A lithium battery operation state evaluation system and method based on data analysis

Abstract

The application discloses a lithium battery operation state evaluation system and method based on data analysis, relates to the technical field of lithium battery operation state evaluation, collects electric excitation response data of a battery under different measurement parameters, generates an impedance spectrum tensor, carries out normalization processing on the impedance spectrum tensor along a temperature dimension, outputs a temperature decoupling impedance tensor, decomposes the temperature decoupling impedance tensor along a rate dimension, outputs a rate separation impedance matrix, extracts static contact impedance from the rate separation impedance matrix, carries out trend deduction along a state of charge dimension, outputs a charge decoupling impedance sequence, decomposes the charge decoupling impedance sequence, outputs a contact impedance fluctuation sequence, carries out filtering processing on the rate separation impedance matrix according to the contact impedance fluctuation sequence, outputs a corrected internal resistance spectrum, calculates a real health state of the battery based on the corrected internal resistance spectrum and carries out determination, and outputs a battery health state signal.

Description

Technical Field

[0001] This invention relates to the field of lithium battery operating status assessment technology, specifically a lithium battery operating status assessment system and method based on data analysis. Background Technology

[0002] Currently, the health status assessment of lithium batteries mainly relies on parameters such as voltage, current, and temperature monitored in real time by the battery management system. Among these, changes in internal resistance are often used as an important indicator to judge the degree of battery aging. However, in actual operation, changes in battery internal resistance are not solely caused by electrochemical aging. As charge and discharge cycles proceed, electrode materials undergo volume expansion and contraction, which may lead to fluctuations in pressure at the battery's internal contact points, resulting in dynamic changes in contact internal resistance.

[0003] In existing technologies, battery management systems typically attribute internal resistance changes primarily to electrochemical aging and correct them using temperature compensation or empirical lookup tables, with little consideration given to the interference of contact internal resistance fluctuations on impedance measurements. This approach has several drawbacks: first, it's difficult to distinguish the contribution of contact state changes from actual chemical aging, and the estimated health status may drift significantly with varying operating conditions; second, there's a lack of effective online separation methods, leading to uncertainty for operators when determining whether battery replacement is necessary; and third, relying on additional sensors increases cost and system complexity, hindering widespread adoption in certain application scenarios. Furthermore, existing technologies have not yet provided a health status assessment method that can identify and isolate mechanical contact interference from total impedance online without adding sensors.

[0004] Therefore, a method is needed that can effectively isolate mechanical contact interference from the total impedance and achieve a more accurate assessment of health status. Summary of the Invention

[0005] The purpose of this invention is to provide a lithium battery operating status evaluation system and method based on data analysis to solve the problems raised in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for evaluating the operating status of a lithium battery based on data analysis, comprising: The battery's electrical excitation response data under different measurement parameters are collected to generate an impedance spectrum tensor; the impedance spectrum tensor is normalized along the temperature dimension to output a temperature-decoupled impedance tensor. The temperature decoupling impedance tensor is decomposed along the scaling factor dimension to output a scaling factor separation impedance matrix; the static contact impedance is extracted from the scaling factor separation impedance matrix and trend subtraction is performed along the charge state dimension to output a charge decoupling impedance sequence. The charged decoupling impedance sequence is decomposed to separate the mechanical relaxation component and the random noise component, and the contact impedance fluctuation sequence is output. Based on the contact impedance fluctuation sequence, the multiplier separation impedance matrix is ​​filtered to output a corrected internal resistance spectrum; Based on the corrected internal resistance spectrum, the true health state of the battery is calculated; a judgment is made based on the true health state, and a battery health state signal is output.

[0007] In conjunction with the first aspect, in the first implementation of the first aspect of this application, the step of collecting electrical excitation response data of the battery under different measurement parameters and generating an impedance spectrum tensor includes: During battery operation, AC excitation currents of a preset frequency are injected sequentially at preset time intervals, and the injected frequency f and current amplitude I(f) are recorded. The battery terminal voltage response is collected synchronously, and the voltage response amplitude U(f) at the corresponding frequency is extracted from the voltage response. The electrical excitation response data includes the frequency f, the current amplitude I(f), and the voltage response amplitude U(f). Based on the electrical excitation response data, the impedance magnitude Z(f) at each frequency f is calculated. Record the measurement parameters during the measurement, including the battery surface temperature T, state of charge S, and instantaneous current rate C. Arrange the impedance modulus Z(f) under the same measurement parameter conditions according to the frequency f, construct and output the impedance spectrum tensor Z. r .

