Lithium battery life prediction method and system, electronic equipment and product

Abstract

The invention belongs to the technical field of battery management, and aims to provide a lithium battery life prediction method and system, electronic equipment and a product. According to the method, the target data type with the maximum association degree with the lithium battery aging characteristics is extracted in advance, so that the characteristic data highly related to the lithium battery aging can be effectively identified, the model is ensured to use the data with the most information amount during training and prediction, and the service life prediction precision of the target lithium battery can be improved; in the process of predicting the service life of a target lithium battery based on the trained lithium battery service life prediction model, target working data of a specified type of target lithium battery is collected in real time, and feature construction processing can be performed through physical model target working data in the trained lithium battery service life prediction model to obtain reconstructed feature data. And then comprehensive prediction is carried out on the target working data and the reconstructed feature data in combination with a deep learning model, so that the prediction accuracy and robustness can be further improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery management, and specifically relates to a lithium battery life prediction method, system, electronic equipment and product. Background Art

[0002] With the rapid development of technologies such as new energy vehicles and portable electronic devices, lithium batteries have been widely used as a high-efficiency and long-life energy storage device. However, during long-term use, lithium batteries will experience performance degradation due to complex internal reactions and external interference, which is manifested as reduced capacity, increased internal resistance, and reduced charging and discharging efficiency, which eventually leads to the end of battery life. Reliable prediction results of the remaining useful life of lithium batteries can be used to improve battery management and optimize battery maintenance strategies. Therefore, accurately estimating the remaining useful life (RUL) of lithium batteries is crucial to ensuring the safety and reliability of equipment.

[0003] In the prior art, life prediction of lithium batteries mainly relies on the following two methods: data-driven method and physical model method. Among them, the data-driven method analyzes a large amount of historical data and uses machine learning or deep learning algorithms to build a prediction model, which can capture complex nonlinear relationships and has strong adaptability. The physical model method is based on a deep understanding of the internal chemical reactions and electrochemical characteristics of lithium batteries, and establishes a mathematical model to provide an interpretable description of the internal state of the battery, such as charging and discharging behavior and aging process.

[0004] However, in the process of using the prior art, the inventors found that the prior art has at least the following problems: Data-driven methods rely heavily on large-scale and high-quality historical data sets. At the same time, pure data-driven methods cannot effectively utilize the physical and chemical knowledge of batteries, resulting in limited prediction accuracy and generalization capabilities of the model. Single physical model methods often rely on specific experimental conditions and are difficult to accurately characterize battery performance under different working conditions in actual applications. In addition, physical models are difficult to cover all influencing factors during battery use, resulting in limited prediction accuracy under complex working conditions. Summary of the invention

[0005] The present invention aims to solve the above technical problems at least to a certain extent. The present invention provides a lithium battery life prediction method, system, electronic equipment and product.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a lithium battery life prediction method, comprising: Obtaining historical aging characteristic time series data and a historical sample time series data set of a specified type of sample lithium battery in a specified historical period; wherein the historical sample time series data set includes a plurality of sub-historical sample time series data of different data types; Respectively obtaining the correlation between each sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data; The data type corresponding to the sub-historical sample time series data whose correlation is greater than a preset correlation threshold value among the multiple sub-historical sample time series data is used as the target data type; Constructing an initial lithium battery life prediction model, and using the sub-historical sample time series data of the target data type and the historical aging feature time series data to train the initial lithium battery life prediction model to obtain a trained lithium battery life prediction model; wherein the trained lithium battery life prediction model includes a physical model and a deep learning model connected in sequence; Collecting target operating data of the target lithium battery of the specified type in real time; wherein the data type of the target operating data is the target data type; The target operating data is input into the trained lithium battery life prediction model so as to obtain the life prediction result of the target lithium battery based on the trained lithium battery life prediction model.

[0007] In one possible design, the historical aging characteristic time series data adopts historical capacity time series data.

[0008] In a possible design, obtaining the correlation between any sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data includes: Performing data alignment processing on the historical aging feature time series data and any sub-historical sample time series data to obtain aligned historical aging feature time series data and any aligned sub-historical sample time series data; Performing standardization processing on the aligned historical aging feature time series data and any aligned sub-historical sample time series data respectively to obtain the standardized historical aging feature time series data and any standardized sub-historical sample time series data; Taking the standardized historical aging feature time series data as a reference sequence, taking any standardized sub-historical sample time series data as a comparison sequence, and respectively obtaining the correlation coefficient between the reference sequence and each historical sample data in the comparison sequence; The degree of association between the reference sequence and the comparison sequence is obtained according to the association coefficient between each historical sample data in the reference sequence and the comparison sequence.

