Lithium ion battery health state evaluation method and system based on time decomposition and feature fusion
Through the method of time decomposition and feature fusion, the problem of long-term time scale and local details identification in the health status of lithium-ion batteries is solved, and the accurate evaluation of the health status of lithium-ion batteries is achieved, which improves the accuracy and robustness of the evaluation.
Patent Information
- Application Number
- CN202411011474.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-07-26
AI Technical Summary
The existing lithium-ion battery health status evaluation methods are difficult to accurately process aging signals and local capacity recovery details on long-term scales, resulting in inaccurate evaluation of battery health status.
Using a method based on time decomposition and feature fusion, the voltage, current, temperature and capacity data of lithium-ion batteries are obtained, and the data is preprocessed and feature extraction is performed. The data is decomposed into trend terms and season terms, and the jump connection and MLP-Block are used for Information fusion combines inverse normalization and projection layers to map input sequences to output sequences to build a battery health status evaluation network.
It improves the long-term dependency recognition capability of the battery aging process, accurately identify the details of local capacity recovery during the aging process, improves the accuracy and robustness of battery health status assessment, and reduces the impact of noise and outliers.
Smart Images

Figure CN118897212A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of battery management, and in particular to a lithium-ion battery health status assessment method and system based on time decomposition and feature fusion. Background Art
[0002] As the core component of the current energy storage system, lithium-ion batteries play a vital role in many fields such as electric vehicles, portable electronic devices, photovoltaic energy storage systems, and backup power supplies. With the rapid development of new energy technologies and the expansion of their application scale, the stability and safety of battery performance have become key factors affecting the reliability of the entire system. Therefore, effective health status assessment and management of lithium-ion batteries has become an indispensable part of the battery power system. The state of health (SOH) of lithium-ion batteries is an important parameter for measuring battery performance and aging, which is directly related to the battery's service life, safety performance, and energy storage efficiency. The health status of the battery is affected by many factors, including the number of charge and discharge cycles, operating temperature, charge and discharge current, charge and discharge depth, etc. As the use time increases, the chemical composition, physical structure, and electrochemical properties of the battery will gradually change, resulting in a decrease in battery capacity, an increase in internal resistance, and even possible safety issues such as thermal runaway.
[0003] The current mainstream research methods mainly include experimental measurement methods, model-based methods and data-driven methods. The main experimental measurement methods are coulomb counting method and open circuit voltage method. The principle of coulomb counting method is simple and easy to use, but due to the inevitable current monitoring error, this method has the problem of error accumulation. The open circuit voltage method has the advantages of small calculation amount and no error accumulation, but obtaining the open circuit voltage of the battery requires a long time of static, which is difficult to apply in practice. Model-based methods mainly include equivalent circuit models, electrochemical models and incremental capacity analysis methods. These methods avoid the problems of long experimental measurement time and error accumulation, and obtain better SOH estimation results. However, since it calculates SOH through battery models, there are still problems such as high dependence on model accuracy, difficulty in model construction and solution process, and the generalization is relatively poor when dealing with a variety of battery types and usage scenarios.
[0004] In recent years, data-driven methods have been widely used in battery SOH prediction because they are independent of models, not restricted by any physical and chemical models, and have advantages such as high adaptability and flexibility. Traditional data-driven methods mainly include probability-based methods such as particle filters and Kalman filters, and machine learning-based methods such as support vector machines and decision trees. Weng et al. combined the data-driven method with the model-based method and proposed a support vector regression algorithm based on incremental capacity analysis for SOH estimation of lithium-ion batteries (Weng C, Cui Y, Sun J, et al. On-board state of health monitoring of lithium-ion batteries using incremental capacity analysis with support vector regression [J]. Journal of Power Sources, 2013, 235: 36-44.). Yang et al. proposed a Gaussian process regression model based on grey correlation analysis, using the parameters in the charging curve instead of the number of cycles as the input of the model, and achieved better SOH prediction results (Yang D, Zhang X, Pan R, et al. A novel gaussian process regression model for state-of-health estimation of lithium-ion battery using charging curve [J]. Journal of Power Sources, 2018, 384: 387-395). However, the traditional data-driven method is difficult to cope with the task of health status assessment of lithium-ion batteries over a long time scale due to the relatively simple model used for estimation.