[0008] In conjunction with the first aspect, in the second implementation of the first aspect of this application, the normalization of the impedance spectrum tensor along the temperature dimension to output a temperature-decoupled impedance tensor includes: According to the impedance spectrum tensor Z r For each frequency f, the Arrhenius equation is used, utilizing the battery surface temperature T and the activation energy E at frequency f. a (f) Given a preset ideal gas constant R and a preset reference temperature T0, calculate and output the temperature decoupling impedance tensor Z. t .

[0009] In conjunction with the first aspect, in the third implementation of the first aspect of this application, the step of decomposing the temperature decoupling impedance tensor along the scaling factor dimension and outputting the scaling factor separation impedance matrix includes: According to the temperature decoupling impedance tensor Z t For the same battery surface temperature T and state of charge S, a polynomial fitting of a preset order is performed along the instantaneous current rate C to obtain a static impedance value independent of the instantaneous current rate C. These static impedance values ​​are then arranged into a matrix according to frequency f to obtain the rate-separated impedance matrix Z. s .

[0010] In conjunction with the first aspect, in the fourth implementation of the first aspect of this application, the step of extracting the static contact impedance from the rate separation impedance matrix and performing trend subtraction along the charge state dimension to output the charge decoupling impedance sequence includes: The impedance matrix Z is separated from the multiplier. s In the middle, select the impedance value R within the preset low-frequency band. c (T,S); For each battery surface temperature T, the impedance value R within the preset low-frequency band is calculated with respect to the state of charge S as the independent variable. c Perform local weighted regression on (T,S) to fit the baseline function R. q (S); The impedance value R within the preset low-frequency band c Subtract the baseline function R from (T,S) q (S), to obtain the contact impedance R after deducting the trend. d (T,S); Contact resistance R after deducting the trend at different battery surface temperatures T d (T,S) is expanded in the order of measurement time, outputting the charge decoupling impedance sequence Z. m (t).

[0011] In conjunction with the first aspect, in the fifth implementation of the first aspect of this application, the step of decomposing the charge decoupling impedance sequence to separate the mechanical relaxation component and the random noise component, and outputting the contact impedance fluctuation sequence, includes: According to the charge decoupling impedance sequence Z m (t), the charge decoupling impedance sequence Z is decomposed using variational mode decomposition. m (t) is decomposed into K eigenmode functions u k (t); The intrinsic mode function u1(t) with the lowest center frequency is extracted as the mechanical relaxation component; based on the mechanical relaxation component and the Hilbert transform, the envelope sequence A(t) is calculated, as follows: ; Where j is the imaginary unit, and H[ ] is the Hilbert transform; Calculate the arithmetic mean A of the envelope sequence A(t). c The envelope sequence is normalized to obtain the contact impedance fluctuation sequence R. b (t), the formula is: .

[0012] In conjunction with the first aspect, in the sixth implementation of the first aspect of this application, the step of filtering the multiplier-separated impedance matrix according to the contact impedance fluctuation sequence and outputting a corrected internal resistance spectrum includes: According to the aforementioned impedance matrix Z, the multiplier is separated. s and the contact impedance fluctuation sequence R b (t), the multiplier separation impedance matrix Z s The battery surface temperature T and state of charge S are mapped to the time axis according to the measurement time, and the impedance value R in the preset mid-frequency band is selected. m (f,t); with the contact impedance fluctuation sequence R b (t) is the reference signal, and an adaptive notch filter is constructed with a center frequency W. n (t) Based on the contact impedance fluctuation sequence R b The instantaneous dominant frequency of (t) is determined by the following formula: ; Where F is the Fourier transform and δ is the preset time window; The transfer function H(z) of the adaptive notch filter is given by: ; Where ρ is the preset notch depth coefficient, and z is the Z-transform variable; The impedance value R within the preset mid-frequency band m (f,t) is passed through the adaptive notch filter to obtain the filtered impedance value R. v (f,t); Iterate through each frequency f and calculate based on the filtered impedance value R. v The corrected internal resistance spectrum Z is obtained by calculating (f,t) and Hilbert transform. c (f,t), the formula is: .

[0013] In conjunction with the first aspect, in the seventh implementation of the first aspect of this application, the step of calculating the true health state of the battery based on the corrected internal resistance spectrum includes: According to the modified internal resistance spectrum Z c (f,t), extract the preset low-frequency point f L impedance value R l (t) and preset high-frequency point f H impedance value R h (t), calculate the impedance spectrum slope S in the preset mid-frequency band. m (t); R l (t), R h (t) and S m (t) are respectively compared with the corresponding characteristic value R in the initial state of the battery. l0 R h0 and S m0 By comparison, the true health state SOH(t) of the battery is calculated through weighted summation.

