Abnormality identification method and device for power battery
By acquiring electrochemical impedance spectroscopy data and battery external characteristic data, and using principal component analysis and swarm intelligence optimization algorithms to construct an equivalent circuit model, the problem of misjudgment and missed judgment in traditional power battery anomaly identification methods has been solved, and high-precision power battery anomaly identification has been achieved.
Patent Information
- Application Number
- CN202511745242.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-25
- Publication Date
- 2026-03-13
AI Technical Summary
Traditional power battery anomaly identification methods rely on simple structural models and input data such as external battery characteristics like voltage, current, and temperature. These data are difficult to match with the internal mechanisms of the battery, leading to false positives and false negatives, and failing to meet the requirements for high-precision diagnosis.
By acquiring electrochemical impedance spectroscopy data and battery external characteristic data, key frequency bands are extracted using principal component analysis, a geometric deformation fingerprint spectrum is established, an equivalent circuit model is constructed, and abnormal power batteries are identified using swarm intelligence optimization algorithms and spatial clustering algorithms.
It improves the accuracy of component parameter identification and the precision of power battery anomaly identification, and can reflect changes in the internal mechanism of the battery, reducing false positives and false negatives.
Smart Images
Figure CN121659552A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vehicle battery management technology, and in particular to a method and apparatus for identifying anomalies in power batteries. Background Technology
[0002] New energy vehicles have become a key force in the green and low-carbon transformation of the transportation sector. Pure electric vehicles are rapidly gaining popularity due to their advantages such as zero emissions, high energy efficiency, and renewability. As the core energy unit, the performance and safety of the onboard power battery directly determine the vehicle's range, power output, and operational reliability. During long-term service, power batteries are frequently subjected to high-rate charging and discharging, frequent temperature differences, environmental stress, and aging, which can easily induce complex faults such as capacity degradation, electrode polarization, lithium dendrite puncture, and thermal runaway. These anomalies not only reduce system performance but may also lead to serious safety accidents such as fires and explosions. Traditional monitoring methods based on single sensors such as voltage, current, and temperature are often sluggish in complex operating conditions and early hazard identification, resulting in misjudgments and missed reports, making it difficult to support high-precision early warning and fault classification.
[0003] Traditional methods for identifying anomalies in vehicle-mounted power batteries primarily rely on simple models and input data consisting mainly of single sensor signals such as voltage, current, and temperature, combined with fixed thresholds for judgment. This heavily depends on prior knowledge. Furthermore, voltage, current, and temperature data represent external battery characteristics and do not reflect the internal mechanisms of the battery. When faced with complex and variable real-world operating conditions, these methods are prone to misjudgments and missed detections, failing to meet the demands for high-precision diagnostics.
[0004] Therefore, there is an urgent need to provide a new method for anomaly identification. Summary of the Invention
[0005] To address the problem that traditional anomaly identification methods rely on simple models and whose input data are mostly external battery characteristics such as voltage, current, and temperature, which do not conform to the internal mechanism of the battery and make it difficult to detect anomalies in battery circuit components, leading to false positives and false negatives, this invention provides an anomaly identification method for power batteries.
[0006] On the one hand, a method for anomaly identification of power batteries is provided, the method comprising: The electrochemical impedance spectroscopy data and external characteristic data of the vehicle power battery are acquired online to form the original data matrix for each power battery. Principal component analysis is used to extract key frequency bands from the original data matrix to characterize changes in the internal state of the battery, and a geometric deformation fingerprint spectrum is established based on the key frequency bands. Construct an equivalent circuit model of the power battery corresponding to the geometric deformation fingerprint spectrum, and derive the relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint spectrum to determine the geometric parameters; The geometric parameters are searched using a swarm intelligence optimization algorithm, and the relation is fitted with parameters using the data matrix of the key frequency band to identify the parameters of each component of each power battery. Spatial clustering algorithm is used to cluster the component parameters of each identified power battery in order to identify abnormal power batteries.
