Method and device for optimizing battery health state prediction model based on federal continuous learning

Abstract

The invention discloses a method and equipment for optimizing a battery health state prediction model based on federal continuous learning, and relates to the technical field of battery health management. The method comprises the following steps: S1, locally training a battery state of health (SOH) prediction model and a data monitoring model, uploading local parameters to a central server, integrating the local parameters by the central server, issuing and repeatedly training until a loss function is converged, and waiting for a local training and federal learning retraining instruction; s2, the SOH prediction model outputs a battery health degree, and the data monitoring model outputs a data drift anomaly degree; and S3, judging the accuracy of an output result of the SOH prediction model according to the data drift anomaly, and continuously optimizing the SOH prediction model through federal re-learning triggered by parameter fine tuning triggering local training and model re-training. According to the method, joint optimization of the multiple vehicle-mounted computer terminal battery prediction models is achieved, the model performance is continuously optimized in the use process, the models are made to continuously adapt to new data, and lifelong learning of the models is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery health management, in particular to a method and device for optimizing a battery health state prediction model based on federated continuous learning. Background Art

[0002] With the rapid development of electric vehicles and renewable energy systems, the performance of the Battery Management System (BMS) has been required to be higher and higher. The performance of the BMS is directly related to the vehicle's endurance, charging efficiency, and safety. Therefore, accurately evaluating the State of Health (SOH) of the battery is of great significance for improving the performance of the BMS. At the same time, in the current era, data privacy has received increasing attention, and it is urgent to protect data privacy and solve the data silo problem during the data processing.

[0003] Most traditional battery SOH evaluation methods rely on single measurement indicators, such as voltage, capacity, and internal resistance. These methods often ignore the comprehensive impact of multiple factors on the battery health state and have high requirements for the accuracy of the measurement equipment, resulting in limited estimation accuracy. In addition, the equivalent circuit model constructed based on partial differential equations with physical constraints is also applied to the health state evaluation of lithium batteries. Although this method is simple and practical with low computational complexity, its reliability and accuracy still need to be further improved due to the interference of multiple uncertain factors.

[0004] With the continuous progress of artificial intelligence technology, data-driven algorithms have been widely used in the field of battery health state evaluation. The emergence of federated continuous learning mainly stems from the consideration of data privacy protection and data silo problems. Federated learning provides a distributed training solution to address these issues, and continuous learning endows the model with the ability to continuously learn in a continuous data stream. Federated continuous learning combines the advantages of both, and through the knowledge sharing between in-vehicle computer terminals, it realizes the joint optimization of the battery health state SOH prediction model while protecting data privacy, reduces data requirements and communication volume, enhances universality, and enables lifelong learning. By combining a deep autoencoder with a Gaussian mixture model, a powerful unsupervised anomaly detection method is also provided. Its end-to-end training mode and the ability to retain key information make it perform excellently in multiple fields. Summary of the Invention

[0005] The object of the present invention is to propose a method and device for optimizing a battery health state prediction model based on federated continuous learning. Through a specific model structure, data processing method, parameter update, and optimization strategy, the joint optimization of the battery health state SOH prediction model under data privacy protection is achieved, the data requirements and communication volume are reduced, the universality is enhanced, and lifelong learning is enabled.

[0006] To achieve the above object, the present invention proposes a method for optimizing a battery health state prediction model based on federated continuous learning, and the specific steps are as follows: Step S1, Federated learning model training: The in-vehicle computer terminal and the central server build a battery health state SOH prediction model and a data monitoring model with the same architecture. Each in-vehicle computer terminal collects and preprocesses the local battery charge and discharge data, trains the model locally, and then uploads the local model parameters to the central server. The central server integrates the parameters and distributes them back to each in-vehicle computer terminal, and repeats the training until the loss functions all converge; Step S2, Use the battery health state SOH prediction model to output the battery health degree, and output the data drift abnormality degree through the data monitoring model to evaluate the performance of the battery health state SOH prediction model; Step S3, Select the optimization method and continuous optimization: According to the data drift abnormality degree, select the optimization method for the battery health state SOH prediction model and perform continuous optimization according to the optimization method.