[0009] In one possible design, the reference sequence and the comparison sequence are obtained. i The correlation coefficient between historical sample data includes: Obtaining an absolute difference value between each historical aging feature in the reference sequence and each historical sample data corresponding to each historical aging feature in the comparison sequence; Obtain the maximum absolute difference value and the minimum absolute difference value among the absolute difference values; Based on the comparison sequence i The historical sample data and the reference sequence i The absolute difference value between the historical aging characteristics, the maximum absolute difference value and the minimum absolute difference value are obtained by comparing the reference sequence with the comparison sequence. i The correlation coefficient between historical sample data; wherein the correlation coefficient is: ; In the formula, is the minimum absolute difference value; is the maximum absolute difference value; is the absolute difference between the i-th historical sample data in the comparison sequence and the i-th historical aging feature in the reference sequence; is the preset resolution factor.

[0010] In one possible design, the degree of association between the reference sequence and the comparison sequence is: ; In the formula, I is the total number of historical sample data in the comparison sequence.

[0011] In a possible design, the initial lithium battery life prediction model is trained using the sub-historical sample time series data of the target data type and the historical aging feature time series data to obtain a trained lithium battery life prediction model, including: Initializing the parameters of the initial lithium battery life prediction model; Propagate the sub-historical sample time series data of the target data type forward through the initial lithium battery life prediction model to obtain the life prediction time series data of the sample lithium battery; Calculating the error between the life prediction time series data and the historical aging characteristic time series data according to the joint loss function of the initial lithium battery life prediction model; The back-propagation algorithm is used to calculate the gradient of the joint loss function relative to the parameters of the initial lithium battery life prediction model, and the parameters of the initial lithium battery life prediction model are updated according to the gradient descent method. Then, the sub-historical sample time series data of the target data type is repeatedly forward-propagated through the initial lithium battery life prediction model until the joint loss function converges or reaches a preset number of training rounds, at which time the trained lithium battery life prediction model is obtained.

[0012] In a second aspect, the present invention provides a lithium battery life prediction system for implementing any of the lithium battery life prediction methods described above; the lithium battery life prediction system comprises: A historical data acquisition module, which acquires historical aging characteristic time series data and a historical sample time series data set of a specified type of sample lithium battery in a specified historical period; wherein the historical sample time series data set includes a plurality of sub-historical sample time series data of different data types; A correlation calculation module, which is in communication with the historical data acquisition module and is used to respectively obtain the correlation between each sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data; A target data type acquisition module is connected to the correlation calculation module for taking the data type corresponding to the sub-historical sample time series data whose correlation is greater than a preset correlation threshold value among the multiple sub-historical sample time series data as the target data type; A model training module, which is in communication connection with the target data type acquisition module, is used to construct an initial lithium battery life prediction model, and train the initial lithium battery life prediction model using the sub-historical sample time series data of the target data type and the historical aging feature time series data to obtain a trained lithium battery life prediction model; wherein the trained lithium battery life prediction model includes a physical model and a deep learning model connected in sequence; A life prediction module is communicatively connected to the model training module, and is used to collect target operating data of the target lithium battery of the specified type in real time, and input the target operating data into the trained lithium battery life prediction model, so as to obtain the life prediction result of the target lithium battery based on the trained lithium battery life prediction model; wherein the data type of the target operating data is the target data type.

[0013] In a third aspect, the present invention provides an electronic device, comprising: a memory for storing computer program instructions; and, A processor is used to execute the computer program instructions to complete the operation of the lithium battery life prediction method as described in any one of the above.

[0014] In a fourth aspect, the present invention provides a computer-readable storage product, characterized in that instructions are stored on the computer-readable storage product, and when the instructions are run on a computer, a lithium battery life prediction method as described in any one of the above is executed.

[0015] In a fifth aspect, the present invention provides a computer program product, including a computer program or instructions, which, when executed by a computer, implements a lithium battery life prediction method as described in any one of the above.

