Battery aging prediction method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a battery aging prediction method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring sample aging data of a sample battery; performing data enhancement processing on the sample aging data to obtain the sample aging data after data enhancement; extracting features in the sample aging data after data enhancement; inputting the features into a to-be-trained model, comparing predicted aging data output by the to-be-trained model with the sample aging data after data enhancement, and performing training according to a comparison result to obtain a trained battery aging prediction model; obtaining aging data of the to-be-tested battery; extracting features in the aging data of the to-be-tested battery; and inputting the characteristics into a battery aging prediction model to obtain a future change trend of the aging data of the to-be-detected battery. By adopting the method, the cost of the battery aging prediction method can be reduced, and the effects of low energy consumption, short training time and acceptable prediction precision are achieved.

Description

Technical Field

[0001] The present application relates to the field of battery technology, and in particular to a battery aging prediction method, device, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the massive promotion of new energy vehicles and various energy storage systems with power batteries as the core, the problem of predicting power battery aging information has gradually become a top priority. Since the aging of power batteries is a nonlinear process, most of the current methods use data-driven methods to predict battery aging information.

[0003] At present, the aging information of batteries is mainly predicted by stacking or nesting multiple neural networks. As the degree of neural network stacking or nesting becomes deeper and deeper, it is now possible to achieve a high-precision prediction of battery aging information. However, the continuous stacking and nesting of neural networks will lead to the gradual increase in the complexity of the method, the increasing requirements for computing power, and the increasing cost of battery aging prediction, resulting in the high cost of existing battery aging prediction methods. Summary of the invention

[0004] Based on this, it is necessary to provide a battery aging prediction method, device, computer equipment, computer-readable storage medium and computer program product that can reduce the cost of battery aging prediction in order to address the above technical problems.

[0005] In a first aspect, the present application provides a battery aging prediction method, comprising:

[0006] Obtaining sample aging data of a sample battery;

[0007] Performing data enhancement processing on the sample aging data to obtain data-enhanced sample aging data;

[0008] Extract features from sample aging data after data augmentation;

[0009] Inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison results to obtain a trained battery aging prediction model;

[0010] Obtain aging data of the battery to be tested;

[0011] Extracting features from the aging data of the battery to be tested;

[0012] The features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

[0013] In one embodiment, performing data enhancement processing on the sample aging data to obtain the sample aging data after data enhancement includes:

[0014] Build a data augmentation network;

[0015] The sample aging data is input into the data enhancement network to obtain the sample aging data after data enhancement.

[0016] In one embodiment, constructing a data enhancement network includes:

[0017] Constructing an initial data enhancement network; the initial data enhancement network includes a generator network and a discriminator network;

[0018] The parameters of the generator network and the parameters of the discriminator network are adjusted by using an optimization algorithm so that the generator network and the discriminator network meet preset conditions; the preset conditions are that when the parameters in the generator network are fixed, the loss function of the discriminator network is maximum; when the parameters in the discriminator network are fixed, the loss function of the generator network is minimum.

[0019] In one embodiment, extracting features from the sample aging data after data enhancement includes:

[0020] Construct feature extraction network;

[0021] The sample aging data after data enhancement is input into the feature extraction network to extract features.

[0022] In one embodiment, constructing a feature extraction network includes:

[0023] Construct the initial feature extraction network;

[0024] The weight parameters and bias parameters in the initial feature extraction network are continuously iterated using an optimization algorithm until the loss function of the initial feature extraction network is minimized, thereby obtaining a feature extraction network.

[0025] In one embodiment, the feature is input into the model to be trained, the predicted aging data of the sample battery output by the model to be trained is compared with the sample aging data after data enhancement, and training is performed according to the comparison result to obtain a trained battery aging prediction model, including:

[0026] The features are input into the model to be trained, and the model to be trained is trained using an optimization algorithm until the gap between the predicted aging data of the sample battery output by the model to be trained and the sample aging data after data enhancement is minimized, thereby obtaining a battery aging prediction model.

