Methods, devices and computer equipment for predicting the health status of power batteries

Abstract

This application relates to a method, apparatus, and computer device for predicting the health status of a power battery. The method includes: acquiring target multivariate data of the power battery; processing the target multivariate data to obtain target grayscale data corresponding to the target multivariate data; and further, analyzing the target grayscale data based on a well-trained prediction model to obtain a predicted value of the power battery's health status. This application solves the problem that the strong coupling relationship between multivariate data makes it impossible to accurately predict the health status of a power battery, thereby improving the accuracy of the power battery's health status prediction results.

Description

Technical Field

[0001] This application relates to the field of electric vehicle technology, and in particular to methods, devices and computer equipment for predicting the health status of power batteries. Background Technology

[0002] As the practicality and diversity of electric vehicles increase, their market share is also gradually increasing. Among them, lithium-ion power batteries are widely used as power sources and energy storage devices in the production of electric vehicles due to their advantages such as high energy density and low self-discharge rate. However, lithium-ion power batteries still face the problem of battery performance degradation. Long-term use will lead to a decrease in battery capacity or power. Therefore, in order to ensure the safe operation of electric vehicles, it is necessary to accurately predict the health status of power batteries.

[0003] Current prediction methods rely on data acquisition devices to collect battery health status data from lithium batteries. Based on the health status values ​​corresponding to each batch of data, multiple relevant datasets are generated. Different health status prediction models are then established for each dataset to predict the battery's health status. However, these methods fail to consider the strong coupling relationships between the various battery variables. Such strong coupling can interfere with the prediction process, leading to reduced accuracy in predicting the battery's health status.

[0004] There is currently no effective solution to the problem that the health status of power batteries cannot be accurately predicted due to the strong coupling between multivariate data in related technologies. Summary of the Invention

[0005] This embodiment provides a method, apparatus, and computer device for predicting the health status of a power battery, in order to solve the problem in related technologies that the health status of a power battery cannot be accurately predicted due to the strong coupling relationship between multivariate data.

[0006] Firstly, this embodiment provides a method for predicting the health status of a power battery, the method comprising:

[0007] Obtain target multivariate data for the power battery;

[0008] The target multivariate data is processed to obtain the target grayscale image data corresponding to the target multivariate data;

[0009] Based on a well-trained prediction model, the target grayscale data is analyzed to obtain the predicted health status value of the power battery.

[0010] In some embodiments, processing the target multivariate data to obtain target grayscale image data corresponding to the target multivariate data includes:

[0011] The voltage, current, and temperature data of the power battery are obtained from the target multivariate data.

[0012] The voltage, current, and temperature data of the power battery are mapped to a preset numerical range, and the mapped voltage, current, and temperature data are spliced ​​together to obtain the target grayscale image data corresponding to the target multivariate data.

[0013] In some embodiments, before analyzing the target grayscale data based on a fully trained prediction model to obtain the predicted health status value of the power battery, the method further includes:

[0014] Based on a fully trained multivariate conditional generative adversarial network model, data augmentation is performed on the first sample multivariate data to generate the training dataset corresponding to the first sample multivariate data.

[0015] The prediction model is trained based on the training dataset to obtain the fully trained prediction model.

[0016] In some embodiments, before processing the first sample multivariate data based on a fully trained multivariate conditional generative adversarial network model to obtain the training dataset corresponding to the first sample multivariate data, the method further includes:

[0017] Obtain second sample multivariate data from the battery test database, and train the multivariate conditional generative adversarial network model using the second sample multivariate data to obtain the initial multivariate conditional generative adversarial network model.

[0018] In some embodiments, training the prediction model based on the training dataset to obtain the fully trained prediction model includes:

[0019] The augmented multivariate data of the first sample in the training dataset is processed to generate a grayscale dataset corresponding to the training dataset.

[0020] The prediction model is trained based on the grayscale image dataset to obtain the fully trained prediction model.