[0007] On the other hand, an anomaly identification device for a power battery based on the steps described in any method embodiment of the specification is provided, the device comprising: The acquisition unit is used to acquire electrochemical impedance spectroscopy data and external characteristic data of the vehicle power battery online, forming the original data matrix of each power battery; The extraction unit is used to extract key frequency bands from the original data matrix to characterize changes in the internal state of the battery using principal component analysis, and to establish a geometric deformation fingerprint spectrum based on the key frequency bands. The derivation unit is used to construct an equivalent circuit model of the power battery corresponding to the geometric deformation fingerprint spectrum, and derive the relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint spectrum to determine the geometric parameters. The identification unit is used to search for the geometric parameters using a swarm intelligence optimization algorithm and to fit the parameters of the relation using the data matrix of the key frequency band, so as to identify the parameters of each component of each power battery. The identification unit is used to cluster the component parameters of each identified power battery using a spatial clustering algorithm in order to identify abnormal power batteries.
[0008] On the other hand, a computer device is provided, the computer device including a memory and a processor, the memory for storing a computer program, and the processor for executing the computer program stored in the memory to implement the steps of the method described above.
[0009] On the other hand, a computer-readable storage medium is provided, wherein a computer program is stored therein, and when the computer program is executed by a processor, it implements the steps of the method described above.
[0010] On the other hand, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method described above.
[0011] The technical solution provided by this invention can bring at least the following beneficial effects: The system adds online-acquired electrochemical impedance spectroscopy (EIS) data of power batteries to generate an original data matrix for each power battery. Geometric deformation fingerprints are established from the data matrix at key frequency bands, constructing an equivalent circuit model of the power battery. The relationships between the parameters of each component in the equivalent circuit model and the geometric features in the EIS are derived to determine the geometric parameters. A swarm intelligence optimization algorithm is then used to identify the parameters of each component in each power battery. Spatial clustering algorithms are employed to cluster the parameters of each component in each power battery to identify abnormal power batteries. The system adds EIS data that reflects changes in the internal mechanisms of the battery, and the construction of the geometric deformation fingerprint and equivalent circuit model closely matches the internal mechanisms of the battery, improving the accuracy of component parameter identification and the precision of identifying abnormal power batteries at the vehicle end. Attached Figure Description
[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. 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.
[0013] Figure 1 This is a flowchart of an anomaly identification method for a power battery according to an embodiment of the present invention; Figure 2 This is a structural diagram of an abnormality identification device for a power battery provided in an embodiment of the present invention; Figure 3 This is a hardware architecture diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation
[0014] 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 some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0015] The following describes the specific implementation of the above concept.
[0016] Please refer to Figure 1 The present invention provides a method for anomaly identification of a power battery, the method comprising: Step 100: Acquire the electrochemical impedance spectroscopy data and external characteristic data of the vehicle power battery online to form the original data matrix for each power battery; Step 102: Use principal component analysis to extract key frequency bands from the original data matrix to characterize changes in the internal state of the battery, and establish a geometric deformation fingerprint spectrum based on the key frequency bands; Step 104: Construct an equivalent circuit model of the power battery corresponding to the geometric deformation fingerprint spectrum, derive the relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint spectrum, and determine the geometric parameters; Step 106: Use swarm intelligence optimization algorithm to search for geometric parameters, and use the data matrix of key frequency bands to fit the parameters of the relational formula in order to identify the parameters of each component of each power battery. Step 108: Use a spatial clustering algorithm to cluster the component parameters of each identified power battery in order to identify abnormal power batteries.
[0017] In this embodiment of the invention, newly acquired online electrochemical impedance spectroscopy data of power batteries are used to generate the original data matrix for each power battery. A geometric deformation fingerprint spectrum is established from the data matrix of key frequency bands, and an equivalent circuit model of the power battery is constructed. The relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint spectrum is derived to determine the geometric parameters. A swarm intelligence optimization algorithm is then used to identify the parameters of each component in each power battery. A spatial clustering algorithm is then used to cluster the parameters of each component in each power battery to identify abnormal power batteries. The addition of electrochemical impedance spectroscopy data that reflects changes in the internal mechanism of the battery, and the construction of the geometric deformation fingerprint spectrum and the equivalent circuit model, closely align with the internal mechanism of the battery, improving the accuracy of component parameter identification and the precision of identifying abnormal power batteries at the vehicle end.