[0027] In a second aspect, the present application also provides a battery aging prediction device, comprising:

[0028] An acquisition module, used for acquiring sample aging data of a sample battery;

[0029] A data enhancement module, used for performing data enhancement processing on the sample aging data to obtain the sample aging data after data enhancement;

[0030] An extraction module, used to extract features from sample aging data after data enhancement;

[0031] A training module, used for inputting the features into a model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison result to obtain a trained battery aging prediction model;

[0032] The acquisition module is also used to acquire aging data of the battery to be tested;

[0033] The extraction module is further used to extract features from the aging data of the battery to be tested;

[0034] The prediction module is used to input the features in the aging data of the battery to be tested into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

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

[0036] Obtaining sample aging data of a sample battery;

[0037] Performing data enhancement processing on the sample aging data to obtain data-enhanced sample aging data;

[0038] Extract features from sample aging data after data augmentation;

[0039] Inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison results to obtain a trained battery aging prediction model;

[0040] Obtain aging data of the battery to be tested;

[0041] Extracting features from the aging data of the battery to be tested;

[0042] The features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

[0043] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0044] Obtaining sample aging data of a sample battery;

[0045] Performing data enhancement processing on the sample aging data to obtain data-enhanced sample aging data;

[0046] Extract features from sample aging data after data augmentation;

[0047] Inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison results to obtain a trained battery aging prediction model;

[0048] Obtain aging data of the battery to be tested;

[0049] Extracting features from the aging data of the battery to be tested;

[0050] The features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

[0051] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0052] Obtaining sample aging data of a sample battery;

[0053] Performing data enhancement processing on the sample aging data to obtain data-enhanced sample aging data;

[0054] Extract features from sample aging data after data augmentation;

[0055] Inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison results to obtain a trained battery aging prediction model;

[0056] Obtain aging data of the battery to be tested;

[0057] Extracting features from the aging data of the battery to be tested;

[0058] The features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

[0059] The above-mentioned battery aging prediction method, device, computer equipment, computer-readable storage medium and computer program product first obtain sample aging data of a sample battery, perform enhancement processing on the sample aging data, obtain sample aging data after data enhancement, and enhance the training amount of the subsequent battery aging prediction model; extract features from the sample aging data after data enhancement, input the features into the model to be trained, compare the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, perform training according to the comparison results, and obtain a trained battery aging prediction model; obtain aging data of the battery to be tested; extract features from the aging data of the battery to be tested; input the features from the aging data of the battery to be tested into the battery aging prediction model, and obtain the future change trend of the aging data of the battery to be tested. In this way, the cost of the battery aging prediction method can be reduced, and the effects of low energy consumption, short training time and acceptable prediction accuracy can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0061] Figure 1 A schematic diagram of a flow chart of a battery aging prediction method in one embodiment;

[0062] Figure 2 A structural diagram of a data enhancement network in one embodiment;

[0063] Figure 3 A schematic diagram of a flow chart of a battery aging prediction method in another embodiment;

[0064] Figure 4 is a structural block diagram of a battery aging prediction device in one embodiment;

[0065] Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with 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.

[0067] In one embodiment, Figure 1 As shown, a battery aging prediction method is provided. This embodiment uses the method applied to a terminal as an example. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps 102 to 114:

[0068] Step 102, obtaining sample aging data of a sample battery.

[0069] Among them, the sample aging data includes the sample battery's charge and discharge cycle data, battery capacity attenuation data, voltage and current data, battery health status data, etc.

[0070] Specifically, the terminal obtains sample aging data of the sample battery, wherein the sample aging data may be battery capacity attenuation data, or may be other data related to battery aging, which is not limited in the embodiments of the present application.

[0071] Step 104: perform data enhancement processing on the sample aging data to obtain the sample aging data after data enhancement.

[0072] Among them, the data enhancement processing is to obtain more battery aging data that is similar to the sample aging data.