[0005] Recently, with the help of the powerful learning ability of deep networks, data-driven methods based on deep learning have gradually become popular and have greatly improved the accuracy of battery SOH estimation. Kaur et al. used three different deep network models, FNN, CNN and LSTM, to estimate battery SOH, and explored the weight of different features for battery SOH estimation (Kaur K, Garg A, Cui X, et al. Deep learning networks for capacity estimation for monitoring SOH of Li-ion batteries for electric vehicles [J]. International Journal of Energy Research, 2021, 45 (2): 3113-3128.). Xu et al. proposed an improved CNN-LSTM network, which improves the prediction accuracy of SOH by adding residual connections to the model and removes useless features to reduce the computational complexity (Xu H, Wu L, Xiong S, et al. An improved CNN-LSTM model-based state-of-health estimation approach for lithium-ion batteries [J]. Energy, 2023, 276: 127585.). Li et al. proposed a SOH and RUL prediction framework called AST-LSTM, which obtains new and old information simultaneously by coupling the input and forget gates with fixed connections, and then multiplies the new input with the historical unit state unit by unit to filter out more useful information (Li P, Zhang Z, Xiong Q, et al. State-of-health estimation and remaining useful life prediction for the lithium-ion battery based on a variant long shortterm memory neural network [J]. Journal of Power Sources, 2020, 459: 228069.). However, the above method mainly estimates the battery SOH through deep networks such as LSTM and its variants. With the help of the LSTM network's ability to learn the time dependency in the features, it has achieved higher accuracy than other traditional SOH estimation methods. However, the LSTM network itself still has some room for improvement when processing long-term data, and is easily affected by noise and sudden fluctuations, resulting in error accumulation.
[0006] Compared with the LSTM series models, the Transformer model can achieve the learning of longer-period data features through the multi-head self-attention mechanism of parallel computing. Some scholars have begun to apply the Transformer model to battery SOH estimation. Gu et al. combined CNN with Transformer for battery SOH estimation, and used principal component analysis to eliminate redundant features to reduce the model calculation burden. (Gu X, See KW, Li P, et al. A novel state-of-health estimation for the lithium-ion battery using a convolutional neural network and transformer model [J]. Energy, 2023, 262: 125501.). Bai et al. proposed a SOH estimation method based on a multi-view information perception framework (MVIP-Trans) based on convolutional transformation, which enhances valuable information and suppresses useless noise through a parallel multi-scale attention local information sensor. (Bai T, Wang H. Convolutional Transformer-Based Multiview Information Perception Framework for Lithium-Ion Battery State-of-Health Estimation[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 1-12.). Although the Transformer model can process long-term series of battery data, parallel computing and self-attention mechanisms also cause it to require more computing resources, and the self-attention mechanism inevitably leads to local information loss due to its arrangement-invariant nature. Therefore, how to simultaneously consider the long-term dependence of the battery aging process and accurately identify the local capacity recovery details during the aging process is crucial to accurately assess the health status of lithium-ion batteries. Summary of the invention
[0007] Therefore, the purpose of the present invention is to provide a lithium-ion battery health status assessment method and system based on time decomposition and feature fusion, which can process battery aging signals on a long time scale while solving the problem that the details of local capacity recovery during battery aging are difficult to accurately predict, thereby achieving accurate assessment of the health status of lithium-ion batteries and ensuring the safe operation of lithium-ion batteries.
[0008] In order to achieve the above object, the present invention provides a lithium-ion battery health status assessment method based on time decomposition and feature fusion, comprising the following steps:
[0009] S1. Obtain the voltage, current, temperature and capacity data of the lithium-ion battery aging process; perform preprocessing and extract key features to form a battery aging data set;
[0010] S2. Serialize the data in the battery aging dataset, and divide the sequence into trend items and seasonal items based on time decomposition; input the trend items and seasonal items as two components into the battery health status assessment network;
[0011] S3, exchange the time and channel dimensions of the sequence, and use skip connections to fuse the time dimension information and channel feature information;
[0012] The fused data is denormalized and then connected to the projection layer consisting of a single linear layer to map the input sequence to the output sequence;
[0013] S4. Acquire real-time aging data of lithium-ion batteries, iteratively train a lithium-ion battery health status assessment network, and use the trained network to predict the health status of lithium-ion batteries.
