A lithium battery energy storage monitoring method and system

Abstract

The application discloses a lithium battery energy storage monitoring method and system, which comprises the following steps: obtaining a historical solid-phase diffusion coefficient sequence and a historical liquid-phase impedance sequence, and establishing a coupling correlation matrix therebetween; obtaining a real-time solid-phase diffusion coefficient and a real-time liquid-phase impedance through online identification; selecting a first candidate period through sliding window correlation analysis; selecting a second candidate period through impedance spectrum shape similarity calculation; calculating a current capacity loss amount by using a nonlinear diffusion-impedance coupling model; determining a current calibration capacity based on an initial calibration capacity and an aging factor; and determining a remaining available capacity based on the current calibration capacity and the current capacity loss amount. By extracting the diffusion-impedance coupling correlation matrix from historical data and deeply embedding the whole process of parameter identification, candidate period screening, capacity loss calculation and remaining capacity evaluation, the accuracy and robustness of the estimation of the remaining available capacity of the lithium battery are significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of battery management technology, and in particular relates to a lithium battery energy storage monitoring method and system. Background Technology

[0002] Lithium-ion batteries are widely used in energy storage systems, and accurate monitoring of their health status and remaining usable capacity is crucial for the safe and efficient operation of these systems. During the cyclic charging and discharging process of lithium-ion batteries, the solid-phase diffusion coefficient and liquid-phase impedance are two core physical parameters reflecting the degradation of electrode material structure and changes in electrolyte transport characteristics, and there is a correlation between the two due to electrochemical coupling effects.

[0003] In existing technologies, online identification methods such as extended Kalman filtering are typically used to estimate these two parameters separately, and then empirical models are used to calculate capacity loss. However, existing methods neglect the coupling characteristics of the solid-phase diffusion coefficient and liquid-phase impedance during the degradation process, resulting in the capacity estimation accuracy being greatly affected by operating condition fluctuations, and lacking the mining and utilization of coupling patterns in historical operating data. Therefore, how to extract diffusion-impedance coupling characteristics from historical data and deeply embed them into various stages of condition identification and capacity assessment is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0004] This invention provides a lithium battery energy storage monitoring method and system to solve the problem of insufficient capacity estimation accuracy in the prior art.

[0005] In a first aspect, the present invention provides a lithium battery energy storage monitoring method, comprising: Obtain the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of lithium battery in each historical charge-discharge cycle, and establish the coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence in the same historical charge-discharge cycle. The real-time terminal voltage sequence and real-time current sequence between the target time and the current time are determined. The real-time terminal voltage sequence and real-time current sequence are used as the observation input of the preset extended Kalman filter to identify the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance online. The real-time solid-phase diffusion coefficient is compared with the historical solid-phase diffusion coefficient sequence in each historical charge-discharge cycle by a sliding window correlation analysis, and the historical charge-discharge cycle with a correlation coefficient greater than a preset first threshold is selected as the first candidate cycle. Within the first candidate period, the impedance spectrum shape similarity between the real-time liquid phase impedance and each historical liquid phase impedance sequence is calculated, and the historical charge-discharge period with a similarity greater than a preset second threshold is selected as the second candidate period. Based on the historical health status parameters corresponding to the second candidate period, combined with the ratio of the current real-time solid-phase diffusion coefficient to the historical solid-phase diffusion coefficient, and the ratio of the current real-time liquid phase impedance to the historical liquid phase impedance, a nonlinear diffusion-impedance coupling model is used to calculate the current capacity loss of the lithium battery between the target time and the current time. The current rated capacity is determined based on the initial rated capacity of the lithium battery and the aging factor in the current charge-discharge cycle, wherein the aging factor is determined according to the ratio of the real-time solid-phase diffusion coefficient to the initial solid-phase diffusion coefficient. Based on the current calibrated capacity and the current capacity loss, the remaining usable capacity of the lithium battery at the current moment is determined.

[0006] In a second aspect, the present invention provides a lithium battery energy storage monitoring system, comprising: The acquisition module is configured to acquire the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of the lithium battery in each historical charge-discharge cycle, and establish a coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence in the same historical charge-discharge cycle. The identification module is configured to determine the real-time terminal voltage sequence and real-time current sequence between the target time and the current time, and use the real-time terminal voltage sequence and real-time current sequence as the observation input of the preset extended Kalman filter to identify the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance online. The first selection module is configured to perform sliding window correlation analysis on the real-time solid-phase diffusion coefficient and the historical solid-phase diffusion coefficient sequence in each historical charge-discharge cycle, and select the historical charge-discharge cycle with a correlation coefficient greater than a preset first threshold as the first candidate cycle. The second selection module is configured to perform impedance spectrum shape similarity calculation between the real-time liquid phase impedance and each historical liquid phase impedance sequence within the first candidate period, and select the historical charge-discharge period with a similarity greater than a preset second threshold as the second candidate period. The calculation module is configured to calculate the current capacity loss of the lithium battery between the target time and the current time based on the historical health status parameters corresponding to the second candidate period, combined with the ratio of the current real-time solid-phase diffusion coefficient to the historical solid-phase diffusion coefficient, and the ratio of the current real-time liquid phase impedance to the historical liquid phase impedance, using a nonlinear diffusion-impedance coupling model. The determination module is configured to determine the current calibration capacity based on the initial calibration capacity of the lithium battery and the aging factor in the current charge-discharge cycle, wherein the aging factor is determined according to the ratio of the real-time solid-phase diffusion coefficient to the initial solid-phase diffusion coefficient. The output module is configured to determine the remaining usable capacity of the lithium battery at the current moment based on the current calibrated capacity and the current capacity loss.

[0007] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the lithium battery energy storage monitoring method of any embodiment of the present invention.

[0008] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the steps of the lithium battery energy storage monitoring method of any embodiment of the present invention.

[0009] The lithium battery energy storage monitoring method and system of this application acquires historical solid-phase diffusion coefficient sequences and historical liquid-phase impedance sequences and establishes a coupling correlation matrix between them; identifies real-time solid-phase diffusion coefficients and real-time liquid-phase impedances online; selects a first candidate period through sliding window correlation analysis; selects a second candidate period through impedance spectrum shape similarity calculation; calculates the current capacity loss using a nonlinear diffusion-impedance coupling model; determines the current calibration capacity based on the initial calibration capacity and aging factor; and determines the remaining usable capacity based on the current calibration capacity and current capacity loss. By extracting the diffusion-impedance coupling correlation matrix from historical data and deeply embedding it into the entire process of parameter identification, candidate period selection, capacity loss calculation, and remaining capacity assessment, the accuracy and robustness of lithium battery remaining usable capacity estimation are significantly improved. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A flowchart of a lithium battery energy storage monitoring method provided in an embodiment of the present invention; Figure 2 This is a structural block diagram of a lithium battery energy storage monitoring system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] Please see Figure 1 The diagram shows a flowchart of a lithium battery energy storage monitoring method according to this application.

[0014] like Figure 1 As shown, the lithium battery energy storage monitoring method specifically includes the following steps: Step S101: Obtain the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of the lithium battery in each historical charge-discharge cycle, and establish the coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence in the same historical charge-discharge cycle.