[0016] The beneficial effects of the present invention are: The present invention discloses a lithium battery life prediction method, system, electronic device and product. By comprehensively utilizing physical models and data-driven methods, the accuracy and adaptability of lithium battery life prediction can be significantly improved, and the reliability and economy of energy storage systems can be ensured. Specifically, during the implementation of the present invention, by pre-extracting the target data type with the highest correlation with the lithium battery aging characteristics, the feature data highly related to the lithium battery aging can be effectively identified, and the model can be ensured to use the most informative data during training and prediction, thereby improving the accuracy of the life prediction of the target lithium battery; in the process of life prediction of the target lithium battery based on the trained lithium battery life prediction model, the target working data of the target lithium battery of the specified type is collected in real time, and the target working data is input into the trained lithium battery life prediction model, and the target working data can be processed by feature construction through the physical model in the trained lithium battery life prediction model to obtain reconstructed feature data, and then the target working data and the reconstructed feature data are comprehensively predicted in combination with the deep learning model, thereby further improving the accuracy and robustness of the prediction.

[0017] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of the lithium battery life prediction method in Example 1; Figure 2 is a module block diagram of the lithium battery life prediction system in Example 2; Figure 3 This is a module block diagram of the electronic device in Example 3. DETAILED DESCRIPTION

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in combination with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0020] Embodiment 1: This embodiment discloses a lithium battery life prediction method, which can be executed by, but is not limited to, a computer device or a virtual machine with certain computing resources, such as a personal computer, a smart phone, a personal digital assistant, or a wearable device, or by a virtual machine.

[0021] like Figure 1 As shown, a lithium battery life prediction method may include but is not limited to the following steps: S1. Obtain historical aging characteristic time series data and a historical sample time series data set of a specified type of sample lithium battery in a specified historical period; wherein the historical sample time series data set includes a plurality of sub-historical sample time series data of different data types.

[0022] Specifically, in step S1, the historical aging characteristic time series data adopts historical capacity time series data. It should be noted that for lithium batteries, capacity can best indicate its aging degree. Based on this, this embodiment sets the aging characteristic as the capacity of the lithium battery. In this embodiment, the data type of the historical sample time series data set includes voltage, current, operating temperature, charge and discharge rate, number of cycles and / or open circuit voltage, etc., which can reflect the real-time working status and environmental conditions of the lithium battery, which are not limited here.

[0023] S2. Obtain the correlation between each sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data respectively.

[0024] Specifically, in step S2, obtaining the correlation between any sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data includes: S201. Perform data alignment processing on the historical aging feature time series data and any sub-historical sample time series data to obtain aligned historical aging feature time series data and any aligned sub-historical sample time series data; it should be noted that in the process of collecting time series data, due to the time asynchrony of different data sources, data misalignment may occur, so as to ensure the accuracy and reliability of the data. To this end, in this embodiment, by performing data alignment processing on the historical aging feature time series data and any sub-historical sample time series data, that is, matching the data in the two according to the timestamps, so that the two are consistent on the time axis, and then facilitate subsequent data analysis and processing.

[0025] S202. Standardize the aligned historical aging feature time series data and any aligned sub-historical sample time series data respectively to obtain the standardized historical aging feature time series data and any standardized sub-historical sample time series data; it should be noted that the standardization process is used to eliminate the influence of the dimension in the time series data so as to obtain the accurate correlation between the two time series data.

[0026] S203. Use the standardized historical aging feature time series data as a reference sequence, and use any standardized sub-historical sample time series data as a comparison sequence, and obtain the correlation coefficients between the reference sequence and each historical sample data in the comparison sequence respectively; it should be noted that the correlation coefficient between any historical sample data in the comparison sequence and the reference sequence is used to evaluate the similarity or correlation between any historical sample data and the reference sequence, and based on the correlation coefficients between each historical sample data in the comparison sequence and the correlation coefficients between the reference sequences, the correlation degree between the reference sequence and the comparison sequence can be obtained.

[0027] Specifically, in step S203, the first i The correlation coefficient between historical sample data includes: S2031. Obtain absolute difference values ​​between each historical aging feature in the reference sequence and each historical sample data corresponding to each historical aging feature in the comparison sequence; it should be understood that the number of absolute difference values ​​is the same as the total number of historical aging features in the reference sequence, that is, the total number of historical sample data in the comparison sequence.

