold of the position to be predicted according to a preset fault threshold table and the position information of the position to be predicted;

[0030] Determine whether a fault occurs at the position to be predicted according to the internal state data and the fault threshold; the fault includes at least one of an oxygen deficiency fault, a proton exchange membrane drying fault, a cathode catalyst layer flooding fault, and a flow channel flooding fault.

[0031] In a second aspect, the present application further provides an identification device for the internal state of a battery, including:

[0032] A first acquisition module, configured to acquire first measurement data of a first battery;

[0033] A second acquisition module, configured to input the first measurement data into a state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model;

[0034] Wherein, the state reconstruction model is trained by using second measurement data and internal state data corresponding to the second measurement data as a training set, the second measurement data and the internal state data corresponding to the second measurement data are obtained based on a simulation model of a second battery, the internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer, and the algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0035] In a third aspect, the present application further provides a computer device, including a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0036] Obtain the first measurement data of the first battery;

[0037] Input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model;

[0038] Wherein, the state reconstruction model is trained with the second measurement data and the internal state data corresponding to the second measurement data as the training set. The second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery. The internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer. The algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0039] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0040] Obtain the first measurement data of the first battery;

[0041] Input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model;

[0042] Wherein, the state reconstruction model is trained with the second measurement data and the internal state data corresponding to the second measurement data as the training set. The second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery. The internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer. The algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0043] In a fifth aspect, the present application also provides a computer program product, including a computer program. When the computer program is executed by a processor, the following steps are implemented:

[0044] Obtain the first measurement data of the first battery;

[0045] Input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model;

[0046] Among them, the state reconstruction model is obtained by training the second measurement data and the internal state data corresponding to the second measurement data as a training set. The second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery. The internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer. The algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0047] The above method, device, computer device, and storage medium for identifying the internal state of a battery obtain the first measurement data of the first battery; input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model. Among them, the state reconstruction model is obtained by training the second measurement data and the internal state data corresponding to the second measurement data as a training set. The second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery. The internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer. The algorithm architecture of the state reconstruction model is set corresponding to the internal state data. By obtaining the second measurement data of the second battery through the simulation model, the accuracy and reliability of the second measurement data are improved, thereby improving the accuracy and reliability of the state reconstruction model trained using the second measurement data. Further, by inputting the easily collected first measurement data into the state reconstruction model, the internal state data is reconstructed and identified, improving the accuracy of the internal state data. Description of the Drawings

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments of the present application or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0049] Figure 1 It is an application environment diagram of the method for identifying the internal state of a battery in an embodiment;

[0050] Figure 2 It is a flowchart of the method for identifying the internal state of a battery in an embodiment;

[0051] Figure 3 It is a schematic diagram of the state reconstruction model in an embodiment;

[0052] Figure 4 It is a flowchart of the method for identifying the internal state of a battery in another embodiment;

[0053] Figure 5 Schematic flowchart of the method for identifying the internal state of the battery in another embodiment;

[0054] Figure 6 Schematic flowchart of the method for identifying the internal state of the battery in another embodiment;

[0055] Figure 7 Schematic flowchart of the fault discrimination in one embodiment;

[0056] Figure 8 Schematic flowchart of the method for identifying the internal state of the battery in another embodiment;

[0057] Figure 9 Schematic flowchart of the method for identifying the internal state of the battery in another embodiment;

[0058] Figure 10 Block diagram of the structure of the device for identifying the internal state of the battery in one embodiment;

[0059] Figure 11 Internal structure diagram of a computer device in one embodiment. Detailed implementation manners

[0060] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0061] Currently, the methods for reconstructing the internal state of large-area fuel cells are generally divided into three categories: direct measurement methods, model-based reconstruction methods, and AI-based reconstruction methods.

[0062] Among them, for the direct measurement method of the internal state, in the prior art, methods such as gas sampling, transparent cells, and neutron imaging can be used to directly measure physical quantities such as the gas content and water content inside the battery; or, methods such as pressure drop and AC impedance can be used for indirect measurement of the water content. For the model-based internal state reconstruction method, the modeling and reconstruction methods for large-area fuel cells include CFD numerical models and discrete dynamic models. The models are refined and calibrated through test results such as polarization curves and high-frequency impedance, and then the internal state can be calculated; the artificial intelligence methods can be divided into two categories based on experimental data and simulation data. At the experimental level, in some studies, transparent channels have been tried to predict the liquid water saturation in the cathode of the channel through deep learning algorithms; while using model simulation data to train the AI algorithm, data can be obtained at low cost and the operation efficiency is relatively fast.

[0063] However, for the direct measurement method of internal states, only a few internal states can be directly measured. Moreover, the tests for the gas and water content inside the battery usually cause great interference to the performance of the battery under test, slow signal acquisition, and high experimental costs. For the indirect measurement methods, such as voltage drop and AC impedance, it is difficult to directly obtain the water content information inside the membrane electrode. Usually, only qualitative relationships can be obtained, and it is difficult to reflect the differences in gas and water content at different positions within the large-area battery surface. For the modeling and reconstruction methods, the CFD model can consider spatial details and has high accuracy, but the calculation time is long. Compared with the CFD model, the discrete dynamic model has lower model accuracy and faster calculation speed, and it cannot ensure both accuracy and efficiency at the same time. In addition, the main limitation of the large-area battery state reconstruction method based on the model is that the model calibration is complex and the calculation time is relatively long, making it difficult to meet the requirements of rapid state reconstruction and diagnosis. When training the AI algorithm with model simulation data, the internal state space distribution of the large-scale fuel cell and the input signal are multi-variable input or output problems, with highly nonlinear phenomena.