[0015] In this step, outliers are removed from the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence within a certain historical charge-discharge cycle. The historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence after outlier removal are aligned by timestamp and then z-score normalized to obtain the first normalized sequence and the second normalized sequence. For each integer delay step d from -L to +L, the second normalized sequence is shifted d steps in the positive direction of the time axis. The Pearson correlation coefficient between the shifted second normalized sequence and the first normalized sequence is calculated, thus obtaining the curve of the Pearson correlation coefficient changing with the delay step d, where L is the maximum delay step. Extract the global maximum correlation coefficient from the curve, and the optimal delay step corresponding to the global maximum correlation coefficient, and determine whether the global maximum correlation coefficient is less than a preset minimum confidence threshold; If the global maximum correlation coefficient is less than the minimum confidence threshold, the coupling correlation matrix is ​​directly defined as a 2×2 identity matrix, where the main diagonal elements of the identity matrix are 1 and the secondary diagonal elements are 0. If the global maximum correlation coefficient is not less than the minimum confidence threshold, then the width of the curve at the half-peak height is calculated, and the ratio of the width to the maximum delay range 2L+1 is used as the coupling stability index, and the direction of influence is determined according to the sign of the optimal delay step. The product of the global maximum correlation coefficient and the coupling stability index is defined as the weighted coupling strength. A 2×2 coupling correlation matrix is ​​constructed based on the weighted coupling strength and the influence direction. The first row and first column of the coupling correlation matrix are fixed to 1, the first row and second column are equal to the weighted coupling strength multiplied by the sign value of the influence direction, the second row and first column are equal to the weighted coupling strength multiplied by the opposite sign value of the influence direction, and the second row and second column are fixed to 1.

[0016] In one specific embodiment, the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of the lithium battery are first obtained through offline parameter identification over multiple complete historical charge-discharge cycles. For the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence within a certain historical charge-discharge cycle, outliers in the sequence are identified and removed using the 3σ criterion or interquartile range method. The two sequences after outlier removal are aligned by timestamps and z-score normalization is performed on each sequence. That is, the mean of the sequence is subtracted from each element in the sequence and then divided by the standard deviation of the sequence to obtain the first normalized sequence and the second normalized sequence with a mean of 0 and a standard deviation of 1.

[0017] Set a maximum delay step number L, for example, L=50. For each integer delay step number d from -L to +L, shift the second normalized sequence as a whole d steps in the positive direction of the time axis, so that the i-th element of the second normalized sequence is aligned with the (i+d)-th element of the first normalized sequence. Calculate the Pearson correlation coefficient between the shifted second normalized sequence and the first normalized sequence within the overlap interval, obtaining a correlation coefficient value. After traversing all integer delay steps d, a correlation coefficient curve is obtained with the delay step number d as the horizontal axis and the Pearson correlation coefficient as the vertical axis.

[0018] The global maximum correlation coefficient and the optimal delay step corresponding to the global maximum correlation coefficient are extracted from the correlation coefficient curve. If the global maximum correlation coefficient is less than the preset minimum confidence threshold, such as 0.3, it indicates that there is no significant linear coupling relationship between the solid phase diffusion coefficient and the liquid phase impedance within this historical charge-discharge cycle. The coupling correlation matrix is ​​directly defined as a 2×2 identity matrix, that is, a matrix with 1s on the main diagonal and 0s on the secondary diagonal.

[0019] If the global maximum correlation coefficient is not less than the minimum confidence threshold, then the half-peak height position is determined on the correlation coefficient curve, which consists of 2L+1 discrete data points. Composition, among which, The Pearson correlation coefficient corresponding to the delay step d; Let the global maximum correlation coefficient be . The half-peak height value is On the correlation coefficient curve, look to the left and right respectively for the first time the correlation coefficient value is lower than... The adjacent integer delay step pairs are as follows: The number of delay steps corresponding to the peak On the left side, from Iterate through the delay steps in the negative direction one by one to find the first one that satisfies the condition. and Let d be the integer delay steps. ; The number of delay steps corresponding to the peak On the right side, from Iterate through the delay steps in the positive direction one by one to find the first one that satisfies the condition. and Let d be the integer delay steps. ; In the interval and The correlation coefficient is obtained by performing linear interpolation using two adjacent discrete points. The corresponding delay steps and And calculate the width of the correlation coefficient curve at half-peak height. ; Width The ratio of the maximum delay range 2L+1 to the maximum delay range is used as a coupling stability index. The closer the index is to 0, the more concentrated the coupling relationship is and the higher the stability.

[0020] Simultaneously, the direction of influence is determined based on the sign of the optimal delay step number: if the optimal delay step number is positive, it indicates that the change in liquid phase impedance lags behind the change in solid phase diffusion coefficient, and the sign value of the direction of influence... Add 1; if the optimal delay step is negative, then Take -1.

[0021] The product of the global maximum correlation coefficient and the coupling stability index is defined as the weighted coupling strength. This value comprehensively reflects the strength and stability of the coupling. A 2×2 coupling correlation matrix is ​​constructed based on the weighted coupling strength and influence direction. The element in the first row and first column of the matrix is ​​fixed at 1, and the element in the first row and second column is equal to... The element in the second row and first column equals The element in the second row and second column is fixed to 1.

[0022] In summary, the coupling strength, coupling stability, and causal influence direction between the solid-phase diffusion coefficient and the liquid-phase impedance were quantitatively extracted from historical operating data and encapsulated into a 2×2 coupling correlation matrix. This matrix not only characterizes the dynamic dependence of the two degradation parameters over time but also filters out the interference of random fluctuations on the extraction of coupling features through the coupling stability index, resulting in higher robustness and repeatability of the extracted coupling features. The Kalman gain is directly modified in the extended Kalman filter, forcing the correction direction of the state estimation to satisfy the historical coupling relationship, thereby effectively suppressing the parameter estimation divergence problem caused by measurement noise and operating condition fluctuations. Simultaneously, the weighted coupling strength and influence direction provided by this coupling correlation matrix also provide unified and self-consistent coupling prior information for the subsequent definition of impedance spectrum shape similarity, the construction of nonlinear capacity loss driving forces, and the adaptive adjustment of remaining capacity assessment confidence. This enables all aspects of the monitoring method to coordinate and consistently utilize the inherent physical laws of battery degradation, significantly improving the accuracy and robustness of capacity estimation.

[0023] Step S102: Determine the real-time terminal voltage sequence and real-time current sequence between the target time and the current time. Use the real-time terminal voltage sequence and real-time current sequence as the observation input of the preset extended Kalman filter to identify the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance online.

[0024] In this step, a state-space model is established with the solid-phase diffusion coefficient and liquid-phase impedance as state variables, the real-time current sequence as input, the real-time terminal voltage sequence as observation, and the state vector and state error covariance matrix are initialized. At each sampling time, the Jacobian matrix of the state transition matrix and the observation matrix is ​​calculated, and the time update step of the extended Kalman filter is performed to obtain the prior state estimate and the prior error covariance matrix. The standard Kalman gain is calculated using the residual between the real-time terminal voltage sequence and the observed predicted value. Multiplying the standard Kalman gain by the coupling correlation matrix yields a modified gain matrix that satisfies the diffusion-impedance coupling constraint. The prior state estimate is updated using the corrected gain matrix and the residual to obtain the posterior state estimate; The solid-phase diffusion coefficient component in the posterior state estimate is used as the real-time solid-phase diffusion coefficient, and the liquid-phase impedance component in the posterior state estimate is used as the real-time liquid-phase impedance.

[0025] In one specific embodiment, the target time for the start of monitoring and the current real-time time are first determined, and the real-time terminal voltage sequence and real-time current sequence are collected at equal time intervals between the target time and the current time. The first-order RC equivalent circuit model of the lithium battery is used as the physical terminal voltage response model, and an extended Kalman filter framework is constructed around the online identification of the solid-phase diffusion coefficient and the liquid-phase impedance.