[0073] In some embodiments, the terminal performs data enhancement processing on the sample aging data to obtain more battery aging data similar to the sample aging data and the initial sample aging data, and use them together as the sample aging data after data enhancement.

[0074] Step 106, extracting features from the sample aging data after data enhancement;

[0075] In some embodiments, the terminal extracts features from the sample aging data after data enhancement, wherein the features are matrices representing characteristics of the sample aging data after data enhancement.

[0076] Step 108, input the features into the model to be trained, compare the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, perform training based on the comparison results, and obtain a trained battery aging prediction model.

[0077] Among them, the model to be trained is an SNN (Spiking Neural Network) network model.

[0078] Specifically, the terminal will input the features into the pulse neural network for training to obtain predicted aging data, and then compare the predicted aging data with the aging data after data enhancement, and continuously train based on the comparison results. After the training is completed, the final battery aging prediction model is obtained.

[0079] Step 110, obtaining aging data of the battery to be tested.

[0080] Specifically, the terminal will obtain the battery aging data that it wants to predict, where the sample aging data can be battery capacity decay data, or other data related to battery aging, which is not limited in this embodiment of the present application.

[0081] Step 112: extracting features from the aging data of the battery to be tested.

[0082] In some embodiments, the terminal inputs the aging data of the battery to be tested into a feature extraction network to extract features from the aging data of the battery to be predicted, wherein the features represent a matrix of characteristics of the aging data of the battery to be predicted.

[0083] Step 114 , input the features in the aging data of the battery to be tested into the battery aging prediction model to obtain the future trend of the aging data of the battery to be tested.

[0084] In some embodiments, the terminal inputs the characteristics of the battery aging data to be predicted into the battery aging prediction model to obtain the future change trend of the battery aging data to be predicted.

[0085] In the above-mentioned battery aging prediction method, sample aging data of a sample battery is obtained, the sample aging data is enhanced to obtain sample aging data after data enhancement, and the training amount of the subsequent battery aging prediction model is enhanced; the features in the sample aging data after data enhancement are extracted, and the features are input into the model to be trained, and the predicted aging data of the sample battery output by the model to be trained is compared with the sample aging data after data enhancement, and training is performed according to the comparison results to obtain a trained battery aging prediction model; the aging data of the battery to be tested is obtained; the features in the aging data of the battery to be tested are extracted; the features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested. In this way, the cost of the battery aging prediction method can be reduced, and the effects of low energy consumption, short training time and acceptable prediction accuracy can be achieved.

[0086] In some exemplary embodiments, data enhancement processing is performed on the sample aging data to obtain the data-enhanced sample aging data, including: constructing a data enhancement network; inputting the sample aging data into the data enhancement network to obtain the data-enhanced sample aging data.

[0087] In actual implementation, the terminal will build a data enhancement network and input the sample aging data into the data enhancement network to obtain more battery aging data similar to the sample aging data and the initial sample aging data as the sample aging data after data enhancement.

[0088] In some exemplary embodiments, constructing a data augmentation network includes: constructing an initial data augmentation network; the initial data augmentation network includes a generator network and a discriminator network; using an optimization algorithm to adjust the parameters of the generator network and the parameters of the discriminator network so that the generator network and the discriminator network meet preset conditions; the preset conditions are that when the parameters in the generator network are fixed, the loss function of the discriminator network is maximized; when the parameters in the discriminator network are fixed, the loss function of the generator network is minimized.

[0089] In some embodiments, the terminal will build an initial data enhancement network, namely a GAN (Generative Adversarial Network) network, which includes a generator network and a discriminator network. The data enhancement network includes but is not limited to a GAN network, and may also be a VAE (Variational Autoencoder) network or other networks that can perform data enhancement, which is not limited in this embodiment. The specific structure is as follows Figure 2 As shown in Figure 2, the generator network and the discriminator are continuously confronted to generate data that is close to reality.

