A lithium battery life prediction method and system based on Mamba model

Abstract

The present application discloses a lithium battery life prediction method and system based on the Mamba model, which relates to the field of energy storage of new energy power systems, including: obtaining aging feature matrices of multiple lithium batteries, substituting them into the Mamba model for training, and extracting the hidden state matrix sequence of each lithium battery according to the trained model, and obtaining the cycle number sequence corresponding to the target cycle life prediction point of each lithium battery through norm calculation and variational mode decomposition; based on the cycle number sequence corresponding to the target health state of all lithium batteries, and the cycle number sequence corresponding to the target cycle life prediction point of all lithium batteries, a linear relationship model for lithium battery life prediction is constructed; according to the linear relationship model for lithium battery life prediction, the cycle number sequence corresponding to the target health state of the lithium battery to be predicted is calculated. The present application improves the interpretability and computational efficiency of lithium battery life prediction and expands the scope of application of the prediction model.

Description

Technical Field

[0001] The present application relates to the field of energy storage of new energy power systems, and in particular to a method and system for predicting the life of a lithium battery based on a Mamba model. Background Art

[0002] With the growth of global energy demand and the improvement of environmental protection awareness, new energy technologies, especially lithium battery energy storage technology, are developing rapidly. As the core component of the energy storage system, the performance of lithium batteries is directly related to the stability and economy of the entire energy storage system. In application scenarios such as new energy power generation and electric vehicles, the cycle life of lithium batteries is a key factor in determining the operating cost and safety and stability of the system. However, the cycle life of lithium batteries is affected by many complex factors, including but not limited to charge and discharge current, temperature, state of charge, and number of cycles. These factors not only make the life prediction of lithium batteries complicated, but also increase the difficulty of battery status monitoring and management by the battery management system. Traditional lithium battery cycle life prediction methods mainly rely on empirical models or linear models based on a single variable. These methods are often difficult to accurately reflect the attenuation characteristics of batteries under actual complex working conditions. With the rise of artificial intelligence and big data technologies, lithium battery life prediction methods based on machine learning have gradually become a research hotspot. Through the analysis and modeling of a large amount of actual operation data, the battery life attenuation trend can be predicted more accurately, thereby realizing the scientific management and optimization of the entire life cycle of the battery. Accurate lithium battery cycle life prediction methods can not only help optimize the design and operation strategy of energy storage systems and extend battery life, but also effectively reduce system operating costs, improve the efficiency of new energy utilization, and provide important support for building a more stable and reliable energy system. Research and technological progress in this field are of great significance to promoting the development of new energy and energy storage industries.

[0003] The current lithium battery cycle life prediction methods cannot accurately reflect the correspondence between the aging characteristics and health status of lithium batteries, the lithium battery life prediction has poor interpretability, inaccurate prediction results, and a small scope of application of the prediction model. Summary of the invention

[0004] The purpose of this application is to provide a lithium battery life prediction method and system based on the Mamba model, which can improve the interpretability and computational efficiency of lithium battery life prediction and expand the scope of application of the prediction model.

[0005] To achieve the above objectives, this application provides the following solutions.

[0006] In a first aspect, the present application provides a lithium battery life prediction method based on the Mamba model, the method comprising: obtaining an aging feature matrix of multiple lithium batteries; the aging feature matrix comprising: a health state sequence and an aging feature sequence; the health state sequence is a sequence consisting of the health state of each cycle number of each lithium battery; substituting the aging feature matrices of multiple lithium batteries into the Mamba model for training, and extracting the hidden state matrix sequence of each lithium battery according to the trained Mamba model; obtaining the cycle number sequence corresponding to the target cycle life prediction point of each lithium battery through norm calculation and variational mode decomposition according to the hidden state matrix sequence of each lithium battery; the target cycle life prediction point is the one with the largest correlation among the candidate cycle life prediction points prediction point; the candidate cycle life prediction point is a prediction point in the norm sequence curve of the latent state matrix norm of the lithium battery obtained after norm calculation, which is correlated with the cycle number sequence corresponding to the target health state of the lithium battery; the cycle number sequence corresponding to the target health state of the lithium battery is obtained according to the health state sequence index; based on the cycle number sequence corresponding to the target health state of all lithium batteries and the cycle number sequence corresponding to the target cycle life prediction point of all lithium batteries, a linear relationship model for lithium battery life prediction is constructed; the cycle number sequence corresponding to the target cycle life prediction point of the lithium battery to be predicted is obtained, and the cycle number sequence is substituted into the linear relationship model for lithium battery life prediction to calculate the cycle number sequence corresponding to the target health state of the lithium battery to be predicted.

