Lithium ion battery pack internal short circuit fault diagnosis method based on relaxation time and unsupervised learning

Through the method based on relaxation time and unsupervised learning, the relaxation voltage characteristics of the battery pack are extracted and the local outlier factor algorithm is used to solve the problem that threshold cannot be adapted in the prior art, and short circuits are realized in the lithium-ion battery pack Accurate detection and evaluation of faults, suitable for complex working conditions.

CN118914901APending Publication Date: 2024-11-08CHONGQING UNIV +1

Patent Information

Application Number
CN202411004524.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-25
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing short-circuit fault diagnosis technology in the lithium-ion battery pack has the threshold that cannot be adapted and is not suitable for complex and changeable actual working conditions, making it difficult to effectively monitor the risk of short-circuit in the battery pack.

Method used

Using a method based on relaxation time and unsupervised learning, the operating data of the battery pack is collected, the relaxation voltage of the static stage after charging is selected, the characteristic relaxation time segment is extracted, and the fault characteristics are used by local outlier factor algorithm. Conduct identification and scoring to realize detection and severity assessment of internal short circuit faults.

Benefits of technology

It realizes effective detection and fault degree scoring of short-circuit faults in each single unit in the battery pack, is suitable for complex actual working conditions, improves the applicability and accuracy of diagnosis, and is more practical than the traditional fixed threshold method.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118914901A_ABST
    Figure CN118914901A_ABST
Patent Text Reader

Abstract

The invention relates to a lithium ion battery pack internal short circuit fault diagnosis method based on relaxation time and unsupervised learning, and belongs to the technical field of batteries. The method comprises the following steps: collecting operation data of a battery pack of the electrified equipment and monomer operation data of each monomer in the battery pack, and establishing a battery operation database; according to the collected battery data, relaxation voltage of each monomer in a standing stage after charging of the battery pack is finished is screened out; extracting a relaxation time change curve from the relaxation voltage of each monomer, and intercepting a segment of a transient process stability degree coefficient alpha in a preset range as a characteristic relaxation time segment; extracting fault features on the basis of the feature relaxation time slice according to the principle of statistics; and performing internal short-circuit fault identification on each monomer by using a local outlier factor algorithm, and scoring the severity of the fault. According to the invention, internal short-circuit fault detection and fault degree scoring can be effectively carried out on the battery pack.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of batteries and relates to a method for diagnosing short-circuit faults in a lithium-ion battery pack based on relaxation time and unsupervised learning. Background Art

[0002] At present, as the low-carbon process is further accelerated, battery packs are widely used as core components in electrified equipment such as new energy vehicles, energy storage power stations and military machinery. The battery pack itself involves complex electrochemical reactions and has many monomers, which are easily disturbed by the external environment. In actual operation, there is a greater risk of internal short circuits due to manufacturing defects and abuse. Therefore, real-time monitoring of the internal short circuit risk of the battery pack and ensuring timely inspection and maintenance are the key to ensuring the safe operation of electrified equipment.

[0003] For example, the prior art with publication number CN113848495A proposes an internal micro-short circuit fault diagnosis method based on a charging curve, the method includes charging the battery, obtaining a battery charging voltage curve, and obtaining an IC curve on this basis; after charging is completed, continuing to collect the battery terminal voltage, and judging the battery fault according to the recovery of the battery terminal voltage; analyzing the attenuation rate between two charging capacities of the battery to judge the battery fault; and comprehensively diagnosing the battery fault according to the IC curve analysis results, the terminal voltage recovery and the capacity attenuation rate.