[0021] In some embodiments, after training the multivariate conditional generative adversarial network model using the second sample multivariate data to obtain an initial multivariate conditional generative adversarial network model, the method further includes:

[0022] Based on the first sample multivariate data, transfer learning is performed on the initial multivariate conditional generative adversarial network model to obtain the fully trained multivariate conditional generative adversarial network model.

[0023] In some embodiments, before obtaining second sample multivariate data from a battery test database and training the multivariate conditional generative adversarial network model using the second sample multivariate data, the method further includes:

[0024] The conditional generative adversarial network model is optimized by a gated recurrent unit network to generate the multivariate conditional generative adversarial network model.

[0025] Secondly, this embodiment provides a power battery health status prediction device, the device comprising:

[0026] The acquisition module acquires the target multivariate data of the power battery;

[0027] The processing module processes the target multivariate data to obtain the target grayscale image data corresponding to the target multivariate data;

[0028] The prediction module analyzes the target grayscale data based on a fully trained prediction model to obtain a predicted value of the health status of the power battery.

[0029] Thirdly, this embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the power battery health status prediction method described in the first aspect above.

[0030] Fourthly, this embodiment provides a storage medium storing a computer program that, when executed by a processor, implements the power battery health status prediction method described in the first aspect above.

[0031] Compared with related technologies, the power battery health status prediction method, device, and computer equipment provided in this embodiment acquire target multivariate data of the power battery, process the target multivariate data to obtain target grayscale data corresponding to the target multivariate data, and further analyze the target grayscale data based on a well-trained prediction model to obtain the predicted value of the power battery health status. This solves the problem that the health status of the power battery cannot be accurately predicted due to the strong coupling relationship between multivariate data, and improves the accuracy of the power battery health status prediction results.

[0032] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0033] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0034] Figure 1 This is a hardware structure block diagram of a terminal device for a power battery health status prediction method provided in an embodiment of this application;

[0035] Figure 2 This is a flowchart of a power battery health status prediction method provided in an embodiment of this application;

[0036] Figure 3 This is a model framework diagram of a power battery health status prediction method provided in an embodiment of this application;

[0037] Figure 4 This is a flowchart illustrating a method for predicting the health status of a power battery according to an embodiment of this application.

[0038] Figure 5 This is a preferred flowchart of a power battery health status prediction method provided in an embodiment of this application;

[0039] Figure 6 This is a structural block diagram of a power battery health status prediction device provided in an embodiment of this application.

[0040] In the diagram: 102, processor; 104, memory; 106, transmission device; 108, input / output device; 10, acquisition module; 20, processing module; 30, prediction module. Detailed Implementation

[0041] To better understand the purpose, technical solution, and advantages of this application, the application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0042] Unless otherwise defined, the technical or scientific terms used in this application shall have the general meaning as understood by one of ordinary skill in the art to which this application pertains. Words such as “a,” “an,” “an,” “the,” “the,” and “these,” used in this application, do not indicate quantitative limitation and may be singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or modules (units) is not limited to the listed steps or modules (units) but may include steps or modules (units) not listed, or may include other steps or modules (units) inherent to such processes, methods, products, or devices. The terms “connected,” “linked,” and “coupled,” used in this application, are not limited to physical or mechanical connections but may include electrical connections, whether direct or indirect. The term “multiple” used in this application refers to two or more. The "and / or" operator describes the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: A alone, A and B simultaneously, and B alone. Typically, the character " / " indicates that the objects before and after it are in an "or" relationship. The terms "first," "second," and "third," etc., used in this application are merely for distinguishing similar objects and do not represent a specific ordering of the objects.

[0043] The method embodiments provided in this example can be executed on a terminal, computer, or similar computing device. For example, it can run on a terminal. Figure 1 This is a hardware structure block diagram of the terminal for the power battery health status prediction method in this embodiment. For example... Figure 1 As shown, a terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 and a memory 104 for storing data are also included. The processor 102 may be, but is not limited to, a microprocessor (MCU) or a programmable logic device (FPGA). The terminal may also include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that… Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the terminal described above. For example, the terminal may also include components that are larger than... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown are illustrated.