[0064] Therefore, there is an urgent need for a method for identifying the internal state of a battery with high accuracy and high identification efficiency.

[0065] The method for identifying the internal state of a battery provided in the embodiments of the present application can be applied to an application environment such as Figure 1 shown. Among them, the electronic device 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or placed in the cloud or other network servers. The electronic device 102 may include a simulation model of the second battery. The server 104 can obtain the second measurement data from the electronic device 102, and thus train the initial state reconstruction model according to the second measurement data to obtain a state reconstruction model. Further, the first measurement data of the first battery obtained is input into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model. Among them, the server 104 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0066] In one embodiment, as Figure 2 shown, a method for identifying the internal state of a battery is provided. Taking the method applied to the Figure 1 server as an example for illustration, it includes:

[0067] S201, obtain the first measurement data of the first battery.

[0068] Among them, the first measurement data may include the average current density of the first battery, the cathode inlet oxygen concentration, the cathode flow rate, the anode flow rate, the cathode inlet pressure, the anode inlet pressure, the cathode inlet humidity, the anode inlet humidity, and the battery operating temperature.

[0069] In the embodiments of the present application, the first measurement data can be obtained through a state monitoring device connected to the first battery; alternatively, the first measurement data of the first battery can be obtained through data input by a user to the server. Optionally, the state monitoring device may include a temperature and humidity sensor, a flow rate monitoring device, a pressure sensor, etc.

[0070] S202, input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model.

[0071] Among them, the state reconstruction model is obtained by training the second measurement data and the internal state data corresponding to the second measurement data as a training set. The second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery. The internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content in the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer. The algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0072] In the embodiments of the present application, the first measurement data is input into a pre-trained state reconstruction model to perform reconstruction processing on the first measurement data to obtain the internal state data of the first battery. Optionally, the internal state data of the first battery can be the internal state data of the entire battery, or the internal state at each level of the first battery can be the internal state data at a certain position in the battery.

[0073] Optionally, in the embodiments of the present application, a simulation model of the second battery is established in advance, and the first battery and the second battery are of the same battery type. By collecting the simulation data of the simulation model multiple times, a sample set corresponding to the initial state reconstruction model is obtained. Part of the data in the sample set is used as the training set, and the rest of the data is used as the test set. The initial state reconstruction model is cyclically trained, cross-validated, and the internal parameters of each algorithm are optimized to avoid underfitting and overfitting, and the internal parameters of the optimal state reconstruction model are determined. Optionally, the initial state reconstruction model is constructed according to at least one of the Support Vector Regression (SVR) algorithm and the Random Forest (RF) algorithm.

[0074] Optionally, in the embodiments of the present application, as Figure 3 shown, the state reconstruction model includes at least one of the following sub-models:

[0075] (1) The first support vector machine is used to reconstruct the first measurement data to obtain the oxygen concentration of the cathode catalyst layer of the first battery;

[0076] (2) The first random forest model is used to reconstruct the first measurement data to obtain the water content of the proton exchange membrane of the first battery;

[0077] (3) The second support vector machine is used to reconstruct the first measurement data to obtain the water vapor concentration in the cathode flow channel of the first battery;

[0078] (4) The second random forest model is used to reconstruct the first measurement data to obtain the liquid water saturation of the cathode catalyst layer of the first battery.

[0079] Optionally, in the embodiments of the present application, since the correlation relationships between the measurement data and the respective internal state data are not the same, using different models to reconstruct the measurement data can improve the accuracy of the reconstructed internal state data. The state reconstruction model may also include any deep learning model that can perform identification and prediction.

[0080] In the above method for identifying the internal state of the battery, the first measurement data of the first battery is obtained; the first measurement data is input into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model; wherein, the state reconstruction model is trained using the second measurement data and the internal state data corresponding to the second measurement data as the training set, the second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery, the internal state data includes at least one of the oxygen concentration of the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation of the cathode catalyst layer, and the algorithm architecture of the state reconstruction model is set corresponding to the internal state data. By using the simulation model to obtain the second measurement data of the second battery, the accuracy and reliability of the second measurement data are improved, thereby improving the accuracy and reliability of the state reconstruction model trained using the second measurement data. Further, by inputting the first measurement data that is easy to collect into the state reconstruction model, the internal state data is reconstructed and identified, improving the accuracy of the internal state data.

[0081] In one embodiment, as Figure 4 shown, the training process of the above state reconstruction model includes:

[0082] S203, obtaining the second measurement data according to the simulation model; the second measurement data is the measurement data of multiple battery partitions of the simulation model, and the battery partitions are obtained by partitioning the second battery based on the active area of the battery.

[0083] Among them, the second measurement data may include the average battery current density of the simulation model, the cathode inlet oxygen concentration, the cathode flow rate, the anode flow rate, the cathode inlet pressure, the anode inlet pressure, the cathode inlet humidity, the anode inlet humidity, and the battery operating temperature.

[0084] In the embodiment of the present application, the second battery is partitioned according to the battery active area to obtain a plurality of battery partitions. Exemplarily, the battery active area is 349 cm 2 , and the plane of the battery is evenly divided into 384 (24×16) partitions. Further, the simulation model is partitioned according to the partitions of the second battery to obtain a plurality of battery partitions of the simulation model that match the plurality of battery partitions of the second battery. Each battery partition is monitored respectively to obtain the second measurement data, and the second test data is calibrated to obtain the internal state data corresponding to the second measurement data.