[0028] S2032. Obtain the maximum absolute difference value and the minimum absolute difference value among the absolute difference values; S2033. Based on the comparison sequence i The historical sample data and the reference sequence i a historical aging feature (i.e., a feature in the reference sequence that is consistent with the iThe absolute difference value between the historical aging characteristics corresponding to the historical sample data), the maximum absolute difference value and the minimum absolute difference value are obtained by: i The correlation coefficient between historical sample data; wherein the correlation coefficient is: ; In the formula, is the minimum absolute difference value; is the maximum absolute difference value; is the i-th historical sample data in the comparison sequence and the i-th historical aging feature in the reference sequence The absolute difference between ; is a preset resolution factor. It should be understood that the resolution factor [0,1], in this embodiment, its value is 0.5.

[0029] S204. According to the correlation coefficient between the reference sequence and each historical sample data in the comparison sequence, the correlation degree between the reference sequence and the comparison sequence is obtained, which is also the correlation degree between any sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data. It should be noted that, in this embodiment, the correlation coefficient ranges from 0 to 1, and the closer the value is to 1, the higher the correlation degree between the reference sequence and the historical sample data corresponding to the comparison sequence. Based on this, the correlation degree between the reference sequence and the comparison sequence can be evaluated by the correlation coefficient between the reference sequence and each historical sample data in the comparison sequence.

[0030] Specifically, in step S204, the degree of association between the reference sequence and the comparison sequence is: ; In the formula, I is the total number of historical sample data in the comparison sequence, that is, the total number of historical sample data in the comparison sequence; it should be understood that [0,1].

[0031] S3. Take the data type corresponding to the sub-historical sample time series data whose correlation is greater than a preset correlation threshold value among the multiple sub-historical sample time series data as the target data type.

[0032] S4. Construct an initial lithium battery life prediction model, and use the sub-historical sample time series data of the target data type and the historical aging feature time series data to train the initial lithium battery life prediction model to obtain a trained lithium battery life prediction model; wherein the trained lithium battery life prediction model includes a physical model and a deep learning model connected in sequence.

[0033] Specifically, in step S4, the initial lithium battery life prediction model is trained using the sub-historical sample time series data of the target data type and the historical aging feature time series data to obtain a trained lithium battery life prediction model, including: S401. Initializing the parameters of the initial lithium battery life prediction model; S402. Forward propagating the sub-historical sample time series data of the target data type through the initial lithium battery life prediction model to obtain the life prediction time series data of the sample lithium battery; S403. Calculate the error between the life prediction time series data and the historical aging characteristic time series data according to the joint loss function of the initial lithium battery life prediction model; S404. Utilize the back-propagation algorithm to calculate the gradient of the joint loss function relative to the parameters of the initial lithium battery life prediction model, and update the parameters of the initial lithium battery life prediction model according to the gradient descent method, and then repeatedly forward-propagate the sub-historical sample time series data of the target data type through the initial lithium battery life prediction model, that is, return to step S402 until the joint loss function converges or reaches the preset number of training rounds, at which time the latest parameters are used as the model parameters of the initial lithium battery life prediction model to obtain the trained lithium battery life prediction model.

[0034] Specifically, in this embodiment, the initial lithium battery life prediction model includes an initial physical model and an initial deep learning model connected in sequence; correspondingly, the joint loss function of the initial lithium battery life prediction model can be defined as: L = α · L phys + β · L LSTM ; In the formula, L phys is a loss function of the initial physical model, used to measure the deviation between the output value of the initial physical model and the reconstructed data value obtained based on the sub-historical sample time series data of the target data type; L LSTMis the loss function of the initial deep learning model, which is used to measure the deviation between the life prediction value and the actual aging characteristic value of the sample lithium battery; in this embodiment, the loss function of the initial physical model is L phys And the loss function of the initial deep learning model L LSTM The mean square error can be used and is not restricted here; α and β is a preset weight coefficient used to balance the influence of the initial physical model and the initial deep learning model, α ∈(0,1), β ∈(0,1), and α + β =1, during implementation, α and β Cross-validation or Bayesian optimization methods can be used for adjustment and determination, which will not be described in detail here.