[0014] Further preferably, in S1, the voltage, current, temperature and capacity data of the lithium-ion battery aging process are obtained and preprocessed, including:
[0015] Select the time for the battery voltage to drop to the discharge threshold voltage, the voltage slope, the time required for the temperature to reach the maximum value, the temperature slope, the maximum temperature, the constant voltage charging time, the constant current charging time and the internal resistance as input characteristics;
[0016] Calculate the current battery health status based on the battery capacity data, and use the battery health status value as the label of the data set;
[0017] The input features and the data set labels are processed for outliers to form a battery aging data set.
[0018] Further preferably, the preprocessing also includes removing outliers from the acquired data based on the 3-sigma principle before extracting the input features, and standardizing the data so that it has zero mean and variance.
[0019] Further preferably, in S2, the data in the battery aging data set is serialized, and the sequence is divided into trend items and seasonal items based on time decomposition; including:
[0020] Divide the battery aging dataset into multiple data series Where X represents an input sequence with a length of and a number of channels (features) of A collection of;
[0021] For any input sequence at any time Average pooling is used to eliminate periodic fluctuations and highlight long-term trends, while padding operations are used to keep the length of the input sequence unchanged. The specific formula is:
[0022]
[0023] in, are the trend term and seasonal term obtained by decomposition respectively; Express Perform a padding operation, which represents average pooling.
[0024] Further preferably, the step of inputting the trend term and the season term as two components into the battery health status assessment network comprises respectively inputting the two decomposed components into the MLP-Block of the lithium-ion battery health status assessment network for time feature extraction, wherein the MLP-Block comprises two fully connected layers and is expressed by the following formula:
[0025]
[0026] Among them, are the weights and biases corresponding to the trend term MLP-Block, are the weights and biases corresponding to the seasonal term MLP-Block, It is the output of the time decomposition module.
[0027] Further preferably, the activation function GELU is expressed by the following formula (4):
[0028]
[0029] Among them, represents the input of the network at the current layer.
[0030] Further preferably, S2 also includes serializing the data in the battery aging data set, using instance normalization to remove non-stationary information, and obtaining a stationary sequence and non-stationary attribute parameters and affine transformation parameters, specifically including the following steps:
[0031] The battery aging data set is divided into multiple data sequences, the mean and standard deviation of each data sequence are calculated, and the instances in each sequence are separated by the mean and standard deviation. Perform normalization processing;
[0032]
[0033] in, is the affine parameter vector, is the normalized instance, In order to obtain the average value and standard deviation of the data, a fixed minimum value is used to prevent the denominator from being 0.
[0034] Further preferably, in S3, the fused data is inversely normalized and then connected to the projection layer consisting of a single linear layer; wherein the fused data is inversely normalized, and the fused result is combined with the normalization-related parameters of S2 and then inverse instance normalization is performed, including:
[0035] In the battery health status assessment network, the symmetrical network output layer outputs Perform a denormalization operation; explicitly remove non-stationary properties from the input data And the affine transformation parameters learned by the network are returned to the model output, that is,
[0036]
[0037] Will replace as the output of the network.
[0038] The present invention also provides a lithium-ion battery health status assessment system based on time decomposition and feature fusion, which is used to implement the steps of the lithium-ion battery health status assessment method based on time decomposition and feature fusion, including a data acquisition module, a data processing module and a battery health status assessment network;
[0039] The data acquisition module is used to acquire the voltage, current, temperature and capacity data of the lithium-ion battery aging process; perform preprocessing and extract key features to form a battery aging data set;
[0040] The data processing module is used to serialize the data in the battery aging data set, and divide the sequence into trend items and seasonal items based on time decomposition; and input the trend items and seasonal items as two components into the battery health status assessment network;
[0041] The battery health status assessment network includes a time decomposition module and a channel feature fusion device; the time decomposition module is used to exchange the time and channel dimensions of the sequence, and the channel feature fusion device fuses the time dimension information and the channel feature information by skip connection; the fused data is inversely normalized and then connected to a projection layer composed of a single linear layer to map the input sequence to the output sequence;
[0042] Acquire real-time aging data of lithium-ion batteries, iteratively train the lithium-ion battery health status assessment network, and use the trained network to predict the health status of lithium-ion batteries.