[0014] In conjunction with the first aspect, in the eighth implementation of the first aspect of this application, the step of determining the battery health status based on the actual health status and outputting a battery health status signal includes: The battery health status is determined based on the actual health status SOH(t). When the actual health status SOH(t) is greater than or equal to a preset health threshold, a good battery health status signal is output. When the actual health status SOH(t) is less than the preset health threshold but greater than or equal to a preset fault threshold, a normal battery health status signal is output. When the actual health status SOH(t) is less than the preset fault threshold, a low battery health status signal is output, and maintenance personnel are prompted to replace the battery.

[0015] Secondly, the present invention provides a lithium battery operating status evaluation system based on data analysis, comprising: Data acquisition and preprocessing module: includes a data acquisition unit and a normalization processing unit; wherein, the data acquisition unit acquires the battery's electrical excitation response data and measurement parameters, and generates an impedance spectrum tensor; the normalization processing unit performs temperature normalization on the impedance spectrum tensor and outputs a temperature-decoupled impedance tensor. Operating condition decoupling and trend subtraction module: includes a rate decomposition unit and a trend subtraction unit; wherein, the rate decomposition unit performs rate dimension decomposition on the temperature decoupling impedance tensor and outputs a rate separation impedance matrix; the trend subtraction unit extracts the static contact impedance from the rate separation impedance matrix and subtracts the charge state trend, and outputs a charge decoupling impedance sequence. Fluctuation extraction and filtering correction module: includes a fluctuation extraction unit and a filtering correction unit; wherein, the fluctuation extraction unit decomposes the charge decoupling impedance sequence, separates the mechanical relaxation component, and outputs the contact impedance fluctuation sequence; the filtering correction unit filters the multiplier separation impedance matrix according to the contact impedance fluctuation sequence and outputs the corrected internal resistance spectrum. Health calculation and judgment module: includes a health calculation unit and a judgment output unit; wherein, the health calculation unit calculates the true health status of the battery based on the corrected internal resistance spectrum; the judgment output unit makes a judgment based on the true health status and outputs the battery health status signal.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention eliminates the interference of operating parameters on impedance step by step through temperature normalization, rate polynomial fitting, and state of charge trend subtraction, solving the problem of existing technology misjudging changes in operating conditions as battery aging, and obtaining pure impedance data.

[0017] 2. This invention extracts the contact internal resistance fluctuation characteristics caused by the expansion and contraction of electrode materials from the impedance sequence through variational mode decomposition and envelope normalization, thereby realizing online identification of mechanical interference.

[0018] 3. This invention uses contact fluctuation characteristics to guide adaptive filtering to correct the impedance spectrum, and calculates the true health status after removing mechanical interference. This solves the problem of health status estimation drift caused by contact internal resistance fluctuation in the prior art, and improves the accuracy of the assessment. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the steps of a data analysis-based method for evaluating the operating status of a lithium battery according to the present invention. Figure 2 This is a flowchart of the fluctuation extraction and filtering correction steps in a data analysis-based lithium battery operating status assessment method of the present invention. Figure 3 This is a system structure diagram of a lithium battery operating status evaluation system based on data analysis according to the present invention. Detailed Implementation

[0020] 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, and 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.

[0021] Example: Figures 1-3 As shown, the present invention provides a technical solution. like Figure 1 A schematic diagram illustrating the steps of a data analysis-based method for assessing the operational status of a lithium battery illustrates this invention. The invention provides a data analysis-based method for assessing the operational status of a lithium battery, comprising: Step S100: Collect electrical excitation response data of the battery under different measurement parameters, and generate an impedance spectrum tensor; normalize the impedance spectrum tensor along the temperature dimension, and output a temperature-decoupled impedance tensor. Specifically, during battery operation, AC excitation currents of a preset frequency are injected sequentially at preset time intervals, and the injected frequency f and current amplitude I(f) are recorded. The battery terminal voltage response is collected simultaneously, and the voltage response amplitude U(f) at the corresponding frequency is extracted from the voltage response. The electrical excitation response data includes the frequency f, current amplitude I(f), and voltage response amplitude U(f). Based on the electrical excitation response data, the impedance magnitude Z(f) at each frequency f is calculated using the following formula: ; Record the measurement parameters during the measurement, including the battery surface temperature T, state of charge S, and instantaneous current rate C. Arrange the impedance modulus Z(f) under the same measurement parameter conditions according to the frequency f, construct and output the impedance spectrum tensor Z. r .

[0022] According to the impedance spectrum tensor Z r For each frequency f, the Arrhenius equation is used, utilizing the battery surface temperature T and the activation energy E at frequency f. a (f) Given a preset ideal gas constant R and a preset reference temperature T0, calculate and output the temperature decoupling impedance tensor Z. t The formula is: ; Among them, E a (f) is the activation energy at frequency f, which is obtained by measuring and fitting the impedance of the same type of battery at different battery surface temperatures in advance. R is the preset ideal gas constant, and T0 is the preset reference temperature.