[0004] However, the existing internal short-circuit fault diagnosis technology and the above-mentioned existing technologies have problems such as the inability to adapt the threshold value when analyzing the fault, failure to consider the inconsistency of the normal operation of the battery pack, and unsuitability for complex and changeable actual working conditions. Summary of the invention

[0005] In view of this, an object of the present invention is to provide a short circuit fault diagnosis method for a lithium ion battery pack based on relaxation time and unsupervised learning.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A method for diagnosing short circuit faults in a lithium-ion battery pack based on relaxation time and unsupervised learning, comprising the following steps:

[0008] S1: Collect the operating data of the battery pack of the electrified equipment and the single-cell operating data of each single cell in the battery pack, and establish a battery operating database;

[0009] S2: based on the collected battery data, screening out the relaxation voltage of each cell in the static stage after the battery pack is charged;

[0010] S3: extracting a relaxation time variation curve from the relaxation voltage of each monomer, and intercepting a segment where the transient process stability coefficient α is within a preset range as a characteristic relaxation time segment;

[0011] S4: Extract fault features based on characteristic relaxation time segments according to statistical principles, and use the fault features as five-dimensional feature vectors for fault identification;

[0012] S5: Based on the five-dimensional feature vector, the local outlier factor algorithm is used to identify the internal short circuit fault of each monomer and score its fault severity.

[0013] Further, step S1 includes the following steps:

[0014] S11: Collecting operating data parameters of the battery pack, wherein the operating data parameters at least include time, current, and voltage;

[0015] S12: Collecting the cell operation data in the battery pack, including the cell voltages of all cells;

[0016] S13: Establish a battery pack operation database.

[0017] Further, step S2 includes the following steps:

[0018] S21: Preliminarily screening out characteristic data related to the single cell voltage theory, wherein the characteristic data related to the single cell voltage theory at least includes: total current, battery pack voltage, average single cell voltage, state of charge (SOC), internal resistance, state of health (SOH), and discharge rate;

[0019] S22: By analyzing the total current, battery pack voltage and SOC of a battery pack in a charging segment with a certain charging time and charging power, the relaxation voltage of each cell in the static stage after the battery pack is charged is screened out.

[0020] Further, step S3 includes the following steps:

[0021] S31: According to the characteristics of lithium-ion batteries, they are equivalent to nonlinear RC equivalent circuits. When the battery is fully charged, due to its internal polarization phenomenon, its terminal voltage exhibits a transient process with a certain regular relaxation time. The voltage expression at time t is:

[0022]

[0023] Among them, U0 is the starting voltage of the transient process, U OCV This is the open circuit voltage of the battery. is the average time constant in the time period from 0 to t;

[0024] Transforming the above formula, we get the following expression:

[0025]

[0026] The coefficient α indicates the degree to which the transient process tends to a steady state;

[0027] S32: Describe the degree to which the battery transient process tends to a steady state as a coefficient α, and calculate the average time constant of each stage according to the preset step size within the preliminary range. And by further analyzing the average time constant within the preliminary range The segments with the stability coefficient of the transient process within a preset range are intercepted as characteristic relaxation time segments.

[0028] Further, in step S32, the preliminary range is α=0-5, and the preset step size is 0.1; the segments with obvious fault characteristics are concentrated in the range of α=0-1, and this segment should be included as much as possible when extracting the fault characteristics.

[0029] Further, step S4 includes the following steps:

[0030] S41: The average time constant of the transient stability coefficient α of each monomer within the preset range is recorded in sequence according to the preset step length as Where n represents the number of relaxation segments cut out with a preset step size;

[0031] S42: Calculation The statistical characteristics mean Mean, variance Var, maximum value Max, skewness Ske and kurtosis Kur are used as fault features, where:

[0032]

[0033]

[0034]

[0035]

[0036]

[0037] The above fault characteristics are used as five-dimensional feature vectors for subsequent fault identification.

[0038] Further, step S5 includes the following steps:

[0039] The local abnormal outlier factor describes the degree of abnormality of the parameter by evaluating the local deviation of the observed data relative to the adjacent data; each feature of the fault feature mean Mean, variance Var, maximum value Max, skewness Ske and kurtosis Kur is regarded as a feature point, and each complete relaxation voltage is calculated to obtain a feature point;

[0040] The characteristic points calculated by the relaxation voltage of the cycle to be detected and the reference points calculated by several normal cycles are used as the input of LOF, and the outliers and the degree of outliers of each characteristic point are output; wherein the reference points refer to the mean Mean, variance Var, maximum value Max, skewness Ske and kurtosis Kur under the normal cycle;

[0041] If the characteristic points of the cycle to be detected do not show outliers, it is determined that the battery has no internal short circuit fault. If the characteristic points of the cycle to be detected are outliers, it is determined that the battery has an internal short circuit fault, and the severity of the fault is determined based on the outlier degree of the outlier.