[0043] Compared with the prior art, the lithium-ion battery health status assessment method and system based on time decomposition and feature fusion disclosed in this application have the following beneficial effects:
[0044] 1. The present invention processes and extracts features of the battery aging process data, extracts features that can be used to estimate the battery health status from different angles such as voltage, current, temperature, and time, and reduces the interference of abnormal value data and noise;
[0045] 2. The time decomposition module perceives global and local information from a time perspective, which can solve the long-term dependence problem of the battery aging process and accurately identify the local capacity recovery details during the aging process, thereby improving the prediction accuracy. Subsequently, the common influence of different features on the aging data is further integrated from the perspective of channel features. Compared with the use of a single feature, the robustness and generalization of battery health status assessment are improved. Finally, by adding reversible instance normalization at both the input and output ends, the non-stationary problem of the statistical attributes of the data changing over time during the battery aging process is alleviated, the distribution differences between the data are reduced, and the accuracy of the health assessment is further improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 It is a flowchart of the steps of the lithium-ion battery health status assessment method based on time decomposition and feature fusion of the present invention.
[0047] Figure 2 Battery voltage, current, temperature and capacity curves during battery aging.
[0048] Figure 3 Comparison of multi-method health status assessment results for batteries with different numbers in the NASA dataset.
[0049] Figure 4 Comparison of multi-method health status assessment results for batteries with different numbers in the CALCE dataset.
[0050] Figure 5 Comparison of error distribution of multi-method health status assessment for batteries with different numbers in the NASA dataset.
[0051] Figure 6 Comparison of error distribution of multi-method health status assessment for batteries with different numbers in the CALCE dataset. DETAILED DESCRIPTION
[0052] The present invention is further described in detail below through the accompanying drawings and specific embodiments.
[0053] like Figure 1 As shown, a lithium-ion battery health status assessment method based on time decomposition and feature fusion provided by an embodiment of one aspect of the present invention includes the following steps:
[0054] S1. Obtain the voltage, current, temperature and capacity data of the lithium-ion battery aging process; perform preprocessing and extract key features to form a battery aging data set;
[0055] The data is preprocessed by first removing outliers based on the 3-sigma principle, then the data is standardized to have zero mean and variance, and then the aging characteristics are extracted based on this.
[0056] Extract the battery voltage, current, temperature and capacity data of the lithium-ion battery during charging and discharging during the entire battery life cycle. The entire charging and discharging cycle includes multiple charging and discharging cycles. Each cycle contains a sequence of battery voltage, current and temperature changes over time. The capacity is the battery capacity value measured after the discharge is completed. The time for the battery voltage to drop to the discharge threshold voltage, the voltage slope, the time required for the temperature to reach the maximum value, the temperature slope, the maximum temperature, the constant voltage charging time, the constant current charging time and the internal resistance are selected as input features. Calculate the current battery health status based on the battery capacity data, and use the battery health status value as the label of the data set. Combined with the obtained input features, remove outliers based on the 3-sigma principle, and standardize the data to have zero mean and variance to form a battery aging data set that can be used for network training. The battery health status is defined by the capacity of the battery's current cycle. The specific formula is:
[0057]
[0058] Among them, and are the capacity and nominal capacity of the battery in the current state, respectively.
[0059] In the preferred embodiment of the present invention, the application object comes from the accelerated aging experiment of lithium-ion batteries conducted by NASA Ames Center of Excellence in Prediction and the Center for Advanced Lifecycle Engineering (CALCE) of the University of Maryland. Two groups of experiments are set up, and the experimental data sets are NASA data sets and CALCE data sets. Each group of experiments has four batteries. The batteries in the NASA data set are numbered B05, B06, B07, and B18, and the batteries in the CALCE data set are numbered CS2_35, CS2_36, CS2_37, and CS2_38. The battery voltage, current, temperature, and capacity curves during battery aging are as follows: Figure 2As shown, the colors of all curves gradually deepen as the number of battery cycles increases, that is, the worse the battery health state (SOH), the darker the color. In the capacity-cycle number scatter plot, the capacity data at each cycle number corresponds to the voltage, current and temperature change curves of the same color. It can be seen from the figure that with the increase in the number of battery cycles, the overall capacity of the battery shows a downward trend, and there is a local capacity recovery phenomenon. At the same time, as the degree of aging deepens, the time for the battery to reach the discharge threshold voltage during discharge gradually decreases, the time change rate gradually increases, and the maximum temperature and the time when the temperature reaches the peak value also gradually change.