[0018] The following description Figure 1 The execution method of each step is shown.
[0019] For step 100: Battery external characteristic data, including common data such as voltage, current, and temperature, are difficult to quickly and accurately reflect the internal mechanism of the battery. Therefore, this step adds online acquisition of electrochemical impedance spectroscopy data of vehicle power batteries. This not only increases the data input dimension, but also allows electrochemical impedance spectroscopy data to more accurately and sensitively reflect changes in the internal state of the battery, thereby improving the accuracy of vehicle identification.
[0020] Regarding step 102: In this embodiment, principal component analysis is used to reduce the dimensionality of the original data matrix with multiple frequency points, and the frequency points with the largest contribution of each principal component are identified in order to extract the data matrix of key frequency bands that represent the trend of battery state changes.
[0021] To closely approximate the battery diffusion mechanism and more accurately identify the parameters of various components in the power battery, the extracted key frequency bands include the mid-frequency band located on the Nyquist curve semi-ellipse and the low-frequency band used to characterize the charge diffusion process. Furthermore, a geometric deformation fingerprint spectrum is established based on the key frequency bands. This spectrum includes at least the charge transfer impedance arc shrinkage rate, the ohmic internal resistance shift, and the diffusion process phase angle. Therefore, the key frequency bands also include the low-frequency band used to characterize the charge diffusion process, and the geometric deformation fingerprint spectrum also includes the diffusion process phase angle. By considering the battery's diffusion mechanism, this approach enables more accurate analysis and identification of battery state changes, improving identification accuracy.
[0022] Regarding step 104: In some implementations, the step "constructing an equivalent circuit model of the power battery corresponding to the geometric deformation fingerprint" includes: The equivalent circuit model of the power battery is composed of an ohmic internal resistance, two RC rings and a diffusion impedance connected in series. The RC ring is composed of a charge transfer impedance and a CPE element connected in parallel. Each RC ring corresponds to a half ellipse of the Nyquist curve, and the complex impedance of each RC ring is calculated based on the ohmic internal resistance, the charge transfer impedance of the RC ring and the parameters of the CPE element. The diffusion impedance corresponds to a sloping line of the Nyquist curve.
[0023] In this embodiment, the geometric deformation fingerprint spectrum includes at least the charge transfer impedance arc shrinkage rate, the ohmic internal resistance offset, and the diffusion process phase angle. The corresponding equivalent circuit model of the power battery consists of an ohmic internal resistance, two RC rings, and a diffusion impedance connected in series. The RC ring is composed of a charge transfer impedance and a CPE element connected in parallel. Each RC ring corresponds to a semi-ellipse of the Nyquist curve. The frequency points of the mid-frequency data are components of the semi-elliptical curve on the Nyquist curve. It can be seen that the Nyquist curve has two semi-ellipses, each corresponding to one of the two RC rings.
[0024] The complex impedance of each RC ring is calculated using the following formula: In the formula, For ohmic internal resistance, The charge transfer impedance of the RC ring is... Here are the CPE element parameters for this RC ring, where j is the imaginary unit. ω is the angular frequency.
[0025] In this embodiment, in order to closely approximate the battery mechanism, the equivalent circuit model of the power battery not only includes ohmic internal resistance, charge transfer impedance and CPE element, but also adds diffusion impedance. A sloping line of the Nyquist curve is used to fit low-frequency data, simulate and analyze the battery diffusion mechanism, and more accurately identify the parameters of each component of the power battery.
[0026] In some implementations, the step "deriving the relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint, and determining the geometric parameters" includes: For each RC ring, an elliptic relationship is derived based on the ohmic internal resistance, the charge transfer impedance of the RC ring, and the parameters of the CPE element. A slant-line relationship is derived based on diffusion impedance; the parameters of each component in the equivalent circuit model include ohmic internal resistance, diffusion impedance, charge transfer impedance of the two RC rings, and CPE element. The major axis, minor axis, and center coordinates of the two elliptic relationships, as well as the inclination angle and intercept of the oblique line relationship, are used as geometric parameters.