[0023] In one specific embodiment, a lithium iron phosphate power battery with a nominal capacity of 100Ah was selected and tested at an ambient temperature of 25℃. AC excitation currents with frequencies of 0.1Hz, 1Hz, 10Hz, 100Hz, and 1000Hz were injected sequentially at preset time intervals of 10 seconds, with a current amplitude of 0.5A at each frequency. Under the conditions of battery surface temperature T=28℃, state of charge S=85%, and instantaneous current ratio C=0.5C, the battery terminal voltage response was synchronously acquired, and the voltage response amplitude at the corresponding frequencies was extracted: 0.52mV at 0.1Hz, 0.51mV at 1Hz, 0.50mV at 10Hz, 0.48mV at 100Hz, and 0.45mV at 1000Hz. The impedance moduli at each frequency were calculated using the formula to be 1.04mΩ, 1.02mΩ, 1.00mΩ, 0.96mΩ, and 0.90mΩ, respectively. The impedance moduli under the same battery surface temperature T=28℃, state of charge S=85%, and instantaneous current rate C=0.5C conditions are arranged by frequency to construct a three-dimensional impedance spectrum tensor Z. r Z r Impedance data at five frequency points are stored in three dimensions: temperature, state of charge, and magnification.

[0024] Step S200: Decompose the temperature decoupling impedance tensor along the scaling factor dimension to output the scaling factor separation impedance matrix; extract the static contact impedance from the scaling factor separation impedance matrix and perform trend subtraction along the charge state dimension to output the charge decoupling impedance sequence. Specifically, based on the temperature decoupling impedance tensor Z tFor the same battery surface temperature T and state of charge S, a polynomial fitting of a preset order is performed along the instantaneous current rate C to calculate the fitting coefficients W. k The static impedance values, independent of the instantaneous current multiplier C, are obtained. These static impedance values ​​are then arranged into a matrix according to frequency f to obtain the multiplier-separated impedance matrix Z. s The formula is: ; Where p is a preset order, k is a summation index, and the multiplier dimension is represented by the instantaneous current multiplier C.

[0025] The impedance matrix Z is separated from the multiplier. s In the middle, select the impedance value R within the preset low-frequency band. c (T,S); For each battery surface temperature T, the impedance value R within the preset low-frequency band is calculated with respect to the state of charge S as the independent variable. c Perform local weighted regression on (T,S) to fit the baseline function R. q (S); The impedance value R within the preset low-frequency band c Subtract the baseline function R from (T,S) q (S), to obtain the contact impedance R after deducting the trend. d (T,S), the formula is: ; Contact resistance R after deducting the trend at different battery surface temperatures T d (T,S) are expanded in chronological order of measurement time to form a one-dimensional sequence Z. m The time index is t, and the output charge decoupling impedance sequence Z is... m (t).

[0026] In one specific embodiment, the received impedance spectrum tensor Z r Temperature normalization was performed for each frequency point. Taking 100Hz as an example, the activation energy E was obtained by first measuring the impedance of the battery model at different battery surface temperatures of 10℃, 20℃, 30℃, and 40℃ and then fitting the data. a (100Hz) = 20kJ / mol. Taking the preset ideal gas constant R = 8.314J / (mol·K), and the preset reference temperature T0 = 25℃, 298.15K. The current battery surface temperature T = 28℃, 301.15K, and the calculated exponent term is 0.924. Therefore, the temperature-normalized impedance Z... t =1.039mΩ. After normalization over all frequencies and operating points, the output temperature decoupling impedance tensor Z is... t For Z tPolynomial fitting was performed along the rate dimension. For the same battery surface temperature T=28℃ and state of charge S=85%, a set of impedance values ​​were measured at different instantaneous current rates C of 0.2C, 0.5C, 0.8C, and 1.0C. Polynomial fitting with a preset order p=2 was used to obtain fitting coefficients W0=1.02, W1=0.05, and W2=-0.03, thus obtaining the static impedance value Z, which is independent of the instantaneous current rate C. s =1.02mΩ, and the impedance matrix Z obtained after arranging by frequency is the multiplier-separated impedance matrix. s .