[0044] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the power battery health status prediction method in this embodiment. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0045] The transmission device 106 is used to receive or send data via a network. This network includes a wireless network provided by the terminal's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 can be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0046] This embodiment provides a method for predicting the health status of a power battery. Figure 2 This is a flowchart of the power battery health status prediction method in this embodiment, as shown below. Figure 2 As shown, the process includes the following steps:

[0047] Step S210: Obtain the target multivariate data of the power battery.

[0048] Step S220: Process the target multivariate data to obtain the target grayscale image data corresponding to the target multivariate data.

[0049] It should be noted that the target multivariate data in this embodiment includes feature data such as voltage data, current data, and temperature data of the power battery. Different target multivariate data can be dynamically selected as training data and corresponding prediction model input values ​​according to the actual situation. Furthermore, during the conversion of target multivariate data, the width and length of the generated grayscale image are dynamically adjusted according to the scale of the selected target multivariate data.

[0050] Step S230: Based on the fully trained prediction model, analyze the target grayscale image data to obtain the predicted health status value of the power battery.

[0051] It is important to know that the prediction model used in this embodiment is a two-dimensional convolutional neural network model. The multivariate data in the training dataset is transformed to generate a corresponding grayscale dataset, and the two-dimensional convolutional neural network model is trained based on the grayscale dataset, so that the trained two-dimensional convolutional neural network model can be applied to predict the state of health (SOH) of the power battery.

[0052] Current prediction methods rely on data acquisition devices to collect battery health status data from lithium batteries. Based on the health status values ​​corresponding to each batch of data, multiple relevant datasets are generated. Different health status prediction models are then established for each dataset to predict the battery's health status. However, these methods fail to consider the strong coupling relationships between the various battery variables, which can interfere with the prediction process and reduce the accuracy of the predicted health status. This application, building upon existing technology, converts the target multivariate data into corresponding target grayscale data and analyzes this grayscale data using a prediction model. This fully considers the coupling relationships between the multivariate data during the prediction process. Specifically, it acquires the target multivariate data of the power battery, processes it to obtain the corresponding target grayscale data, and further analyzes the target grayscale data based on a well-trained prediction model to obtain the predicted health status value of the power battery. This solves the problem of inaccurate prediction of the power battery's health status due to the strong coupling relationships between the multivariate data, thus improving the accuracy of the predicted health status.

[0053] In some embodiments, the target multivariate data is processed to obtain target grayscale image data corresponding to the target multivariate data, including the following steps:

[0054] Step S221: Obtain the voltage, current, and temperature data of the power battery from the target multivariate data;

[0055] Step S222: Map the voltage, current and temperature data of the power battery to a preset numerical range, and stitch the mapped voltage, current and temperature data together to obtain the target grayscale data corresponding to the target multivariate data.

[0056] Specifically, the voltage, current, and temperature data of the power battery are mapped to a range of 0 to 255, and the mapped voltage, current, and temperature data are then stitched together to generate target grayscale data corresponding to the target multivariate data.

[0057] It is important to know that the grayscale value range of a grayscale image is 0 to 255, where the grayscale value refers to the brightness of each pixel. Each pixel in a grayscale image contains only one sampled color. Therefore, in this embodiment, the voltage, current, and temperature data of the power battery are mapped to the range of 0 to 255, so that each target multivariate data is converted into a single pixel in the target grayscale image. Then, the target grayscale image can be analyzed through the grayscale value distribution. Based on the grayscale image, feature information such as the coupling relationship between the target multivariate data can be further obtained.

[0058] In this embodiment, voltage, current, and temperature data of the power battery are obtained from target multivariate data. The voltage, current, and temperature data of the power battery are mapped to a preset numerical range, and the mapped voltage, current, and temperature data are stitched together to obtain target grayscale data corresponding to the target multivariate data. This enables the analysis of the health status of the power battery based on multivariate data in grayscale form, thereby improving the accuracy of health status prediction results.

[0059] In some embodiments, before analyzing the target grayscale data based on a fully trained prediction model to obtain the predicted health status value of the power battery, the following steps are also included:

[0060] Based on a fully trained multivariate conditional generative adversarial network model, data augmentation is performed on the first sample multivariate data to generate the training dataset corresponding to the first sample multivariate data.