[0007] In a second aspect, the present application provides a computer system, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-mentioned lithium battery life prediction methods based on the Mamba model.

[0008] According to the specific embodiments provided in this application, this application discloses the following technical effects.

[0009] The present application substitutes the aging feature matrix of the lithium battery into the Mamba model for training, and uses the trained model to extract the hidden state matrix sequence of each lithium battery, learns the relationship between the battery aging characteristics and the health state from the historical time series data, improves the model calculation efficiency, and further improves the interpretability of the lithium battery life prediction; converts the three-dimensional hidden state matrix sequence into a one-dimensional hidden state matrix norm sequence through norm calculation, amplifies the change law of the internal hidden state of the lithium battery aging in the time series, and improves the representativeness of the life prediction; extracts the main mode sequence of the hidden state matrix norm of each lithium battery through variational mode decomposition, reduces the interference error caused by noise, and improves the robustness of the model; determines the cycle life prediction point by analyzing the cycle number sequence corresponding to the target health state of all lithium batteries and the correlation of the cycle number sequence corresponding to the target cycle life prediction point of all lithium batteries, takes into account the commonalities of different lithium batteries in the cycle aging process, and constructs a linear relationship model with lithium battery life prediction, which expands the scope of application of the prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0011] Figure 1 A schematic diagram of a flow chart of a lithium battery life prediction method based on the Mamba model provided in an embodiment of the present application.

[0012] Figure 2 A graph showing the number of lithium battery cycles and health status provided in an embodiment of the present application.

[0013] Figure 3 A scatter plot of the relationship between the aging characteristics and health status of the lithium battery provided in the embodiment of the present application.

[0014] Figure 4 A graph of the hidden state matrix norm provided in an embodiment of the present application.

[0015] Figure 5 This is a main mode curve diagram of the hidden state matrix norm provided in an embodiment of the present application.

[0016] Figure 6 A heat map showing the correlation between the number of cycles at the candidate cycle life prediction point and the number of cycles corresponding to the target health state provided in an embodiment of the present application.

[0017] Figure 7This is a fitting result diagram of the number of cycles corresponding to the target cycle life prediction point provided in the embodiment of the present application and the number of cycles corresponding to the target health state of 0.8.

[0018] Figure 8 This is a diagram of prediction results of different health states of a lithium battery to be predicted provided in an embodiment of the present application.

[0019] Fig. 9 A schematic diagram of the structure of a computer system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0021] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0022] Embodiment 1, as Figure 1 As shown, this embodiment provides a lithium battery life prediction method based on the Mamba model, comprising the following steps.

[0023] S1. Obtain an aging feature matrix of multiple lithium batteries; the aging feature matrix includes: a health state sequence and an aging feature sequence; the health state sequence is a sequence consisting of the health state of each lithium battery for each cycle. Figure 2 The graph shows the relationship between the number of lithium battery cycles and health status.

[0024] Step S1 specifically includes the following steps.

[0025] S11. Obtaining cycle operation data of multiple lithium batteries; the cycle operation data at least includes: voltage, temperature, power change value and cycle number at different sampling times.

[0026] Optionally, in this embodiment, the number of lithium batteries .

[0027] S12. Calculate the health status of each lithium battery for each cycle according to the cycle operation data of each lithium battery, and construct a health status sequence for each lithium battery.

[0028] S13. Extracting the aging characteristic sequence of each lithium battery according to the cycle operation data of each lithium battery.

[0029] S14. Figure 3 As shown, based on the aging feature sequences of multiple lithium batteries and the health state sequences of multiple lithium batteries, an aging feature matrix of multiple lithium batteries is constructed.

[0030] Furthermore, the aging characteristics specifically include: the characteristic sequence of the coefficient of variation of the power change value , the Euclidean distance feature sequence of the power change value , Capacity increment curve peak height characteristic sequence , the voltage characteristic sequence corresponding to the peak of the capacity increment curve , Temperature increment curve peak height characteristic sequence The voltage characteristic sequence corresponding to the peak value of the temperature increment curve One or more of .

[0031] Optionally, the aging characteristic matrix of lithium batteries is as follows.