[0027] Step S300: Decompose the charge decoupling impedance sequence to separate the mechanical relaxation component and random noise component, and output the contact impedance fluctuation sequence. Specifically, based on the charge decoupling impedance sequence Z m (t), the charge decoupling impedance sequence Z is decomposed using variational mode decomposition. m (t) is decomposed into K eigenmode functions u k (t), the formula is: ; Where K is the number of preset modes; The intrinsic mode function u1(t) with the lowest center frequency is extracted as the mechanical relaxation component. Specifically, this mechanical relaxation component represents the slowly varying contact resistance fluctuation caused by the expansion and contraction of the electrode material. The remaining modes represent random noise components. Based on the mechanical relaxation component and the Hilbert transform, the envelope sequence A(t) is calculated using the following formula: ; Where j is the imaginary unit, and H[ ] is the Hilbert transform; Calculate the arithmetic mean A of the envelope sequence A(t). c The envelope sequence is normalized to obtain the contact impedance fluctuation sequence R. b (t), the formula is: .

[0028] In one specific embodiment, the obtained multiplier-separated impedance matrix Z is... s In the middle, select the impedance value R within the preset low-frequency range of 0.1Hz to 1Hz. c (T,S). R was measured at a battery surface temperature T=28℃ and states of charge S of 30%, 50%, 70%, 85%, and 100%. c The values ​​were 1.12 mΩ, 1.08 mΩ, 1.05 mΩ, 1.02 mΩ, and 0.99 mΩ, respectively. Using the state of charge S as the independent variable, a locally weighted regression was employed for fitting, with a preset bandwidth of 0.3, to obtain the baseline function R. q(S) = 1.15 - 0.0016S. When S = 85%, R... q (85) = 1.014 mΩ. R is obtained after subtracting the trend. d (28,85)=0.006mΩ. RΩ is calculated for different battery surface temperatures T at 20℃, 25℃, 30℃, and 35℃. d (T,S) are expanded in chronological order of measurement time to form a one-dimensional sequence Z. m (t), Z corresponding to time indices t=1, 2, 3, 4 m The values ​​are 0.005mΩ, 0.006mΩ, 0.007mΩ, and 0.008mΩ, respectively, and the output charge decoupling impedance sequence Z m (t).

[0029] Step S400: Based on the contact impedance fluctuation sequence, filter the multiplier-separated impedance matrix to output a corrected internal resistance spectrum; Specifically, based on the aforementioned multiplier separation impedance matrix Z s and the contact impedance fluctuation sequence R b (t), the multiplier separation impedance matrix Z s The battery surface temperature T and state of charge S are mapped to the time axis according to the measurement time, and the impedance value R in the preset mid-frequency band is selected. m (f,t); with the contact impedance fluctuation sequence R b (t) is the reference signal, and an adaptive notch filter is constructed with a center frequency W. n (t) Based on the contact impedance fluctuation sequence R b The instantaneous dominant frequency of (t) is determined by the following formula: ; Where F is the Fourier transform and δ is the preset time window; The transfer function H(z) of the adaptive notch filter is given by: ; Where ρ is the preset notch depth coefficient, and z is the Z-transform variable; The impedance value R within the preset mid-frequency band m (f,t) is passed through the adaptive notch filter to obtain the filtered impedance value R. v (f,t); Iterate through each frequency f, repeat the above process for each frequency f, and calculate based on the filtered impedance value R. v The corrected internal resistance spectrum Z is obtained by calculating (f,t) and Hilbert transform. c (f,t), the formula is: .

[0030] In one specific embodiment, the output charge decoupling impedance sequence Z is used. m (t) is the input, and variational mode decomposition is used, with a preset number of modes of 3, a quadratic penalty factor of 2000, and a convergence tolerance of 1×10⁻⁶. -7 , will Z m The signal (t) is decomposed into three intrinsic mode functions, where u1(t) = [0.0051, 0.0059, 0.0068, 0.0077]mΩ, u2(t) = [-0.0001, 0.0001, -0.0002, 0.0003]mΩ, and u3(t) = [0.0000, 0.0000, 0.0004, -0.0000]mΩ. The lowest center frequency u1(t) is extracted as the mechanical relaxation component. A Hilbert transform is applied to u1(t) to obtain the envelope sequence of the analytic signal A(t) = [0.0051, 0.0060, 0.0069, 0.0078]mΩ. The arithmetic mean A is calculated. c =0.00645mΩ, normalized to obtain the contact impedance fluctuation sequence R b (t)=[0.791,0.930,1.070,1.209]. Using this sequence as a reference signal, the multiplier-separated impedance matrix Z is... s Filtering is performed, and an impedance value R within a preset mid-frequency band is selected at frequency f=100Hz. m (100,t)=[1.02,1.03,1.02,1.04]mΩ; according to R b The instantaneous dominant frequency of (t) is approximately 0.05 Hz obtained through Fourier transform. An adaptive notch filter is constructed with a notch depth coefficient ρ = 0.95 and a time window δ = 1 s. The filtered result is R. v (100,t)=[1.019,1.029,1.019,1.038]mΩ. Repeat the above process for each frequency f to finally reconstruct the corrected internal resistance spectrum Z. c (100,t)=[1.019,1.029,1.019,1.038]mΩ.

