Methods, equipment and media for detecting and locating short circuit faults in lithium-ion battery packs

Abstract

This invention discloses a method, device, and medium for detecting and locating internal short-circuit faults in lithium-ion battery packs. The method includes: designing a single-cell voltage anomaly index based on local resultant force correlation analysis of the individual cell voltage signals; extracting battery pack temperature features based on kernel principal component analysis to obtain the squared prediction error of the individual cell temperature; constructing a multi-scale fault detection statistic based on the voltage anomaly index and the squared temperature prediction error to achieve fault detection; and constructing a fault contribution function based on the squared temperature prediction error to achieve fault location. This invention designs a single-cell voltage anomaly index, extracts the squared prediction error of the individual cell temperature, and finally constructs a multi-scale fault detection statistic and a fault contribution function to achieve fault detection and location. This method is a purely data-driven approach with low equipment cost, simple operation, and is highly sensitive to changes in the temperature and voltage signals of individual cells in the early stages of an internal short circuit.

Description

Technical Field

[0001] This invention belongs to the field of battery pack fault diagnosis technology, specifically relating to a method, device and medium for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis. Background Technology

[0002] As a new type of green energy, lithium-ion batteries are widely used in electric vehicles, large-scale energy storage systems, and various electronic devices due to their excellent characteristics such as high energy density, long cycle life, low self-discharge rate, and environmental friendliness. However, safety accidents caused by lithium-ion battery systems occur frequently. For example, internal short circuits are one of the most common lithium-ion battery failures, which can lead to excessive heat generation in the battery system and thermal runaway, causing not only economic losses but also endangering people's lives. Therefore, developing effective methods for rapid detection and accurate location of early internal short circuit faults in lithium-ion battery packs is crucial.

[0003] Currently, several data-based and model-based methods have been proposed for detecting and locating short-circuit faults within battery packs. Examples include fault diagnosis methods based on the correlation coefficient of individual cell voltage signals, methods for detecting and locating short-circuit faults within lithium-ion battery packs based on principal component analysis of voltage signals, and model-based methods that estimate the current electrochemical state of the battery using measurements of individual cell temperature and voltage. However, in the early stages of internal short circuits, changes in individual cell voltage and temperature are not significant, which limits the performance of traditional data-based and model-based detection and location methods.

[0004] To improve diagnostic accuracy, many advanced technologies have been introduced into the detection and localization of short circuits within lithium-ion battery packs. For example, electrochemical impedance spectroscopy, infrared thermography, and ultrasonic testing methods are used in the diagnosis of short circuit faults within lithium-ion battery packs. While these methods offer good diagnostic results, their high equipment cost and complexity limit their widespread application in practice. Furthermore, some current diagnostic methods can only detect faults but cannot accurately locate them, hindering targeted maintenance of the battery pack. Therefore, it is necessary to propose a method based on multi-scale variable analysis for rapid detection and precise localization of early-stage internal short circuits in battery packs. Summary of the Invention

[0005] To address the problems of existing data- and model-based methods for diagnosing internal short-circuit faults in battery packs being insensitive to changes in voltage and temperature signals during the early stages of internal short circuits, and the high cost and complexity of methods based on advanced technologies, this invention provides a method, device, and medium for detecting and locating internal short-circuit faults in lithium-ion battery packs. This method uses multi-scale variable analysis to rapidly detect and accurately locate early internal short-circuit phenomena in battery packs.

[0006] To solve the above-mentioned technical problems, the present invention employs the following technical means:

[0007] The first aspect of the present invention provides a method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis, the method comprising:

[0008] S1. Based on the local resultant force correlation analysis of the single cell voltage signal of lithium-ion battery pack, design the single cell voltage anomaly index;

[0009] S2. Based on the extraction of battery pack temperature features by kernel principal component analysis, the squared prediction error of battery cell temperature is obtained.

[0010] S3. Construct a multi-scale fault detection statistic based on the voltage anomaly index and the temperature squared prediction error to achieve fault detection, and construct a fault contribution function based on the temperature squared prediction error to achieve fault location.