[0060] S2. Serialize the data in the battery aging dataset, and divide the sequence into trend items and seasonal items based on time decomposition; input the trend items and seasonal items as two components into the battery health status assessment network;
[0061] The details are as follows:
[0062] Divide the dataset into multiple input sequences as input to the lithium-ion battery health status assessment network That is, it contains an input sequence of length and number of channels (features) A collection of, for a certain moment of input sequence Average pooling is used to eliminate periodic fluctuations and highlight long-term trends, while padding operations are used to keep the length of the input sequence unchanged. The specific formula is:
[0063]
[0064] in The trend term and seasonal term are decomposed, and the overall aging trend of the battery is learned from the trend term, and the local fluctuation characteristics and capacity recovery phenomenon of the battery during the aging process are learned from the seasonal term. Finally, the two decomposed components are input into the lithium-ion battery health status assessment network respectively.
[0065] Among them, the time feature extraction is performed in the MLP-Block of the battery health status assessment network. The MLP-Block consists of two fully connected layers and a GELU activation function, namely
[0066]
[0067] Among them, are the weights and biases corresponding to the trend term MLP-Block, are the weights and biases corresponding to the seasonal term MLP-Block, is the output of the time decomposition module, and the activation function GELU is defined as
[0068]
[0069] In the preferred embodiment of the present invention, the sliding window of average pooling is 25, and the number of hidden neurons in the MLP-Block of trend term and seasonal term is 64. The default length of the input sequence is 96. It is worth noting that the above parameters are non-restrictive and preferred choices and can be adjusted according to the number of samples in the data set. The above time decomposition and feature extraction are both operations on the time dimension of the sequence, that is, the sequence The subsequent channel feature fusion operates on the channel dimension of the sequence, that is, the sequence Columns.
[0070] S3, exchange the time and channel dimensions of the sequence, and use skip connections to fuse the time dimension information and channel feature information;
[0071] The fused data is denormalized and then connected to the projection layer consisting of a single linear layer to map the input sequence to the output sequence;
[0072] The result of the time dimension information perception of the time decomposition module is jump-connected with the input of the time decomposition module, and the result of the jump connection is normalized through the layer. Then the time and channel dimensions of the sequence are exchanged, and the channel feature fusion device composed of a single MLP-Block is used to learn the relationship between multiple different aging features of the battery in the spatial dimension. The learning result of the channel feature fusion device is jump-connected with the learning result of the time decomposition module to fuse the perceived time dimension information and channel feature information. Finally, after the fusion result is further normalized through the layer, it is subsequently connected to a projection layer composed of a single linear layer to complete the mapping learning from the input sequence to the output sequence.
[0073] In the preferred embodiment of the present invention, the MLP-Block structure of the channel feature fusion device is the same as that in step 2, but the number of hidden layer neurons is different. The number of hidden layer neurons in the MLP-Block of the channel feature fusion device is 32. At the same time, the entire encoder module that integrates the time dimension information and the channel feature information is not a single one, but is composed of four identical structures in parallel. The weights obtained by training each individual encoder module are different. This parallel structure can learn the differences in the battery aging process from different angles and improve the accuracy of the prediction.
[0074] It should be noted that the instance normalization in step 2 and the inverse instance normalization in step 3 are as follows:
[0075] In step 2, for an input sequence at a time To solve the corresponding prediction sequence distribution problem, where the input sequence length and the network prediction length can be different.
[0076] First, the input data is normalized using its instance mean and standard deviation. Every instance of The mean and standard deviation of
[0077]
[0078] Then use these statistics to Normalized to
[0079]
[0080] in, is the network learnable affine parameter vector, is a normalized instance. The sequence of It has more consistent means and variances at different times, and reduces the differences between different data distributions by reducing non-stationary information.
[0081] In step 3, the output Perform a denormalization operation to explicitly remove non-stationary attributes from the input data And the affine transformation parameters learned by the network are returned to the model output, that is,
[0082]
[0083] Will replace As the output of the network, it effectively alleviates the distribution difference of the data while ensuring that the original distribution information is not lost, so that the battery health status assessment network can better learn the mapping relationship between the battery aging sequence and the health status sequence, and improve the accuracy of the battery health status assessment by correcting the model prediction error caused by the difference in battery data distribution.
[0084] S4. Acquire real-time aging data of lithium-ion batteries, iteratively train a lithium-ion battery health status assessment network, and use the trained network to predict the health status of lithium-ion batteries.