[0026] The terminal voltage equation of the first-order RC equivalent circuit model is: ; in, Let be the terminal voltage at time t. Let be the current at time t. The open-circuit voltage associated with the state of charge (SOC). For ohmic internal resistance, It is the RC polarization voltage; Liquid phase resistance Z and ohmic internal resistance There is a definite mapping relationship between the polarization parameter and the solid-phase diffusion coefficient D, which determines the solid-phase diffusion rate of lithium ions inside the electrode active material particles, thereby affecting the dynamic change of the state of charge (SOC) and the transient response of the polarization voltage.

[0027] To identify the solid-phase diffusion coefficient D and liquid-phase impedance Z online, a state-space model is constructed with D and Z as the parameters to be identified. The state vector is defined as follows:

[0028] ; in, For state vectors, The first component of the state vector. , representing the natural logarithm of the solid-phase diffusion coefficient D. It is the second component of the state vector. , representing the natural logarithm of the liquid phase impedance; The real-time current sequence is used as the deterministic input to the state-space model, and the real-time terminal voltage sequence is used as the observation. Since the solid-phase diffusion coefficient D and the liquid-phase impedance Z change extremely slowly within a single sampling interval, the state transition equation employs an identity mapping:

[0029] ; in, Let k+1 be the state vector at the (k+1)th sampling time. Let k be the state vector at the k-th sampling time. The process noise at the k-th sampling time follows a Gaussian distribution with zero mean and covariance matrix Q. Since the state transition equation is an identity mapping, the state transition matrix is ​​a 2×2 identity matrix.

[0030] The observation equation, based on the first-order RC equivalent circuit model described above, is expressed as a nonlinear mapping of the terminal voltage under the current state vector: ; in, This represents the real-time terminal voltage observation at the k-th sampling time. Let be the real-time current input value at the k-th sampling time. The observation noise at the k-th sampling time follows a one-dimensional Gaussian distribution with zero mean and variance R. This is the terminal voltage response function; When filtering starts, an initial guess value is assigned to the state vector. Its first component The second component is the natural logarithm of the initial guess of the solid diffusion coefficient. Let be the natural logarithm of the initial guessed value of the liquid phase impedance. And assign an initial diagonal matrix to the state error covariance matrix:

[0031] ; in, The initial state error covariance matrix, i.e., the initial diagonal matrix. The element in the first row and first column of the initial state error covariance matrix represents the first component of the state vector at the initial time step. The estimated variance, The element in the second row and second column of the initial state error covariance matrix represents the estimated variance of the second component of the state vector at the initial time. At each sampling time, perform the following iterative steps: Calculate the Jacobian matrix of the observation matrix Jacobian matrix Terminal voltage response function For the state vector The partial derivatives in prior state estimation The value at that location is expressed as: ; in, For terminal voltage response pair The partial derivatives, For terminal voltage response pair The partial derivatives; Execution time update: The prior state estimate equals the posterior state estimate of the previous time step, expressed as: ; in, For prior state estimation at the k-th sampling time, i.e., based on the information at the (k-1)-th time, the state vector at the k-th time is... The predicted value, The posterior state estimate at the k-1 sampling time is the optimal state estimate obtained after updating the terminal voltage observation at the k-1 time. The prior error covariance matrix is ​​updated as follows: ; in, Let be the prior error covariance matrix at the k-th sampling time, representing the prior state estimation. Quantification of uncertainty Let be the posterior error covariance matrix at the (k-1)th sampling time, representing the posterior state estimate. Quantification of uncertainty Let be the process noise covariance matrix, representing the state transition equation. Mid-process noise Statistical characteristics; Calculate the observation residuals and standard Kalman gain: Determine the observation residuals based on the observed predictions, as expressed by the following expressions: ; ; in, For the predicted value of the observation at the k-th sampling time, The observation residual at the k-th sampling time; The standard Kalman gain is calculated using the following expression: ; in, For the standard Kalman gain at the k-th sampling time, Let be the prior error covariance matrix at the k-th sampling time. It is the transpose symbol. To observe the noise variance, Invert a matrix; Apply coupling constraints: Obtain the coupling correlation matrix established in step S101. The expression is: ; Standard Kalman gain Coupling correlation matrix Multiplying these matrices yields the corrected gain matrix that satisfies the diffusion-impedance coupling constraint. The expression is: ; Perform measurement updates: Updating the state estimate using the corrected gain matrix and observed residuals yields the posterior state estimate, expressed as: ; The posterior error covariance matrix is ​​updated synchronously, and the expression is: ; in, It is a 2×2 identity matrix; Extracting physical parameters: posterior state estimation First component Second component They are respectively and The optimal estimate at time k. For the first component. Second component By taking the exponents respectively, we can obtain the real-time parameter values ​​in physical space:

[0032] ; ; in, This represents the real-time solid-phase diffusion coefficient at the current moment. This represents the real-time liquid phase impedance at the current moment. At the next sampling time, the process returns to continue, enabling continuous online identification of the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance throughout the entire monitoring period.

[0033] Step S103: Perform sliding window correlation analysis on the real-time solid-phase diffusion coefficient and the historical solid-phase diffusion coefficient sequence in each historical charge-discharge cycle, and select the historical charge-discharge cycle with a correlation coefficient greater than a preset first threshold as the first candidate cycle.

[0034] In this step, it is assumed that during the monitoring period from the target time to the current time, a total of M real-time solid diffusion coefficient values ​​are output, forming a real-time solid diffusion coefficient sequence. ,in, Let be the real-time solid-phase diffusion coefficient at the i-th sampling time. , This represents the total number of sampling points during the current monitoring period. For the j-th historical charge-discharge cycle, its historical solid-phase diffusion coefficient sequence is: ; in, Let j be the historical solid-phase diffusion coefficient sequence for the j-th historical charge-discharge cycle. Let be the historical solid-phase diffusion coefficient at the t-th sampling time in the j-th historical period. , , The total number of historical cycles, The total number of sampling points in the j-th historical period is usually... That is, the length of the historical sequence is much greater than the length of the real-time sequence; The real-time solid-phase diffusion coefficient sequence of length M As a window, with a length of historical sequence Starting from the initial position s=1, slide the window backward with a step size of 1 until the end of the window aligns with the end of the history sequence. That is, the range of values ​​for s is: ; For each sliding starting position s, extract a subsequence from the s-th element to the (s+M-1)-th element in the historical sequence. ; Calculate the subsequence and the real-time solid-phase diffusion coefficient sequence. Pearson correlation coefficient between ; After traversing all possible starting positions s, a set of correlation coefficient values ​​is obtained. The maximum value among them is taken as the correlation coefficient between the j-th historical charge-discharge cycle and the current real-time diffusion degradation mode; Historical charge-discharge cycles with a correlation coefficient greater than a preset first threshold are selected as the first candidate cycles.

[0035] In summary, this embodiment utilizes sliding window correlation analysis, using the real-time solid-phase diffusion coefficient sequence as a template, to search for the best matching segment in the long sequence of each historical cycle. The maximum Pearson correlation coefficient is used as the similarity measure, fully considering the potential for phased changes in the diffusion degradation process within historical charge-discharge cycles, thus avoiding misjudgments of similarity due to time axis misalignment. Simultaneously, using the solid-phase diffusion coefficient as the first-level screening criterion efficiently filters out historical cycles whose diffusion degradation modes are significantly different from the current state. This provides a prerequisite for subsequent steps to perform fine-grained impedance spectrum shape matching within a smaller candidate range, reducing computational load and improving overall monitoring efficiency.

[0036] Step S104: Within the first candidate period, the impedance spectrum shape similarity between the real-time liquid phase impedance and each historical liquid phase impedance sequence is calculated, and the historical charge-discharge period with a similarity greater than a preset second threshold is selected as the second candidate period.