[0090] The dataset of sample aging data is used as real data. At the same time, the generator network generates fake data close to the real data from random noise, and continuously learns to generate fake data closer to the real data. Both the real data and the fake data are input into the discriminator network. The discriminator network determines the authenticity of the input data and outputs the real data and the fake data of the battery close to the sample aging data as the sample aging data after data enhancement.

[0091] Furthermore, through continuous adversarial learning, the parameters in the generator network and the discriminator network are continuously modified, so that when the parameters in the generator network are fixed, the loss function of the discriminator network is maximized; when the parameters in the discriminator network are fixed, the loss function of the generator network is minimized. Among them, the loss function of the discriminator network is The calculation formula is shown in (1), and the loss function of the generator network is The calculation formula is shown in (2).

[0092]

[0093]

[0094] in, represents any real data, Represents the result judged by the discriminator network on the real data, represents fake data generated by the generator network with the same structure as the real data, Represents the result of the discriminator network on the fake data, data is the real data, z is the noise, and g is the distribution of the generator.

[0095] Furthermore, the terminal will continuously train the generative adversarial network, using the Adam (adaptive optimization algorithm) optimizer to continuously optimize the parameters in the generator network and the discriminator network, so that when the parameters in the generator network are fixed, the loss function of the discriminator network is maximized; when the parameters in the discriminator network are fixed, the loss function of the generator network is minimized.

[0096] In the above embodiment, by continuously training the GAN network, the fake data generated by the GAN network is made closer to the real data, so that the model result finally trained with the data is more accurate.

[0097] In the above embodiment, by performing enhancement processing on the sample aging data, more training data is obtained, so that the model prediction results obtained after subsequent training are more accurate.

[0098] In some exemplary embodiments, extracting features from the sample aging data after data enhancement includes: constructing a feature extraction network; inputting the sample aging data after data enhancement into the feature extraction network to extract features.

[0099] In some embodiments, the terminal will establish a multi-layer CNN (convolutional neural network) network and continuously train it. After the training is completed, the trained convolutional neural network is used as a feature extraction network. The feature extraction network includes but is not limited to a CNN network, and can also be other networks that can perform feature extraction, such as an RNN (recurrent neural network) network, which is not limited in this embodiment. The sample aging data after data enhancement is input into the feature extraction network, and the features of the sample aging data after data enhancement are extracted, and the features are output in the form of a matrix.

[0100] In some exemplary embodiments, constructing a feature extraction network includes: constructing an initial feature extraction network; and continuously iterating weight parameters and bias parameters in the initial feature extraction network using an optimization algorithm until the loss function of the initial feature extraction network is minimized, thereby obtaining a feature extraction network.

[0101] In some embodiments, the terminal will establish an initial multi-layer CNN network, the structure of which uses the network structure of LeNet-5 (convolutional neural network). Among them, the network structure of the CNN network can also be the network structure of AlexNet (convolutional neural network), which is not limited in this embodiment. The specific network structure includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a third convolutional layer, a fully connected layer, and an output layer. Among them, each convolutional layer has a corresponding convolution kernel, and the corresponding convolution kernel is used for convolution; each pooling layer also has a corresponding convolution kernel for pooling. Each convolution kernel has a corresponding bias and weight.

[0102] Furthermore, the terminal uses a back propagation algorithm to iterate the parameters in the CNN network, namely, the bias parameters and weight parameters. Among them, the iterative method for the parameters in the CNN network can also be other methods that have an iterative effect on the parameters, which is not limited in this embodiment. The error is back-propagated from the output layer to the input layer layer by layer through the chain rule, and the weight parameters and bias parameters of each layer are continuously adjusted, so that the network output gradually approaches the true value.

[0103] Among them, the loss function of the CNN network is minimized based on the gradient descent algorithm. The gradient descent algorithm uses the Adam algorithm to update the corresponding weights and biases according to the gradient of each layer. The specific calculation formula is shown in (3).

[0104]

[0105] in, is the weight from layer i to layer j, is the learning rate, L is the loss function, is the gradient of loss with respect to weights.