[0085] Optionally, the second measurement data and the internal state data corresponding to the second measurement data can be normalized to eliminate the scale differences between features, accelerate optimization convergence, and ensure numerical stability.

[0086] In the embodiment of the present application, the simulation model is simulated multiple times under different oxygen concentrations and different flow rates, and the current density simulation values of the simulation model during each simulation process are collected. Exemplarily, the simulation of the current density distribution at 1 A / cm 2 is carried out under different oxygen concentrations and flow rates, and a total of 33 groups of simulations are carried out. The experimental conditions are shown in Table 1 below. During the test, data such as the high-frequency impedance and voltage of the battery are collected.

[0087] Table 1

[0088]

[0089] S204, input the second measurement data into the initial state reconstruction model to obtain the initial internal state data of each battery partition inside the battery.

[0090] In the embodiment of the present application, a part of the second measurement data and the internal state data corresponding to the second measurement data is used as the training set of the model, and the other part is used as the test set of the model. For example, 75% of the data in the second measurement data is used as the training set, and 25% of the data is used as the test set. Any set of data in the training set is input into the initial state reconstruction model, and the support vector regression algorithm, the random forest algorithm, the support vector regression algorithm, and the random forest algorithm are respectively used to determine the oxygen concentration of the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation of the cathode catalyst layer in the corresponding battery partition in the plane, so as to obtain the initial internal state data of each battery partition inside the battery.

[0091] S205. Optimize the initial state reconstruction model according to the internal state data corresponding to the second measurement data and the initial internal state data to obtain a state reconstruction model.

[0092] Among them, the model parameters that the random forest algorithm mainly needs to determine include the number of trees, the depth of the trees, the minimum number of samples for splitting, etc.; the model parameters that the support vector regression algorithm mainly needs to determine include the regularization parameter, the threshold of the loss function, the type of kernel function, etc.

[0093] In the embodiment of the present application, optimize the model parameters in the initial state reconstruction model according to the similarity between the internal state data corresponding to the second measurement data and the initial internal state data to obtain an optimized initial state reconstruction model, and return to execute the above step S204. When the cut-off condition for training is met, determine the initial state reconstruction model as the state reconstruction model. Optionally, the cut-off condition may include that the number of training times reaches a preset number of times, the similarity between the internal state data corresponding to the second measurement data and the initial internal state data is less than a preset similarity, etc.

[0094] Exemplarily, the coefficient of determination (R 2 ), root mean square error (RMSE) of the prediction results of the validation set after the algorithm training for each output physics are shown in Table 2, where R 2 are all above 0.96 and RMSE are all below 0.1.

[0095] Table 2

[0096]

[0097] In the above embodiment of the application, the initial state reconstruction model is trained using the second measurement data of the simulation model, which improves the accuracy and reliability of the state reconstruction model.

[0098] In one embodiment, an implementation manner of the above S203 is provided. As Figure 5 shown, the above "obtain the second measurement data according to the simulation model" includes:

[0099] S301. Construct a simulation model corresponding to the internal state of the second battery.

[0100] In the embodiment of the present application, construct a simulation model corresponding to the second battery according to the internal state of the second battery. For example, the PCB board integrated with the partition current sensor is regarded as the simulation model, and the number of partitions of the simulation model is the same as that of the second battery.

[0101] S302. Obtain the simulated current density value through the simulation model.

[0102] In the embodiments of the present application, a spatial discrete dynamic model is used to simulate the internal state. Before the simulation, the model needs to be refined and calibrated in combination with the current density test results. Therefore, the current density of each partition in the simulation model is collected to obtain the simulated value of the current density.

[0103] S303. Determine the current density error based on the simulated current density value and the experimental current density value corresponding to the second battery.

[0104] In the embodiments of the present application, the current density error is determined according to the simulated current density value and the known experimental current density value. Optionally, the difference between the simulated current density value and the experimental current density value can be used as the current density error.

[0105] S304. When the current density error is less than the error threshold, obtain the second measurement data of the second battery based on the simulation model.

[0106] In the embodiments of the present application, when the current density error is less than the error threshold, it is determined that the simulation accuracy of the simulation model is relatively high, and the second measurement data of the second battery can be obtained by measuring the simulation model.

[0107] In the above embodiments of the application, the simulation accuracy of the simulation model is first verified. When the simulation accuracy meets the requirements, the second measurement data of the second battery is obtained based on the simulation model, ensuring the accuracy of the second measurement data.

[0108] In one embodiment, an implementation manner of the above S303 is provided. The simulated current density value includes the average simulated current density value and the simulated current density value of each partition, and the experimental current density value includes the average experimental current density value and the experimental current density value of each partition. As Figure 6 shown, the above "determine the current density error based on the simulated current density value and the experimental current density value corresponding to the second battery" includes:

[0109] S401. Determine the first current density error according to the average simulated current density value and the average experimental current density value.

[0110] In the embodiments of the present application, for each set of working conditions, the average simulated current density value is the average of the simulated current density values of each battery partition under this working condition, and the average experimental current density value is the average of the experimental current density values of each battery partition under this working condition. The first current density error can be as shown in Equation 1:

[0111] (Equation 1)

[0112] where err is the first current density error, is the average experimental current density value, is the average simulated current density value.

[0113] S402. Determine the second current density error based on the simulated value and the experimental value of the partition current density.

[0114] In the embodiment of the present application, for each set of working conditions, the simulated value of the partition current density is the simulated value of the current density of each partition, and the experimental value of the partition current density can be the experimental value of the current density of each partition. The second current density error can be as shown in Equation 2:

[0115] (Equation 2)

[0116] where MAE is the second current density error, is the experimental value of the current density of the j-th partition, is the simulated value of the current density of the j-th partition.