[0037] In this step, the real-time solid-phase diffusion coefficient sequence and the real-time liquid-phase impedance sequence are used as inputs. Based on the maximum delay step L, the Pearson correlation coefficient between the real-time liquid-phase impedance sequence and the real-time solid-phase diffusion coefficient sequence is calculated at each integer delay step d to obtain the real-time correlation coefficient curve. The real-time half-peak width of the real-time correlation coefficient curve is then extracted. For each first candidate period, extract the historical correlation coefficient curve corresponding to the first candidate period when establishing the coupling correlation matrix of the first candidate period, and obtain the historical half-peak width of the historical correlation coefficient curve. The real-time correlation coefficient curve and the historical correlation coefficient curve are normalized to their maximum values, so that the peak values ​​of the real-time correlation coefficient curve and the historical correlation coefficient curve are scaled to 1, thus obtaining the normalized real-time correlation coefficient curve and the normalized historical correlation coefficient curve. On a coordinate system with integer delay steps d as the horizontal axis and Pearson correlation coefficient as the vertical axis, calculate the area of ​​overlap between the normalized real-time correlation coefficient curve and the normalized historical correlation coefficient curve, and calculate the area of ​​union of the areas enclosed by the normalized real-time correlation coefficient curve and the normalized historical correlation coefficient curve. The ratio of the overlapping area to the union area is used as the degree of overlap of the coupling mode between the real-time liquid phase impedance sequence and the historical liquid phase impedance sequence within the first candidate period. Calculate the absolute value of the difference between the real-time half-peak width and the historical half-peak width, and determine the maximum value between the real-time half-peak width and the historical half-peak width. Subtract the ratio of the absolute value to the maximum value from 1 to obtain the coupling stability similarity. The product of the coupling mode overlap and the coupling stability similarity is taken as the impedance spectrum shape similarity. The historical charge-discharge cycles in which the impedance spectrum shape similarity is greater than the preset second threshold are selected as the second candidate cycles.

[0038] In one specific embodiment, the real-time solid-phase diffusion coefficient sequence obtained from the online identification in step S102 is acquired. and real-time liquid phase impedance sequence ,in, Let be the real-time liquid phase impedance at the i-th sampling point.

[0039] Using the same maximum delay step number L as in step S101 (e.g., L=50), calculate the Pearson correlation coefficient between the real-time liquid phase impedance sequence and the real-time solid phase diffusion coefficient sequence for each integer delay step d∈[−L,L]. For a given delay step number d, shift the real-time liquid phase impedance sequence relative to the real-time solid phase diffusion coefficient sequence by d sampling points, calculate the Pearson correlation coefficient between the two sequences in the overlapping interval, and obtain a correlation coefficient value. After traversing 2L+12L+1 integer delay steps, obtain a real-time correlation coefficient curve with the delay step number d as the horizontal axis and the Pearson correlation coefficient as the vertical axis, denoted as... .

[0040] In the real-time correlation coefficient curve Find its maximum value, i.e., the global maximum correlation coefficient. Take the half-peak height value on the vertical axis. Find the correlation coefficient value equal to on the real-time correlation coefficient curve. The number of delay steps corresponding to the two points and ( < The real-time half-peak width is then... .

[0041] For the j-th historical charge-discharge cycle in the first candidate cycle set, the historical correlation coefficient curve of that cycle has already been calculated and stored when establishing the coupling correlation matrix of that cycle in step S101. This refers to the curve showing the Pearson correlation coefficient between the historical liquid phase impedance sequence and the historical solid phase diffusion coefficient sequence as a function of the delay step d. This curve, along with the historical half-peak width of the j-th historical charge-discharge cycle, is retrieved directly from the stored data. .

[0042] For real-time correlation coefficient curves The historical correlation coefficient curve for the j-th historical charge-discharge cycle. The maximum value is normalized to obtain the normalized real-time correlation coefficient curve. The normalized historical correlation coefficient curve of the j-th historical charge-discharge cycle ; On a two-dimensional plane with integer delay steps d as the horizontal axis and the normalized Pearson correlation coefficient as the vertical axis, the normalized real-time correlation coefficient curve is plotted. The normalized historical correlation coefficient curve of the j-th historical charge-discharge cycle Plotted in the same coordinate system; The area of ​​overlap between two curves is calculated using the following expression: , in, For the normalized real-time correlation coefficient curve The normalized historical correlation coefficient curve of the j-th historical charge-discharge cycle The overlapping area between them; The area of ​​the union of the regions enclosed by the two curves is expressed as: ; in, For the normalized real-time correlation coefficient curve The normalized historical correlation coefficient curve of the j-th historical charge-discharge cycle The area of ​​the union of the enclosed regions; The ratio of the overlapping area to the union area is defined as the coupling mode overlap degree. ; Calculate the absolute value of the difference between the real-time half-peak width and the historical half-peak width of the j-th historical charge-discharge cycle. Then, the maximum value between the real-time half-peak width and the historical half-peak width of the j-th historical charge-discharge cycle is determined. The ratio of the absolute value to the maximum value is then subtracted from 1 to obtain the coupling stability similarity. ; Coupling mode overlap Similarity with coupling stability The product is defined as the similarity in impedance spectrum shape between the real-time liquid phase impedance sequence and the historical liquid phase impedance sequence within the j-th first candidate period. ; Historical charge-discharge cycles with impedance spectrum shape similarity greater than a preset second threshold are selected as the second candidate cycles.

[0043] In summary, in this embodiment, the shape characteristics of the correlation coefficient curves between liquid phase impedance and solid phase diffusion coefficient are compared. Quantitative shape matching of the two correlation coefficient curves is performed using two complementary indices: coupling mode overlap and coupling stability similarity. The coupling mode overlap effectively eliminates interference from the absolute magnitude of coupling strength by using the ratio of the overlapping area to the union area under the normalized curve, focusing on the consistency of the delay structure itself. The coupling stability similarity captures the similarity of the degree of coupling concentration through the relative difference in the half-peak width. Multiplying the two yields the impedance spectrum shape similarity. Only when both the coupling delay mode and coupling stability are highly similar is the historical period included in the second candidate period, thus ensuring that the selected historical reference conditions are highly consistent with the current battery state at the aging degradation mechanism level. This fine matching at the mechanism level provides a high-quality reference sample for calculating the capacity loss using the nonlinear diffusion-impedance coupling model in subsequent step S105, effectively improving the accuracy and reliability of capacity loss estimation.

[0044] Step S105: Based on the historical health status parameters corresponding to the second candidate period, combined with the ratio of the current real-time solid-phase diffusion coefficient to the historical solid-phase diffusion coefficient, and the ratio of the current real-time liquid-phase impedance to the historical liquid-phase impedance, a nonlinear diffusion-impedance coupling model is used to calculate the current capacity loss of the lithium battery between the target time and the current time.

[0045] In this step, the logarithmic decay rate of the solid diffusion coefficient is calculated based on the ratio of the real-time solid diffusion coefficient to the initial solid diffusion coefficient, and the logarithmic growth rate of the liquid phase impedance is calculated based on the ratio of the real-time liquid phase impedance to the initial liquid phase impedance. The product of the optimal delay step number and the data sampling time interval of the historical charge and discharge cycle is used as the characteristic time delay of the change in the solid phase diffusion coefficient relative to the change in the liquid phase impedance; Based on the characteristic time delay, calculate the coupling degradation driving force after time delay alignment, and input the coupling degradation driving force into the pre-constructed capacity loss rate function; Using the target time as the lower limit of integration and the current time as the upper limit of integration, the capacity loss rate function is numerically integrated to obtain the current capacity loss.