[0007] Preferably, in step S1, the structure of the battery health state SOH prediction model is as follows: The preprocessed data passes through a feature extraction layer including multiple fully connected layers or convolutional layers to extract high-level features from it. Subsequently, a mapping layer composed of fully connected layers, with the help of a non-linear activation function, captures the complex relationships in the data, further transforms the feature data into specific feature representations and outputs the prediction results. Finally, the degradation simulation layer simulates the battery attenuation rate by adding physical constraints; The structure of the data monitoring model is as follows: Input the same data as the battery health state SOH prediction model and copy the feature extraction layer parameters to the fitting layer so that the two models have the same feature data output at the front end. The feature data outputs the feature vector through the mapping layer, and then the data is restored by the demapping layer to calculate the reconstruction error, and finally the model performance is calculated through the Gaussian mixture model.

[0008] Preferably, in step S1, the specific steps of the federated learning model training are as follows: Step S11, The in-vehicle computer terminal downloads the current battery health state SOH prediction model and data monitoring model from the central server. The parameters include the number of hidden layers and the number of neurons H, the activation function, and decompose the local model parameters into globally shared parameters and specific parameters , parameter matrix , is the bias term; Step S12, The in-vehicle computer terminal sorts out and prepares the charging data stored locally, performs data cleaning and preprocessing, and divides it according to the ratio of 6:2:2 to determine the training set, validation set and test set; Step S13: Use the training set data to train the downloaded battery state of health (SOH) prediction model and data monitoring model. Calculate the model output through forward propagation, calculate the loss function for backpropagation, calculate the gradient of the model parameters, and use the gradient descent method to update the global shared parameters of the model. Specific parameters Bias term Iterate multiple times until the loss functions of both models converge. Step S14: The in-vehicle computer terminal sends the updated global shared parameters and bias term back to the central server through the homomorphic encryption algorithm, and wait for the central server to aggregate all the uploaded model parameters and issue new model parameters for the next round of training iteration until the loss functions of both models converge. Among them, are the aggregated parameter matrix, global parameters, and bias term respectively.

[0009] Preferably, in step S11, the federated learning model parameters include global shared parameters and specific parameters . After the local training of the in-vehicle computer terminal is completed, the global shared parameters are uploaded to the central server, and the specific parameters are left locally to adapt to the personalization of the in-vehicle computer terminal. The formula for the federated learning model parameters is as follows: ; where represents the element-wise multiplication of the corresponding elements of each network unit, is the federated parameter, is the global shared parameter, is the migration vector, is the specific parameter of the in-vehicle computer terminal.

[0010] Preferably, in step S12, the data collection steps are as follows: Collect data during battery charging. Based on the in-vehicle terminal in the federated framework , record the monitoring data every 0.1 s. The data form is time-series streaming data, that is, the input feature of the th in-vehicle terminal is , where R is the resistance of the in-vehicle terminal, is the voltage of the in-vehicle terminal, is the current of the in-vehicle terminal, S is the mean square error of the in-vehicle terminal, P k is the power of the in-vehicle terminal, D e is the slope of the in-vehicle terminal. is the charging time of the in - vehicle terminal, is the cumulative power of the in - vehicle terminal, is the curve slope of the in - vehicle terminal, is the temperature of the in - vehicle terminal; Based on criterion, the in - vehicle computer terminal performs data cleaning and pre - processing on the local charge - discharge data, and at the same time performs data normalization. The formula is as follows: ; Among them, is the inter - quartile range, and 0.7413 is a constant multiple of the inter - quartile range of, is the processed data, is the original data point, is the median of the data set median.