[0027] In this embodiment of the invention, the elliptic relation is expressed as follows: in, In the formula, Let be the real part of the impedance. This is the imaginary part of the impedance. Let be the coordinates of the center of the i-th ellipse. Let b be the major axis of the i-th ellipse. i Let Ri be the minor axis of the i-th ellipse, R0 be the ohmic internal resistance, and Ri be the minor axis of the ellipse. i This represents the charge transfer impedance of the RC ring corresponding to the ellipse. The peak value is arc-shaped, C i The parameters of the CPE element corresponding to the RC ring of the ellipse are given.
[0028] In this embodiment, the major axis of the ellipse corresponds to the charge transfer resistor of the RC ring, and the minor axis of the ellipse is related to the CPE element of the corresponding RC ring. The center coordinates are... and These correspond to the starting point and the maximum arc position of the total series resistance, respectively, with the peak value of the arc corresponding to the frequency point. The offset of the center real axis is related to the ohmic internal resistance, and the peak frequency of the semi-ellipse is used to identify the CPE component parameters.
[0029] In the low-frequency region, the diffusion process is dominant, and the impedance tends to have a sloping characteristic. The formula for calculating the diffusion impedance is: in, is the diffusion coefficient; the larger the coefficient, the stronger the diffusion ability. f is the frequency, and j is the imaginary unit.
[0030] In the Nyquist plot, there is a corresponding diagonal line. The relationship between this diagonal line and the low-frequency points is as follows: in, Let be the real part of the impedance. θ is the imaginary part of the impedance, θ is the tilt angle, the ideal value for semi-infinite diffusion is θ≈45°, and c is the intercept, which reflects the polarization shift when extrapolating at low frequencies.
[0031] Therefore, the geometric parameters are the major axis, minor axis, and center coordinates of the two elliptic relations, and the inclination angle and intercept of the oblique line relation. Regarding step 106: In some implementations, step 106 may include: For each power battery, the following is performed: S1 generates an initial population based on geometric parameters; the initial population contains several sets of geometric parameters.
[0032] It is understandable that a set of geometric parameters It is a random solution.
[0033] S2, input the initial population into the Fireworks Algorithm to generate the initial optimized population.
[0034] In this step, the initial population is input into the fireworks algorithm. Each firework represents a random solution. The better-performing solutions will produce a small number of fireworks and perform a local search around them. The worse-performing solutions will produce more fireworks and cover a larger area to escape the local optimum.
[0035] The formulas for calculating the explosion range A and the number of sparks S are as follows: in, Represents the fitted function. It is the minimum fitness among the potential optimal solutions; It represents the maximum fitness in the potential optimal solution; A is a constant used to adjust the explosion radius; M is used to describe the number of sparks.
[0036] The diverse initial optimization populations generated in the parameter space using the fireworks algorithm are input into three optimizers. By controlling the number and amplitude of explosions to regulate the density and distribution of solutions, each optimizer can be provided with a high-quality reference and search starting point for further optimization identification. This reduces the number of iterations for each optimizer, saving time and computational resources.
[0037] S3. Input the initial optimization population into three optimizers with different optimization algorithms; where the optimization algorithm is either the firefly algorithm, the light particle swarm optimization algorithm, or the sunflower algorithm.
[0038] Next, the optimization process of the three optimization algorithms will be explained.
[0039] The Firefly Algorithm finds local optima by leveraging the attraction between fireflies. Using the solution set of the Fireworks Algorithm as the initial population, each individual corresponds to a set of solutions for search and optimization. A fitness function is established to adjust the brightness of individuals; lower fitness corresponds to higher brightness. During iteration, fireflies move based on the brightness and position information of other individuals. If an individual is attracted to a brighter (better) solution, it updates its position according to the position update formula, guiding other solutions closer to the brighter solution, creating a gradient descent effect. Finally, after iterating to the maximum number of iterations or when the overall fitness of the population converges to a stable region, the locally optimal solution found by the fireflies is output.