[0061] The prediction model is trained based on the training dataset to obtain a fully trained prediction model.

[0062] Specifically, in the power battery data cloud platform, market vehicles whose battery aging level has reached the upper limit are screened out, and the corresponding battery aging data is extracted as the first sample multivariate data. The first sample multivariate data includes voltage data, current data, temperature data and SOH value during the historical charging process of the market vehicles.

[0063] Furthermore, based on the specified SOH value, corresponding battery aging data is generated by training a fully-fledged Multivariate Conditional Generative Adversarial Network (MV-CGAN) model to achieve data augmentation. Simultaneously, when aging data of market vehicles is missing, corresponding battery aging data is generated based on the specified SOH value using the fully-fledged MV-CGAN model, and data from the corresponding time period is selected to complete the aging data of market vehicles. Finally, the training dataset is composed of the battery aging data after data augmentation and data completion.

[0064] In this embodiment, based on a fully trained multivariate conditional generative adversarial network model, data augmentation is performed on the first sample multivariate data to generate a training dataset corresponding to the first sample multivariate data. The prediction model is then trained based on the training dataset to obtain a fully trained prediction model. This allows for the acquisition of a large amount of high-quality training data through data augmentation, making the prediction model more reliable.

[0065] In some embodiments, before processing the first sample multivariate data based on a fully trained multivariate conditional generative adversarial network model to obtain the training dataset corresponding to the first sample multivariate data, the following steps are also included:

[0066] Second sample multivariate data is obtained from the battery test database. The multivariate conditional generative adversarial network model is trained using the second sample multivariate data to obtain the initial multivariate conditional generative adversarial network model.

[0067] Specifically, cycle life test data of power batteries are obtained from the battery test database, including voltage data, current data, temperature data and corresponding SOH values ​​during each test charging process, and these are used as the second sample of multivariate data.

[0068] Furthermore, based on the second sample multivariate data, the multivariate conditional generative adversarial network model is trained until the model reaches the optimization termination condition, thus obtaining the initial multivariate conditional generative adversarial network model.

[0069] In this embodiment, second sample multivariate data is obtained from the battery test database, and the multivariate conditional generative adversarial network model is trained using the second sample multivariate data to obtain an initial multivariate conditional generative adversarial network model, thereby achieving the initial training of the multivariate conditional generative adversarial network model.

[0070] In some embodiments, the prediction model is trained based on a training dataset to obtain a fully trained prediction model, including the following steps:

[0071] Step S231: Process the first sample multivariate data in the training dataset after data augmentation to generate the grayscale dataset corresponding to the training dataset.

[0072] Step S232: Train the prediction model based on the grayscale image dataset to obtain a fully trained prediction model.

[0073] Specifically, in the training dataset, the first sample of multivariate data after data augmentation includes voltage, current, and temperature data of the power battery. The voltage, current, and temperature data of the power battery are mapped to values ​​between 0 and 255, and the mapped voltage, current, and temperature data are concatenated to generate a corresponding grayscale dataset. For example, if the length of voltage data x1, current data x2, and temperature data x3 is preset to T, the following formula is used to assign the j-th voltage data element x1 to x3. 1,j Perform the conversion:

[0074]

[0075] Where, x 1,max This represents the voltage data element with the largest value in x1, x 1,min Let x1 represent the voltage data element with the smallest value, and the right side of the above formula represents the voltage data element x1 after normalization. 1,j Mapping to the range of 0 to 1, the normalized result is multiplied by 255 to obtain the mapped voltage data. The voltage data x1, current data x2, and temperature data x3 are then transformed using the formula described above. The transformed data are then concatenated to generate a grayscale dataset with a width of 3 and a length of T.

[0076] Furthermore, the prediction model is trained based on the grayscale dataset until the mean squared error loss of the prediction model is less than a preset threshold, thus obtaining a fully trained prediction model. This allows the prediction model to fully incorporate the coupling relationship between multivariate data when analyzing target multivariate data.