[0011] In the above scheme, step S1 includes:

[0012] S11. Calculate the local resultant force of the battery cell voltage;

[0013] S12. Based on the correlation coefficient, perform local resultant force correlation analysis on the voltage local resultant force of the battery cell to obtain the voltage local resultant force correlation coefficient.

[0014] S13. Design a voltage anomaly index based on local resultant force, including:

[0015] Calculate the average value of the correlation coefficient between the local resultant voltage of each individual cell and all individual cells;

[0016] The absolute value of the maximum value of the average value of the correlation coefficients of the local resultant voltage of all individual units at each moment is taken as the voltage anomaly index.

[0017] In the above scheme, step S11 includes:

[0018] The voltage signal of a single battery cell is standardized to obtain a standard fraction of the voltage:

[0019]

[0020] Among them, V i ′ (t) is the standard fraction of the voltage of battery cell i at second t, V i (t) is the voltage signal of battery cell i at second t. σ(t) and σ(t) represent the average value and standard deviation of the voltage signal of battery cell i within a given time window ending at t, respectively;

[0021] Based on this, the local resultant force of the battery cell voltage is calculated:

[0022]

[0023] Among them, F i (t) is the local resultant force generated by battery cell i due to the influence of n′ other cells, V j ′ (t) and V i ′ (t) represents the standard voltage fractions of monomers j and i at time t, respectively, and n ′ It is the number of monomers other than monomer i;

[0024] Step S12 includes:

[0025] Based on the correlation coefficient, a local resultant force correlation analysis is performed on the voltage local resultant force of the battery cell:

[0026]

[0027] Among them, F i and F j These are the local resultant voltage forces acting on individual cells i and j, respectively, ρ i,j (t) represents the correlation coefficient of the local resultant voltage of individual cells i and j within the time window from t0 to t, cov i,j (t) represents F within this time window i and F j covariance, σ i (t) and σ j (t) represents F within the time window. i and F j The variance, μ i (t) and μ j (t) represents F within the time window. i and F j The mean;

[0028] Step S13 includes:

[0029] Calculate the average value of the correlation coefficient between the local resultant voltage of each individual cell and all individual cells:

[0030]

[0031] in, It is the average of the correlation coefficients between the voltage local resultant force of each individual cell and all individual cells, where n is the number of individual cells, and ρ is the average value. i,j (t) is the correlation coefficient of the local resultant force of voltage for individual cells i and j;

[0032] Take n individuals at each time step The absolute value of the maximum value is used as the voltage anomaly index:

[0033]

[0034] Where h(t) represents the voltage anomaly index.

[0035] In the above scheme, step S2 includes:

[0036] S21. Calculate the kernel matrix of the temperature signal;

[0037] S22. Based on the nonlinear feature extraction of the temperature signal kernel matrix, the temperature data after feature extraction is obtained;

[0038] S23. Based on the temperature data after feature extraction, derive the temperature squared prediction error.

[0039] In the above scheme, step S21 includes:

[0040] Based on the battery cell temperature data measured by the sensor, a temperature signal kernel matrix is ​​constructed:

[0041]

[0042] Where K is the temperature signal kernel matrix, l represents the number of training samples, and x(t) = [x1(t), ..., x... m (t)] T This represents the temperature data of each battery cell at second t, where m is the number of cells, φ(x(t)) represents a mapping from m dimensions to higher dimensions, and k(x,x′) is the radial basis function, specifically defined as follows:

[0043]

[0044] Radial basis functions are used to measure the difference between two vectors x and x'. ′ The similarity between them, where σ is called the bandwidth of the function;

[0045] Step S22 includes:

[0046] Solve for the eigenvalues ​​of the kernel matrix K of the temperature signal:

[0047] λv=Kv

[0048] Where λ is the eigenvalue of the temperature signal kernel matrix K, and v is the eigenvector of matrix K;

[0049] Arrange the eigenvectors into a feature matrix V according to the eigenvalues ​​in descending order:

[0050] V = [v1, v2, ..., v l ]

[0051] Where, v1, v2, ..., vl The corresponding eigenvalues ​​are λ1, λ2, ..., λ n And λ1>λ2>…>λ n ;

[0052] Based on the temperature signal kernel matrix K, nonlinear feature extraction of temperature data is performed:

[0053] G d =KV d

[0054] Among them, G d =[g′(1),…,g′(l)] T It is the temperature data after feature extraction, V d Let g′(t) = [g′1(t), ..., g′] be the first d columns of V. d (t)] T ;

[0055] Step S23 includes:

[0056] Derivation of temperature squared prediction error:

[0057]

[0058] Where SPE(t) is the squared prediction error of the temperature at time t, and g(t) = [g1(t), ..., g l (t)] T It is the t-th row of G=KV.