[0032] .

[0033] in, represents the transpose of a sequence; is a matrix containing seven sequences.

[0034] In the specific implementation process, the health status of the lithium battery = the remaining power of the lithium battery / the nominal capacity of the lithium battery; in this embodiment, the nominal capacity of the lithium battery An Shi.

[0035] S2. Substitute the aging feature matrices of multiple lithium batteries into the Mamba model for training, and extract the hidden state matrix sequence of each lithium battery according to the trained Mamba model.

[0036] Step S2 specifically includes the following steps.

[0037] S21. Substitute the aging feature matrices of multiple lithium batteries into the Mamba model, use the aging feature sequence as a training set, and use the health status sequence as a label to train the Mamba model to obtain a trained Mamba model.

[0038] S22. Extract the hidden state matrix of each cyclic charge and discharge segment of the lithium battery in the trained Mamba model, and construct a hidden state matrix sequence for each lithium battery; the cyclic charge and discharge segment is the operation data of the voltage within a preset voltage range in the cyclic operation data of each lithium battery; the hidden state matrix sequence is a sequence composed of hidden state matrices of different cycle numbers of multiple lithium batteries.

[0039] Specifically, the preset voltage range [ ]=[ ], the unit is volt.

[0040] In practical applications, the Mamba model is an emerging sequence modeling architecture, which is essentially a state space model (Structured Space Model, SSM). By introducing a selection mechanism, the traditional time-invariant state space model is transformed into a time-varying model that changes with the input. Mamba overcomes the limitations of Transformer in computational efficiency and long sequence processing. By introducing the concept of selective state space models, combined with hardware-aware algorithms and simpler architectures, it achieves higher computational efficiency and stronger sequence modeling capabilities.

[0041] Further, the Mamba model is as follows.

[0042] .

[0043] .

[0044] in, , , As the model input, output and hidden state respectively, Indicates lithium battery The aging characteristics of the number of cycles, Indicates lithium battery The health status of the number of cycles, Indicates lithium battery The hidden state of the number of cycles, Indicates lithium battery -1 cycle of hidden states, , , represents the learnable parameter matrix, , It is the total cycle number of the lithium battery.

[0045] Optionally, the training process of the Mamba model is as follows.

[0046] Extract The hidden state matrix of the cyclic discharge segment .

[0047] Initialize each model parameter randomly.

[0048] For the number of cycles (time steps) , calculated recursively according to the formula , set the loss function.

[0049] .

[0050] is the aging feature matrix No. The element in the row and column 7, that is, The SOH value of the cycle is used to minimize the loss function through back propagation and update the model parameters ( , , , ).

[0051] Extract the last cycle number (time step) matrix , which is the The hidden state matrix of the cyclic discharge segment .

[0052] To extract the hidden state matrix of all cycles , forming a hidden state matrix sequence .

[0053] S3. Obtain the cycle number sequence corresponding to the target cycle life prediction point of each lithium battery through norm calculation and variational mode decomposition according to the hidden state matrix sequence of each lithium battery; the target cycle life prediction point is the prediction point with the greatest correlation among the candidate cycle life prediction points; the candidate cycle life prediction point is the prediction point in the norm sequence curve of the hidden state matrix of the lithium battery obtained after norm calculation that is correlated with the cycle number sequence corresponding to the target health state of the lithium battery; the cycle number sequence corresponding to the target health state of the lithium battery is obtained according to the health state sequence index.

[0054] Optionally, the target health status is as follows.

[0055] .

[0056] Step S3 specifically includes the following steps.

[0057] S31. Figure 4 As shown, the norm of the hidden state matrix of each cycle number in the hidden state matrix sequence of each lithium battery is calculated respectively to obtain the norm sequence of the hidden state matrix of each lithium battery.

[0058] Furthermore, the calculation formula of the hidden state matrix norm sequence is as follows.

[0059] .

[0060] .

[0061] in, is the hidden state matrix norm sequence, For the The norm of the hidden state matrix of the number of cycles, For lithium battery The hidden state matrix of the number of cycles Line The hidden state of the elements of the column, , It is the total cycle number of the lithium battery.

[0062] S32. Figure 5 As shown, the variational modal decomposition is performed on the norm sequence of the hidden state matrix of each lithium battery to extract the main modal sequence of the norm of the hidden state matrix of each lithium battery.