[0106] In the above embodiment, the CNN network is continuously trained to minimize the loss function of the network, so that the extracted features are more in line with the sample aging data after the input data enhancement.

[0107] In the above embodiment, features in the data, i.e., a matrix capable of representing data characteristics, are extracted through a feature extraction network. In subsequent processing, the matrix capable of representing data characteristics only needs to be directly input into the model for training, thereby reducing the time for data processing.

[0108] In some exemplary embodiments, features are input into the model to be trained, the predicted aging data of the sample battery output by the model to be trained is compared with the sample aging data after data enhancement, and training is performed based on the comparison results to obtain a trained battery aging prediction model, including: inputting features into the model to be trained, training the model to be trained using an optimization algorithm until the gap between the predicted aging data of the sample battery output by the model to be trained and the sample aging data after data enhancement is minimized, thereby obtaining a battery aging prediction model.

[0109] In some embodiments, the terminal constructs a model to be trained, which is a three-layer SNN (spiking neural network) network. The number of layers of the SNN network is not limited, and this embodiment does not limit this.

[0110] Exemplarily, the three-layer SNN network structure includes an input layer, three hidden layers and an output layer. Each hidden layer contains a group of spiking neurons that simulate the behavior of biological neurons. Furthermore, each hidden layer receives the pulse output emitted by the previous layer as input, accumulates the membrane potential, and emits pulses to the next layer. If the membrane potential of the neurons exceeds a certain threshold, they will emit pulses and reset the membrane potential to the initial value.

[0111] Furthermore, the STBP (space-time back propagation) algorithm is used to train the model to be trained in the time dimension and the space dimension. In the time dimension, the gradient of the loss to the membrane potential and weight of each neuron at each time step is calculated, and the parameters are continuously changed to reduce the error; among them, the gradient of the non-differentiable pulse emission function is represented by a differentiable smooth approximate function. In the space dimension, the error is propagated from the input layer to the input layer, and the error is reduced by continuously changing the parameters.

[0112] Furthermore, the parameters in the SNN network are continuously changed, and the mean square error between the predicted aging data output by the SNN network and the sample aging data after data enhancement is calculated. When the mean square error between the two is the smallest, the training is completed and the battery aging prediction model is obtained.

[0113] In the above embodiment, the training model is continuously trained by the STBP algorithm, which can realize double back propagation in the time dimension and the space dimension, so that the prediction result in the actual implementation process is more accurate.

[0114] In a specific embodiment, the battery aging prediction method includes the following steps:

[0115] The terminal obtains sample aging data of the sample battery; constructs an initial data enhancement network; the initial data enhancement network includes a generator network and a discriminator network; uses an optimization algorithm to adjust the parameters of the generator network and the parameters of the discriminator network so that the generator network and the discriminator network meet preset conditions; the preset conditions are that when the parameters in the generator network are fixed, the loss function of the discriminator network is the largest; when the parameters in the discriminator network are fixed, the loss function of the generator network is the smallest.

[0116] Input the sample aging data into the data enhancement network to obtain the sample aging data after data enhancement. Construct the initial feature extraction network; use the optimization algorithm to continuously iterate the weight parameters and bias parameters in the initial feature extraction network until the loss function of the initial feature extraction network is minimized, and obtain the feature extraction network.

[0117] The sample aging data after data enhancement is input into the feature extraction network to extract features. The features are input into the model to be trained, and the model to be trained is trained using the optimization algorithm until the gap between the predicted aging data of the sample battery output by the model to be trained and the sample aging data after data enhancement is minimized, thus obtaining a battery aging prediction model.

[0118] Obtain aging data of the battery to be tested; extract features from the aging data of the battery to be tested; input the features from the aging data of the battery to be tested into a battery aging prediction model to obtain a future change trend of the aging data of the battery to be tested.

[0119] To explain the battery aging prediction method in the present application in detail, an embodiment is used for explanation below. For example, the present application uses a battery aging prediction method based on battery capacity attenuation data of a specific aging battery for explanation. The specific process is as follows: Figure 3 shown.