[0059] In the above scheme, step S3 includes:

[0060] S31. Construct a multi-scale fault detection statistic based on the voltage anomaly index and the temperature squared prediction error. If the multi-scale fault detection statistic is greater than or equal to the threshold of the multi-scale statistic, then there is a short circuit fault in the lithium-ion battery pack.

[0061] S32. Based on the correlation coefficient of local voltage resultant force and the squared prediction error of temperature, a fault contribution function is constructed, and the battery cell with the largest fault contribution is considered as the fault cell.

[0062] In the above scheme, step S31 includes:

[0063] Based on the voltage anomaly index h(t) and the temperature squared prediction error SPE(t), a multi-scale fault detection statistic is designed:

[0064]

[0065] Where M(t) is the multi-scale fault detection statistic, t is time, and h is the time interval. r and SPE rThese are the reference signals for the voltage anomaly index and the temperature squared prediction error, respectively, with α being the weighting parameter;

[0066] Calculate the threshold of the multiscale statistic based on the weight parameter α:

[0067] M r =αh r +(1-α)SPE r

[0068] Among them, M r It is the threshold of multi-scale statistics;

[0069] Internal short-circuit faults in lithium-ion battery packs are detected using multi-scale fault detection statistics and their threshold values.

[0070]

[0071] Among them, the first time M(t)≥M r The time record is t d ;

[0072] Step S32 includes:

[0073] Based on the correlation coefficient of local voltage resultant force and the squared prediction error of temperature, a fault contribution function is designed:

[0074]

[0075] Among them, C j (t) is the contribution function of internal short-circuit fault in battery cell j, and β is the weighting parameter. It is the average of the correlation coefficients between the voltage local resultant force of each individual cell and all individual cells, sgn() is the sign function, x j (t) represents the temperature data of battery cell j at second t; in t d At any given moment, the battery cell contributing the most to the failure is considered the faulty cell:

[0076] z f =arg maxC z (t d )

[0077] Among them, z f It is the serial number of the faulty unit.

[0078] In the above scheme, the reference signals for the voltage anomaly index and the temperature squared prediction error are calculated using the kernel probability density estimation method:

[0079]

[0080] Where p is the confidence level, f(s) is the probability density function of variable s, s∈{h(t),SPE(t)}, and β r ∈{h r SPE r}

[0081] According to a second aspect of the present invention, a wind turbine bearing service life prediction device is provided, comprising: a processor and a memory, wherein the memory stores a program or instructions executable on the processor, and the program or instructions, when executed by the processor, implement the steps of the lithium-ion battery pack short circuit fault detection and location method based on multi-scale variable analysis described above.

[0082] According to a third aspect of the present invention, a readable storage medium is provided having a program or instructions stored thereon, which, when executed by a processor, implement the steps of the method for detecting and locating short-circuit faults in a lithium-ion battery pack based on multi-scale variable analysis as described in any one of the preceding claims.

[0083] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0084] This invention provides a method for detecting and locating internal short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis. It utilizes local resultant force correlation analysis of individual cell voltage signals to design an abnormal voltage index for each cell, and extracts the squared prediction error of individual cell temperature based on battery pack temperature characteristics derived from kernel principal component analysis. Finally, it constructs multi-scale fault detection statistics and a fault contribution function to achieve fault detection and location. This method is purely data-driven, with low equipment cost and simple operation, making it suitable for detecting and locating internal short-circuit faults in actual battery packs. Furthermore, this method is highly sensitive to changes in the temperature and voltage signals of individual cells in the early stages of an internal short circuit. Attached Figure Description

[0085] Figure 1 This is a schematic diagram of the method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis according to the present invention.