[0046] In one specific embodiment, the initial solid-phase diffusion coefficient of the lithium battery at the time of manufacture is obtained. and initial liquid phase impedance And step S102 at the current time Real-time solid-phase diffusion coefficient obtained online and real-time liquid phase impedance Define the logarithmic decay rate of the solid-phase diffusion coefficient. Logarithmic growth rate of liquid phase impedance They are respectively:

[0047] , ; Extract the optimal delay step number and weighted coupling strength corresponding to the coupling correlation matrix established in step S101 for each historical charge-discharge cycle from the second candidate cycle set. Take the average value of the parameters corresponding to all second candidate cycles as the input to the current coupling degradation model:

[0048] ; ; in, The average optimal delay steps, The average weighted coupling strength, The number of historical charge-discharge cycles in the second candidate cycle set. Let be the optimal delay step number for the j-th historical charge-discharge cycle in the second candidate cycle set. The weighted coupling strength of the j-th historical charge-discharge cycle in the second candidate cycle set; Average optimal delay steps Data sampling time interval Multiplying these yields the characteristic time delay of the change in solid diffusion coefficient relative to the change in liquid impedance. ; To incorporate the temporal coupling relationship between diffusion degradation and impedance growth into the calculation of capacity loss rate, a time-delay aligned coupling degradation driving force is constructed. The coupling degradation driving force at the current time t is... Defined as:

[0049] ; in, Let be the logarithmic decay rate of the solid-phase diffusion coefficient at the current time t. For backtracking feature time delay The logarithmic growth rate of the liquid phase impedance after that, i.e., using The logarithmic growth rate is calculated from the real-time liquid phase impedance value at time t. The coupling degradation driving force at the current time t Input the pre-constructed capacity loss rate function, expressed as: ; in, For capacity loss rate, The baseline loss rate constant was pre-calibrated through accelerated battery aging tests. To compensate for the nonlinearity exponent, which is dimensionless, the preset value is typically taken as 1.0~2.0. For a sign function, when When +1 is taken, When -1 is taken, Take 0 at the time; With target time The lower limit of integration, the current time. As the upper limit of integration, numerical integration is performed on the capacity loss rate function to obtain the current capacity loss. .

[0050] In summary, the model in this step starts from the inherent physical coupling mechanism of battery degradation, makes full use of the diffusion-impedance causal relationship information contained in historical operating data, significantly improves the accuracy and physical interpretability of capacity loss estimation, and provides a more reliable basis for subsequent assessment of remaining usable capacity.

[0051] Step S106: Determine the current rated capacity based on the initial rated capacity of the lithium battery and the aging factor in the current charge-discharge cycle, wherein the aging factor is determined according to the ratio of the real-time solid-phase diffusion coefficient to the initial solid-phase diffusion coefficient.

[0052] In this step, the final value of the historical solid-phase diffusion coefficient, the final value of the historical liquid-phase impedance, and the historical capacity retention rate are extracted from all historical charge-discharge cycles at the end of each historical charge-discharge cycle. A three-dimensional historical aging state point set is constructed with the final value of the historical solid-phase diffusion coefficient and the final value of the historical liquid-phase impedance as coordinates and the historical capacity retention rate as the response value. The real-time solid-phase diffusion coefficient and the real-time liquid-phase impedance at the current moment constitute the current aging state point. In the three-dimensional historical aging state point set, the K historical aging state points that are closest to the current aging state point are selected using Euclidean distance as a metric, where K is a preset integer. Calculate the inverse distance weighted average of the historical capacity retention rates corresponding to the K historical aging state points, and use the inverse distance weighted average as the benchmark aging factor; Obtain the equivalent full cycle count completed up to the current time of the current charge / discharge cycle, and the average equivalent full cycle count completed for the K historical charge / discharge cycles corresponding to the K historical aging state points, and calculate the time scale correction factor; Multiply the baseline aging factor by the time scale correction factor to obtain the aging factor for the current charge / discharge cycle; The current calibrated capacity is obtained by multiplying the initial calibrated capacity by the aging factor of the current charge / discharge cycle.

[0053] In one specific embodiment, for the j-th historical charge-discharge cycle, the final value of the historical solid-phase diffusion coefficient at the end of the j-th historical charge-discharge cycle is extracted. Historical liquid phase impedance final value and historical capacity retention rate ,in The actual usable capacity of the battery at the end of the j-th historical charge-discharge cycle. Initial calibration capacity; by For two-dimensional coordinates, with To determine the response value, a three-dimensional historical aging state point set is constructed. ; Step S102 at the current time The real-time solid-phase diffusion coefficient obtained by identification and real-time liquid phase impedance This constitutes the current aging state point. ; In the three-dimensional historical aging state point set In this process, the current aging state point is calculated using Euclidean distance as the metric. Two-dimensional Euclidean distance between each historical aging state point The points are sorted by distance from smallest to largest, and the K historical aging state points with the smallest distance are selected to form the nearest neighbor set. ; For the nearest neighbor set The m-th nearest neighbor in the array, Its corresponding historical capacity retention rate is The Euclidean distance from the current aging state point is Define the weight of the nearest neighbor. Normalized value of the reciprocal of its distance:

[0054] ; Based on the capacity retention rate of each nearest neighbor point, a weighted average is performed according to their respective weights to obtain the baseline aging factor. ; Get the equivalent number of full cycles completed up to the current time in the current charge / discharge cycle. The equivalent total number of cycles is obtained by statistically analyzing the charge and discharge process within the current cycle using the ampere-hour integration method. For the nearest neighbor set Given K historical charge-discharge cycles, obtain the total equivalent full cycle count at the end of each historical charge-discharge cycle. Calculate its arithmetic mean ; Define time scale correction factor for: ; in, This is a preset time acceleration index, dimensionless, typically between 0.5 and 1.0; Base aging factor With time scale correction factor Multiply to obtain the aging factor for the current charge / discharge cycle. The initial rated capacity of the lithium battery is multiplied by the aging factor of the current charge-discharge cycle to obtain the current rated capacity. .

[0055] In summary, by using the solid-phase diffusion coefficient and liquid-phase impedance at the end of the historical cycle as two-dimensional feature coordinates and the capacity retention rate as the response value, a three-dimensional historical aging state point set is formed, which completely preserves the multi-dimensional information of the battery aging state. Using the K-nearest neighbor inverse distance weighted interpolation method, based on the position of the currently identified diffusion-impedance parameters in the feature space, several historical samples with the closest aging trajectories are adaptively selected, and the capacity retention rate is weighted according to their distance, fully capturing the local nonlinear mapping relationship between the aging state and capacity degradation, avoiding model bias caused by a preset fixed function form. Furthermore, a time-scale correction factor based on the equivalent full cycle count is introduced, effectively correcting the time-scale differences in the aging process caused by different charge / discharge depths and rates, enabling the aging factor to accurately reflect the current battery's true aging degree relative to historical samples. Finally, the initial calibrated capacity is multiplied by the aging factor to obtain the current calibrated capacity, significantly improving the accuracy and robustness of the calibrated capacity estimation, and providing a reliable benchmark for the accurate calculation of the remaining usable capacity.

[0056] Step S107: Based on the current calibrated capacity and the current capacity loss, determine the remaining usable capacity of the lithium battery at the current moment.