[0063] S33. Calculate the number of cycles when the main mode value reaches each candidate cycle life prediction point according to the main mode sequence of the hidden state matrix norm of each lithium battery.

[0064] S34. Select the cycle number sequence corresponding to the prediction point with the greatest correlation among the candidate cycle life prediction points as the cycle number sequence corresponding to the target cycle life prediction point of each lithium battery.

[0065] Furthermore, the calculation formula for the cycle number sequence corresponding to the target cycle life prediction point is as follows.

[0066] .

[0067] in, represents the cycle number sequence corresponding to the target cycle life prediction point, represents the index of the maximum value in the sequence, Indicates lithium battery The target health state cycle number matrix No. List, Indicates lithium battery Cycle life prediction point cycle number matrix No. List, is a matrix No. The mean of the column, is a matrix No. The mean of the column, , N Indicates the total amount used for lithium batteries, Represents the number of candidate cycle life prediction points.

[0068] In the actual application process, for all the training For each battery, follow steps 1 to 4 to extract the corresponding cycle sequence when each battery cycles to the preset health state (SOH) The cycle number sequence corresponding to the candidate cycle life prediction point , forming a cycle number matrix corresponding to different batteries when they cycle to a preset health state (SOH) The cycle number sequence corresponding to the candidate cycle life prediction point ;in No. Line Column Elements Indicates lithium battery Loop to the set of health states (SOH) to be predicted Middle The number of circles corresponding to the number of elements is No. Line Column Elements Indicates battery Loop to the candidate life prediction point set No. The number of circles corresponding to the number of elements is , , , For collection The number of elements, For collection Number of elements.

[0069] In practical applications, this method uses variational mode decomposition to perform mode separation on the latent state matrix norm time series, extracts the main mode as the analysis benchmark between the internal hidden state of battery aging and the number of cycles, reduces noise interference errors, and improves robustness.

[0070] The variational mode decomposition method is used to perform mode decomposition on the hidden state matrix norm sequence to obtain a finite number of inherent modes of the hidden state matrix norm sequence and a residual sequence.

[0071] .

[0072] in represents the number of intrinsic mode sequences obtained by variational mode decomposition, represents the first The intrinsic mode sequence, , represents the residual sequence, where It is called the main mode.

[0073] The main mode is calculated according to the following formula The number of cycles when the value reaches the candidate cycle life prediction point.

[0074] .

[0075] in, Represents the set of candidate cycle life prediction points No. elements, Representation calculation The elements in the sequence The index position in ; for all Calculate separately , get the cycle number sequence corresponding to the candidate cycle life prediction point ,in Representing a collection The number of elements; in this embodiment , .

[0076] S4. Based on the cycle number sequence corresponding to the target health status of all lithium batteries and the cycle number sequence corresponding to the target cycle life prediction point of all lithium batteries, a linear relationship model for lithium battery life prediction is constructed.

[0077] Step S4 specifically includes the following steps.

[0078] S41. Construct a target health status cycle number matrix based on the cycle number sequence corresponding to the target cycle life prediction points of all lithium batteries.

[0079] S42. Construct a target cycle life prediction point cycle number matrix based on the cycle number sequences corresponding to the target cycle life prediction points of all lithium batteries.

[0080] S43. Figure 6-Figure 7 As shown, based on the target health status cycle number matrix and the target cycle life prediction point cycle number matrix, the least squares method is used to fit the linear relationship model for lithium battery life prediction.

[0081] Furthermore, the linear relationship model for lithium battery life prediction is as follows.

[0082] ;

[0083] in, Represents the target health state cycle number matrix No. Column, represents the target cycle life prediction point cycle number matrix of sequence, is the cycle number sequence corresponding to the target cycle life prediction point, and They are and The slope and intercept of the linear relationship model.

[0084] S5. Figure 8As shown, the cycle number sequence corresponding to the target cycle life prediction point of the lithium battery to be predicted is obtained, and is substituted into the linear relationship model of lithium battery life prediction to calculate the cycle number sequence corresponding to the target health state of the lithium battery to be predicted.

[0085] Specifically, in the process of obtaining the cycle number sequence corresponding to the target cycle life prediction point of the lithium battery to be predicted in the prediction stage, when extracting the main mode sequence of the hidden state matrix norm of each lithium battery After that, we need to determine if the hidden state matrix norm main mode sequence The last element value is less than , indicating that the current lithium battery to be predicted has not reached the cycle life prediction point, then the previous lithium battery to be predicted has not reached the cycle life prediction point, and more cycle data needs to be input until the main mode sequence of the hidden state matrix norm The last element value is greater than or equal to When , the hidden state matrix norm main mode sequence value reaches the life prediction point according to the formula The specific formula is as follows.