[0086] Figure 2 This is a schematic diagram of the battery pack structure used in the experiment of this invention;

[0087] Figure 3 This is a waveform diagram of UDDS according to an embodiment of the present invention;

[0088] Figure 4 This is a waveform diagram of the voltage and temperature signals of each battery cell in fault 1 of this embodiment of the invention; wherein, Figure 4 (a) in the diagram is a voltage waveform. Figure 4 (b) in the diagram is a temperature waveform;

[0089] Figure 5 This is a diagram showing the internal short-circuit fault detection results in an embodiment of the present invention; wherein, Figure 5 (a), (b), (c) and (d) are the internal short-circuit fault detection results under fault conditions 6, 8, 9 and 10 respectively;

[0090] Figure 6 This is a diagram showing the location result of an internal short-circuit fault in an embodiment of the present invention; wherein, Figure 6 (a), (b), (c) and (d) are the results of locating the internal short circuit fault under the conditions of fault 1, fault 2, fault 3 and fault 4, respectively. Detailed Implementation

[0091] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0092] This invention provides a method, device, and medium for detecting and locating internal short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis. It utilizes local resultant force correlation analysis of individual cell voltage signals to design an abnormal voltage index for each cell, and extracts the squared prediction error of individual cell temperature based on battery pack temperature characteristics derived from kernel principal component analysis. Finally, it constructs multi-scale fault detection statistics and a fault contribution function to achieve fault detection and location. This method is a purely data-driven approach with low equipment cost and simple operation, making it suitable for detecting and locating internal short-circuit faults in actual battery packs. Furthermore, this method is highly sensitive to changes in the temperature and voltage signals of individual cells in the early stages of an internal short circuit.

[0093] like Figure 1 As shown, according to an embodiment of the present invention, a method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis is proposed, including the following steps:

[0094] S1. Based on the local resultant force correlation analysis of the single cell voltage signal of lithium-ion battery pack, design the single cell voltage anomaly index;

[0095] S2. Based on the extraction of battery pack temperature features by kernel principal component analysis, the squared prediction error of the battery cell temperature is obtained.

[0096] S3. Construct multi-scale fault detection statistics and fault contribution functions to achieve fault detection and localization.

[0097] The specific process of step S1 includes:

[0098] S11. Calculation of the local resultant force of the voltage signal;

[0099] S12, Correlation analysis of local resultant forces;

[0100] S13. Voltage anomaly index design based on local resultant force.

[0101] Specifically, step S11 is as follows:

[0102] The voltage signal of a single battery cell is standardized to obtain a standard fraction of the voltage:

[0103]

[0104] Among them, V i ′ (t) is the standard fraction of the voltage of battery cell i at second t, V i (t) is the voltage signal of battery cell i at second t, measured by the sensor. σ and σ(t) represent the average value and standard deviation of the voltage signal of a single cell i within a given time window ending at t, respectively.

[0105] Based on this, the local resultant force of the battery cell voltage is calculated:

[0106]

[0107] Among them, F i (t) is the local resultant force generated by monomer i due to the influence of n′ other monomers, V j ′ (t) and V i ′ (t) represents the standard fractions of monomers j and i at time t, respectively, and n ′ It represents the number of monomers other than monomer i.

[0108] Specifically, step S12 is as follows:

[0109] Based on the correlation coefficient, a local resultant force correlation analysis is performed on the voltage local resultant force of the battery cell:

[0110]

[0111] Among them, F i and F j These are the local resultant forces received by monomers i and j, respectively, ρ i,j (t) represents the correlation coefficient of the local resultant voltage of individual cells i and j within the time window from t0 to t, cov i,j (t) represents F within this time windowi and F j covariance, σ i (t) and σ j (t) represents F within the time window. i and F j The variance, μ i (t) and μ j (t) represents F within the time window. i and F j The mean.

[0112] Specifically, step S13 is as follows:

[0113] Calculate the average of the correlation coefficients between the local resultant forces of each individual unit and all individual units:

[0114]

[0115] in, It is the average of the correlation coefficients between the local resultant forces of each individual and all individual individuals, where n is the number of individual individuals, and ρ is the average of the correlation coefficients between the local resultant forces of each individual and all individual individuals. i,j (t) is the correlation coefficient of the local resultant voltage force of individual cells i and j.