[0011] Preferably, in step S14, the parameter update of the in - vehicle computer terminal is realized by stochastic gradient descent. The central server aggregates all uploaded model parameters in two parts, including training the local data in batches on the in - vehicle computer terminal, calculating the gradient of each batch, accumulating gradients for multiple rounds, and sending them to the server at one time for aggregation and adding random noise to protect privacy; when retraining, the in - vehicle computer terminal determines the weight of the updated parameter according to the degree of abnormality of the outlier. When the degree of abnormality is large, the weight is increased, and when the degree of abnormality is small, the weight is decreased; among them, the calculation formula for the central server parameter aggregation is as follows: ; Among them, is the global parameter, is the learning rate, N is the number of in - vehicle computer terminals, is the th gradient of the in - vehicle computer terminal, is the in - vehicle computer terminal data set, is the weight of the in - vehicle computer terminal, is the random noise.

[0012] Preferably, in step S3, for the optimization method selection and continuous optimization, the data monitoring model outputs the data drift abnormality as a performance index for judging the performance of the state of health (SOH) prediction model of the battery. According to the data drift abnormality, the update method of the SOH prediction model of the battery is selected. The central server records the number of times with a high degree of abnormality and the normal number of times. When the output result has a low abnormality degree, the model fine - tunes the local in - vehicle computer terminal parameters according to the input data; when the output result has a high abnormality degree, the central server records an abnormal number of times and calculates the F1 score. If the F1 score is lower than the threshold, all models are retrained using a new data set, and the global parameters are updated.

[0013] Preferably, the loss function formula of the data monitoring model is as follows: ; Wherein, is the loss function of the data monitoring model, is the reconstruction loss, is the true value of the th vehicle-mounted computer terminal, is the predicted value of the model for the th vehicle-mounted computer terminal, is the weight hyperparameter, is the feature combiner loss, is the th vehicle-mounted computer terminal passes through the feature extractor to obtain the feature representation, is the regularization term; The loss function formula of the battery state of health (SOH) prediction model is as follows: ; Wherein, is the total loss function, , , are the weights of each loss function, is the predicted value of the model for the charging time , is the predicted value of the model for the spatial position, is the true value of the battery health, is the predicted value of the battery health, , is the predicted value of the adjacent time step, H (·) is the solution function of the partial differential equation of the prediction of the output and input features of the model, M is the number of time steps, is the number of battery cycles; The calculation formula for feature extraction is as follows: ; Wherein, T is the output feature vector, is the start time for intercepting the charging curve, is the end time for intercepting the charging curve, is the current value, is the starting voltage value, is the ending voltage value, is the peak value of the IC curve, F (·) is the feature solution function.

[0014] The present invention also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method for optimizing the battery health state prediction model based on federated continuous learning are implemented.

[0015] Therefore, the present invention proposes a method and device for optimizing the battery health state prediction model based on federated continuous learning, and its beneficial effects are as follows: (1) The method for optimizing the battery health state prediction model based on federated continuous learning proposed by the present invention uses a federated learning architecture. Only the global parameters are uploaded at the in-vehicle computer end, and the local data does not leave the end, effectively protecting data privacy. At the same time, it reduces the data collection cost and communication volume, realizes the joint optimization of multi-end data for the model, and improves the data utilization efficiency.

[0016] (2) The method for optimizing the battery health state prediction model based on federated continuous learning proposed by the present invention, through a unique model structure design and training method, combined with multi-dimensional data collection and cleaning, enhances the model's prediction ability for battery health; the continuous learning mechanism enables the model to continuously optimize according to new data, improving accuracy and stability.

[0017] (3) The method and device for optimizing the battery health state prediction model based on federated continuous learning proposed by the present invention can dynamically adjust the model according to the performance prediction results, adapt to different battery state changes, and avoid catastrophic forgetting; the central server and the in-vehicle computer end work together, can be flexibly extended, adapt to different scales of vehicle data, and improve the overall adaptability and scalability of the system.