[0031] Step S500: Calculate the true health status of the battery based on the corrected internal resistance spectrum; determine the true health status based on the true health status and output the battery health status signal.

[0032] Specifically, according to the modified internal resistance spectrum Z c (f,t), extract the preset low-frequency point f L impedance value R l (t) and preset high-frequency point f H impedance value R h (t), calculate the impedance spectrum slope S in the preset mid-frequency band. m (t), the formula is: ; R l (t), R h (t) and S m (t) are respectively compared with the corresponding characteristic value R in the initial state of the battery. l0 R h0 and S m0 By comparison, the true state of health (SOH) of the battery is calculated using a weighted summation method, as shown in the formula: ; Where α is the preset low-frequency empirical coefficient, β is the preset high-frequency empirical coefficient, and γ is the preset mid-frequency empirical coefficient.

[0033] The battery health status is determined based on the actual health status SOH(t). When the actual health status SOH(t) is greater than or equal to a preset health threshold, a good battery health status signal is output. When the actual health status SOH(t) is less than the preset health threshold but greater than or equal to a preset fault threshold, a normal battery health status signal is output. When the actual health status SOH(t) is less than the preset fault threshold, a low battery health status signal is output, and maintenance personnel are prompted to replace the battery.

[0034] In one specific embodiment, based on the modified internal resistance spectrum Z c (f,t), extract the preset low-frequency point f L Impedance value R = 0.1Hz l (t) and preset high-frequency point f H Impedance value R at 1000Hz h (t). At time index t=1, R l (1) = 1.12 mΩ, R h (1) = 0.88 mΩ. Calculate the impedance spectrum slope S within the preset mid-frequency band of 1 Hz to 10 Hz. m (1) = -0.0078 mΩ / Hz. The characteristic value of the battery in its initial state is R. l0 =1.05mΩ, R h0 =0.85mΩ, S m0 =-0.0072mΩ / Hz. Preset empirical coefficients α=0.5, β=0.3, γ=0.2. According to the formula SOH(1)=99.94%. Taking the preset health threshold of 95% and the preset fault threshold of 75%, since SOH(1)=99.94%≥95%, the battery health status is determined to be good, and a signal is output. The SOH values ​​at subsequent time points t=2, 3, and 4 are 99.81%, 99.65%, and 99.42% respectively, all greater than 95%, therefore a good signal is continuously output and sent.

[0035] like Figure 2The flowchart illustrates the fluctuation extraction and filtering correction steps in a data analysis-based lithium battery operating status assessment method. This invention provides a data analysis-based lithium battery operating status assessment method, comprising: First, the output charge decoupling impedance sequence Z m (t) Enter this step for Z m (t) Perform variational mode decomposition to decompose it into multiple eigenmode functions u k (t). The mode u1(t) with the lowest center frequency is extracted as the mechanical relaxation component to characterize the contact resistance fluctuation caused by the expansion and contraction of the electrode material. Subsequently, a Hilbert transform is performed on u1(t) to calculate the envelope sequence A(t) of its analytic signal. The envelope sequence is then normalized by dividing it by its arithmetic mean to obtain the contact impedance fluctuation sequence R. b (t).

[0036] At the same time, the obtained multiplier separation impedance matrix Z s Enter this step. Select the impedance value R within the preset mid-frequency band. m (f,t). Using the obtained contact impedance fluctuation sequence R... b Using (t) as the reference signal, an adaptive notch filter is constructed, and the center frequency of the filter dynamically tracks R. b The instantaneous dominant frequency of (t). The selected impedance value R... m (f,t) is passed through the filter to obtain the filtered impedance value R. v (f,t). Finally, a Hilbert transform is performed on the filtered impedance, and the modulus is taken to reconstruct the corrected internal resistance spectrum Z. c (f,t) and output.

[0037] like Figure 3 The system architecture diagram of a data analysis-based lithium battery operating status assessment system is shown. This invention provides a data analysis-based lithium battery operating status assessment system, comprising: Data acquisition and preprocessing module: includes a data acquisition unit and a normalization processing unit; wherein, the data acquisition unit acquires the battery's electrical excitation response data and measurement parameters, and generates an impedance spectrum tensor; the normalization processing unit performs temperature normalization on the impedance spectrum tensor and outputs a temperature-decoupled impedance tensor. Operating condition decoupling and trend subtraction module: includes a rate decomposition unit and a trend subtraction unit; wherein, the rate decomposition unit performs rate dimension decomposition on the temperature decoupling impedance tensor and outputs a rate separation impedance matrix; the trend subtraction unit extracts the static contact impedance from the rate separation impedance matrix and subtracts the charge state trend, and outputs a charge decoupling impedance sequence. Fluctuation extraction and filtering correction module: includes a fluctuation extraction unit and a filtering correction unit; wherein, the fluctuation extraction unit decomposes the charge decoupling impedance sequence, separates the mechanical relaxation component, and outputs the contact impedance fluctuation sequence; the filtering correction unit filters the multiplier separation impedance matrix according to the contact impedance fluctuation sequence and outputs the corrected internal resistance spectrum. Health calculation and judgment module: includes a health calculation unit and a judgment output unit; wherein, the health calculation unit calculates the true health status of the battery based on the corrected internal resistance spectrum; the judgment output unit makes a judgment based on the true health status and outputs the battery health status signal.