[0116] Take n individuals at each time step The absolute value of the maximum value is used as the voltage anomaly index:

[0117]

[0118] Where h(t) represents the voltage anomaly index.

[0119] The specific process of step S2 includes:

[0120] S21. Temperature signal kernel matrix calculation;

[0121] S22. Nonlinear feature extraction based on kernel matrix;

[0122] S23, Derivation of temperature squared prediction error.

[0123] Specifically, step S21 is as follows:

[0124] Based on the battery cell temperature data measured by the sensors, a temperature kernel matrix is ​​constructed:

[0125]

[0126] Where K is the kernel matrix of the temperature signal, l represents the number of training samples, and x(t) = [x1(t), ..., x... m (t)] THere, φ(x(t)) represents the temperature data of each battery cell at second t, m is the number of cells, φ(x(t)) represents a mapping from m dimensions to higher dimensions, and κ(x,x′) is the radial basis function, which is defined as follows:

[0127]

[0128] Radial basis functions are used to measure the difference between two vectors x and x'. ′ The similarity between the functions is given by σ, where σ is called the bandwidth of the function.

[0129] Specifically, step S22 is as follows:

[0130] Solving for the eigenvalues ​​of the temperature kernel matrix K:

[0131] λv=Kv

[0132] Where λ is the eigenvalue of matrix K, and v is the eigenvector of matrix K.

[0133] Arrange the eigenvectors into a feature matrix V according to the eigenvalues ​​in descending order:

[0134] V = [v1, v2, ..., v l ]

[0135] Where, v1, v2, ..., v l The corresponding eigenvalues ​​are λ1, λ2, ..., λ n And λ1>λ2>…>λ n .

[0136] Based on the kernel matrix K, nonlinear feature extraction of temperature data is performed.

[0137] G d =KV d

[0138] Among them, G d =[g′(1),…,g′(l)] T It is the temperature data after feature extraction, V d Let g′(t) = [g′1(t), ..., g′] be the first d columns of V. d (t)] T .

[0139] Specifically, step S23 is as follows:

[0140] Based on the above results, the temperature squared prediction error is derived as follows:

[0141]

[0142] Where SPE(t) is the squared prediction error at time t, and g(t) = [g1(t), ..., gl (t)] T It is the t-th row of G=KV.

[0143] The specific process of step S3 includes:

[0144] S31. Design of multi-scale statistics based on voltage anomaly index and temperature squared prediction error;

[0145] S32. Design of a fault contribution function based on the correlation coefficient of local voltage resultant force and the squared prediction error of temperature.

[0146] Specifically, step S31 is as follows:

[0147] Based on the voltage anomaly index and the squared temperature prediction error, a multi-scale statistic is designed:

[0148]

[0149] Where M(t) is a multiscale statistic, h r and SPE r These are the reference signals for the voltage anomaly index and the temperature squared prediction error, respectively, with α being the weighting parameter. The reference signals can be calculated using the kernel probability density estimation method.

[0150]

[0151] Where p is the confidence level, usually set to 0.95 or 0.99, f(s) is the probability density function of variable s, which can be calculated by kernel density estimation, s∈{h(t),SPE(t)}, β r ∈{h r SPE r}

[0152] Calculate the threshold of the multiscale statistic based on the weight parameter α:

[0153] M r =αh r +(1-α)SPE r

[0154] Among them, M r This refers to the threshold of a multi-scale statistic. Using multi-scale statistics and their thresholds, internal short-circuit faults in lithium-ion battery packs can be detected.

[0155]

[0156] Among them, the first time M(t)≥M r The time record is t d .

[0157] Specifically, step S32 is as follows:

[0158] Based on the correlation coefficient of local voltage resultant force and the squared prediction error of temperature, a fault contribution function is designed:

[0159]

[0160] Among them, C j (t) is the contribution function of internal short-circuit fault in battery cell j, and β is the weighting parameter. In t... d At any given moment, the battery cell contributing the most to the failure is considered the faulty cell:

[0161] z f =arg maxC z (t d )

[0162] Among them, z f It is the serial number of the faulty unit.