[0018] The technical solution of the present invention will be further described in detail below through the accompanying drawings and embodiments. Description of the Drawings

[0019] Figure 1 is the overall implementation flowchart of the method for optimizing the battery health state prediction model based on federated continuous learning of the present invention; Figure 2 is the flowchart of the federated continuous learning process; Figure 3 is the flowchart of the model optimization method in the present invention. Detailed Embodiments

[0020] To make the technical solution, advantages, and objectives of the present invention clearer, the technical solution of the embodiments of the present invention will be described clearly and completely below. The described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the described embodiments of the present invention without creative efforts belong to the protection scope of this application.

[0021] Unless otherwise defined, the technical terms or scientific terms used in this invention shall have the ordinary meanings as understood by those of ordinary skill in the field to which this invention belongs.

[0022] Embodiment 1 As Figure 1 shown, the present invention provides a method for optimizing a battery health state prediction model based on federated continuous learning. The specific steps are as follows: Step S1: Federated learning model training. The on-vehicle computer terminal and the central server build the same architecture of the battery health state (SOH) prediction model and data monitoring model. Each on-vehicle computer terminal collects and preprocesses the local battery charge and discharge data, trains the model locally, and then uploads the local model parameters to the central server. The central server integrates the parameters and distributes them back to each on-vehicle computer terminal, and repeats the training until the loss functions all converge; The structure of the battery health state (SOH) prediction model is as follows: The preprocessed data passes through a feature extraction layer containing multiple fully connected layers or convolutional layers to extract high-level features. Subsequently, a mapping layer composed of fully connected layers, with the help of a non-linear activation function, captures the complex relationships in the data, further transforms the feature data into specific feature representations and outputs the prediction results. Finally, the degradation simulation layer simulates the battery attenuation rate by adding physical constraints; The structure of the data monitoring model is as follows: The same data as the battery health state (SOH) prediction model is input, and the parameters of the feature extraction layer are copied to the fitting layer so that the two models have the same feature data output at the front end. The feature data outputs a feature vector through the mapping layer, and then the data is restored by the demapping layer to calculate the reconstruction error. Finally, the model performance is calculated through the Gaussian mixture model.

[0023] As Figure 2 shown, the specific steps of the federated learning model training are as follows: Step S11: The on-vehicle computer terminal downloads the current battery health state (SOH) prediction model and data monitoring model from the central server. The parameters include the number of hidden layers and the number of neurons H, the activation function, and decomposes the local model parameters into global shared parameters and specific parameters , the parameter matrix , is the bias term; The federated learning model parameters include global shared parameters and specific parameters . After the local training of the on-vehicle computer terminal is completed, the global shared parameters are uploaded to the central server, and the specific parameters remain local to adapt to the personalization of the on-vehicle computer terminal. The formula for the federated learning model parameters is as follows: ; Among them, represents the element-wise multiplication of each network unit, is a federal parameter, is a globally shared parameter, is a migration vector, is a specific parameter for the in-vehicle computer side.

[0024] Step S12: The in-vehicle computer side organizes and prepares the charging data stored locally, performs data cleaning and preprocessing, and divides it according to the ratio of 6:2:2 to determine the training set, validation set, and test set; The data collection step is: collect the data during battery charging, based on the in-vehicle terminal in the federated framework , record the monitoring data every 0.1 s, and the data form is time-series streaming data, that is, the input feature of the th in-vehicle terminal is , where R is the resistance of the in-vehicle terminal, is the voltage of the in-vehicle terminal, is the current of the in-vehicle terminal, S is the mean square error of the in-vehicle terminal, P k is the power of the in-vehicle terminal, D e is the slope of the in-vehicle terminal, is the charging time of the in-vehicle terminal, is the cumulative power of the in-vehicle terminal, is the curve slope of the in-vehicle terminal, is the temperature of the in-vehicle terminal; The in-vehicle computer side performs data cleaning and preprocessing on the local charge and discharge data based on the criterion, and at the same time performs data normalization. The formula is as follows: ; where is the interquartile range, and 0.7413 is a constant multiple of the interquartile range of is the processed data, is the original data point, is the median of the data set median.