[0085] The battery health status assessment network selects 70% of the data as the training set and 30% of the data as the test set. MSE is selected as the loss function and optimized using the Adam optimizer. The parameters of Adam and are the default values of 0.9 and 0.999 respectively. The initial learning rate is set to 0.005 and halved at each epoch. The batch size is set to 16, and the early stopping strategy is used to prevent overfitting. The input sequence length of the model is 96, the prediction sequence length is 3, the hyperparameters are optimized using grid search, and all data are Z-Score standardized. The battery health status assessment network iterates 50 epochs in total, compares the predicted battery health status value with the actual battery health status value, calculates the error loss, and updates the network parameters according to the loss value. Finally, the trained network is used to predict the health status of the lithium-ion battery.
[0086] It is worth noting that in the preferred embodiment of the present invention, the network training method and the parameter configuration of the learning rate optimization strategy are a non-restrictive and preferred choice. Those skilled in the art can select the network training method and parameter configuration based on various indicators such as the effect and accuracy of health assessment.
[0087] In a preferred embodiment of the present invention, the early stopping strategy is determined by the validation set loss. When the validation set loss does not decrease for 10 consecutive epochs, it is determined that the network has been overfitted, and the network with the smallest validation set loss is saved as the optimal network for prediction.
[0088] The present invention also provides a lithium-ion battery health status assessment system based on time decomposition and feature fusion, which is used to implement the steps of the lithium-ion battery health status assessment method of the above claim, including a data acquisition module, a data processing module and a battery health status assessment network;
[0089] The data acquisition module is used to acquire the voltage, current, temperature and capacity data of the lithium-ion battery aging process; perform preprocessing and extract key features to form a battery aging data set;
[0090] The data processing module is used to serialize the data in the battery aging data set, and divide the sequence into trend items and seasonal items based on time decomposition; and input the trend items and seasonal items as two components into the battery health status assessment network;
[0091] The battery health status assessment network includes a time decomposition module and a channel feature fusion device; the time decomposition module is used to exchange the time and channel dimensions of the sequence, and the channel feature fusion device fuses the time dimension information and the channel feature information by skip connection; the fused data is inversely normalized and then connected to a projection layer composed of a single linear layer to map the input sequence to the output sequence;
[0092] Acquire real-time aging data of lithium-ion batteries, iteratively train the lithium-ion battery health status assessment network, and use the trained network to predict the health status of lithium-ion batteries.
[0093] In the preferred embodiment of the present invention, the software configuration for model building and testing is python3.9, pytorch2.1.1, cuda12.1, and visual studio code2023; the hardware configuration is inter core i5-8300H 8-core 16-thread CPU, Nvidia GTX-1050Ti 4GB GPU, and 16G memory.
[0094] The lithium-ion battery health status assessment network obtained in the preferred embodiment of the present invention is tested using the test set data divided by the NASA data set and the CALCE data set. In order to verify the effectiveness of the method proposed in the present invention, the method is compared with other existing mainstream methods such as CNN-LSTM, CNN-Transformer, and MVIP-Trans. The experimental results of the NASA data set are as follows: Figure 3 As shown, the experimental results of the CALCE dataset Figure 4 shown.
[0095] from Figure 3 and Figure 4 It can be seen that in the battery experiments with different numbers, the proposed method is better than other mainstream methods, which shows that the proposed method can be well adapted to different battery health status assessment tasks. Figure 3 In the NASA dataset, this method can well predict the details of local capacity recovery and avoid the overall battery aging trend caused by local mutations. However, CNN-LSTM and CNN-Transformer cannot accurately predict the details of capacity recovery when performing battery health assessment, resulting in error accumulation, which leads to large fluctuations at 150-170 cycles for batteries with numbers B05 and B06. For the CALCE dataset with more dense data, such as Figure 4As shown, the local details of the aging process are richer. The traditional CNN-LSTM model can only determine the general aging trend of the battery. Although the Transformer-based CNN-Transformer and MVIP-Trans methods can roughly predict the local change details, the interference between position encoding and CNN in capturing time dependencies leads to inaccurate prediction of local changes. Compared with other existing methods, the method proposed in the present invention benefits from its time decomposition and feature fusion architecture, which can learn trends, local details and different features as a whole, and thus achieve better prediction results in battery health status assessment tasks over long time scales.