[0057] In this step, the historical capacity loss corresponding to all the second candidate periods is obtained, and the impedance spectrum shape similarity of the second candidate periods is used as a weight to construct the weighted empirical cumulative distribution function of the historical capacity loss. Calculate the adaptive confidence level based on the coupling stability index; On the weighted empirical cumulative distribution function, the upper quantile corresponding to the adaptive confidence level is extracted as the basic robust capacity loss value; Calculate the safety margin expansion factor, and multiply the basic robust capacity loss value by the safety margin expansion factor to obtain the robust capacity loss value; Determine the baseline remaining capacity based on the current calibrated capacity and the robust capacity loss value; Obtain the current battery surface temperature and average discharge rate, query the pre-stored temperature-rate capacity compensation two-dimensional mapping table to obtain the capacity compensation coefficient, and multiply the baseline remaining capacity by the capacity compensation coefficient to obtain the remaining usable capacity of the lithium battery at the current time.

[0058] In one specific embodiment, the historical capacity loss recorded in the j-th historical charge-discharge cycle in the second candidate cycle set determined in step S104 is obtained. And the impedance spectrum shape similarity corresponding to the j-th historical charge-discharge cycle. (Calculated by step S104), the historical capacity loss is the cumulative capacity loss from the beginning to the end of the historical charge and discharge cycle. impedance spectrum shape similarity The historical capacity loss recorded during the j-th historical charge-discharge cycle. We construct a weighted empirical cumulative distribution function for the historical capacity loss of all second candidate cycles based on the weights, expressed as: , in, The weighted empirical cumulative distribution function with independent variable is The function value at time, The subscript filtering criteria for the summation operation, The value of the i-th historical capacity loss does not exceed , The normalized weight for the historical capacity loss recorded in the i-th historical charge-discharge cycle; Define adaptive confidence level for: ; in, The preset confidence level adjustment coefficient. This is a metric for coupling stability. Extracting the corresponding adaptive confidence level The upper quantile, that is, satisfying The smallest Value, as the basic robust capacity loss value The expression is: ; According to the coupling stability index Calculate the safety margin expansion factor , This is a preset margin adjustment coefficient, with a value range of [value range missing]. ; Multiplying the basic robust capacity loss value by the safety margin expansion factor yields the robust capacity loss value. ; The current calibrated capacity and the robust capacity loss value obtained using step S106 are used. Calculate the baseline remaining capacity ; Get the current battery surface temperature value and average discharge rate Query the pre-stored temperature-rate capacity compensation two-dimensional mapping table in the battery management system. This mapping table is pre-built using battery performance test data at different temperatures and discharge rates. It is a two-dimensional lookup table with battery surface temperature as the row index and average discharge rate as the column index. The capacity compensation coefficient under the current operating conditions is obtained by looking up the table.

[0059] Multiplying the baseline remaining capacity by the capacity compensation factor yields the final remaining usable capacity of the lithium battery at the current moment.

[0060] In summary, a weighted empirical cumulative distribution function is constructed using impedance spectrum shape similarity as weight, ensuring that historical samples most similar to the current battery aging coupling mode dominate statistical inference, thus improving the relevance of the statistical reference benchmark. Secondly, the confidence level is adaptively adjusted using coupling stability indicators: the more stable the coupling, the higher the confidence level and the more accurate the estimation; the less stable the coupling, the lower the confidence level and the more conservative the estimation. Thirdly, a safety margin extension factor linked to coupling stability is introduced, adding an additional physical conservatism buffer on top of the statistical quantile, forming a two-layer robust mechanism of "statistical confidence + physical margin". Finally, temperature-rate capacity compensation maps the benchmark remaining capacity under standard operating conditions to the current actual operating conditions, ensuring the direct usability of the output results. Compared with existing technologies that use fixed confidence levels or simple means as capacity loss estimates, this entire method allows the estimation of remaining usable capacity to automatically adjust the degree of conservatism based on the quality of historical battery coupling data while ensuring accuracy, significantly improving the robustness and engineering practicality of the estimation results, and providing a more reliable decision-making basis for battery management system safety warnings and charging / discharging strategy optimization.

[0061] Please see Figure 2 The diagram shows a structural block diagram of a lithium battery energy storage monitoring system according to this application.

[0062] like Figure 2 As shown, the lithium battery energy storage monitoring system 200 includes an acquisition module 210, an identification module 220, a first selection module 230, a second selection module 240, a calculation module 250, a determination module 260, and an output module 270.

[0063] The acquisition module 210 is configured to acquire the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of the lithium battery in each historical charge-discharge cycle, and establish a coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence in the same historical charge-discharge cycle; the identification module 220 is configured to determine the real-time terminal voltage sequence and real-time current sequence between the target time and the current time, and use the real-time terminal voltage sequence and real-time current sequence as the observation input of a preset extended Kalman filter to identify the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance online; the first selection module 230 is configured to perform sliding window correlation analysis on the real-time solid-phase diffusion coefficient and the historical solid-phase diffusion coefficient sequence in each historical charge-discharge cycle, and select historical charge-discharge cycles with a correlation coefficient greater than a preset first threshold as the first candidate cycle; the second selection module 240 is configured to, within the first candidate cycle, compare the real-time liquid-phase impedance with each historical charge-discharge cycle... The impedance spectrum shape similarity of the historical liquid phase impedance sequence is calculated, and the historical charge-discharge cycle with a similarity greater than a preset second threshold is selected as the second candidate cycle; the calculation module 250 is configured to calculate the current capacity loss of the lithium battery between the target time and the current time based on the historical health state parameters corresponding to the second candidate cycle, combined with the ratio of the current real-time solid phase diffusion coefficient to the historical solid phase diffusion coefficient, and the ratio of the current real-time liquid phase impedance to the historical liquid phase impedance, using a nonlinear diffusion-impedance coupling model; the determination module 260 is configured to determine the current calibration capacity based on the initial calibration capacity of the lithium battery and the aging factor in the current charge-discharge cycle, wherein the aging factor is determined according to the ratio of the real-time solid phase diffusion coefficient to the initial solid phase diffusion coefficient; the output module 270 is configured to determine the remaining usable capacity of the lithium battery at the current time based on the current calibration capacity and the current capacity loss.

[0064] It should be understood that Figure 2 The modules and references described in the document Figure 1 The steps described in the text correspond to those in the method described above. Therefore, the operations, features, and corresponding technical effects described above also apply to the method described in the text. Figure 2 The various modules in the document will not be described in detail here.

[0065] In other embodiments, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the lithium battery energy storage monitoring method in any of the above method embodiments. In one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, which are configured as follows: Obtain the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of lithium battery in each historical charge-discharge cycle, and establish the coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence in the same historical charge-discharge cycle. The real-time terminal voltage sequence and real-time current sequence between the target time and the current time are determined. The real-time terminal voltage sequence and real-time current sequence are used as the observation input of the preset extended Kalman filter to identify the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance online. The real-time solid-phase diffusion coefficient is compared with the historical solid-phase diffusion coefficient sequence in each historical charge-discharge cycle by a sliding window correlation analysis, and the historical charge-discharge cycle with a correlation coefficient greater than a preset first threshold is selected as the first candidate cycle. Within the first candidate period, the impedance spectrum shape similarity between the real-time liquid phase impedance and each historical liquid phase impedance sequence is calculated, and the historical charge-discharge period with a similarity greater than a preset second threshold is selected as the second candidate period. Based on the historical health status parameters corresponding to the second candidate period, combined with the ratio of the current real-time solid-phase diffusion coefficient to the historical solid-phase diffusion coefficient, and the ratio of the current real-time liquid phase impedance to the historical liquid phase impedance, a nonlinear diffusion-impedance coupling model is used to calculate the current capacity loss of the lithium battery between the target time and the current time. The current rated capacity is determined based on the initial rated capacity of the lithium battery and the aging factor in the current charge-discharge cycle, wherein the aging factor is determined according to the ratio of the real-time solid-phase diffusion coefficient to the initial solid-phase diffusion coefficient. Based on the current calibrated capacity and the current capacity loss, the remaining usable capacity of the lithium battery at the current moment is determined.