[0040] The Particle Swarm Optimization (PSO) algorithm simulates a luminous swarm to perceive brightness and migrate, thus optimizing the handling of multi-peak problems and the coexistence of multiple optimal regions. Using the solution set of the Fireworks Algorithm as the initial population, an initial brightness is set to identify solutions in multiple regions. Each individual has a dynamically adjustable perception radius. Within this radius, individuals only focus on brighter individuals (better solutions), avoiding ineffective movements. Once an individual finds a brighter individual in its neighborhood, it begins to move towards that individual, achieving adaptive clustering. After each iteration, the individual's brightness is updated based on its new fitness. By dynamically adjusting the perception radius, the self-organization of the population is enhanced; the perception radius expands when there are too few neighbors and shrinks when there are too many, ultimately outputting the brightest individual in the current population.
[0041] The Sunflower Optimization Algorithm is based on the phenomenon of sunflowers continuously moving towards the sun during their growth. Through deterministic growth and random perturbation, it aims to guide individual sunflowers towards the globally brightest direction while simultaneously increasing perturbation to enhance diversity. The initial position of each sunflower is determined by the spark solution set provided by the Fireworks Algorithm. Once the globally optimal individual (the sun) is identified, the sunflower grows in that direction. Unlike the Firefly Algorithm and the Particle Swarm Optimization Algorithm, the Sunflower Algorithm does not rely on the population brightness structure; it only follows a unique global guiding point. The growth and Gaussian perturbation process is repeated until the maximum number of iterations is reached, at which point the optimal individual is output as the optimal solution.
[0042] S4. The optimal solutions output by the three optimizers are weighted and fused based on brightness to obtain the fused solution.
[0043] In the collaborative adaptive swarm optimization structure, the optimal solutions output by the three sub-optimizers are fused based on brightness to obtain the fused solution. : in, The weighting factor is the optimal solution for each sub-optimizer; the smaller the error, the higher the weight.
[0044] S5. By substituting the coordinates of the major axis, minor axis, and center of the circle in the fusion solution into the elliptic relation, the estimated value of the imaginary part of the impedance is obtained. The data matrix of the mid-frequency band in the key frequency range of the power battery is used to perform parameter fitting on the elliptic relation to determine the mid-frequency band loss function.
[0045] In this embodiment, the mid-frequency loss function is: in, This is the imaginary part of the actual measured value. It is an estimate of the imaginary part of the impedance obtained by solving the elliptic formula. These are the coordinates of the major axis, minor axis, and center of the semi-ellipse, respectively, and N represents the number of frequency points on the two ellipses.
[0046] S6. The diffusion impedance is solved by substituting the tilt angle and intercept in the fused solution into the slant line relationship. The slant line relationship is then parameterized using the data matrix of the low-frequency band in the key frequency range to determine the low-frequency band loss function. Finally, the total loss function is determined based on the mid-frequency band loss function and the low-frequency band loss function.
[0047] Low-frequency loss function: in, This is a low-frequency loss function, dominated by diffusion, and its Nyquist trajectory is approximated by a straight line. N represents the number of frequency points. This is the imaginary part of the actual measured value. It is an estimated value of the imaginary part of the impedance obtained by solving the slant formula.
[0048] Total loss function: in, , Weights are set to avoid uneven sample size.
[0049] S7, with the total loss function as the objective, the fireworks algorithm and three optimizers calculate the fitness of candidate solutions and update the population based on minimizing the loss function; a new generation of initial population is generated with the previous round of weighted fusion solution as the center, and the process returns to step S2 for iteration until the total loss is lower than the set threshold or the maximum number of iterations is reached, and the final weighted fusion solution is used as the predicted value of geometric parameters.
[0050] S8 maps the predicted geometric parameters back to the component parameters to obtain the component parameters of the power battery.
[0051] Based on the relationships between the geometric parameters of the Nyquist curve and the parameters of each component in the equivalent circuit model, the predicted values of the geometric parameters can be obtained. Mapping back to component parameters enables the identification of parameters in the equivalent circuit model, that is, the identification of the parameters of each component in the corresponding power battery.
[0052]
[0053] Therefore, it is possible to base it on the optimal geometric parameters. The component parameters of the power battery are obtained through mapping. .
[0054] Regarding step 108: In this step, the parameters of each component of the identified power battery are input into the spatial clustering model, and an unsupervised density clustering algorithm is used to identify the abnormal parameter distribution in the battery group. Based on the density reachability criterion, the clusters to which each sample belongs are divided, and abnormal groups that deviate from the main cluster structure are marked as abnormal batteries, thereby realizing the identification of abnormal batteries at the vehicle end.