[0077] It is important to know that the preset threshold is 0.01, and the mean squared error loss of the prediction model is defined as follows:

[0078] loss cnn =avg(||SOH pre -SOH real || 2 );

[0079] Where, loss cnn To predict the mean squared error loss of the model, SOH pre Predicted values ​​for health status, and SOH real This represents the true health status.

[0080] In this embodiment, the first sample multivariate data after data augmentation in the training dataset is processed to generate a grayscale dataset corresponding to the training dataset. The prediction model is then trained based on the grayscale dataset to obtain a fully trained prediction model. In this way, the health status of the power battery is analyzed by combining the coupling relationship between voltage data, current data and temperature data during the prediction process, thereby improving the accuracy of the prediction.

[0081] In some embodiments, after training the multivariate conditional generative adversarial network model with second sample multivariate data to obtain an initial multivariate conditional generative adversarial network model, the following steps are also included:

[0082] Based on the first sample of multivariate data, transfer learning is performed on the initial multivariate conditional generative adversarial network model to obtain a fully trained multivariate conditional generative adversarial network model.

[0083] Specifically, in this embodiment, the first sample multivariate data is battery aging data of market vehicles. Based on the first sample multivariate data, transfer learning is performed on the initial multivariate conditional generative adversarial network model, that is, the parameters of the initial multivariate conditional generative adversarial network model are fine-tuned until the optimization training of the model is completed. In this case, only the training configuration items are adjusted, such as the learning rate and batch size.

[0084] It is important to understand that transfer learning refers to using a model pre-trained on task A as the starting point for developing a model for task B, and then fine-tuning the model to adapt it to task B. In this embodiment, a multivariate conditional generative adversarial network model trained on the second sample multivariate data is used as the starting point for transfer learning, and the model parameters are fine-tuned based on the first sample multivariate data to make the multivariate conditional generative adversarial network model applicable to the actual first sample multivariate data, thus ensuring the reliability of the data generated based on the model.

[0085] In this embodiment, based on the first sample multivariate data, the initial multivariate conditional generative adversarial network model is transferred to obtain a fully trained multivariate conditional generative adversarial network model, thereby improving the reliability and usability of the generated data during the data augmentation process.

[0086] In some embodiments, before obtaining second sample multivariate data from a battery test database and training the multivariate conditional generative adversarial network model using the second sample multivariate data, the following steps are also included:

[0087] The conditional generative adversarial network model is optimized by gated recurrent unit network to generate a multivariate conditional generative adversarial network model.

[0088] It is important to know that, such asFigure 3 As shown, in the multivariate conditional generative adversarial network model built based on the Gated Recurrent Unit (GRU) network, the generator consists of three GRU networks, and the discriminator consists of three bidirectional GRU networks. The cycle life test data refers to the voltage data, current data, temperature data, and corresponding SOH values ​​during each test charging process. The data generated during each charging process is defined as a sample x = [x1, x2, x3], where x1, x2, and x3 represent the voltage data, current data, and temperature data during charging, respectively. In this embodiment, in addition to the GRU network, the multivariate conditional generative adversarial network model can be built based on any deep learning model.

[0089] Specifically, in the generator of the model, three sets of noise data z1, z2 and z3 are randomly generated. Each set of noise and the corresponding SOH value are input into the corresponding GRU network. The three networks then generate voltage data G(z1), current data G(z2) and temperature data G(z3) respectively. In the discriminator of the model, taking bidirectional GRU network 1 as an example, the real voltage data x1 and the generated voltage data G(z1) and the corresponding SOH value are input into the corresponding GRU network to obtain the probability that the input data is real, thereby determining whether the input data is real or generated.