[0025] Step S13: Use the training set data to train the downloaded model, calculate the model output through forward propagation, calculate the loss function for backpropagation, calculate the model parameter gradient, and use the gradient descent method to update the globally shared parameters of the model , specific parameter , bias term , and iterate multiple times until the loss functions of both models converge; Step S14: The in-vehicle computer side will update the globally shared parameters and the bias term Send them back to the central server through the homomorphic encryption algorithm, wait for the central server to aggregate all the uploaded model parameters, and distribute the new model parameters , and iterate for the next round of training until the loss functions of both models converge; among them, are the aggregated parameter matrix, global parameters, and bias term respectively.

[0026] The parameter update on the in-vehicle computer side is achieved through stochastic gradient descent. The central server aggregates all the uploaded model parameters in two parts, including batch training of local data on the in-vehicle computer side, calculating the gradient of each batch, accumulating gradients over multiple rounds, and sending them to the server for aggregation and adding random noise to protect privacy at one time; when retraining, the in-vehicle computer side determines the weight of the updated parameters according to the degree of anomaly of the outliers, increasing the weight when the degree of anomaly is large and decreasing the weight when the degree of anomaly is small; the calculation formula for the central server parameter aggregation is as follows: ; Among them, are the global parameters, is the learning rate, N is the number of in-vehicle computer sides, is the th gradient of the in-vehicle computer side, is the in-vehicle computer side 's dataset, is the weight of the in-vehicle computer side, is the random noise.

[0027] Step S2: Use the battery health state SOH prediction model to output the battery health degree, and output the data drift anomaly degree through the data monitoring model to evaluate the performance of the battery health state SOH prediction model; Step S3: Select the optimization method and continuously optimize. According to the data drift anomaly degree, select the optimization method of the battery health state SOH prediction model and continuously optimize according to the optimization method.

[0028] As Figure 3 shown, select the optimization method and continuously optimize. The data monitoring model outputs the data drift anomaly degree as the performance index for judging the battery health state SOH prediction model. According to the data drift anomaly degree, select the update method of the battery health state SOH prediction model. The central server records the number of times with a high degree of anomaly and the number of normal times. When the output result has a low anomaly degree, the model fine-tunes the parameters of the local in-vehicle computer side according to the input data; when the output result has a high anomaly degree, the central server records an anomaly count and calculates the F1 score. If the F1 score is lower than the threshold, retrain all models with a new dataset and update the global parameters.

[0029] The loss function formula of the data monitoring model is as follows: ; Among them, is the loss function of the data monitoring model, is the reconstruction loss, is the true value of the th in-vehicle computer terminal, is the predicted value of the model for the th in-vehicle computer terminal, is the weight hyperparameter, is the weight hyperparameter, is the feature combiner loss, is the th feature representation obtained by the in-vehicle computer terminal through the feature extractor , is the regularization term; The loss function formula of the battery state of health (SOH) prediction model is as follows: ; Among them, is the total loss function, , , are the weights of each loss function, is the predicted value of the model for the charging time , is the predicted value of the model for the spatial position, is the true value of the battery health, is the predicted value of the battery health, , is the predicted value of adjacent time steps, H (·) is the solution function of the partial differential equation of the output and input feature predictions of the model, M is the number of time steps, is the number of battery cycles; The calculation formula for feature extraction is as follows: ; Among them, T is the output feature vector, is the start time for intercepting the charging curve, is the end time for intercepting the charging curve, is the current value, is the starting voltage value, is the ending voltage value, is the peak value of the IC curve, F (·) is the feature solution function.

[0030] Embodiment 2 The present invention also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements a method for optimizing a battery health state prediction model based on federated continuous learning.