[0055] Please refer to Figure 2 This invention provides an anomaly identification device for a power battery, used to implement the steps of any method embodiment in the specification. The device includes: The acquisition unit 201 is used to acquire electrochemical impedance spectroscopy data and external characteristic data of the vehicle power battery online, forming the original data matrix of each power battery. Extraction unit 202 is used to extract key frequency bands from the original data matrix to characterize changes in the internal state of the battery using principal component analysis, and to establish a geometric deformation fingerprint spectrum based on the key frequency bands. The derivation unit 203 is used to construct an equivalent circuit model of the power battery corresponding to the geometric deformation fingerprint spectrum, and derive the relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint spectrum to determine the geometric parameters. The identification unit 204 is used to search for geometric parameters using a swarm intelligence optimization algorithm and to fit the relational formula using a data matrix of key frequency bands in order to identify the parameters of each component of each power battery. The identification unit 205 is used to cluster the component parameters of each identified power battery using a spatial clustering algorithm in order to identify abnormal power batteries.
[0056] It should be noted that the above device embodiments and method embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.
[0057] Embodiments of this application also provide a computer device, please refer to... Figure 3The computer device includes a processor and a memory, the memory storing at least one instruction, at least one program, code set or instruction set, the at least one instruction, at least one program, code set or instruction set being loaded and executed by the processor to implement the abnormal identification method for power batteries provided in the above-described method embodiments.
[0058] The embodiments of this application also provide a computer-readable storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, at least one program, code set, or instruction set is loaded and executed by a processor to implement the abnormal identification method for power batteries provided in the above-described method embodiments.
[0059] Embodiments of this application also provide a computer program product, which includes a computer program. A processor of a computer device reads the computer program from a computer-readable storage medium and executes the computer program, causing the computer device to perform any of the power battery anomaly identification methods described in the above embodiments.
[0060] For ease of description, the above devices or apparatuses are described separately according to their functions, divided into various modules or units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.
[0061] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, 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 storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes 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 the embodiments of this application.
[0062] Finally, it should be noted that in this document, relational terms such as first, second, third, and fourth are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.
[0063] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.
Claims
1. A method for anomaly identification of a power battery, characterized in that, include: The electrochemical impedance spectroscopy data and external characteristic data of the vehicle power battery are acquired online to form the original data matrix for each power battery. Principal component analysis is used to extract key frequency bands from the original data matrix to characterize changes in the internal state of the battery, and a geometric deformation fingerprint spectrum is established based on the key frequency bands. Construct an equivalent circuit model of the power battery corresponding to the geometric deformation fingerprint spectrum, and derive the relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint spectrum to determine the geometric parameters; The geometric parameters are searched using a swarm intelligence optimization algorithm, and the relation is fitted with parameters using the data matrix of the key frequency band to identify the parameters of each component of each power battery. Spatial clustering algorithm is used to cluster the component parameters of each identified power battery in order to identify abnormal power batteries.
2. The method as described in claim 1, characterized in that, The key frequency bands include the mid-frequency band located on the Nyquist curve semi-ellipse and the low-frequency band used to characterize the charge diffusion process; the geometric deformation fingerprint spectrum includes at least the charge transfer impedance arc shrinkage rate, ohmic internal resistance offset, and diffusion process phase angle.
3. The method as described in claim 1, characterized in that, The construction of the equivalent circuit model of the power battery corresponding to the geometric deformation fingerprint includes: The equivalent circuit model of the power battery is composed of an ohmic internal resistance, two RC rings and a diffusion impedance connected in series; the RC ring is composed of a charge transfer impedance and a CPE element connected in parallel, each RC ring corresponds to a half ellipse of the Nyquist curve, and the complex impedance of each RC ring is calculated based on the charge transfer impedance of the RC ring and the parameters of the CPE element; the diffusion impedance corresponds to a sloping line of the Nyquist curve.