[0042] The beneficial effects of the present invention are:

[0043] The present invention can effectively detect the internal short-circuit faults of each cell in the battery pack and score the fault degree. In addition, the present invention adopts an unsupervised learning method, so that the different actual operating conditions and working environments of the vehicle have less impact on the effect of fault detection. Compared with the traditional fixed threshold method, the present invention has better applicability and practicality for complex real vehicle scenarios.

[0044] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0046] Figure 1 It is an overall flow chart of the internal short circuit fault diagnosis method of the power battery of the present invention;

[0047] Figure 2 are cell voltage curves of a normal cell and a faulty cell in an embodiment of the present invention;

[0048] Figure 3 It is the internal short circuit fault evaluation result of the normal monomer and the faulty monomer in the embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0050] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0051] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0052] See also Figure 1 to Figure 3 , which is a short-circuit fault diagnosis method for lithium-ion battery packs based on relaxation time and unsupervised learning.

[0053] like Figure 1 The overall flow chart of the safety risk assessment method of the power battery of the present invention is shown, which specifically includes the following steps:

[0054] S1: Collect the operating data of the battery pack of the electrified equipment and the single-cell operating data of each single cell in the battery pack, and establish a battery operating database;

[0055] S2: based on the collected battery data, screening out the relaxation voltage of each cell in the static stage after the battery pack is charged;

[0056] S3: Extract the relaxation time curve from the relaxation voltage of each monomer and extract the fragment with obvious characteristics;

[0057] S4: Extract fault features based on relaxation time segments according to statistical principles;

[0058] S5: Use the local outlier factor algorithm to identify the internal short circuit fault of each monomer and score its fault severity.

[0059] Example

[0060] As a preferred implementation, the above step S1 specifically includes S11-S13:

[0061] S11: Collect the operating data of a battery pack of an electric vehicle, including battery parameters such as time, current, and voltage;

[0062] S12: Collecting the cell operation data in the battery pack, including the cell voltages of all cells;

[0063] S13: Establish a battery pack operation database.

[0064] As a preferred implementation, the above step S2 specifically includes S21-S22:

[0065] S21: Preliminary screening of data theoretically related to monomer voltage as features;

[0066] S22: By analyzing the total current, battery pack voltage and SOC, the relaxation voltage of each cell in the static stage after the battery pack is charged is screened out.

[0067] As a preferred implementation manner, the data theoretically related to the cell voltage in step S21 include total current, battery pack voltage, average cell voltage, SOC, internal resistance, SOH, discharge rate, etc.

[0068] As a preferred implementation, in step S22, it is necessary to screen out charging segments with a certain charging time and charging power, and on this basis, screen out the relaxation voltage of each cell in the static stage after the battery pack is charged.

[0069] As a preferred implementation, the above step S3 specifically includes S31-S32:

[0070] S31: According to the characteristics of lithium-ion batteries, they can be equivalent to nonlinear RC equivalent circuits. When the battery is fully charged, due to its internal polarization phenomenon, its terminal voltage exhibits a transient process with a certain regular relaxation time. The voltage expression at time t is:

[0071]

[0072] Among them, U0 is the starting voltage of the transient process, U OCV This is the open circuit voltage of the battery. is the average time constant in the time period from 0 to t;

[0073] Transforming the above formula, we can get the following expression:

[0074]

[0075] The coefficient α can be described as the degree to which the transient process tends to a steady state. That is, when α = 1, the voltage drop is approximately 63.2% of the total voltage drop in the transient process. At this time, the average time constant When α tends to infinity, it indicates that the transient process reaches a steady state. It is generally believed that when α reaches 5, that is, When the transient process is considered to have ended, that is, when the voltage drop reaches about 99.33% of the total voltage drop, the terminal voltage U0 at this time can be considered as the open circuit voltage U OCV .