[0086] .

[0087] in, Representation sequence Value equal to Time The index results of is the target cycle life prediction point, For its collection The corresponding index in .

[0088] The cycle number sequence corresponding to the target health state of the lithium battery to be predicted is calculated according to the formula.

[0089] .

[0090] .

[0091] in, is the cycle number sequence corresponding to the target health state of the lithium battery to be predicted, are the parameters of the linear relationship model, The health status of the target to be targeted (set) No. elements, for all .

[0092] In actual application, the technical effects of this application are as follows.

[0093] In the training phase, the present application extracts the battery aging characteristics from the actual battery charge and discharge cycle data fragments, and constructs the battery aging characteristic time series based on the number of cycles; then extracts the internal hidden state matrix of battery aging from the battery aging characteristic time series through the Mamba model; amplifies the change law of the battery hidden state in the number of cycles by calculating the hidden state matrix norm sequence; uses the variational mode decomposition method to modally decompose the hidden state matrix norm sequence and extract the main mode sequence; determines the battery cycle life prediction point through correlation analysis; establishes a linear relationship model, uses the least squares method to learn the number of cycles corresponding to the battery reaching the cycle life prediction point, and the linear relationship model parameters between the target health state, and stores the learned parameters. In the prediction phase, input the battery cycle data to be predicted, and use the same method as the training phase to calculate the hidden state matrix norm main mode sequence. By judging whether the last value of the hidden state matrix norm main mode sequence is greater than the cycle life prediction point determined in the training phase, it is determined whether the cycle life prediction condition is met. If the cycle life prediction condition is met, the number of cycles when the cycle life prediction point is reached is brought into the linear relationship model of the training phase, and the corresponding parameters are matched to complete the health state (SOH) prediction.

[0094] Embodiment 2, the present application also provides a computer system, which may be a server or a terminal, and its internal structure diagram may be as follows: Fig. 9 As shown. The computer system includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer system is used to provide computing and control capabilities. The memory of the computer system includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer system is used to store video tag processing data. The input / output interface of the computer system is used to exchange information between the processor and an external device. The communication interface of the computer system is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the lithium battery life prediction method based on the Mamba model described above is implemented.

[0095] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer system to which the scheme of the present application is applied. The specific computer system may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0096] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0097] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0098] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A lithium battery life prediction method based on the Mamba model, characterized in that: The lithium battery life prediction method based on the Mamba model includes: Acquire an aging characteristic matrix of multiple lithium batteries; the aging characteristic matrix includes: a health state sequence and an aging characteristic sequence; the health state sequence is a sequence composed of the health state of each cycle number of each lithium battery; Substituting the aging feature matrices of multiple lithium batteries into the Mamba model for training, and extracting the hidden state matrix sequence of each lithium battery according to the trained Mamba model, specifically including: substituting the aging feature matrices of multiple lithium batteries into the Mamba model, taking the aging feature sequence as a training set, taking the health state sequence as a label, training the Mamba model, and obtaining a trained Mamba model; extracting the hidden state matrix of each cyclic charge and discharge segment of the lithium battery in the trained Mamba model, and constructing a hidden state matrix sequence of each lithium battery; the cyclic charge and discharge segment is the operation data of the voltage in the preset voltage range in the cyclic operation data of each lithium battery; the hidden state matrix sequence is a sequence composed of the hidden state matrices of different cycle numbers of multiple lithium batteries; the Mamba model is as follows: ； ； in, , , As the model input, output and hidden state respectively, Indicates lithium battery The aging characteristics of the number of cycles, Indicates lithium battery The health status of the number of cycles, Indicates lithium battery The hidden state of the number of cycles, Indicates lithium battery -1 cycle of hidden states, , , represents the learnable parameter matrix, , is the total cycle number of the lithium battery; According to the norm calculation and variational mode decomposition of the hidden state matrix sequence of each lithium battery, the cycle number sequence corresponding to the target cycle life prediction point of each lithium battery is obtained, which specifically includes: calculating the norm of the hidden state matrix of each cycle number in the hidden state matrix sequence of each lithium battery respectively to obtain the hidden state matrix norm sequence of each lithium battery; performing variational mode decomposition on the hidden state matrix norm sequence of each lithium battery to extract the main mode sequence of the hidden state matrix norm of each lithium battery; and calculating the main mode value when the main mode value reaches each candidate cycle life prediction point according to the main mode sequence of the hidden state matrix norm of each lithium battery. The number of cycles; the cycle number sequence corresponding to the prediction point with the largest correlation among the candidate cycle life prediction points is selected as the cycle number sequence corresponding to the target cycle life prediction point of each lithium battery; the target cycle life prediction point is the prediction point with the largest correlation among the candidate cycle life prediction points; the candidate cycle life prediction point is a prediction point in the norm sequence curve of the hidden state matrix of the lithium battery obtained after norm calculation that has a correlation with the cycle number sequence corresponding to the target health state of the lithium battery; the cycle number sequence corresponding to the target health state of the lithium battery is obtained according to the health state sequence index; Based on the cycle number sequence corresponding to the target health status of all lithium batteries and the cycle number sequence corresponding to the target cycle life prediction point of all lithium batteries, a linear relationship model for lithium battery life prediction is constructed; Obtain the cycle number sequence corresponding to the target cycle life prediction point of the lithium battery to be predicted, and substitute it into the linear relationship model of lithium battery life prediction to calculate the cycle number sequence corresponding to the target health state of the lithium battery to be predicted.