[0163] More specifically, the implementation of the scheme was carried out using the internal short-circuit fault detection and location of a battery pack composed of 24 individual battery cells as an example. The structure of the battery pack used in the experiment is as follows: Figure 2 As shown, four battery cells are first connected in parallel, and then the six parallel cells are connected in series. The main parameters of the model are listed in Table 1.

[0164] Table 1 Main Model Parameters

[0165]

[0166] The following four performance metrics are used for performance evaluation and comparison of internal short-circuit fault detection and location methods:

[0167] Recall rate:

[0168] False alarm rate:

[0169] Fault detection delay: FDD = t d -t f

[0170] Here, recall represents the number of fault samples N detected. d N% of the total number of fault samples af The proportion, the false alarm rate represents the number of false alarms N. f N% of the total normal sample size an The proportion, t d and t f These represent the moment when the fault was first detected and the actual time when the fault occurred, respectively.

[0171] Ten fault scenarios were designed for this battery pack, and detailed fault parameters are shown in Table 2. Here, 1C represents the constant current required for discharge from 100% to 0% in one hour. Under 2C conditions, the current doubles, and theoretically, a full discharge can be completed in just half an hour. The waveform of UDDS is shown below. Figure 3 As shown. Temperature and voltage sensors measured the temperature and voltage data of 24 individual battery cells from 0 to 2000 seconds. The curves showing the voltage and temperature changes of the 24 individual battery cells over time in fault 1 are shown below. Figure 4 As shown, where Figure 4 (a) in the diagram is a voltage waveform. Figure 4 (b) in the diagram is a temperature waveform.

[0172] Table 2 Fault Setting Table

[0173]

[0174] Using the proposed multi-scale variable analysis method, short-circuit fault detection and location experiments were conducted on faults 1 to 10 in Table 2 for lithium-ion battery packs.

[0175] The detection results for faults 6, 8, 9, and 10 are as follows: Figure 5 As shown. Fault 6 occurs at 2C rate on battery #23 at the 1000th second, and is an internal short-circuit fault with an internal short-circuit resistance of 10Ω. Figure 5 As shown in (a), the proposed multi-scale variable analysis method captures the internal short-circuit fault promptly and clearly. The parameters of fault 8 and fault 6 differ only in their internal short-circuit resistance; the former has a smaller internal short-circuit resistance, indicating that the internal short-circuit phenomenon is more pronounced in fault 8. Figure 5 (a) and Figure 5 In (b), it can be seen that under fault 8, the changes in the values ​​of the multi-scale statistics after the fault occurs are more significant. Compared with fault 6, the fault occurrence time of fault 9 is delayed by 500 seconds. Figure 5 (a) and Figure 5 As shown in (c), the proposed method detects fault 9 approximately 1500 seconds in time. Fault 10 is an internal short-circuit fault occurring under dynamic operating conditions. The proposed method's detection performance for fault 10 is as follows: Figure 5 As shown in (d) in the figure, the short circuit fault occurring within the 1000th second is detected quickly.

[0176] The proposed method achieves the following results in locating faults 1, 2, 3, and 4: Figure 6 As shown, where Figure 6Figures (a), (b), (c), and (d) show the results of locating internal short-circuit faults under fault conditions 1, 2, 3, and 4, respectively. The figures show that these internal short-circuit faults occurred in battery cells 4, 5, 11, and 16, respectively. Figure 6 As can be seen from the bar chart, there is a significant difference in the contribution of faulty battery cells and normal battery cells, indicating that this method can accurately locate faulty battery cells.

[0177] The proposed method was applied to 10 groups of faults, and the recall rate, false alarm rate, and fault detection delay for different faults are shown in Table 3. It can be seen that the proposed method exhibits high recall, low false alarm rate, and short fault detection delay for all faults, indicating that it can quickly detect internal short-circuit faults in lithium-ion battery packs. Furthermore, the location results obtained by this method are consistent with the parameters set in the fault configuration, demonstrating that the method can accurately locate the faulty individual cell.

[0178] Table 3 shows the effectiveness of the proposed method on 10 groups of faults.