[0076] S32: Describe the degree to which the battery transient process tends to a steady state as a coefficient α, and calculate the average time constant of each stage in a preliminary range of α=0-5 and a preset step size of 0.1 And through further analysis, segments with obvious features within a preset range are intercepted. In this embodiment, the preset range refers to α=0-1.

[0077] As a preferred implementation, the above step S4 specifically includes S41-S42:

[0078] S41: Based on the principle in step S32, the average time constant of each monomer whose α is in the range of 0 to 1 with a step length of 0.1 is obtained n represents the number of relaxation segments cut out with a preset step size of 0.1, that is, n=10 in this embodiment.

[0079] S42: Calculation The statistical characteristics mean, variance, maximum value, skewness and kurtosis are used as the five-dimensional input feature vector of the unsupervised learning algorithm in step S5.

[0080] As a preferred implementation, the statistical feature mean (Mean), variance (Var), maximum value (Max), skewness (Ske) and kurtosis (Kur) in step S42 can be expressed as:

[0081]

[0082]

[0083]

[0084]

[0085]

[0086] As a preferred implementation, the above step S5 specifically includes:

[0087] The local abnormal outlier factor describes the degree of abnormality of the parameter by evaluating the local deviation of the observed data relative to the adjacent data; each feature of the fault feature mean Mean, variance Var, maximum value Max, skewness Ske and kurtosis Kur is regarded as a feature point, and each complete relaxation voltage is calculated to obtain a feature point;

[0088] The characteristic points calculated from the relaxation voltage of the cycle to be detected and the reference points calculated from several normal cycles are used as the input of LOF, and the outliers and the degree of outlier of each characteristic point are output; wherein, the reference points refer to the mean Mean, variance Var, maximum value Max, skewness Ske and kurtosis Kur under normal cycles; through the above scheme, the internal short-circuit fault evaluation of each monomer in the battery pack can be effectively realized.

[0089] In order to illustrate the effectiveness of the present invention, this embodiment prepares laboratory operation data of a 6-cell commercial 18650 series battery pack. The battery is charged and discharged at a constant temperature of 25°C, wherein each set of cycles includes a dynamic stress test (DST) discharge condition and a 1C constant current charge condition. In the first 8 cycles, each cell is normal, and in the last 8 cycles, a 1000Ω resistor is connected in parallel to cell No. 2 to simulate the electrical characteristics of an internal short circuit. Figure 2 The operating data of the voltage of each cell in the series battery pack is shown in black, where the voltage curve of the faulty cell is shown in black. After testing, the present invention can accurately evaluate the internal short circuit fault of the battery and its severity. The evaluation results are as follows: Figure 3 As shown in the figure, the black solid circle indicates that an internal short circuit fault is detected, the black hollow circle indicates that an internal short circuit fault is not detected, and the gray solid circle indicates a false alarm. The final accuracy rate reaches more than 95%. Therefore, an effective evaluation of battery internal short circuit faults based on unsupervised learning is achieved.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A method for diagnosing short-circuit faults in lithium-ion battery packs based on relaxation time and unsupervised learning, characterized in that: It includes the following steps: S1: Collect the operating data of the battery pack of the electrified equipment and the single-cell operating data of each single cell in the battery pack, and establish a battery operating database; S2: based on the collected battery data, screening out the relaxation voltage of each cell in the static stage after the battery pack is charged; S3: extracting a relaxation time variation curve from the relaxation voltage of each monomer, and intercepting a segment where the transient process stability coefficient α is within a preset range as a characteristic relaxation time segment; S4: Extract fault features based on characteristic relaxation time segments according to statistical principles, and use the fault features as five-dimensional feature vectors for fault identification; S5: Based on the five-dimensional feature vector, the local outlier factor algorithm is used to identify the internal short-circuit fault of each monomer and score its fault severity.

2. The method for diagnosing short circuit faults in lithium-ion battery packs based on relaxation time and unsupervised learning according to claim 1, characterized in that: Step S1 includes the following steps: S11: Collecting operating data parameters of the battery pack, wherein the operating data parameters at least include time, current, and voltage; S12: Collecting the cell operation data in the battery pack, including the cell voltages of all cells; S13: Establish a battery pack operation database.