[0031] Therefore, the present invention provides a method and a device for optimizing a battery health state prediction model based on federated continuous learning, which realizes the joint optimization of multiple types of battery prediction models on multiple in-vehicle computer terminals, requires a smaller dataset, has a lower data communication volume, and effectively improves the universality of the model. At the same time, in the actual application process, the model can continuously self-optimize, continuously improve its performance, avoid catastrophic forgetting, further enhance the universality, achieve the lifelong learning of the model, and ensure its long-term stable and efficient operation.

[0032] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for optimizing a battery health status prediction model based on federated continuous learning, characterized in that: The specific steps are as follows: Step S1, federated learning model training. The on-board computer terminal and the central server build a battery health status SOH prediction model and data monitoring model with the same architecture. Each on-board computer terminal collects and pre-processes local battery charging and discharging data. After training the model locally, the local model parameters are uploaded to the central server, and the central server integrates the parameters and sends them back to each on-board computer terminal. The training is repeated until the loss function converges. Step S2: outputting the battery health degree using the battery health state SOH prediction model, and outputting the data drift abnormality degree through the data monitoring model to evaluate the performance of the battery health state SOH prediction model; Step S3: Select an optimization method and perform continuous optimization. According to the data drift abnormality, select an optimization method for the battery health status SOH prediction model and perform continuous optimization according to the optimization method.

2. The method for optimizing the battery health status prediction model based on federated continuous learning according to claim 1 is characterized in that: In step S1, the structure of the battery health state SOH prediction model is as follows: the preprocessed data is passed through a feature extraction layer including multiple fully connected layers or convolutional layers to extract high-level features, and then the mapping layer composed of fully connected layers uses nonlinear activation functions to capture the complex relationships in the data, and further converts the feature data into specific feature representations and outputs prediction results. Finally, the degradation simulation layer simulates the battery attenuation rate by adding physical constraints; the data monitoring model structure is as follows: the same data as the battery health state SOH prediction model is input and the feature extraction layer parameters are copied to the fitting layer so that the two models have the same feature data output at the front end. The feature data outputs a feature vector through the mapping layer, and then the demapping layer restores the data to calculate the reconstruction error, and finally the model performance is calculated through the Gaussian mixture model.

3. The method for optimizing the battery health status prediction model based on federated continuous learning according to claim 1 is characterized in that: In step S1, the specific steps of federated learning model training are as follows: Step S11: The onboard computer downloads the current battery health status SOH prediction model and data monitoring model from the central server. The parameters include the number of hidden layers and the number of neurons H, the activation function, and decomposes the local model parameters into global shared parameters. With specific parameters , parameter matrix , is the bias term; Step S12: The onboard computer prepares the locally stored charging data, cleans and preprocesses the data, and divides the data into a training set, a validation set, and a test set according to a ratio of 6:2:2; Step S13: Use the training set data to train the downloaded battery health status SOH prediction model and data monitoring model, calculate the model output through forward propagation, perform back propagation by calculating the loss function, calculate the model parameter gradient, and use the gradient descent method to update the global shared parameters of the model. , specific parameters , bias term , multiple iterations until both model loss functions converge; Step S14: The vehicle computer updates the global shared parameters and the bias term The data is sent back to the central server through the homomorphic encryption algorithm, and the central server aggregates all uploaded model parameters and sends new model parameters. , iterate the next round of training until both model loss functions converge; where, They are the aggregated parameter matrix, global parameters and bias terms respectively.

4. The method for optimizing the battery health status prediction model based on federated continuous learning according to claim 3 is characterized in that: In step S11, the federated learning model parameters include global shared parameters With specific parameters After the local training is completed on the vehicle computer, the global shared parameters are uploaded to the central server, specific parameters Staying local to adapt to the personalization of the on-board computer, the parameter formula of the federated learning model is as follows: ； in, Represents the multiplication of the corresponding elements of each network unit, is the federation parameter, is a globally shared parameter, is the migration vector, These are specific parameters for the vehicle computer.