[0090] Furthermore, during the training process of the above model, for the discriminator, the label corresponding to the real data is set to 1, indicating that the data is true, while the label corresponding to the generated data is set to 0, indicating that the data is false. The sum of the three network losses is then obtained, and the GRU network parameters in the discriminator are updated using the Adaptive Moment Estimation (Adam) optimizer. At this time, the generator parameters are not updated. Taking any real sample voltage data x1 as an example, the calculation process in the bidirectional GRU network is as follows:

[0091]

[0092]

[0093]

[0094] Where, x 1, For the j-th voltage data element in x1, and These represent the forward and reverse GRU network units in a bidirectional GRU network, respectively. The difference is that the reverse GRU network unit inputs data elements from back to front. jThe j-th element represents the output vector, and the output vector of the bidirectional GRU network is then input into the fully connected layer to obtain the probability that the sample voltage data is true. For the generator of the model, the generated data is input into the discriminator, and the label corresponding to the generated data is set to 1, which means that the data is true. Then, based on the output result of the discriminator and the corresponding label, the corresponding loss value is calculated, and the parameters of the three GRU networks in the generator are updated by the Adam optimizer. At this time, the discriminator parameters are not updated. The above training process of the discriminator and generator is repeated alternately until the corresponding loss values ​​of the two reach a balance state, thereby enhancing the authenticity of the generated data.

[0095] In this embodiment, a multivariate conditional generative adversarial network model is constructed based on a gated recurrent unit network. This model can then be used to augment multivariate sample data, resulting in a complete training dataset.

[0096] Figure 4 This is a flowchart illustrating the power battery health status prediction method of this embodiment, as shown below. Figure 4 As shown, the specific process of this power battery health status prediction method is as follows:

[0097] Cycle life test data (S410) is obtained from the battery test database. A multivariate conditional generative adversarial network (MCGAN) model is trained using this data to obtain an initial MCGAN model (S420). Battery aging data from market vehicles is obtained from the power battery data cloud platform (S430). Based on this data, transfer learning is performed on the initial MCGAN model, i.e., parameter fine-tuning (S440) is applied. Based on the fully trained MCGAN model, data augmentation is performed on the market vehicle battery aging data to generate a corresponding training dataset (S450). The augmented battery aging data in the training dataset is processed to generate a corresponding grayscale dataset (S460). A convolutional neural network (CNN) model is trained on this grayscale dataset to obtain a fully trained CNN model (S470). Finally, the fully trained CNN model is used to predict the state of health (SOH) of the power battery, yielding the SOH value (S480).

[0098] The present embodiment will now be described and illustrated through preferred embodiments.

[0099] Figure 5 This is a preferred flowchart of the power battery health status prediction method in this embodiment, as follows: Figure 5 As shown, the method for predicting the health status of a power battery includes the following steps:

[0100] Step S510: Optimize the conditional generative adversarial network model through a gated recurrent unit network to generate a multivariate conditional generative adversarial network model.

[0101] Step S520: Obtain second sample multivariate data from the battery test database, and train the multivariate conditional generative adversarial network model using the second sample multivariate data to obtain the initial multivariate conditional generative adversarial network model.

[0102] Step S530: Based on the first sample multivariate data, transfer learning is performed on the initial multivariate conditional generative adversarial network model to obtain a fully trained multivariate conditional generative adversarial network model.

[0103] Step S540: Based on the fully trained multivariate conditional generative adversarial network model, perform data augmentation on the first sample multivariate data to generate the training dataset corresponding to the first sample multivariate data.

[0104] Step S550: Process the first sample multivariate data in the training dataset after data augmentation to generate the grayscale dataset corresponding to the training dataset.

[0105] Step S560: Train the prediction model based on the grayscale image dataset to obtain a fully trained prediction model.

[0106] Step S570: Obtain the target multivariate data of the power battery;

[0107] Step S580: Process the target multivariate data to obtain the target grayscale image data corresponding to the target multivariate data;

[0108] Step S590: Based on the fully trained prediction model, analyze the target grayscale data to obtain the predicted health status value of the power battery.