2. The lithium battery life prediction method based on the Mamba model according to claim 1 is characterized in that: Obtain the aging characteristic matrix of multiple lithium batteries, including: Acquire the cycle operation data of multiple lithium batteries; the cycle operation data at least includes: voltage, temperature, power change value and cycle number at different sampling times; The health status of each lithium battery at each cycle is calculated based on the cycle operation data of each lithium battery, and a health status sequence of each lithium battery is constructed; Extracting the aging characteristic sequence of each lithium battery according to the cycle operation data of each lithium battery; Based on the aging feature sequences of multiple lithium batteries and the health state sequences of multiple lithium batteries, an aging feature matrix of multiple lithium batteries is constructed.

3. The lithium battery life prediction method based on the Mamba model according to claim 1 is characterized in that: The calculation formula of the hidden state matrix norm sequence is as follows: ； ； in, is the hidden state matrix norm sequence, For the The norm of the hidden state matrix of the number of cycles, For lithium battery The hidden state matrix of the number of cycles Line The hidden state of the elements of the column, , It is the total cycle number of the lithium battery.

4. The lithium battery life prediction method based on the Mamba model according to claim 1, characterized in that: The calculation formula of the cycle number sequence corresponding to the target cycle life prediction point is as follows: ； in, represents the cycle number sequence corresponding to the target cycle life prediction point, represents the index of the maximum value in the sequence, Indicates lithium battery The target health state cycle number matrix No. List, Indicates lithium battery Cycle life prediction point cycle number matrix No. List, is a matrix No. The mean of the column, is a matrix No. The mean of the column, , N Indicates the total amount used for lithium batteries, Represents the number of candidate cycle life prediction points.

5. The lithium battery life prediction method based on the Mamba model according to claim 1, characterized in that: Based on the cycle number sequence corresponding to the target health status of all lithium batteries and the cycle number sequence corresponding to the target cycle life prediction point of all lithium batteries, a linear relationship model for lithium battery life prediction is constructed, which specifically includes: According to the cycle number sequence corresponding to the target cycle life prediction points of all lithium batteries, the target health state cycle number matrix is ​​constructed. According to the cycle number sequence corresponding to the target cycle life prediction point of all lithium batteries, a target cycle life prediction point cycle number matrix is ​​constructed; Based on the target health status cycle number matrix and the target cycle life prediction point cycle number matrix, the least squares method is used to fit the linear relationship model for lithium battery life prediction.

6. The lithium battery life prediction method based on the Mamba model according to claim 1, characterized in that: The linear relationship model for predicting the life of the lithium battery is as follows: ； in, Represents the target health state cycle number matrix No. List, Represents the target cycle life prediction point cycle number matrix of sequence, is the cycle number sequence corresponding to the target cycle life prediction point, and They are and The slope and intercept of the linear relationship model.

7. A computer system comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the lithium battery life prediction method based on the Mamba model described in any one of claims 1 to 6.

Images