[0117] Optionally, when the first current density error is less than the first error threshold and the second current density error is less than the second error threshold, the second measurement data of the second battery is obtained based on the simulation model. Optionally, the first error threshold can be 5%, and the second error threshold can be .

[0118] Exemplarily, at 33 working condition points in the embodiment of the present application, the maximum is 4.9% and the minimum is 0.5%; the maximum is 19.81% and the minimum is 10.54%. Furthermore, the oxygen concentration in the cathode catalyst layer, the water content in the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer within 384 partitions inside the battery are calculated at each working condition point using the simulation model.

[0119] In the above application embodiment, the first current density error and the second current density error are respectively determined, so as to determine the accuracy of the simulation model through double determination, thereby improving the accuracy of the second measurement data.

[0120] In one embodiment, an implementation manner of the above S202 is provided. The above "input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model" includes: inputting the first measurement data and the position information of the position to be predicted into the state reconstruction model to obtain the internal state data of the position to be predicted; the position to be predicted is determined according to the order of the risk of failure occurrence.

[0121] In the embodiment of the present application, the input of the state reconstruction model, that is, the first measurement data and the position information of the position to be predicted, can be expressed as , where is the average current density of the battery, is the oxygen concentration of the cathode inlet air, is the cathode flow rate, is the anode flow rate, is the cathode inlet pressure, is the anode inlet pressure, is the cathode inlet humidity, is the anode inlet humidity, is the battery operating temperature, is the abscissa of the partition, is the ordinate of the partition. Exemplarily, in the embodiments of the present application, the range of the abscissa of the partition can be (1 - 24), and the range of the ordinate of the partition can be (1 - 16). The internal state data of the position to be predicted can be expressed as , where is the current density inside the corresponding position, is the oxygen concentration in the cathode catalyst layer at the corresponding position, is the water content of the proton exchange membrane at the corresponding position, is the water vapor concentration in the cathode flow channel at the corresponding position, is the liquid water saturation in the cathode catalyst layer at the corresponding position.

[0122] In the embodiments of the present application, the in-plane area can be divided into three parts: the inlet area corresponds to the in-plane coordinates , the middle area corresponds to the in-plane coordinates , and the outlet area corresponds to the in-plane coordinates . Since the potential positions and risk levels of different types of faults occurring inside the fuel cell are also different, the target fault type can be determined first, and then the area with the greatest fault risk can be determined according to the target fault type. The position information of any battery partition in the area with the greatest fault risk can be used as the position information of the position to be predicted. The position information of the position to be predicted and the first measurement data are input into the state reconstruction model to obtain the internal state data of the position to be predicted.

[0123] Optionally, according to the target fault type, the target sub-model in the state reconstruction model can be determined, and the first measurement data and the position information of the position to be predicted are input into the target sub-model to obtain the internal state data corresponding to the target fault type of the position to be predicted; or, the first measurement data and the position information of the position to be predicted can be input into the state reconstruction model to obtain all the internal state data.

[0124] Exemplarily, the fault types and the area risk rankings corresponding to each fault type are as follows:

[0125] ① Oxygen shortage fault: outlet area, middle area, inlet area;

[0126] ② Proton exchange membrane dry fault: inlet area, outlet area, middle area;

[0127] ③ Flooding fault of the cathode catalyst layer: middle region, outlet region, inlet region;

[0128] ④ Flooding fault of the flow channel: outlet region, middle region, inlet region.

[0129] Exemplarily, as the upper computer for control diagnosis, the server can obtain the fault thresholds of each region of the first battery through the lower computer, and the order in which the lower computer outputs the fault thresholds to the upper computer can be as Figure 7 shown:

[0130] ① Oxygen shortage fault: minimum value of the oxygen concentration in the cathode catalyst layer of the outlet region and the corresponding position, minimum value of the oxygen concentration in the cathode catalyst layer of the middle region and the corresponding position, minimum value of the oxygen concentration in the cathode catalyst layer of the inlet region and the corresponding position;

[0131] ② Proton exchange membrane dryness fault: minimum value of the water content of the proton exchange membrane in the inlet region and the corresponding position, minimum value of the water content of the proton exchange membrane in the outlet region and the corresponding position, minimum value of the water content of the proton exchange membrane in the middle region and the corresponding position;

[0132] ③ Flooding fault of the cathode catalyst layer: maximum value of the liquid water saturation in the cathode catalyst layer of the middle region and the corresponding position, maximum value of the liquid water saturation in the cathode catalyst layer of the outlet region and the corresponding position, maximum value of the liquid water saturation in the cathode catalyst layer of the inlet region and the corresponding position;

[0133] ④ Flooding fault of the flow channel: maximum value of the water vapor concentration in the cathode flow channel of the outlet region and the corresponding position, maximum value of the water vapor concentration in the cathode flow channel of the middle region and the corresponding position, maximum value of the water vapor concentration in the cathode flow channel of the inlet region and the corresponding position.

[0134] Among them, Figure 7 the maximum or minimum value is the fault threshold corresponding to the corresponding fault. Exemplarily, when the oxygen concentration in the cathode catalyst layer in the internal state data is less than the fault threshold of the oxygen shortage fault, there is an oxygen shortage fault; when the water content of the proton exchange membrane in the internal state data is less than the fault threshold of the proton exchange membrane dryness fault, there is a proton exchange membrane dryness fault; when the liquid water saturation in the cathode catalyst layer in the internal state data is greater than the fault threshold of the flooding fault of the cathode catalyst layer, there is a flooding fault of the cathode catalyst layer; when the water vapor concentration in the cathode flow channel in the internal state data is greater than the fault threshold of the flooding fault of the flow channel, there is a flooding fault of the flow channel.