[0038] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for evaluating the operating status of lithium batteries based on data analysis, characterized in that, include: The electrical excitation response data of the battery under different measurement parameters are collected to generate an impedance spectrum tensor. Specifically, the measurement parameters are recorded, including the battery surface temperature T, state of charge S, and instantaneous current rate C. The impedance magnitudes Z(f) under the same measurement parameters are arranged according to frequency f to construct and output the impedance spectrum tensor Z. r The impedance spectrum tensor is normalized along the temperature dimension to output a temperature-decoupled impedance tensor. The temperature decoupling impedance tensor is decomposed along the scaling dimension to output a scaling-separated impedance matrix; The static contact impedance is extracted from the ratio separation impedance matrix and trend subtraction is performed along the charge state dimension to output the charge decoupling impedance sequence. The charged decoupling impedance sequence is decomposed to separate the mechanical relaxation component and the random noise component, and the contact impedance fluctuation sequence is output. Based on the contact impedance fluctuation sequence, the multiplier separation impedance matrix is ​​filtered to output a corrected internal resistance spectrum; Based on the corrected internal resistance spectrum, the true health status of the battery is calculated; The battery health status signal is output based on the actual health status.

2. The lithium battery operating status evaluation method based on data analysis according to claim 1, characterized in that, The acquisition of electrical excitation response data of the battery under different measurement parameters generates an impedance spectrum tensor, including: During battery operation, AC excitation currents of a preset frequency are injected sequentially at preset time intervals, and the injected frequency f and current amplitude I(f) are recorded. The battery terminal voltage response is collected synchronously, and the voltage response amplitude U(f) at the corresponding frequency is extracted from the voltage response. The electrical excitation response data includes the frequency f, the current amplitude I(f), and the voltage response amplitude U(f). Based on the electrical excitation response data, the impedance magnitude Z(f) at each frequency f is calculated.

3. The lithium battery operating status evaluation method based on data analysis according to claim 1, characterized in that, The normalization of the impedance spectrum tensor along the temperature dimension to output a temperature-decoupled impedance tensor includes: According to the impedance spectrum tensor Z r For each frequency f, the Arrhenius equation is used, utilizing the battery surface temperature T and the activation energy E at frequency f. a (f) Given a preset ideal gas constant R and a preset reference temperature T0, calculate and output the temperature decoupling impedance tensor Z. t .

4. The lithium battery operating status evaluation method based on data analysis according to claim 1, characterized in that, The process of decomposing the temperature decoupling impedance tensor along the scaling factor dimension to output a scaling factor-separated impedance matrix includes: According to the temperature decoupling impedance tensor Z t For the same battery surface temperature T and state of charge S, a polynomial fitting of a preset order is performed along the instantaneous current rate C to obtain a static impedance value independent of the instantaneous current rate C. These static impedance values ​​are then arranged into a matrix according to frequency f to obtain the rate-separated impedance matrix Z. s .

5. The lithium battery operating status evaluation method based on data analysis according to claim 1, characterized in that, The step of extracting the static contact impedance from the rate-separated impedance matrix and performing trend subtraction along the charge state dimension to output the charge decoupling impedance sequence includes: The impedance matrix Z is separated from the multiplier. s In the middle, select the impedance value R within the preset low-frequency band. c (T,S); For each battery surface temperature T, the impedance value R within the preset low-frequency band is calculated with respect to the state of charge S as the independent variable. c Perform local weighted regression on (T,S) to fit the baseline function R. q (S); The impedance value R within the preset low-frequency band c Subtract the baseline function R from (T,S) q (S), to obtain the contact impedance R after deducting the trend. d (T,S); Contact resistance R after deducting the trend at different battery surface temperatures T d (T,S) is expanded in the order of measurement time, outputting the charge decoupling impedance sequence Z. m (t).