[0179]

[0180] To demonstrate the superiority of the proposed method, its performance on fault 1 was compared with two other methods (correlation coefficient method based solely on voltage signals and kernel principal component analysis based solely on temperature signals). The results are shown in Table 4. Analysis of the data in the table shows that while the correlation coefficient method based solely on voltage data has a lower false alarm rate, its reduced recall and increased fault detection delay negate its superiority. Furthermore, the kernel principal component analysis method based solely on temperature signals generates false alarms well before the fault occurs, making it very ineffective on fault 1.

[0181] Table 4 Comparison of the effectiveness of the three methods for fault diagnosis 1

[0182]

[0183] According to a second aspect of the present invention, an electronic device is provided, comprising: a processor and a memory, the memory storing a program or instructions executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the lithium-ion battery pack short-circuit fault detection and location method based on multi-scale variable analysis provided in the first aspect.

[0184] According to a third aspect of the present invention, a readable storage medium is provided on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method for detecting and locating short-circuit faults in a lithium-ion battery pack based on multi-scale variable analysis as provided in the first aspect.

[0185] In summary, this invention utilizes local resultant force correlation analysis of individual cell voltage signals in lithium-ion battery packs to design an abnormal voltage index for individual cells. It also extracts the squared prediction error of individual cell temperature based on battery pack temperature characteristics derived from kernel principal component analysis. Finally, it constructs multi-scale fault detection statistics and a fault contribution function to achieve fault detection and location. This method is a purely data-driven approach with low equipment cost and simple operation, making it suitable for detecting and locating internal short-circuit faults in actual battery packs. Furthermore, this method is highly sensitive to changes in the temperature and voltage signals of individual cells in the early stages of an internal short circuit, enabling it to detect anomalies early on. The proposed method was validated in a battery pack consisting of 24 individual cells. Using this method to diagnose 10 groups of internal short-circuit faults in the battery pack, it demonstrated high recall, low false alarm rate, and short fault detection delay, accurately locating the faulty individual cell.

[0186] It should be noted that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0187] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0188] Those skilled in the art will readily understand that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis, characterized in that, The method includes: S1. Based on the local resultant force correlation analysis of the individual cell voltage signals of lithium-ion battery packs, design an abnormal voltage index for individual cells, including: S11. Calculate the local resultant voltage force of a single battery cell, including: The voltage signal of a single battery cell is standardized to obtain a standard fraction of the voltage: in, It is a battery cell In the The standard fraction of voltage per second It is a battery cell In the A voltage signal of seconds, and Representing individual battery cells In The average value and standard deviation of the voltage signal within a given time window, with the endpoint as the endpoint; Based on this, the local resultant force of the battery cell voltage is calculated: in, It is a battery cell because The local resultant force generated by the influence of other individual units. and Representing monomers respectively and monomers In time t voltage standard fraction, It is except monomer The number of other monomers; S12. Based on the correlation coefficient, perform local resultant force correlation analysis on the voltage local resultant force of the battery cell to obtain the voltage local resultant force correlation coefficient. S13. Design a voltage anomaly index based on local resultant force, including: Calculate the average value of the correlation coefficient between the local resultant voltage of each individual cell and all individual cells; The absolute value of the maximum value of the average value of the correlation coefficients of the local resultant voltage of all individual units at each moment is taken as the voltage anomaly index. S2. Based on the extraction of battery pack temperature features by kernel principal component analysis, the squared prediction error of battery cell temperature is obtained. S3. Construct a multi-scale fault detection statistic based on the voltage anomaly index and the temperature squared prediction error to achieve fault detection, and construct a fault contribution function based on the temperature squared prediction error to achieve fault location.

2. The method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis according to claim 1, characterized in that, Step S12 includes: Based on the correlation coefficient, a local resultant force correlation analysis is performed on the voltage local resultant force of the battery cell: in, and They are monomers and monomers The local resultant force of the voltage it receives, Representative at arrive Within the time window, single and monomers The correlation coefficient of the local resultant force of voltage, Represents the time window and covariance, and These represent the time window respectively. and variance and These represent the time window respectively. and The mean; Step S13 includes: Calculate the average value of the correlation coefficient between the local resultant voltage of each individual cell and all individual cells: in, It is the average value of the correlation coefficient between the local resultant voltage of each individual cell and all individual cells. It refers to the number of individual units. It is a monomer and monomers The correlation coefficient of local resultant force of voltage; Pick Each individual at every moment The absolute value of the maximum value is used as the voltage anomaly index: in, This represents the voltage anomaly index.