[0066] Computer-readable storage media may include a stored program area and a stored data area, wherein the stored program area may store an operating system and an application program required for at least one function; the stored data area may store data created based on the use of the lithium battery energy storage monitoring system, etc. Furthermore, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include memory remotely configured relative to a processor, and these remote memories can be connected to the lithium battery energy storage monitoring system via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0067] Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present invention, such as... Figure 3As shown, the device includes a processor 310 and a memory 320. The electronic device may also include an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 can be connected via a bus or other means. Figure 3 Taking a bus connection as an example, the memory 320 is the computer-readable storage medium described above. The processor 310 executes various server functions and data processing by running non-volatile software programs, instructions, and modules stored in the memory 320, thereby implementing the lithium battery energy storage monitoring method described in the above embodiment. The input device 330 can receive input digital or character information and generate key signal inputs related to user settings and function control of the lithium battery energy storage monitoring system. The output device 340 may include a display screen or other display device.

[0068] The aforementioned electronic device can execute the method provided in the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in the embodiments of the present invention.

[0069] In one implementation, the above-described electronic device is applied to a lithium battery energy storage monitoring system for a client application, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to: Obtain the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of lithium battery in each historical charge-discharge cycle, and establish the coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence in the same historical charge-discharge cycle. The real-time terminal voltage sequence and real-time current sequence between the target time and the current time are determined. The real-time terminal voltage sequence and real-time current sequence are used as the observation input of the preset extended Kalman filter to identify the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance online. The real-time solid-phase diffusion coefficient is compared with the historical solid-phase diffusion coefficient sequence in each historical charge-discharge cycle by a sliding window correlation analysis, and the historical charge-discharge cycle with a correlation coefficient greater than a preset first threshold is selected as the first candidate cycle. Within the first candidate period, the impedance spectrum shape similarity between the real-time liquid phase impedance and each historical liquid phase impedance sequence is calculated, and the historical charge-discharge period with a similarity greater than a preset second threshold is selected as the second candidate period. Based on the historical health status parameters corresponding to the second candidate period, combined with the ratio of the current real-time solid-phase diffusion coefficient to the historical solid-phase diffusion coefficient, and the ratio of the current real-time liquid phase impedance to the historical liquid phase impedance, a nonlinear diffusion-impedance coupling model is used to calculate the current capacity loss of the lithium battery between the target time and the current time. The current rated capacity is determined based on the initial rated capacity of the lithium battery and the aging factor in the current charge-discharge cycle, wherein the aging factor is determined according to the ratio of the real-time solid-phase diffusion coefficient to the initial solid-phase diffusion coefficient. Based on the current calibrated capacity and the current capacity loss, the remaining usable capacity of the lithium battery at the current moment is determined.

[0070] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring lithium battery energy storage, characterized in that, include: Obtain the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of lithium battery in each historical charge-discharge cycle, and establish the coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence in the same historical charge-discharge cycle. The real-time terminal voltage sequence and real-time current sequence between the target time and the current time are determined. The real-time terminal voltage sequence and real-time current sequence are used as the observation input of the preset extended Kalman filter to identify the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance online. The real-time solid-phase diffusion coefficient is compared with the historical solid-phase diffusion coefficient sequence in each historical charge-discharge cycle by a sliding window correlation analysis, and the historical charge-discharge cycle with a correlation coefficient greater than a preset first threshold is selected as the first candidate cycle. Within the first candidate period, the impedance spectrum shape similarity between the real-time liquid phase impedance and each historical liquid phase impedance sequence is calculated, and the historical charge-discharge period with a similarity greater than a preset second threshold is selected as the second candidate period. Based on the historical health status parameters corresponding to the second candidate period, combined with the ratio of the current real-time solid-phase diffusion coefficient to the historical solid-phase diffusion coefficient, and the ratio of the current real-time liquid phase impedance to the historical liquid phase impedance, a nonlinear diffusion-impedance coupling model is used to calculate the current capacity loss of the lithium battery between the target time and the current time. The current rated capacity is determined based on the initial rated capacity of the lithium battery and the aging factor in the current charge-discharge cycle, wherein the aging factor is determined according to the ratio of the real-time solid-phase diffusion coefficient to the initial solid-phase diffusion coefficient. Based on the current calibrated capacity and the current capacity loss, the remaining usable capacity of the lithium battery at the current moment is determined.

2. The lithium battery energy storage monitoring method according to claim 1, characterized in that, The establishment of the coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence within the same historical charge-discharge cycle includes: Outliers are removed from the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence within a certain historical charge-discharge cycle. The historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence after outlier removal are aligned by timestamp and then z-score normalized to obtain the first normalized sequence and the second normalized sequence. For each integer delay step d from -L to +L, the second normalized sequence is shifted d steps in the positive direction of the time axis. The Pearson correlation coefficient between the shifted second normalized sequence and the first normalized sequence is calculated, thus obtaining the curve of the Pearson correlation coefficient changing with the delay step d, where L is the maximum delay step. Extract the global maximum correlation coefficient from the curve, and the optimal delay step corresponding to the global maximum correlation coefficient, and determine whether the global maximum correlation coefficient is less than a preset minimum confidence threshold; If the global maximum correlation coefficient is less than the minimum confidence threshold, the coupling correlation matrix is ​​directly defined as a 2×2 identity matrix, where the main diagonal elements of the identity matrix are 1 and the secondary diagonal elements are 0. If the global maximum correlation coefficient is not less than the minimum confidence threshold, then the width of the curve at the half-peak height is calculated, and the ratio of the width to the maximum delay range 2L+1 is used as the coupling stability index, and the direction of influence is determined according to the sign of the optimal delay step. The product of the global maximum correlation coefficient and the coupling stability index is defined as the weighted coupling strength. A 2×2 coupling correlation matrix is ​​constructed based on the weighted coupling strength and the influence direction. The first row and first column of the coupling correlation matrix are fixed to 1, the first row and second column are equal to the weighted coupling strength multiplied by the sign value of the influence direction, the second row and first column are equal to the weighted coupling strength multiplied by the opposite sign value of the influence direction, and the second row and second column are fixed to 1.

3. The lithium battery energy storage monitoring method according to claim 2, characterized in that, The online identification of the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance, using the real-time terminal voltage sequence and real-time current sequence as the observation input of a preset extended Kalman filter, includes: A state-space model with solid-phase diffusion coefficient and liquid-phase impedance as state variables is established, with real-time current sequence as input and real-time terminal voltage sequence as observation, and the state vector and state error covariance matrix are initialized. At each sampling time, the Jacobian matrix of the state transition matrix and the observation matrix is ​​calculated, and the time update step of the extended Kalman filter is performed to obtain the prior state estimate and the prior error covariance matrix. The standard Kalman gain is calculated using the residual between the real-time terminal voltage sequence and the observed predicted value. Multiplying the standard Kalman gain by the coupling correlation matrix yields a modified gain matrix that satisfies the diffusion-impedance coupling constraint. The prior state estimate is updated using the corrected gain matrix and the residual to obtain the posterior state estimate; The solid-phase diffusion coefficient component in the posterior state estimate is used as the real-time solid-phase diffusion coefficient, and the liquid-phase impedance component in the posterior state estimate is used as the real-time liquid-phase impedance.