[0120] First, the terminal obtains the battery capacity attenuation data of the sample battery and constructs an initial data enhancement network, namely the GAN (Generative Adversarial Network) network, which includes a generator network and a discriminator network. The generator network and the discriminator are constantly confronted to generate approximate real data. The data set of sample aging data is used as real data. At the same time, the generator network generates fake data close to the real data from random noise, and continuously learns to generate fake data closer to the real data.

[0121] Furthermore, through continuous adversarial learning, the parameters in the generator network and the discriminator network are continuously modified, so that when the parameters in the generator network are fixed, the loss function of the discriminator network is maximized; when the parameters in the discriminator network are fixed, the loss function of the generator network is minimized. Among them, the loss function of the discriminator network is The calculation formula is shown in (1), and the loss function of the generator network is The calculation formula is shown in (2). The terminal will continuously train the generative adversarial network and use the Adam (adaptive optimization algorithm) optimizer to continuously optimize the parameters in the generator network and the discriminator network, so that when the parameters in the generator network are fixed, the loss function of the discriminator network is maximized; when the parameters in the discriminator network are fixed, the loss function of the generator network is minimized.

[0122] Both the real data and the forged data are input into the discriminator network, which determines the authenticity of the input data and outputs the real data and the forged data of the battery that is similar to the sample aging data as the sample aging data after data enhancement.

[0123] Establish a multi-layer CNN (convolutional neural network) network and continuously train it. After the training is completed, use the trained convolutional neural network as a feature extraction network. Establish an initial multi-layer CNN network, the structure of which uses the network structure of LeNet-5 (convolutional neural network). The specific network structure includes the input layer, the first convolution layer, the first pooling layer, the second convolution layer, the second pooling layer, the third convolution layer, the fully connected layer, and the output layer. Among them, each convolution layer has a corresponding convolution kernel, and the corresponding convolution kernel is used for convolution; each pooling layer also has a corresponding convolution kernel for pooling. Each convolution kernel has a corresponding bias and weight.

[0124] Furthermore, the terminal uses a back propagation algorithm to iterate the parameters in the CNN network, namely, the bias parameters and weight parameters. Among them, the iterative method for the parameters in the CNN network can also be other methods that have an iterative effect on the parameters, which is not limited in this embodiment. The error is back-propagated from the output layer to the input layer layer by layer through the chain rule, and the weight parameters and bias parameters of each layer are continuously adjusted, so that the network output gradually approaches the true value.

[0125] Among them, the loss function of the CNN network is minimized based on the gradient descent algorithm. The gradient descent algorithm uses the Adam algorithm to update the corresponding weights and biases according to the gradient of each layer. The specific calculation formula is shown in (3).

[0126] The sample aging data after data enhancement is input into the feature extraction network, and the features of the sample aging data after data enhancement are extracted, and the features are output in the form of a matrix.

[0127] The terminal will build a model to be trained, which is a three-layer SNN (Spiking Neural Network) network. The three-layer SNN network structure includes an input layer, three hidden layers, and an output layer. Each hidden layer contains a group of spiking neurons that simulate the behavior of biological neurons. Furthermore, each hidden layer receives the pulse output emitted by the previous layer as input, accumulates the membrane potential, and emits pulses to the next layer. If the membrane potential of the neurons exceeds a certain threshold, they will emit pulses and reset the membrane potential to the initial value.

[0128] Furthermore, the STBP (space-time back propagation) algorithm is used to train the model to be trained in the time dimension and the space dimension. In the time dimension, the gradient of the loss to the membrane potential and weight of each neuron at each time step is calculated, and the parameters are continuously changed to reduce the error; among them, the gradient of the non-differentiable pulse emission function is represented by a differentiable smooth approximate function. In the space dimension, the error is propagated from the input layer to the input layer, and the error is reduced by continuously changing the parameters.