[0135] In the above application embodiments, the position to be predicted is determined according to the order of the occurrence risks of the corresponding faults, so as to identify the internal state data of the position to be predicted, and when a fault occurs, the efficiency of obtaining the internal state data of the fault position is improved.

[0136] In one embodiment, as Figure 8As shown, after inputting the first measurement data into the state reconstruction model and obtaining the internal state data of the first battery reconstructed by the state reconstruction model, the above method for identifying the internal state of the battery further includes:

[0137] S206. Determine the fault threshold of the position to be predicted according to the preset fault threshold table and the position information of the position to be predicted.

[0138] In the embodiment of the present application, the in-plane area can be divided into three parts: the entrance area corresponds to the in-plane coordinates , the middle area corresponds to the in-plane coordinates , and the exit area corresponds to the in-plane coordinates . According to the in-plane area corresponding to the position to be predicted, determine the fault thresholds of each fault type in this area from the preset fault threshold table.

[0139] S207. Determine whether a fault occurs at the position to be predicted according to the internal state data and the fault threshold; the fault includes at least one of oxygen deficiency fault, proton exchange membrane dry fault, cathode catalyst layer waterlogging fault, and flow channel waterlogging fault.

[0140] In the embodiment of the present application, the internal state data includes the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation of the cathode catalyst layer. Compare each internal state data with the corresponding fault threshold to determine whether a fault occurs at the position to be predicted. Optionally, the oxygen deficiency fault corresponds to the oxygen concentration in the cathode catalyst layer; the proton exchange membrane dry fault corresponds to the water content concentration of the proton exchange membrane; the cathode catalyst layer waterlogging fault corresponds to the water vapor concentration in the cathode flow channel; the flow channel waterlogging fault corresponds to the liquid water saturation of the cathode catalyst layer.

[0141] Exemplarily, determine whether the oxygen concentration in the cathode catalyst layer is greater than the fault threshold of the oxygen deficiency fault. If it is greater, a oxygen deficiency fault occurs at the position to be predicted. If it is not greater, no oxygen deficiency fault occurs at the position to be predicted; determine whether the water content of the proton exchange membrane is greater than the fault threshold of the proton exchange membrane dry fault. If it is greater, a proton exchange membrane dry fault occurs at the position to be predicted. If it is not greater, no proton exchange membrane dry fault occurs at the position to be predicted; determine whether the water vapor concentration in the cathode flow channel is greater than the fault threshold of the cathode catalyst layer waterlogging fault. If it is greater, a cathode catalyst layer waterlogging fault occurs at the position to be predicted. If it is not greater, no cathode catalyst layer waterlogging fault occurs at the position to be predicted; determine whether the liquid water saturation of the cathode catalyst layer is greater than the fault threshold of the flow channel waterlogging fault. If it is greater, a flow channel waterlogging fault occurs at the position to be predicted. If it is not greater, no flow channel waterlogging fault occurs at the position to be predicted.

[0142] In the above application embodiments, the faults at the position are identified according to the preset fault threshold table and the position information of the position to be predicted, improving the accuracy of fault identification.

[0143] In one embodiment, a method for identifying the complete internal state of a battery is provided. As Figure 9 shown, the method includes:

[0144] S1, constructing a simulation model corresponding to the internal state of the second battery.

[0145] S2, obtaining the simulated current density value through the simulation model; the simulated current density value includes the average simulated current density value and the zonal simulated current density value.

[0146] S3, determining the first current density error according to the average simulated current density value and the average experimental current density value; determining the second current density error according to the zonal simulated current density value and the zonal experimental current density value.

[0147] S4, when the first current density error is less than the first error threshold and the second current density error is less than the second error threshold, obtaining the second measurement data of the second battery based on the simulation model; the second measurement data is the measurement data of multiple battery zones of the simulation model, and the battery zones are obtained by partitioning the second battery based on the battery active area.

[0148] S5, inputting the second measurement data into the initial state reconstruction model to obtain the initial internal state data of each battery zone inside the battery.

[0149] S6, optimizing the initial state reconstruction model according to the internal state data corresponding to the second measurement data and the initial internal state data to obtain the state reconstruction model.

[0150] S7, obtaining the first measurement data of the first battery.

[0151] S8, inputting the first measurement data and the position information of the position to be predicted into the state reconstruction model to obtain the internal state data of the position to be predicted; the position to be predicted is determined according to the order of the risk of fault occurrence; wherein, the state reconstruction model includes at least one of the following sub-models:

[0152] The first support vector machine, which is used to reconstruct the first measurement data to obtain the oxygen concentration in the cathode catalyst layer of the first battery;

[0153] The first random forest model, which is used to reconstruct the first measurement data to obtain the water content of the proton exchange membrane of the first battery;

[0154] The second support vector machine, which is used to reconstruct the first measurement data to obtain the water vapor concentration in the cathode flow channel of the first battery;

[0155] A second random forest model, which is used to reconstruct the first measurement data to obtain the liquid water saturation of the cathode catalyst layer of the first battery.

[0156] S9. According to a preset fault threshold table and the position information of the position to be predicted, determine the fault threshold of the position to be predicted.

[0157] S10. According to the internal state data and the fault threshold, determine whether a fault occurs at the position to be predicted; the fault includes at least one of an oxygen shortage fault, a proton exchange membrane dry fault, a cathode catalyst layer flooding fault, and a flow channel flooding fault.