4. The method as described in claim 3, characterized in that, Deriving the relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint, and determining the geometric parameters, including: For each RC ring, an elliptic relationship is derived based on the ohmic internal resistance, the charge transfer impedance of the RC ring, and the CPE element parameters. Based on the diffusion impedance, a slant-line relationship is derived; the parameters of each component in the equivalent circuit model include ohmic internal resistance, diffusion impedance, charge transfer impedance of the two RC rings, and CPE element; The major axis, minor axis, and center coordinates of the two elliptic relationships, as well as the inclination angle and intercept of the oblique line relationship, are used as geometric parameters.
5. The method as described in claim 4, characterized in that, The elliptic relation is expressed as follows: in, In the formula, Let be the real part of the impedance. This is the imaginary part of the impedance. Let be the coordinates of the center of the i-th ellipse. Let be the major axis of the i-th ellipse. Let be the minor axis of the i-th ellipse. The internal resistance is ohmic. Let be the charge transfer impedance of the RC ring corresponding to the ellipse. It is an arc-shaped peak. The parameters of the CPE element corresponding to the RC ring of the ellipse are given.
6. The method as described in claim 4, characterized in that, The process of using a swarm intelligence optimization algorithm to search for the geometric parameters and using the data matrix of the key frequency band to fit the parameters of the relational expression to identify the parameters of each component of each power battery includes: For each power battery, the following is performed: S1, an initial population is generated based on the geometric parameters; wherein, the initial population contains several sets of the geometric parameters; S2, input the initial population into the fireworks algorithm to generate an initial optimized population; S3, the initial optimized population is input into three optimizers of different optimization algorithms; wherein the optimization algorithm is either the firefly algorithm, the light particle swarm optimization algorithm, or the sunflower algorithm. S4, the optimal solutions output by the three optimizers are weighted and fused based on brightness to obtain the fused solution; S5. Substitute the coordinates of the major axis, minor axis, and center of the circle in the fusion solution into the elliptic relation to solve for the estimated value of the imaginary part of the impedance. Use the data matrix of the mid-frequency band in the key frequency band of the power battery to perform parameter fitting on the elliptic relation to determine the mid-frequency band loss function. S6. Substitute the tilt angle and intercept in the fusion solution into the slant line relationship to solve for the diffusion impedance. Use the data matrix of the low-frequency band in the key frequency band to perform parameter fitting on the slant line relationship to determine the low-frequency band loss function. And determine the total loss function based on the mid-frequency band loss function and the low-frequency band loss function. S7, with the total loss function as the objective, the fireworks algorithm and three optimizers calculate the fitness of candidate solutions and update the population based on minimizing the loss function; a new generation of initial population is generated with the previous round of weighted fusion solution as the center, and the process returns to step S2 for iteration until the total loss is lower than the set threshold or the maximum number of iterations is reached, and the final fusion solution is used as the predicted value of geometric parameters. S8, map the predicted geometric parameters back to the component parameters to obtain the component parameters of the power battery.
7. An anomaly identification device for a power battery, used to implement the steps of the method according to any one of claims 1-6, characterized in that, include: The acquisition unit is used to acquire electrochemical impedance spectroscopy data and external characteristic data of the vehicle power battery online, forming the original data matrix of each power battery; The extraction unit is used to extract key frequency bands from the original data matrix to characterize changes in the internal state of the battery using principal component analysis, and to establish a geometric deformation fingerprint spectrum based on the key frequency bands. The derivation unit is used to construct an equivalent circuit model of the power battery corresponding to the geometric deformation fingerprint spectrum, and derive the relationship between the parameters of each component in the equivalent circuit model and the geometric features in the geometric deformation fingerprint spectrum to determine the geometric parameters. The identification unit is used to search for the geometric parameters using a swarm intelligence optimization algorithm and to fit the parameters of the relation using the data matrix of the key frequency band, so as to identify the parameters of each component of each power battery. The identification unit is used to cluster the component parameters of each identified power battery using a spatial clustering algorithm in order to identify abnormal power batteries.
8. A computer device, characterized in that, The computer device includes a memory and a processor. The memory is used to store computer programs, and the processor is used to execute the computer programs stored in the memory to implement the steps of the method according to any one of claims 1-6.
9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the method described in any one of claims 1-6.
10. A computer program product, characterized in that, Includes a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1-6.