4. The lithium battery energy storage monitoring method according to claim 2, characterized in that, The step of calculating the impedance spectrum shape similarity between the real-time liquid phase impedance and each historical liquid phase impedance sequence, and selecting historical charge-discharge cycles with a similarity greater than a preset second threshold as second candidate cycles includes: Using the real-time solid-phase diffusion coefficient sequence and the real-time liquid-phase impedance sequence as input, and based on the maximum delay step L, calculate the Pearson correlation coefficient between the real-time liquid-phase impedance sequence and the real-time solid-phase diffusion coefficient sequence at each integer delay step d, obtain the real-time correlation coefficient curve, and extract the real-time half-peak width of the real-time correlation coefficient curve. For each first candidate period, extract the historical correlation coefficient curve corresponding to the first candidate period when establishing the coupling correlation matrix of the first candidate period, and obtain the historical half-peak width of the historical correlation coefficient curve. The real-time correlation coefficient curve and the historical correlation coefficient curve are normalized to their maximum values, so that the peak values ​​of the real-time correlation coefficient curve and the historical correlation coefficient curve are scaled to 1, thus obtaining the normalized real-time correlation coefficient curve and the normalized historical correlation coefficient curve. On a coordinate system with integer delay steps d as the horizontal axis and Pearson correlation coefficient as the vertical axis, calculate the area of ​​overlap between the normalized real-time correlation coefficient curve and the normalized historical correlation coefficient curve, and calculate the area of ​​union of the areas enclosed by the normalized real-time correlation coefficient curve and the normalized historical correlation coefficient curve. The ratio of the overlapping area to the union area is used as the degree of overlap of the coupling mode between the real-time liquid phase impedance sequence and the historical liquid phase impedance sequence within the first candidate period. Calculate the absolute value of the difference between the real-time half-peak width and the historical half-peak width, and determine the maximum value between the real-time half-peak width and the historical half-peak width. Subtract the ratio of the absolute value to the maximum value from 1 to obtain the coupling stability similarity. The product of the coupling mode overlap and the coupling stability similarity is taken as the impedance spectrum shape similarity. The historical charge-discharge cycles in which the impedance spectrum shape similarity is greater than the preset second threshold are selected as the second candidate cycles.

5. A lithium battery energy storage monitoring method according to claim 2, characterized in that, The calculation of the current capacity loss of the lithium battery between the target time and the current time using a nonlinear diffusion-impedance coupling model includes: The logarithmic decay rate of the solid diffusion coefficient is calculated based on the ratio of the real-time solid diffusion coefficient to the initial solid diffusion coefficient, and the logarithmic growth rate of the liquid phase impedance is calculated based on the ratio of the real-time liquid phase impedance to the initial liquid phase impedance. The product of the optimal delay step number and the data sampling time interval of the historical charge and discharge cycle is used as the characteristic time delay of the change in the solid phase diffusion coefficient relative to the change in the liquid phase impedance; Based on the characteristic time delay, calculate the coupling degradation driving force after time delay alignment, and input the coupling degradation driving force into the pre-constructed capacity loss rate function; Using the target time as the lower limit of integration and the current time as the upper limit of integration, the capacity loss rate function is numerically integrated to obtain the current capacity loss.

6. The lithium battery energy storage monitoring method according to claim 1, characterized in that, Determining the current rated capacity based on the initial rated capacity of the lithium battery and the aging factor in the current charge-discharge cycle includes: From all historical charge-discharge cycles, extract the final value of the historical solid-phase diffusion coefficient, the final value of the historical liquid-phase impedance, and the historical capacity retention rate at the end of each historical charge-discharge cycle, and construct a three-dimensional historical aging state point set with the final value of the historical solid-phase diffusion coefficient and the final value of the historical liquid-phase impedance as coordinates and the historical capacity retention rate as the response value. The real-time solid-phase diffusion coefficient and the real-time liquid-phase impedance at the current moment constitute the current aging state point. In the three-dimensional historical aging state point set, the K historical aging state points that are closest to the current aging state point are selected using Euclidean distance as a metric, where K is a preset integer. Calculate the inverse distance weighted average of the historical capacity retention rates corresponding to the K historical aging state points, and use the inverse distance weighted average as the benchmark aging factor; Obtain the equivalent full cycle count completed up to the current time of the current charge / discharge cycle, and the average equivalent full cycle count completed for the K historical charge / discharge cycles corresponding to the K historical aging state points, and calculate the time scale correction factor; Multiply the baseline aging factor by the time scale correction factor to obtain the aging factor for the current charge / discharge cycle; The current calibrated capacity is obtained by multiplying the initial calibrated capacity by the aging factor of the current charge / discharge cycle.

7. A lithium battery energy storage monitoring method according to claim 2, characterized in that, Determining the remaining usable capacity of the lithium battery at the current moment based on the current calibrated capacity and the current capacity loss includes: Obtain the historical capacity loss corresponding to all the second candidate periods, and use the impedance spectrum shape similarity of the second candidate periods as weights to construct the weighted empirical cumulative distribution function of the historical capacity loss. Calculate the adaptive confidence level based on the coupling stability index; On the weighted empirical cumulative distribution function, the upper quantile corresponding to the adaptive confidence level is extracted as the basic robust capacity loss value; Calculate the safety margin expansion factor, and multiply the basic robust capacity loss value by the safety margin expansion factor to obtain the robust capacity loss value; Determine the baseline remaining capacity based on the current calibrated capacity and the robust capacity loss value; Obtain the current battery surface temperature and average discharge rate, query the pre-stored temperature-rate capacity compensation two-dimensional mapping table to obtain the capacity compensation coefficient, and multiply the baseline remaining capacity by the capacity compensation coefficient to obtain the remaining usable capacity of the lithium battery at the current time.

8. A lithium battery energy storage monitoring system, characterized in that, include: The acquisition module is configured to acquire the historical solid-phase diffusion coefficient sequence and historical liquid-phase impedance sequence of the lithium battery in each historical charge-discharge cycle, and establish a coupling correlation matrix between the historical solid-phase diffusion coefficient sequence and the historical liquid-phase impedance sequence in the same historical charge-discharge cycle. The identification module is configured to determine the real-time terminal voltage sequence and real-time current sequence between the target time and the current time, and use the real-time terminal voltage sequence and real-time current sequence as the observation input of the preset extended Kalman filter to identify the real-time solid-phase diffusion coefficient and real-time liquid-phase impedance online. The first selection module is configured to perform sliding window correlation analysis on the real-time solid-phase diffusion coefficient and the historical solid-phase diffusion coefficient sequence in each historical charge-discharge cycle, and select the historical charge-discharge cycle with a correlation coefficient greater than a preset first threshold as the first candidate cycle. The second selection module is configured to perform impedance spectrum shape similarity calculation between the real-time liquid phase impedance and each historical liquid phase impedance sequence within the first candidate period, and select the historical charge-discharge period with a similarity greater than a preset second threshold as the second candidate period. The calculation module is configured to calculate the current capacity loss of the lithium battery between the target time and the current time based on the historical health status parameters corresponding to the second candidate period, combined with the ratio of the current real-time solid-phase diffusion coefficient to the historical solid-phase diffusion coefficient, and the ratio of the current real-time liquid phase impedance to the historical liquid phase impedance, using a nonlinear diffusion-impedance coupling model. The determination module is configured to determine the current calibration capacity based on the initial calibration capacity of the lithium battery and the aging factor in the current charge-discharge cycle, wherein the aging factor is determined according to the ratio of the real-time solid-phase diffusion coefficient to the initial solid-phase diffusion coefficient. The output module is configured to determine the remaining usable capacity of the lithium battery at the current moment based on the current calibrated capacity and the current capacity loss.

9. An electronic device, characterized in that, include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method according to any one of claims 1 to 7.

Images