5. The method for optimizing the battery health status prediction model based on federated continuous learning according to claim 3 is characterized in that: In step S12, the data collection step is: collecting data when the battery is charging, based on the vehicle terminal in the federal framework , the monitoring data is recorded every 0.1s, and the data is in the form of streaming data based on time series, that is, The input characteristics of the vehicle terminal are ,in, R is the resistance of the vehicle terminal, is the voltage of the vehicle terminal, is the current of the vehicle terminal, S is the mean square error of the vehicle terminal, P k is the power of the vehicle terminal, D e is the slope of the vehicle terminal, is the charging time of the vehicle terminal, is the accumulated power of the vehicle terminal, is the curve slope of the vehicle terminal, is the temperature of the vehicle terminal; The vehicle computer is based on The criterion cleans and preprocesses the local charging and discharging data, and normalizes the data at the same time. The formula is as follows: ； in, is the interquartile range, 0.7413 is the interquartile range A constant multiple of For the processed data, is the original data point, is the median of the values ​​in the data set.

6. The method for optimizing the battery health status prediction model based on federated continuous learning according to claim 3 is characterized in that: In step S14, the parameter update of the on-board computer is realized by stochastic gradient descent. The central server aggregates all uploaded model parameters in two parts, including training the local data in batches on the on-board computer, calculating the gradient of each batch, accumulating multiple rounds of gradients, and sending them to the server for aggregation at one time and adding random noise to protect privacy; during retraining, the on-board computer determines the weight of the updated parameter according to the abnormality of the abnormal value, and the weight is increased when the abnormality is large, and the weight is decreased when the abnormality is small; wherein, the calculation formula of the central server parameter aggregation is as follows: ； in, is a global parameter, is the learning rate, N is the number of onboard computers, For the The gradient of the on-board computer, For vehicle computer Datasets, is the weight of the onboard computer, is random noise.

7. The method for optimizing the battery health status prediction model based on federated continuous learning according to claim 1 is characterized in that: In step S3, the optimization method is selected and continuously optimized. The data monitoring model outputs the data drift abnormality as a performance indicator for judging the battery health state SOH prediction model. The update method of the battery health state SOH prediction model is selected according to the data drift abnormality. The central server records the number of high abnormalities and the normal number of times. When the output result abnormality is low, the model fine-tunes the parameters of the local on-board computer according to the input data; when the output result abnormality is high, the central server records the number of abnormalities once and calculates the F1 score. If the F1 score is lower than the threshold, all models are retrained using new data sets, and the global parameters are updated.

8. The method for optimizing the battery health status prediction model based on federated continuous learning according to claim 3 is characterized in that: The loss function formula of the data monitoring model is as follows: ； in, is the data monitoring model loss function, is the reconstruction loss, For the The real value of the on-board computer, For the model The predicted value of the on-board computer, is the weight hyperparameter, is the weight hyperparameter, is the feature combiner loss, For the The vehicle computer uses a feature extractor The obtained feature representation is is the regularization term; The loss function formula of the battery health status SOH prediction model is as follows: ； in, is the total loss function, , , is the weight of each loss function, Charging time for the model Predicted value, is the predicted value of the model space position, is the actual value of battery health, is the battery health prediction value, , is the predicted value of the adjacent time step, H (·) is the function that solves the partial differential equations that predict the output and input features of the model, M is the number of time steps, is the number of battery cycles; The calculation formula for feature extraction is as follows: ； in, T is the output feature vector, To capture the start time of the charging curve, To capture the end time of the charging curve, is the current value, is the starting voltage value, is the ending voltage value, is the peak value of the IC curve, F (·) is the characteristic solving function.

9. 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 for optimizing the battery health status prediction model based on federated continuous learning as described in any one of claims 1-8 are implemented.

Images