[0158] In the above method for identifying the internal state of the battery, obtain the first measurement data of the first battery; input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model; wherein, the state reconstruction model is trained by using the second measurement data and the internal state data corresponding to the second measurement data as a training set, the second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery, the internal state data includes at least one of the oxygen concentration of the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration of the cathode flow channel, and the liquid water saturation of the cathode catalyst layer, and the algorithm architecture of the state reconstruction model is set corresponding to the internal state data. By using the simulation model to obtain the second measurement data of the second battery, the accuracy and reliability of the second measurement data are improved, thereby improving the accuracy and reliability of the state reconstruction model trained by using the second measurement data. Further, by inputting the first measurement data that is easy to collect into the state reconstruction model to reconstruct and identify the internal state data, the accuracy of identifying the internal state data is improved.

[0159] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are displayed in sequence according to the indication of the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps does not have a strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0160] Based on the same inventive concept, an embodiment of the present application further provides an apparatus for identifying the internal state of a battery for implementing the method for identifying the internal state of a battery involved above. The solution provided by this apparatus for solving the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the apparatus for identifying the internal state of a battery provided below can refer to the limitations on the method for identifying the internal state of a battery in the above text, and will not be elaborated here.

[0161] In one embodiment, as Figure 10 shown, an apparatus for identifying the internal state of a battery is provided, including: a first acquisition module 10 and a second acquisition module 11, where:

[0162] The first acquisition module 10 is configured to acquire first measurement data of a first battery.

[0163] The second acquisition module 11 is configured to input the first measurement data into a state reconstruction model and acquire the internal state data of the first battery reconstructed by the state reconstruction model.

[0164] Among them, the state reconstruction model is trained by using the second measurement data and the internal state data corresponding to the second measurement data as a training set. The second measurement data and the internal state data corresponding to the second measurement data are obtained based on a simulation model of a second battery. The internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content in the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer. The algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0165] In one embodiment, the above state reconstruction model includes at least one of the following sub-models:

[0166] The first support vector machine is configured to reconstruct the first measurement data to obtain the oxygen concentration in the cathode catalyst layer of the first battery;

[0167] The first random forest model is configured to reconstruct the first measurement data to obtain the water content in the proton exchange membrane of the first battery;

[0168] The second support vector machine is configured to reconstruct the first measurement data to obtain the water vapor concentration in the cathode flow channel of the first battery;

[0169] The second random forest model is configured to reconstruct the first measurement data to obtain the liquid water saturation in the cathode catalyst layer of the first battery.

[0170] In one embodiment, the above method for identifying the internal state of a battery further includes: a third acquisition module, an input module, and an optimization module, where:

[0171] A third acquisition module, configured to acquire second measurement data according to a simulation model; the second measurement data is measurement data of multiple battery partitions of the simulation model, and the battery partitions are obtained by partitioning the second battery based on the battery active area;

[0172] An input module, configured to input the second measurement data into an initial state reconstruction model to obtain initial internal state data of each battery partition inside the battery;

[0173] An optimization module, configured to optimize the initial state reconstruction model according to the internal state data corresponding to the second measurement data and the initial internal state data to obtain a state reconstruction model.

[0174] In one embodiment, the above-mentioned third acquisition module includes: a construction unit, a first acquisition unit, a determination unit, and a second acquisition unit, where:

[0175] The construction unit is configured to construct a simulation model corresponding to the internal state of the second battery.

[0176] The first acquisition unit is configured to acquire a simulated current density value through the simulation model.

[0177] The determination unit is configured to determine a current density error based on the simulated current density value and the experimental current density value corresponding to the second battery.

[0178] The second acquisition unit is configured to acquire the second measurement data of the second battery based on the simulation model when the current density error is less than an error threshold.

[0179] In one embodiment, the above-mentioned determination unit is specifically configured to determine a first current density error according to the average simulated current density value and the average experimental current density value; and determine a second current density error according to the partition simulated current density value and the partition experimental current density value.

[0180] In one embodiment, the above-mentioned second acquisition module 11 includes: a third acquisition unit, configured to input the first measurement data and the location information of the location to be predicted into the state reconstruction model to obtain the internal state data of the location to be predicted; the location to be predicted is determined according to the order of the risk of fault occurrence.

[0181] In one embodiment, the above-mentioned device for identifying the internal state of the battery further includes: a first determination module and a second determination module, where:

[0182] The first determination module is configured to determine a fault threshold for the location to be predicted according to a preset fault threshold table and the location information of the location to be predicted.

[0183] A second determination module, configured to determine whether a failure occurs at a position to be predicted according to internal state data and a failure threshold; the failure includes at least one of an oxygen deficiency failure, a proton exchange membrane drying failure, a cathode catalyst layer flooding failure, and a flow channel flooding failure.

[0184] Each module in the above battery internal state recognition device can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in or independent of a processor in a computer device in the form of hardware, or stored in a memory in the computer device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to each of the above modules.

[0185] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as Figure 11 shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the recognition data of the battery internal state. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for recognizing the internal state of a battery.

[0186] Those skilled in the art can understand that Figure 11 the structure shown in

[0187] merely represents a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have a different component layout.

[0188] Obtain first measurement data of a first battery;

[0189] Input the first measurement data into a state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model;

[0190] Among them, the state reconstruction model is obtained by training the second measurement data and the internal state data corresponding to the second measurement data as a training set. The second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery. The internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content in the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer. The algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0191] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0192] The first support vector machine is used to reconstruct the first measurement data to obtain the oxygen concentration in the cathode catalyst layer of the first battery;

[0193] The first random forest model is used to reconstruct the first measurement data to obtain the water content in the proton exchange membrane of the first battery;

[0194] The second support vector machine is used to reconstruct the first measurement data to obtain the water vapor concentration in the cathode flow channel of the first battery;

[0195] The second random forest model is used to reconstruct the first measurement data to obtain the liquid water saturation in the cathode catalyst layer of the first battery.