[0129] Furthermore, the parameters in the SNN network are continuously changed, and the mean square error between the predicted aging data output by the SNN network and the sample aging data after data enhancement is calculated. When the mean square error between the two is the smallest, the training is completed and the battery aging prediction model is obtained.

[0130] The terminal obtains the battery capacity attenuation data of the battery to be predicted, inputs the aging data of the battery to be tested into the feature extraction network, and extracts the features in the aging data of the battery to be predicted. The features of the aging data of the battery to be predicted are input into the battery aging prediction model to obtain the future change trend of the battery capacity attenuation data of the battery to be predicted.

[0131] The present application obtains sample aging data of a sample battery, performs enhancement processing on the sample aging data, obtains sample aging data after data enhancement, and increases the training amount of the subsequent battery aging prediction model; extracts features from the sample aging data after data enhancement, inputs the features into the model to be trained, compares the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, performs training based on the comparison results, and obtains a trained battery aging prediction model; obtains aging data of the battery to be tested; extracts features from the aging data of the battery to be tested; inputs the features from the aging data of the battery to be tested into the battery aging prediction model, and obtains the future change trend of the aging data of the battery to be tested. In this way, the cost of the battery aging prediction method can be reduced, and the effects of low energy consumption, short training time, and acceptable prediction accuracy can be achieved.

[0132] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0133] Based on the same inventive concept, the embodiment of the present application also provides a battery aging prediction device for implementing the battery aging prediction method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more battery aging prediction device embodiments provided below can refer to the limitations of the battery aging prediction method above, and will not be repeated here.

[0134] In an exemplary embodiment, Figure 4 As shown, a battery aging prediction device is provided, including: an acquisition module 401, a data enhancement module 402, an extraction module 403, a training module 404 and a prediction module 405, wherein:

[0135] An acquisition module, used for acquiring sample aging data of a sample battery;

[0136] A data enhancement module, used for performing data enhancement processing on the sample aging data to obtain the sample aging data after data enhancement;

[0137] An extraction module, used to extract features from sample aging data after data enhancement;

[0138] A training module, used for inputting the features into a model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison result to obtain a trained battery aging prediction model;

[0139] The acquisition module is also used to acquire aging data of the battery to be tested;

[0140] The extraction module is further used to extract features from the aging data of the battery to be tested;

[0141] The prediction module is used to input the features in the aging data of the battery to be tested into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

[0142] In some embodiments, the data enhancement module is further used to construct a data enhancement network; the sample aging data is input into the data enhancement network to obtain the sample aging data after data enhancement.

[0143] In some embodiments, the device also includes a construction module for constructing an initial data enhancement network; the initial data enhancement network includes a generator network and a discriminator network; the parameters of the generator network and the parameters of the discriminator network are adjusted using an optimization algorithm so that the generator network and the discriminator network meet preset conditions; the preset conditions are that when the parameters in the generator network are fixed, the loss function of the discriminator network is maximum; when the parameters in the discriminator network are fixed, the loss function of the generator network is minimum.

[0144] In some embodiments, the extraction module is further used to construct a feature extraction network; the sample aging data after data enhancement is input into the feature extraction network to extract features.

[0145] In some embodiments, the above-mentioned construction module is also used to construct an initial feature extraction network; the weight parameters and bias parameters in the initial feature extraction network are continuously iterated using an optimization algorithm until the loss function of the initial feature extraction network is minimized, thereby obtaining a feature extraction network.

[0146] In some embodiments, the above-mentioned training module is also used to input features into the model to be trained, and use the optimization algorithm to train the model to be trained until the gap between the predicted aging data of the sample battery output by the model to be trained and the sample aging data after data enhancement is minimized, thereby obtaining a battery aging prediction model.

[0147] Each module in the above-mentioned battery aging prediction device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0148] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 5As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. 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 and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a battery aging prediction method is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0149] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure 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 certain components, or have a different arrangement of components.