[0096] In the preferred embodiment of the present invention, mean square error (MSE), mean absolute error (MAE), root mean square error (RMSE) and determination coefficient R 2 The accuracy of the prediction of the health status of lithium-ion batteries is evaluated as an evaluation index, and the results are shown in Tables 1 and 2.
[0097] Table 1 Battery health status evaluation results of NASA dataset
[0098]
[0099]
[0100] As can be seen from Table 1, the battery health status assessment network of this embodiment is tested on the NASA data set. The MSE, MAE, RMSE and R 2 The average values are 0.00003, 0.0039, 0.0047 and 0.9859. Compared with other existing evaluation methods, the average MAE and RMSE of this method are 27.9%, 24.5%, 29.3% and 30.1%, 22.5%, 28.8% of CNN-LSTM, CNN-Transformer and MVIP-Trans. As shown in Table 2, the battery health status evaluation network of this embodiment is tested on the CALCE dataset. The MSE, MAE, RMSE and R 2 The average values are 0.0038, 0.0345, 0.0617 and 0.9956. Compared with other existing evaluation methods, the average MAE and RMSE of this method are 27.4%, 26.2%, 38.2% and 37.3%, 32.0%, 50.0% of CNN-LSTM, CNN-Transformer and MVIP-Trans. The results show that the health status assessment error of the method proposed in the present invention is lower, which is a significant improvement compared with the existing methods, and has a good effect on the global trend and local details of the battery aging process on a long time scale.
[0101] Table 2 Battery health status evaluation results of CALCE dataset
[0102]
[0103] In order to more intuitively compare the error distribution predicted by the method proposed in the present invention with that of other existing methods, Figure 5 and Figure 6 The distribution plots the absolute error box plots of different methods under different numbered batteries in NASA and CALCE data sets, where the black horizontal line is the median error, the upper and lower boundaries of the gray frame line are the first quartile and the third quartile respectively, the black boundary is the maximum and minimum values except for outliers, and the circles are outliers. It can be seen from the figure that compared with other methods, the median error of the method proposed in the present invention is almost the lowest. Although there are a small number of outliers like other methods, the error range is more concentrated and the overall error distribution range is lower. This proves that the method we proposed can adapt to the task of lithium-ion battery health status assessment on a long time scale, has a smaller prediction fluctuation under different battery cycles, and has better robustness and accuracy.
[0104] Obviously, the above embodiments are only examples for clear explanation, and are not intended to limit the implementation methods. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from them are still within the protection scope of the invention.
Claims
1. A lithium-ion battery health status assessment method based on time decomposition and feature fusion, It is characterized in that The following steps are involved: S1. Obtaining voltage, current, temperature and capacity data of lithium-ion battery aging process; Perform preprocessing and extract key features to form a battery aging data set; S2. Serialize the data in the battery aging dataset, and divide the sequence into trend items and seasonal items based on time decomposition; input the trend items and seasonal items as two components into the battery health status assessment network; S3, exchange the time and channel dimensions of the sequence, and use skip connections to fuse the time dimension information and channel feature information; The fused data is denormalized and then connected to the projection layer consisting of a single linear layer to map the input sequence to the output sequence; S4. Acquire real-time aging data of lithium-ion batteries, iteratively train a lithium-ion battery health status assessment network, and use the trained network to predict the health status of lithium-ion batteries.
2. According to claim 1, a lithium-ion battery health status assessment method based on time decomposition and feature fusion, It is characterized in that In S1, the voltage, current, temperature and capacity data of the lithium-ion battery aging process are obtained and pre-processed, including: Select the time for the battery voltage to drop to the discharge threshold voltage, the voltage slope, the time required for the temperature to reach the maximum value, the temperature slope, the maximum temperature, the constant voltage charging time, the constant current charging time and the internal resistance as input characteristics; Calculate the current battery health status based on the battery capacity data, and use the battery health status value as the label of the data set; The input features and the data set labels are processed for outliers to form a battery aging data set.
3. According to claim 1, a lithium-ion battery health status assessment method based on time decomposition and feature fusion, It is characterized in that The preprocessing also includes removing outliers from the acquired data based on the 3-sigma principle and standardizing the data to have zero mean and variance before extracting the input features.