[0109] In this embodiment, a conditional generative adversarial network (CGAN) model is optimized using a gated recurrent unit network (GRN) to generate a multivariate GRN model. Based on first and second sample multivariate data, the GRN model is further optimized and trained to obtain a fully trained GRN model. Then, based on this fully trained GRN model, data augmentation is performed on the first sample multivariate data to generate a corresponding training dataset. The augmented first sample multivariate data in the training dataset is processed to generate a corresponding grayscale dataset. The prediction model is trained on this grayscale dataset to obtain a fully trained prediction model. The target multivariate data is then processed to obtain the corresponding target grayscale data. Finally, based on the fully trained prediction model, the target grayscale data is analyzed to obtain a predicted value for the power battery's health status. This solves the problem of inaccurate prediction of the power battery's health status due to the strong coupling between multivariate data, thus improving the accuracy of the power battery's health status prediction results.

[0110] It should be noted that the steps shown in the above process or in the flowchart of the accompanying figures can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0111] This embodiment also provides a power battery health status prediction device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. The terms "module," "unit," "subunit," etc., used below can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0112] Figure 6 This is a structural block diagram of the power battery health status prediction device in this embodiment, as shown below. Figure 6 As shown, the device includes: an acquisition module 10, a processing module 20, and a prediction module 30;

[0113] Module 10 acquires the target multivariate data of the power battery;

[0114] Processing module 20 processes the target multivariate data to obtain the target grayscale image data corresponding to the target multivariate data;

[0115] The prediction module 30 analyzes the target grayscale data based on a well-trained prediction model to obtain the predicted health status value of the power battery.

[0116] The apparatus provided in this embodiment acquires target multivariate data of the power battery, processes the target multivariate data to obtain target grayscale data corresponding to the target multivariate data, and further analyzes the target grayscale data based on a well-trained prediction model to obtain the predicted health status value of the power battery. This solves the problem that the health status of the power battery cannot be accurately predicted due to the strong coupling relationship between multivariate data, and improves the accuracy of the prediction results of the health status of the power battery.

[0117] In some of these embodiments, Figure 6 Based on this, the device also includes a conversion module, which is used to obtain the voltage data, current data and temperature data of the power battery from the target multivariate data; map the voltage data, current data and temperature data of the power battery to a preset numerical range; and stitch the mapped voltage data, current data and temperature data together to obtain the target grayscale image data corresponding to the target multivariate data.

[0118] In some of these embodiments, Figure 6 Based on this, the device also includes a first training module, which is used to perform data augmentation on the first sample multivariate data based on the fully trained multivariate conditional generative adversarial network model, and generate a training dataset corresponding to the first sample multivariate data; and to train the prediction model based on the training dataset to obtain a fully trained prediction model.

[0119] In some of these embodiments, Figure 6 Based on this, the device also includes a second training module, which is used to obtain second sample multivariate data from the battery test database, and train the multivariate conditional generative adversarial network model using the second sample multivariate data to obtain an initial multivariate conditional generative adversarial network model.

[0120] In some of these embodiments, Figure 6 In addition to the above, the device also includes a third training module, which processes the augmented multivariate data of the first sample in the training dataset to generate a grayscale dataset corresponding to the training dataset; and trains the prediction model based on the grayscale dataset to obtain a fully trained prediction model.

[0121] In some of these embodiments, Figure 6 Based on this, the device also includes an optimization module, which is used to perform transfer learning on the initial multivariate conditional generative adversarial network model based on the first sample multivariate data, so as to obtain a fully trained multivariate conditional generative adversarial network model.

[0122] In some of these embodiments, ​Based on this, the device also includes a generation module, which is used to optimize the conditional generative adversarial network model through a gated recurrent unit network to generate a multivariate conditional generative adversarial network model.

[0123] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.

[0124] This embodiment also provides a computer device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0125] Optionally, the computer device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0126] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated in this embodiment.

[0127] Furthermore, in conjunction with the power battery health status prediction method provided in the above embodiments, this embodiment can also provide a storage medium for implementation. This storage medium stores a computer program; when executed by a processor, the computer program implements any of the power battery health status prediction methods described in the above embodiments.

[0128] It should be understood that the specific embodiments described herein are merely illustrative of the application and not intended to limit it. All other embodiments derived by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0129] Obviously, the accompanying drawings are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar situations based on these drawings without any creative effort. Furthermore, it is understood that although the work done in this development process may be complex and lengthy, for those skilled in the art, certain design, manufacturing, or production modifications made based on the technical content disclosed in this application are merely conventional technical means and should not be considered as insufficient disclosure of this application.