[0035] It should be noted that in the process of training the initial lithium battery life prediction model, the reconstructed data values ​​generated by the initial physical model are used as additional input values ​​of the initial deep learning model to achieve joint training of the initial physical model and the initial deep learning model, so as to help the initial deep learning model better understand the internal mechanism of the specified type of lithium battery and improve the prediction accuracy through the prior knowledge provided by the initial physical model.

[0036] It should also be noted that, in this embodiment, the trained lithium battery life prediction model includes a physical model and a deep learning model connected in sequence. In the process of using the trained lithium battery life prediction model to predict the life of the target lithium battery, the prior knowledge of the physical model can be used to guide the feature extraction and modeling process of the deep learning model, thereby improving the accuracy of the trained lithium battery life prediction model in predicting the life of the target lithium battery.

[0037] S5. Real-time collection of target working data of the target lithium battery of the specified type; wherein the data type of the target working data is the target data type; it should be noted that the frequency and accuracy of the collection of the target working data will directly affect the accuracy and real-time performance of the trained lithium battery life prediction model, and for this reason, high-frequency sampling is usually required in each charge and discharge cycle of the target lithium battery. In addition, in order to ensure the accuracy and consistency of the data input into the trained lithium battery life prediction model, after the target working data is collected, the target working data is pre-processed in advance, such as data cleaning, missing value processing, and data enhancement, so as to ensure the accuracy of subsequent life prediction results.

[0038] S6. Inputting the target working data into the trained lithium battery life prediction model, so as to obtain the life prediction result of the target lithium battery based on the trained lithium battery life prediction model.

[0039] It should be noted that the measurement of aging characteristic data such as lithium battery capacity that can directly indicate its aging degree is usually difficult. Based on this, this embodiment obtains sub-historical sample time series data corresponding to the data type whose correlation with the parameter is greater than the correlation threshold to achieve training of the initial lithium battery life prediction model, and based on the trained lithium battery life prediction model and the target working data of the target lithium battery, realizes life prediction of the target lithium battery, which can effectively reduce the difficulty of predicting the life of the target lithium battery.

[0040] This embodiment can significantly improve the accuracy and adaptability of lithium battery life prediction by comprehensively utilizing physical models and data-driven methods, and ensure the reliability and economy of the energy storage system. Specifically, during the implementation of this embodiment, by pre-extracting the target data type with the highest correlation with the lithium battery aging characteristics, it can effectively identify the feature data that is highly related to the lithium battery aging, and ensure that the model uses the most informative data during training and prediction, thereby improving the accuracy of the life prediction of the target lithium battery; in the process of predicting the life of the target lithium battery based on the trained lithium battery life prediction model, the target working data of the target lithium battery of the specified type is collected in real time, and the target working data is input into the trained lithium battery life prediction model, and the target working data can be processed by feature construction through the physical model in the trained lithium battery life prediction model to obtain reconstructed feature data, and then the target working data and the reconstructed feature data are comprehensively predicted in combination with the deep learning model, thereby further improving the accuracy and robustness of the prediction.

[0041] Embodiment 2: This embodiment discloses a lithium battery life prediction system for implementing the lithium battery life prediction method in embodiment 1; Figure 2 As shown, the lithium battery life prediction system includes: A historical data acquisition module, which acquires historical aging characteristic time series data and a historical sample time series data set of a specified type of sample lithium battery in a specified historical period; wherein the historical sample time series data set includes a plurality of sub-historical sample time series data of different data types; A correlation calculation module, which is in communication with the historical data acquisition module and is used to respectively obtain the correlation between each sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data; A target data type acquisition module is connected to the correlation calculation module for taking the data type corresponding to the sub-historical sample time series data whose correlation is greater than a preset correlation threshold value among the multiple sub-historical sample time series data as the target data type; A model training module, which is in communication connection with the target data type acquisition module, is used to construct an initial lithium battery life prediction model, and train the initial lithium battery life prediction model using the sub-historical sample time series data of the target data type and the historical aging feature time series data to obtain a trained lithium battery life prediction model; wherein the trained lithium battery life prediction model includes a physical model and a deep learning model connected in sequence; A life prediction module is communicatively connected to the model training module, and is used to collect target operating data of the target lithium battery of the specified type in real time, and input the target operating data into the trained lithium battery life prediction model, so as to obtain the life prediction result of the target lithium battery based on the trained lithium battery life prediction model; wherein the data type of the target operating data is the target data type.