[0196] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0197] Obtain the second measurement data according to the simulation model; the second measurement data is the measurement data of multiple battery partitions of the simulation model, and the battery partitions are obtained by partitioning the second battery based on the active area of the battery;

[0198] Input the second measurement data into the initial state reconstruction model to obtain the initial internal state data of each battery partition inside the battery;

[0199] Optimize the initial state reconstruction model according to the internal state data corresponding to the second measurement data and the initial internal state data to obtain the state reconstruction model.

[0200] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0201] Construct a simulation model corresponding to the internal state of the second battery;

[0202] Obtain the simulated value of the current density through the simulation model;

[0203] Based on the simulated value of the current density and the experimental value of the current density corresponding to the second battery, determine the current density error;

[0204] When the current density error is less than the error threshold, second measurement data of the second battery is obtained based on the simulation model.

[0205] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0206] Determine a first current density error according to the average current density simulation value and the average current density experimental value;

[0207] Determine a second current density error according to the partition current density simulation value and the partition current density experimental value.

[0208] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0209] Input the first measurement data and the location information of the location to be predicted into the state reconstruction model to obtain the internal state data of the location to be predicted; the location to be predicted is determined according to the order of the risk of fault occurrence.

[0210] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0211] Determine the fault threshold of the location to be predicted according to the preset fault threshold table and the location information of the location to be predicted;

[0212] Determine whether a fault occurs at the location to be predicted according to the internal state data and the fault threshold; the fault includes at least one of oxygen deficiency fault, proton exchange membrane dry fault, cathode catalyst layer waterlogging fault, and flow channel waterlogging fault.

[0213] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0214] Obtain first measurement data of the first battery;

[0215] Input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model;

[0216] Wherein, the state reconstruction model is trained by using the second measurement data and the internal state data corresponding to the second measurement data as a training set. The second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery. The internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer. The algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0217] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0218] The first support vector machine is used to reconstruct the first measurement data to obtain the oxygen concentration of the cathode catalyst layer of the first battery;

[0219] The first random forest model is used to reconstruct the first measurement data to obtain the water content of the proton exchange membrane of the first battery;

[0220] The second support vector machine is used to reconstruct the first measurement data to obtain the water vapor concentration of the cathode flow channel of the first battery;

[0221] The second random forest model is used to reconstruct the first measurement data to obtain the liquid water saturation of the cathode catalyst layer of the first battery.

[0222] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0223] Obtain second measurement data according to the simulation model; the second measurement data is the measurement data of multiple battery partitions of the simulation model, and the battery partitions are obtained by partitioning the second battery based on the active area of the battery;

[0224] Input the second measurement data into the initial state reconstruction model to obtain the initial internal state data of each battery partition inside the battery;

[0225] Optimize the initial state reconstruction model according to the internal state data corresponding to the second measurement data and the initial internal state data to obtain the state reconstruction model.

[0226] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0227] Construct a simulation model corresponding to the internal state of the second battery;

[0228] Obtain the simulated current density value through the simulation model;

[0229] Determine the current density error based on the simulated current density value and the experimental current density value corresponding to the second battery;

[0230] When the current density error is less than the error threshold, obtain the second measurement data of the second battery based on the simulation model.

[0231] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0232] Determine the first current density error according to the average simulated current density value and the average experimental current density value;

[0233] Determine the second current density error according to the partition simulated current density value and the partition experimental current density value.

[0234] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0235] Input the first measurement data and the position information of the position to be predicted into the state reconstruction model to obtain the internal state data of the position to be predicted; the position to be predicted is determined according to the order of the risk of failure occurrence.

[0236] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0237] Determine the fault threshold of the position to be predicted according to the preset fault threshold table and the position information of the position to be predicted;

[0238] Determine whether a fault occurs at the position to be predicted according to the internal state data and the fault threshold; the fault includes at least one of oxygen deficiency fault, proton exchange membrane drying fault, cathode catalyst layer flooding fault, and flow channel flooding fault.

[0239] In one embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0240] Obtain the first measurement data of the first battery;

[0241] Input the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model;

[0242] Wherein, the state reconstruction model is trained by using the second measurement data and the internal state data corresponding to the second measurement data as a training set, the second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery, the internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer, and the algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

[0243] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0244] The first support vector machine is used to reconstruct the first measurement data to obtain the oxygen concentration in the cathode catalyst layer of the first battery;

[0245] The first random forest model is used to reconstruct the first measurement data to obtain the water content of the proton exchange membrane of the first battery;

[0246] The second support vector machine is used to reconstruct the first measurement data to obtain the water vapor concentration in the cathode flow channel of the first battery;

[0247] A second random forest model is used to reconstruct the first measurement data to obtain the liquid water saturation of the cathode catalyst layer of the first battery.

[0248] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0249] Obtain second measurement data according to the simulation model; the second measurement data is the measurement data of multiple battery partitions of the simulation model, and the battery partitions are obtained by partitioning the second battery based on the battery active area;

[0250] Input the second measurement data into the initial state reconstruction model to obtain the initial internal state data of each battery partition inside the battery;

[0251] Optimize the initial state reconstruction model according to the internal state data corresponding to the second measurement data and the initial internal state data to obtain the state reconstruction model.