[0130] The term "embodiment" in this application refers to a specific feature, structure, or characteristic described in connection with an embodiment that may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily imply the same embodiment, nor does it imply that it is mutually exclusive with or independent of other embodiments. It will be clearly or implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0131] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of patent protection. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the appended claims.

Claims

1. A method for predicting the health status of a power battery, characterized in that, The method includes: Obtain target multivariate data for the power battery; The target multivariate data is processed to obtain target grayscale data corresponding to the target multivariate data, including: obtaining voltage data, current data and temperature data of the power battery from the target multivariate data; mapping the voltage data, current data and temperature data of the power battery to a preset numerical range; and splicing the mapped voltage data, current data and temperature data to obtain target grayscale data corresponding to the target multivariate data. Based on a fully trained multivariate conditional generative adversarial network model, data augmentation is performed on the first sample multivariate data to generate the training dataset corresponding to the first sample multivariate data. The prediction model is trained based on the training dataset to obtain a fully trained prediction model; the multivariate conditional generative adversarial network model includes a generator composed of three sets of gated recurrent unit networks and a discriminator composed of three sets of bidirectional gated recurrent unit networks. Based on the fully trained prediction model, the target grayscale data is analyzed to obtain the predicted health status value of the power battery.

2. The method for predicting the health status of a power battery according to claim 1, characterized in that, Before processing the first sample multivariate data to obtain the training dataset corresponding to the first sample multivariate data based on the fully trained multivariate conditional generative adversarial network model, the method further includes: Obtain second sample multivariate data from the battery test database, and train the multivariate conditional generative adversarial network model using the second sample multivariate data to obtain the initial multivariate conditional generative adversarial network model.

3. The method for predicting the health status of a power battery according to claim 1, characterized in that, The step of training the prediction model based on the training dataset to obtain the fully trained prediction model includes: The augmented multivariate data of the first sample in the training dataset is processed to generate a grayscale dataset corresponding to the training dataset. The prediction model is trained based on the grayscale image dataset to obtain the fully trained prediction model.

4. The method for predicting the health status of a power battery according to claim 2, characterized in that, After training the multivariate conditional generative adversarial network model using the second sample multivariate data to obtain the initial multivariate conditional generative adversarial network model, the method further includes: Based on the first sample multivariate data, transfer learning is performed on the initial multivariate conditional generative adversarial network model to obtain the fully trained multivariate conditional generative adversarial network model.

5. The method for predicting the health status of a power battery according to claim 2, characterized in that, Before obtaining the second sample multivariate data from the battery test database and training the multivariate conditional generative adversarial network model using the second sample multivariate data, the method further includes: The conditional generative adversarial network model is optimized by a gated recurrent unit network to generate the multivariate conditional generative adversarial network model.

6. A power battery health status prediction device, characterized in that, The device includes: The acquisition module acquires the target multivariate data of the power battery; The processing module processes the target multivariate data to obtain the target grayscale image data corresponding to the target multivariate data; The processing module is further configured to obtain the voltage data, current data, and temperature data of the power battery from the target multivariate data; map the voltage data, current data, and temperature data of the power battery to a preset numerical range; and stitch the mapped voltage data, current data, and temperature data together to obtain the target grayscale image data corresponding to the target multivariate data. The prediction module, based on a fully trained multivariate conditional generative adversarial network (GAN) model, performs data augmentation on the first sample multivariate data to generate a training dataset corresponding to the first sample multivariate data; the prediction model is trained based on the training dataset to obtain a fully trained prediction model; the multivariate conditional GAN ​​model includes a generator composed of three sets of gated recurrent unit networks and a discriminator composed of three sets of bidirectional gated recurrent unit networks; based on the fully trained prediction model, the target grayscale image data is analyzed to obtain the predicted health status value of the power battery.

7. A computer device, comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the steps of the power battery health status prediction method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the power battery health status prediction method according to any one of claims 1 to 5.

Images