6. The lithium battery operating status evaluation method based on data analysis according to claim 1, characterized in that, The process of decomposing the charged decoupling impedance sequence to separate the mechanical relaxation component and random noise component, and outputting the contact impedance fluctuation sequence, includes: According to the charge decoupling impedance sequence Z m (t), the charge decoupling impedance sequence Z is decomposed using variational mode decomposition. m (t) is decomposed into K eigenmode functions u k (t); The intrinsic mode function u1(t) with the lowest center frequency is extracted as the mechanical relaxation component; based on the mechanical relaxation component and the Hilbert transform, the envelope sequence A(t) is calculated, as follows: ； Where j is the imaginary unit, and H[ ] is the Hilbert transform; Calculate the arithmetic mean A of the envelope sequence A(t). c The envelope sequence is normalized to obtain the contact impedance fluctuation sequence R. b (t), the formula is: 。 7. The lithium battery operating status evaluation method based on data analysis according to claim 1, characterized in that, The step of filtering the multiplier-separated impedance matrix based on the contact impedance fluctuation sequence and outputting a corrected internal resistance spectrum includes: According to the aforementioned impedance matrix Z, the multiplier is separated. s and the contact impedance fluctuation sequence R b (t), the multiplier separation impedance matrix Z s The battery surface temperature T and state of charge S are mapped to the time axis according to the measurement time, and the impedance value R in the preset mid-frequency band is selected. m (f,t); with the contact impedance fluctuation sequence R b (t) is the reference signal, and an adaptive notch filter is constructed with a center frequency W. n (t) Based on the contact impedance fluctuation sequence R b The instantaneous dominant frequency of (t) is determined by the following formula: ； Where F is the Fourier transform and δ is the preset time window; The formula for the transfer function H(z) of the adaptive notch filter is: ； Where ρ is the preset notch depth coefficient, and z is the Z-transform variable; The impedance value R within the preset mid-frequency band m (f,t) is passed through the adaptive notch filter to obtain the filtered impedance value R. v (f,t); Iterate through each frequency f and calculate based on the filtered impedance value R. v The corrected internal resistance spectrum Z is obtained by calculating (f,t) and Hilbert transform. c (f,t), the formula is: 。 8. The lithium battery operating status evaluation method based on data analysis according to claim 1, characterized in that, The calculation of the battery's true health status based on the corrected internal resistance spectrum includes: According to the modified internal resistance spectrum Z c (f,t), extract the preset low-frequency point f L impedance value R l (t) and preset high-frequency point f H impedance value R h (t), calculate the impedance spectrum slope S in the preset mid-frequency band. m (t); R l (t), R h (t) and S m (t) are respectively compared with the corresponding characteristic value R in the initial state of the battery. l0 R h0 and S m0 By comparison, the true health state SOH(t) of the battery is calculated through weighted summation.

9. The lithium battery operating status evaluation method based on data analysis according to claim 1, characterized in that, The step of determining the battery health status based on the actual health status and outputting a battery health status signal includes: The battery health status is determined based on the actual health status SOH(t). When the actual health status SOH(t) is greater than or equal to a preset health threshold, a good battery health status signal is output. When the actual health status SOH(t) is less than the preset health threshold but greater than or equal to a preset fault threshold, a normal battery health status signal is output. When the actual health status SOH(t) is less than the preset fault threshold, a low battery health status signal is output, and maintenance personnel are prompted to replace the battery.

10. A lithium battery operating status assessment system based on data analysis, using the lithium battery operating status assessment method based on data analysis according to any one of claims 1-9, characterized in that, include: Data acquisition and preprocessing module: includes a data acquisition unit and a normalization processing unit; wherein, the data acquisition unit acquires the battery's electrical excitation response data and measurement parameters, and generates an impedance spectrum tensor; the normalization processing unit performs temperature normalization on the impedance spectrum tensor and outputs a temperature-decoupled impedance tensor. Operating condition decoupling and trend subtraction module: includes a rate decomposition unit and a trend subtraction unit; wherein, the rate decomposition unit performs rate dimension decomposition on the temperature decoupling impedance tensor and outputs a rate separation impedance matrix; the trend subtraction unit extracts the static contact impedance from the rate separation impedance matrix and subtracts the charge state trend, and outputs a charge decoupling impedance sequence. Fluctuation extraction and filtering correction module: includes a fluctuation extraction unit and a filtering correction unit; wherein, the fluctuation extraction unit decomposes the charge decoupling impedance sequence, separates the mechanical relaxation component, and outputs the contact impedance fluctuation sequence; the filtering correction unit filters the multiplier separation impedance matrix according to the contact impedance fluctuation sequence and outputs the corrected internal resistance spectrum. Health calculation and judgment module: includes a health calculation unit and a judgment output unit; wherein, the health calculation unit calculates the true health status of the battery based on the corrected internal resistance spectrum; the judgment output unit makes a judgment based on the true health status and outputs the battery health status signal.

Images