[0042] It should be noted that the working process, working details and technical effects of the lithium battery life prediction system provided in this embodiment 2 can be found in embodiment 1 and will not be repeated here.

[0043] Embodiment 3: Based on Embodiment 1 or 2, this embodiment discloses an electronic device, which may be a smart phone, a tablet computer, a laptop computer, or a desktop computer. The electronic device may be referred to as a user terminal, a portable terminal, a desktop terminal, etc. Figure 3 As shown, the electronic equipment includes: a memory for storing computer program instructions; and, A processor is used to execute the computer program instructions to complete the operation of the lithium battery life prediction method as described in any one of Example 1.

[0044] Specifically, the processor 301 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 301 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 301 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 301 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen.

[0045] The memory 302 may include one or more computer-readable storage media, which may be non-transitory. The memory 302 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 302 is used to store at least one instruction, which is used to be executed by the processor 301 to implement the lithium battery life prediction method provided in Example 1 of the present application.

[0046] In some embodiments, the terminal may further optionally include: a communication interface 303 and at least one peripheral device. The processor 301, the memory 302 and the communication interface 303 may be connected via a bus or a signal line. Each peripheral device may be connected to the communication interface 303 via a bus, a signal line or a circuit board. Specifically, the peripheral device includes: at least one of a radio frequency circuit 304, a display screen 305 and a power supply 306.

[0047] The communication interface 303 may be used to connect at least one peripheral device related to I / O (Input / Output) to the processor 301 and the memory 302. In some embodiments, the processor 301, the memory 302, and the communication interface 303 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 301, the memory 302, and the communication interface 303 may be implemented on a separate chip or circuit board, which is not limited in this embodiment.

[0048] The radio frequency circuit 304 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The radio frequency circuit 304 communicates with a communication network and other communication devices through electromagnetic signals.

[0049] The display screen 305 is used to display a UI (User Interface). The UI may include graphics, text, icons, videos, and any combination thereof.

[0050] The power supply 306 is used to supply power to various components in the electronic device.

[0051] Embodiment 4: Based on any one of embodiments 1 to 3, this embodiment discloses a computer-readable storage product, characterized in that instructions are stored on the computer-readable storage product, and when the instructions are run on a computer, the lithium battery life prediction method described in any one of embodiment 1 is executed. The computer-readable storage product refers to a carrier for storing data, which may include but is not limited to computer-readable storage media such as floppy disks, optical disks, hard disks, flash memories, USB flash drives, and / or memory sticks (MemoryStick), and the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0052] Embodiment 5: Based on any one of Examples 1 to 3, this embodiment discloses a computer program product, including a computer program or instructions, which, when executed by a computer, implements the lithium battery life prediction method as described in any one of Example 1.

[0053] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, and optionally, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A lithium battery life prediction method, characterized in that: include: Obtaining historical aging characteristic time series data and a historical sample time series data set of a specified type of sample lithium battery in a specified historical period; wherein the historical sample time series data set includes a plurality of sub-historical sample time series data of different data types; Respectively obtaining the correlation between each sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data; The data type corresponding to the sub-historical sample time series data whose correlation is greater than a preset correlation threshold value among the multiple sub-historical sample time series data is used as the target data type; Constructing an initial lithium battery life prediction model, and using the sub-historical sample time series data of the target data type and the historical aging feature time series data to train the initial lithium battery life prediction model to obtain a trained lithium battery life prediction model; wherein the trained lithium battery life prediction model includes a physical model and a deep learning model connected in sequence; Collecting target operating data of the target lithium battery of the specified type in real time; wherein the data type of the target operating data is the target data type; The target operating data is input into the trained lithium battery life prediction model so as to obtain the life prediction result of the target lithium battery based on the trained lithium battery life prediction model.

2. A lithium battery life prediction method according to claim 1, characterized in that: The historical aging characteristic time series data adopts historical capacity time series data.

3. A lithium battery life prediction method according to claim 1, characterized in that: Obtaining the correlation between any sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data, including: Performing data alignment processing on the historical aging feature time series data and any sub-historical sample time series data to obtain aligned historical aging feature time series data and any aligned sub-historical sample time series data; Performing standardization processing on the aligned historical aging feature time series data and any aligned sub-historical sample time series data respectively to obtain the standardized historical aging feature time series data and any standardized sub-historical sample time series data; Taking the standardized historical aging feature time series data as a reference sequence, taking any standardized sub-historical sample time series data as a comparison sequence, and respectively obtaining the correlation coefficient between the reference sequence and each historical sample data in the comparison sequence; The degree of association between the reference sequence and the comparison sequence is obtained according to the association coefficient between each historical sample data in the reference sequence and the comparison sequence.