[0252] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0253] Construct a simulation model corresponding to the internal state of the second battery;

[0254] Obtain the simulated current density value through the simulation model;

[0255] Determine the current density error based on the simulated current density value and the experimental current density value corresponding to the second battery;

[0256] When the current density error is less than the error threshold, obtain the second measurement data of the second battery based on the simulation model.

[0257] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0258] Determine the first current density error according to the average simulated current density value and the average experimental current density value;

[0259] Determine the second current density error according to the partition simulated current density value and the partition experimental current density value.

[0260] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0261] Input the first measurement data and the position information of the position to be predicted into the state reconstruction model to obtain the internal state data of the position to be predicted; the position to be predicted is determined according to the order of the risk of failure occurrence.

[0262] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0263] Determine the fault threshold of the position to be predicted according to the preset fault threshold table and the position information of the position to be predicted;

[0264] Determine whether a fault has occurred at the position to be predicted according to the internal state data and the fault threshold; the fault includes at least one of oxygen deficiency fault, proton exchange membrane drying fault, cathode catalyst layer flooding fault, and flow channel flooding fault.

[0265] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0266] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this application.

[0267] The above-described embodiments merely represent several implementation manners of this application, and the description thereof is relatively specific and detailed. However, it should not be construed as a limitation on the patent scope of this application. It should be noted that for those of ordinary skill in the art, without departing from the concept of this application, several modifications and improvements can still be made, and these all belong to the protection scope of this application. Therefore, the protection scope of this application shall be subject to the appended claims.

Claims

1. A method for identifying the internal state of a battery, characterized in that, The method includes: Obtaining first measurement data of a first battery; Inputting the first measurement data into a state reconstruction model to obtain internal state data of the first battery reconstructed by the state reconstruction model; Wherein, the state reconstruction model is trained by using second measurement data and internal state data corresponding to the second measurement data as a training set, the second measurement data and the internal state data corresponding to the second measurement data are obtained based on a simulation model of a second battery, the internal state data includes at least one of oxygen concentration in the cathode catalyst layer, water content in the proton exchange membrane, water vapor concentration in the cathode flow channel, and liquid water saturation in the cathode catalyst layer, and the algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

2. The method according to claim 1, wherein The state reconstruction model includes at least one of the following sub-models: A first support vector machine for reconstructing the first measurement data to obtain the oxygen concentration in the cathode catalyst layer of the first battery; A first random forest model for reconstructing the first measurement data to obtain the water content in the proton exchange membrane of the first battery; A second support vector machine for reconstructing the first measurement data to obtain the water vapor concentration in the cathode flow channel of the first battery; A second random forest model for reconstructing the first measurement data to obtain the liquid water saturation in the cathode catalyst layer of the first battery.

3. The method according to claim 1, characterized in that The training process of the state reconstruction model includes: Obtaining second measurement data according to the simulation model; the second measurement data is measurement data of multiple battery partitions of the simulation model, and the battery partitions are obtained by partitioning the second battery based on the active area of the battery; Inputting the second measurement data into an initial state reconstruction model to obtain initial internal state data of each of the battery partitions inside the battery; Optimizing the initial state reconstruction model according to the internal state data corresponding to the second measurement data and the initial internal state data to obtain the state reconstruction model.

4. The method according to claim 3, characterized in that The obtaining the second measurement data according to the simulation model includes: Constructing a simulation model corresponding to the internal state of the second battery; Obtaining a simulated current density value through the simulation model; Determining a current density error based on the simulated current density value and the experimental current density value corresponding to the second battery; When the current density error is less than an error threshold, obtaining the second measurement data of the second battery based on the simulation model.

5. The method according to claim 4, wherein The simulated current density value includes an average simulated current density value and a partitioned simulated current density value, the experimental current density value includes an average experimental current density value and a partitioned experimental current density value, and the determining the current density error based on the simulated current density value and the experimental current density value corresponding to the second battery includes: Determining a first current density error according to the average simulated current density value and the average experimental current density value; Determining a second current density error according to the partitioned simulated current density value and the partitioned experimental current density value.

6. The method according to claim 1, wherein The inputting the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model includes: Input the first measurement data and the location information of the location to be predicted into the state reconstruction model to obtain the internal state data of the location to be predicted; the location to be predicted is determined according to the order of the risk of fault occurrence.

7. The method according to any one of claims 1-6, characterized in that, After inputting the first measurement data into the state reconstruction model and obtaining the internal state data of the first battery reconstructed by the state reconstruction model, the method further includes: Determine the fault threshold of the location to be predicted according to the preset fault threshold table and the location information of the location to be predicted; Determine whether a fault occurs at the location to be predicted according to the internal state data and the fault threshold; the fault includes at least one of oxygen deficiency fault, proton exchange membrane drying fault, cathode catalyst layer flooding fault, and flow channel flooding fault.

8. An identification device for the internal state of a battery, characterized in that, The device includes: A first acquisition module for acquiring the first measurement data of the first battery; A second acquisition module for inputting the first measurement data into the state reconstruction model to obtain the internal state data of the first battery reconstructed by the state reconstruction model; Wherein, the state reconstruction model is trained with the second measurement data and the internal state data corresponding to the second measurement data as the training set, the second measurement data and the internal state data corresponding to the second measurement data are obtained based on the simulation model of the second battery, the internal state data includes at least one of the oxygen concentration in the cathode catalyst layer, the water content of the proton exchange membrane, the water vapor concentration in the cathode flow channel, and the liquid water saturation in the cathode catalyst layer, and the algorithm architecture of the state reconstruction model is set corresponding to the internal state data.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images