3. The method for diagnosing short-circuit faults in lithium-ion battery packs based on relaxation time and unsupervised learning according to claim 1, characterized in that: Step S2 includes the following steps: S21: Preliminarily screen out characteristic data related to the single cell voltage theory, wherein the characteristic data related to the single cell voltage theory at least includes: total current, battery pack voltage, average single cell voltage, state of charge SOC, internal resistance, state of health SOH, and discharge rate; S22: By analyzing the total current, battery pack voltage and SOC of a battery pack in a charging segment with a certain charging time and charging power, the relaxation voltage of each cell in the static stage after the battery pack is charged is screened out.

4. The method for diagnosing short-circuit faults in lithium-ion battery packs based on relaxation time and unsupervised learning according to claim 1, characterized in that: Step S3 includes the following steps: S31: According to the characteristics of lithium-ion batteries, they are equivalent to nonlinear RC equivalent circuits. When the battery is fully charged, due to its internal polarization phenomenon, its terminal voltage exhibits a transient process with a certain regular relaxation time. The voltage expression at time t is: Among them, U0 is the starting voltage of the transient process, U OCV This is the open circuit voltage of the battery. is the average time constant in the time period from 0 to t; Transforming the above formula, we get the following expression: The coefficient α indicates the degree to which the transient process tends to a steady state; S32: Describe the degree to which the battery transient process tends to a steady state as a coefficient α, and calculate the average time constant of each stage according to the preset step size within the preliminary range. And by further analyzing the average time constant within the preliminary range The segments with the stability coefficient of the transient process within a preset range are intercepted as characteristic relaxation time segments.

5. The method for diagnosing short-circuit faults in lithium-ion battery packs based on relaxation time and unsupervised learning according to claim 4, characterized in that: In step S32, the initial range is α=0-5, and the preset step size is 0.1; The segments with obvious fault characteristics are concentrated in the range of α=0-1, and this segment should be included as much as possible when extracting fault characteristics.

6. The method for diagnosing short circuit faults in lithium-ion battery packs based on relaxation time and unsupervised learning according to claim 1, characterized in that: Step S4 includes the following steps: S41: The average time constant of the transient stability coefficient α of each monomer within the preset range is recorded in sequence according to the preset step length as Where n represents the number of relaxation segments cut out with a preset step size; S42: Calculation The statistical characteristics mean Mean, variance Var, maximum value Max, skewness Ske and kurtosis Kur are used as fault features, where: The above five fault features are combined into a five-dimensional feature vector as the fault detection feature vector of the cycle and used for subsequent fault identification.

7. The method for diagnosing short-circuit faults in lithium-ion battery packs based on relaxation time and unsupervised learning according to claim 1, characterized in that: Step S5 includes the following steps: The local abnormal outlier factor describes the degree of abnormality of the parameter by evaluating the local deviation of the observed data relative to the adjacent data; each feature of the fault feature mean Mean, variance Var, maximum value Max, skewness Ske and kurtosis Kur is regarded as a feature point, and each complete relaxation voltage is calculated to obtain a feature point; The characteristic points calculated by the relaxation voltage of the cycle to be detected and the reference points calculated by several normal cycles are used as the input of LOF, and the outliers and the degree of outliers of each characteristic point are output; wherein the reference points refer to the mean Mean, variance Var, maximum value Max, skewness Ske and kurtosis Kur under the normal cycle; If the characteristic points of the cycle to be detected do not show outliers, it is determined that the battery has no internal short circuit fault. If the characteristic points of the cycle to be detected are outliers, it is determined that the battery has an internal short circuit fault, and the severity of the fault is determined based on the outlier degree of the outlier.

Citation Information

Patent Citations

  • Internal micro short circuit fault diagnosis method based on charging curve

    CN113848495A

Cited By

  • Lithium ion battery pack internal short circuit fault diagnosis method based on relaxation time and adaptive clustering algorithm

    CN120214621A

  • Internal short circuit detection method, device, equipment, medium and product

    CN122063474A

  • A battery micro-short circuit early diagnosis method and system based on voltage dynamic response

    CN122836596A

  • A battery micro-short circuit early diagnosis method and system based on voltage dynamic response

    CN122836596B