4. A lithium battery life prediction method according to claim 3, characterized in that: Get the first i The correlation coefficient between historical sample data includes: Obtaining an absolute difference value between each historical aging feature in the reference sequence and each historical sample data corresponding to each historical aging feature in the comparison sequence; Obtain the maximum absolute difference value and the minimum absolute difference value among the absolute difference values; Based on the comparison sequence i The historical sample data and the reference sequence i The absolute difference value between the historical aging characteristics, the maximum absolute difference value and the minimum absolute difference value are obtained by comparing the reference sequence with the comparison sequence. i The correlation coefficient between historical sample data; wherein the correlation coefficient is: ； In the formula, is the minimum absolute difference value; is the maximum absolute difference value; is the absolute difference between the i-th historical sample data in the comparison sequence and the i-th historical aging feature in the reference sequence; is the preset resolution factor.

5. A lithium battery life prediction method according to claim 4, characterized in that: The degree of association between the reference sequence and the comparison sequence is: ； In the formula, I is the total number of historical sample data in the comparison sequence.

6. A lithium battery life prediction method according to claim 1, characterized in that: The initial lithium battery life prediction model is trained using the sub-historical sample time series data of the target data type and the historical aging feature time series data to obtain a trained lithium battery life prediction model, including: Initializing the parameters of the initial lithium battery life prediction model; Propagate the sub-historical sample time series data of the target data type forward through the initial lithium battery life prediction model to obtain the life prediction time series data of the sample lithium battery; Calculating the error between the life prediction time series data and the historical aging characteristic time series data according to the joint loss function of the initial lithium battery life prediction model; The back-propagation algorithm is used to calculate the gradient of the joint loss function relative to the parameters of the initial lithium battery life prediction model, and the parameters of the initial lithium battery life prediction model are updated according to the gradient descent method. Then, the sub-historical sample time series data of the target data type is repeatedly forward-propagated through the initial lithium battery life prediction model until the joint loss function converges or reaches a preset number of training rounds, at which time the trained lithium battery life prediction model is obtained.

7. A lithium battery life prediction system, characterized in that: Used to implement the lithium battery life prediction method according to any one of claims 1 to 6; the lithium battery life prediction system comprises: A historical data acquisition module, which acquires historical aging characteristic time series data and a historical sample time series data set of a specified type of sample lithium battery in a specified historical period; wherein the historical sample time series data set includes a plurality of sub-historical sample time series data of different data types; A correlation calculation module, which is in communication with the historical data acquisition module and is used to respectively obtain the correlation between each sub-historical sample time series data in the historical sample time series data set and the historical aging feature time series data; A target data type acquisition module is connected to the correlation calculation module for taking the data type corresponding to the sub-historical sample time series data whose correlation is greater than a preset correlation threshold value among the multiple sub-historical sample time series data as the target data type; A model training module, which is in communication connection with the target data type acquisition module, is used to construct an initial lithium battery life prediction model, and train the initial lithium battery life prediction model using the sub-historical sample time series data of the target data type and the historical aging feature time series data to obtain a trained lithium battery life prediction model; wherein the trained lithium battery life prediction model includes a physical model and a deep learning model connected in sequence; A life prediction module is communicatively connected to the model training module, and is used to collect target operating data of the target lithium battery of the specified type in real time, and input the target operating data into the trained lithium battery life prediction model, so as to obtain the life prediction result of the target lithium battery based on the trained lithium battery life prediction model; wherein the data type of the target operating data is the target data type.

8. An electronic device, characterized in that: include: a memory for storing computer program instructions; as well as, A processor is used to execute the computer program instructions to complete the operation of the lithium battery life prediction method as described in any one of claims 1 to 6.

9. A computer readable storage product, characterized in that: The computer-readable storage product stores instructions, and when the instructions are executed on a computer, the lithium battery life prediction method according to any one of claims 1 to 6 is executed.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or the instruction is executed by a computer, the lithium battery life prediction method as described in any one of claims 1 to 6 is implemented.

Images