3. The method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis according to claim 1, characterized in that, Step S2 includes: S21. Calculate the kernel matrix of the temperature signal; S22. Based on the nonlinear feature extraction of the temperature signal kernel matrix, the temperature data after feature extraction is obtained; S23. Based on the temperature data after feature extraction, derive the temperature squared prediction error.

4. The method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis according to claim 3, characterized in that, Step S21 includes: Based on the battery cell temperature data measured by the sensor, a temperature signal kernel matrix is ​​constructed: in, It is a temperature signal kernel matrix. Represents the number of training samples. It is the first Temperature data for each individual battery cell per second. It refers to the number of individual units. Representative from A mapping from 1 to a higher dimension. These are radial basis functions, and their specific definitions are as follows: Radial basis functions are used to measure the difference between two vectors. The similarity between them, among which, This is called the bandwidth of the function; Step S22 includes: Solving for the kernel matrix of the temperature signal eigenvalues: in, It is a temperature signal kernel matrix eigenvalues, It is a matrix eigenvectors; Arrange the eigenvectors into a feature matrix V according to the eigenvalues ​​in descending order: in, The corresponding eigenvalues ​​are ,and ; Based on temperature signal kernel matrix Nonlinear feature extraction of temperature data: in, These are the temperature data after feature extraction. , yes The former List; Step S23 includes: Derivation of temperature squared prediction error: in, yes The predicted error of the squared temperature at time t is yes The OK.

5. The method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis according to claim 1, characterized in that, Step S3 includes: S31. Construct a multi-scale fault detection statistic based on the voltage anomaly index and the temperature squared prediction error. If the multi-scale fault detection statistic is greater than or equal to the threshold of the multi-scale statistic, then there is a short circuit fault in the lithium-ion battery pack. S32. Based on the correlation coefficient of local voltage resultant force and the squared prediction error of temperature, a fault contribution function is constructed, and the battery cell with the largest fault contribution is considered as the fault cell.

6. The method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis according to claim 5, characterized in that, Step S31 includes: Based on voltage anomaly index Sum of temperature squared prediction error Design multi-scale fault detection statistics: in, This is a multi-scale fault detection statistic. For time, and These are the reference signals for the voltage anomaly index and the temperature squared prediction error, respectively. These are weight parameters; Based on weight parameters Calculate the threshold for multiscale statistics: in, It is the threshold of multi-scale statistics; Internal short-circuit faults in lithium-ion battery packs are detected using multi-scale fault detection statistics and their threshold values. Among them, the first time it was used The time record is ; Step S32 includes: Based on the correlation coefficient of local voltage resultant force and the squared prediction error of temperature, a fault contribution function is designed: in, It is a battery cell The contribution function of internal short-circuit faults, These are weight parameters. It is the average value of the correlation coefficient between the local resultant voltage of each individual cell and all individual cells. For symbolic functions, For battery cells No. Temperature data in seconds; At any given moment, the battery cell contributing the most to the failure is considered the faulty cell: in, It is the serial number of the faulty unit.

7. The method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis according to claim 6, characterized in that, The reference signals for the voltage anomaly index and the temperature squared prediction error are calculated using the kernel probability density estimation method: in, For confidence level, For variables s The probability density function, , .

8. An electronic device, characterized in that, include: The processor and memory, wherein the memory stores a program or instructions that can run on the processor, and when the program or instructions are executed by the processor, implement the steps of the method for detecting and locating short-circuit faults in lithium-ion battery packs based on multi-scale variable analysis as described in any one of claims 1 to 7.

9. A readable storage medium, characterized in that, It stores programs or instructions, which, when executed by a processor, implement the steps of the lithium-ion battery pack short-circuit fault detection and location method based on multi-scale variable analysis as described in any one of claims 1 to 7.

Images