[0150] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0151] Obtaining sample aging data of a sample battery;

[0152] Performing data enhancement processing on the sample aging data to obtain data-enhanced sample aging data;

[0153] Extract features from sample aging data after data augmentation;

[0154] Inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison results to obtain a trained battery aging prediction model;

[0155] Obtain aging data of the battery to be tested;

[0156] Extracting features from the aging data of the battery to be tested;

[0157] The features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

[0158] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0159] Obtaining sample aging data of a sample battery;

[0160] Performing data enhancement processing on the sample aging data to obtain data-enhanced sample aging data;

[0161] Extract features from sample aging data after data augmentation;

[0162] Inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison results to obtain a trained battery aging prediction model;

[0163] Obtain aging data of the battery to be tested;

[0164] Extracting features from the aging data of the battery to be tested;

[0165] The features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

[0166] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0167] Obtaining sample aging data of a sample battery;

[0168] Performing data enhancement processing on the sample aging data to obtain data-enhanced sample aging data;

[0169] Extract features from sample aging data after data augmentation;

[0170] Inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison results to obtain a trained battery aging prediction model;

[0171] Obtain aging data of the battery to be tested;

[0172] Extracting features from the aging data of the battery to be tested;

[0173] The features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

[0174] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0175] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and 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-mentioned methods. Among them, any reference to the 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), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0176] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0177] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A battery aging prediction method, characterized in that: The method comprises: Obtaining sample aging data of a sample battery; Performing data enhancement processing on the sample aging data to obtain data-enhanced sample aging data; Extract features from sample aging data after data augmentation; Inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison results to obtain a trained battery aging prediction model; Obtain aging data of the battery to be tested; Extracting features from the aging data of the battery to be tested; The features in the aging data of the battery to be tested are input into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

2. The method according to claim 1, characterized in that The performing data enhancement processing on the sample aging data to obtain the sample aging data after data enhancement includes: Build a data augmentation network; The sample aging data is input into the data enhancement network to obtain the sample aging data after data enhancement.

3. The method according to claim 2, characterized in that The data enhancement network is constructed, including: Constructing an initial data enhancement network; the initial data enhancement network includes a generator network and a discriminator network; The parameters of the generator network and the parameters of the discriminator network are adjusted by using an optimization algorithm so that the generator network and the discriminator network meet preset conditions; the preset conditions are that when the parameters in the generator network are fixed, the loss function of the discriminator network is maximum; when the parameters in the discriminator network are fixed, the loss function of the generator network is minimum.

4. The method according to claim 1, characterized in that: The extracting features of the sample aging data after data enhancement includes: Construct feature extraction network; The sample aging data after data enhancement is input into the feature extraction network to extract features.

5. The method according to claim 4, characterized in that The constructing of the feature extraction network comprises: Build the initial feature extraction network; The weight parameters and bias parameters in the initial feature extraction network are continuously iterated using an optimization algorithm until the loss function of the initial feature extraction network is minimized, thereby obtaining a feature extraction network.

6. The method according to claim 1, characterized in that The step of inputting the features into the model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison result to obtain a trained battery aging prediction model includes: The features are input into the model to be trained, and the model to be trained is trained using an optimization algorithm until the gap between the predicted aging data of the sample battery output by the model to be trained and the sample aging data after data enhancement is minimized, thereby obtaining a battery aging prediction model.

7. A battery aging prediction device, characterized in that: The device comprises: An acquisition module, used for acquiring sample aging data of a sample battery; A data enhancement module, used for performing data enhancement processing on the sample aging data to obtain the sample aging data after data enhancement; An extraction module, used to extract features from sample aging data after data enhancement; A training module, used for inputting the features into a model to be trained, comparing the predicted aging data of the sample battery output by the model to be trained with the sample aging data after data enhancement, and performing training according to the comparison result to obtain a trained battery aging prediction model; The acquisition module is also used to acquire aging data of the battery to be tested; The extraction module is further used to extract features from the aging data of the battery to be tested; The prediction module is used to input the features in the aging data of the battery to be tested into the battery aging prediction model to obtain the future change trend of the aging data of the battery to be tested.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

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

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images