4. According to claim 1, a lithium-ion battery health status assessment method based on time decomposition and feature fusion, It is characterized in that In S2, the data in the battery aging dataset is serialized, and the series are divided into trend items and seasonal items based on time decomposition; including: Divide the battery aging dataset into multiple data series Where X represents an input sequence with N length L and C number of channels A collection of; For any input sequence at any time Average pooling is used to eliminate periodic fluctuations and highlight long-term trends, while padding operations are used to keep the length of the input sequence unchanged. The specific formula is: in are the trend term and seasonal term obtained by decomposition respectively; Express Perform padding operation, AvgPool(·) represents average pooling.
5. According to claim 1, a lithium-ion battery health status assessment method based on time decomposition and feature fusion, It is characterized in that The trend term and the season term are input as two components into the battery health status assessment network, including respectively inputting the two decomposed components into the MLP-Block of the lithium-ion battery health status assessment network for time feature extraction. The MLP-Block includes two fully connected layers, which are expressed by the following formula: Among them, W 1 , W 2 , b 1 , b 2 is the corresponding weight and bias of the trend term MLP-Block, W 3 , W 4 , b 3 , b 4 are the weights and biases corresponding to the seasonal term MLP-Block, It is the output of the time decomposition module.
6. A lithium-ion battery health status assessment method based on time decomposition and feature fusion according to claim 1, It is characterized in that The activation function GELU is also expressed by the following formula (4): Among them, x represents the input of the network at the current layer.
7. A lithium-ion battery health status assessment method based on time decomposition and feature fusion according to claim 1, It is characterized in that S2 also includes serializing the data in the battery aging data set, using instance normalization to remove non-stationary information, and obtaining stationary sequences and non-stationary attribute parameters and affine transformation parameters, which specifically includes the following steps: The battery aging data set is divided into multiple data sequences, the mean and standard deviation of each data sequence are calculated, and the instances in each sequence are separated by the mean and standard deviation. Perform normalization processing; in, is the affine parameter vector, is the normalized instance, To obtain the average value of the data, Var[·] is to obtain the standard deviation of the data, and ∈ is a fixed minimum value to prevent the denominator from being zero.
8. A lithium-ion battery health status assessment method based on time decomposition and feature fusion according to claim 7, It is characterized in that In S3, the fused data is inverse normalized and then connected to a projection layer consisting of a single linear layer; The fused data is denormalized, which is to combine the fused result with the S2 normalization related parameters and then perform denormalization, including: In the battery health status assessment network, the symmetrical network output layer outputs Perform a denormalization operation; explicitly remove non-stationary properties from the input data And the affine transformation parameters γ learned by the network c , β c Returning to the model output, Will replace as the output of the network.
9. A lithium-ion battery health status assessment system based on time decomposition and feature fusion, It is characterized in that The steps for implementing any one of the above-mentioned lithium-ion battery health status assessment methods based on time decomposition and feature fusion in claims 1-8 include a data acquisition module, a data processing module and a battery health status assessment network The data acquisition module is used to acquire the voltage, current, temperature and capacity data of the lithium-ion battery during the aging process; Perform preprocessing and extract key features to form a battery aging data set; The data processing module is used to serialize the data in the battery aging data set, and divide the sequence into trend items and seasonal items based on time decomposition; and input the trend items and seasonal items as two components into the battery health status assessment network; The battery health status assessment network includes a time decomposition module and a channel feature fusion module; The time decomposition module is used to exchange the time and channel dimensions of the sequence, and the channel feature fuser fuses the time dimension information and the channel feature information by means of skip connection; the fused data is inversely normalized and then connected to the projection layer composed of a single linear layer to map the input sequence to the output sequence; Acquire real-time aging data of lithium-ion batteries, iteratively train the lithium-ion battery health status assessment network, and use the trained network to predict the health status of lithium-ion batteries.
Citation Information
Patent Citations
Online prediction method and system for state of charge of lithium ion battery
CN115407207A
Attention and multivariate graph convolution fusion space-time traffic speed prediction method and system
CN116504075A
Prediction method based on irregular time series data
CN118194242A
Cited By
Time-frequency characteristic deep fusion method for secondary power supply of high-temperature aging test equipment
CN119903328A
SOC-SOH joint estimation method based on double-cross physical guidance framework
CN120142959A
Battery life prediction method fusing transfer learning and long short-term memory network
CN120254680A
Lithium battery health state assessment method and system
CN120334784A
Lithium ion battery health state evaluation method and equipment based on deconstruction physical information, medium and product
CN120610170A
