Lithium ion battery pack safety early warning method and device based on enhanced kernel mahalanobis distance

Abstract

This invention relates to the field of battery safety technology and discloses a method and device for safety early warning of lithium-ion battery packs based on enhanced core Markov distance. The invention processes historical monitoring data of lithium-ion battery packs that have experienced thermal runaway using data augmentation and nonlinear transformation methods to obtain a target relationship characterizing the enhanced core Markov distance. Data augmentation better reflects the dynamic characteristics of the lithium-ion battery pack, preserving the correlation between each feature and historical data. Simultaneously, the nonlinear transformation method better describes the nonlinear relationships between monitoring data during the historical charge and discharge processes of the lithium-ion battery. Furthermore, by using historical monitoring data of lithium-ion battery packs that have not experienced thermal runaway and combining it with the target relationship characterizing the enhanced core Markov distance, a control limit for safety early warning can be generated. Finally, by combining the enhanced core Markov distance of the lithium-ion battery pack to be warned, a safety early warning can be achieved for the lithium-ion battery pack to be warned.

Description

Technical Field

[0001] This invention relates to the field of battery safety technology, specifically to a method and device for safety early warning of lithium-ion battery packs based on enhanced nuclear Marsh distance. Background Technology

[0002] Lithium-ion batteries are widely used in transportation electrification fields such as electric vehicles, distributed energy storage, and energy storage power stations due to their advantages such as high specific energy, specific power, long life, and small size. However, in recent years, there have been frequent spontaneous combustion accidents in electric vehicles and fires in energy storage power stations. The main reason for most of these incidents is thermal runaway of the lithium-ion batteries they carry. Therefore, developing a fast and accurate thermal runaway safety early warning algorithm for lithium-ion battery packs is of great significance for ensuring the safe application of lithium-ion batteries.

[0003] Methods for lithium-ion battery safety early warning can be mainly divided into model-based methods and data-driven methods. Model-based methods require the establishment of accurate electrical, thermal, or fractional-order models of the lithium-ion battery, demanding high accuracy, involving complex and difficult modeling processes, and their diagnostic effectiveness is affected by the model's accuracy and robustness. Data-driven methods are divided into machine learning-based and signal-based methods. Machine learning methods require a large amount of training data, while signal-based methods suffer from the problem that a single signal cannot fully characterize the battery state. Therefore, both approaches result in poorer safety early warning performance for lithium-ion batteries. Summary of the Invention

[0004] In view of this, the present invention provides a method and apparatus for safety warning of lithium-ion battery packs based on enhanced nuclear Martens distance, so as to solve the problem of poor safety warning effect of lithium-ion batteries.

[0005] In a first aspect, the present invention provides a safety warning method for lithium-ion battery packs based on enhanced nuclear Markov distance, the method comprising:

[0006] A first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and a second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway are obtained. Based on the first historical monitoring dataset, a target relation characterizing the enhanced core Mahalanobis distance is obtained through data augmentation and nonlinear transformation methods. Based on the second historical monitoring dataset and the target relation, the control limit is determined. An offline monitoring dataset of the lithium-ion battery pack to be warned is obtained, and based on the offline monitoring dataset and the target relation, the first enhanced core Mahalanobis distance of the lithium-ion battery pack to be warned is calculated. Based on the first enhanced core Mahalanobis distance and the control limit, a safety warning is issued for the lithium-ion battery pack to be warned, and the safety warning result of the lithium-ion battery pack to be warned is obtained.

[0007] The present invention provides a safety early warning method for lithium-ion battery packs based on enhanced core Markov distance. First, historical monitoring data of lithium-ion battery packs that have experienced thermal runaway and those that have not are acquired. Second, the historical monitoring data of the thermal runaway lithium-ion battery packs is processed using data augmentation and nonlinear transformation methods to obtain a target relation characterizing the enhanced core Markov distance. Data augmentation better reflects the dynamic characteristics of the lithium-ion battery pack, preserving the correlation between each feature and historical data. Simultaneously, the nonlinear transformation method better describes the nonlinear relationships between monitoring data during the historical charge and discharge processes of the lithium-ion battery. Furthermore, by combining historical monitoring data of lithium-ion battery packs that have not experienced thermal runaway with the target relation characterizing the enhanced core Markov distance, control limits for safety early warning can be generated. Finally, by combining the enhanced core Markov distance of the lithium-ion battery pack to be warned, a safety early warning can be achieved. Therefore, by implementing this invention, a safety early warning for the nonlinear process of multi-feature fusion in lithium-ion battery packs is realized, and the safety early warning process does not require the establishment of a complex model.

[0008] In one optional implementation, acquiring a first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and a second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway includes:

[0009] Obtain the third historical monitoring dataset of the lithium-ion battery pack that has experienced thermal runaway and the second historical monitoring dataset of the lithium-ion battery pack that has not experienced thermal runaway; obtain the fourth historical monitoring dataset of the lithium-ion battery pack that has experienced thermal runaway when it has not experienced thermal runaway; preprocess the third historical monitoring dataset based on the fourth historical monitoring dataset to obtain the first historical monitoring dataset of the lithium-ion battery pack that has experienced thermal runaway.

[0010] The present invention provides a lithium-ion battery pack safety early warning method based on enhanced nuclear Martens distance. By preprocessing the monitoring data of the lithium-ion battery pack that has experienced thermal runaway before thermal runaway occurs, the accuracy of the monitoring data of the lithium-ion battery pack that has experienced thermal runaway is improved.

[0011] In one optional implementation, based on a first historical monitoring dataset, and after processing with data augmentation and nonlinear transformation methods, a target relation characterizing the augmented kernel Mahalanobis distance is obtained, including:

[0012] Select a feature input dataset from the first historical monitoring dataset; establish a multidimensional feature matrix based on the feature input dataset; establish a target augmentation matrix based on the multidimensional feature matrix and data augmentation method; process the target augmentation matrix through a nonlinear transformation method to obtain the target relation representing the augmentation kernel Mahalanobis distance.

[0013] The safety early warning method for lithium-ion battery packs based on enhanced nuclear Markov distance provided by this invention improves the robustness and reliability of subsequent safety early warning results by using multiple historical monitoring data of the lithium-ion battery pack as feature inputs. Furthermore, data augmentation methods better reflect the dynamic characteristics of the lithium-ion battery pack, preserving the correlation between each feature and historical data. Simultaneously, combining nonlinear transformation methods better describes the nonlinear relationships between monitoring data during the historical charge and discharge processes of the lithium-ion battery.

[0014] In one optional implementation, the target enhancement matrix is ​​processed by a nonlinear transformation method to obtain a target relation characterizing the Mahalanobis distance of the enhancement kernel, including:

[0015] The target enhancement matrix is ​​processed by a nonlinear transformation method to obtain the enhancement kernel feature matrix; based on the enhancement kernel feature matrix, the initial relation characterizing the Mahalanobis distance of the enhancement kernel is determined; based on the initial relation and the enhancement kernel feature matrix, the target relation characterizing the Mahalanobis distance of the enhancement kernel is obtained by processing with a Gaussian kernel function.

[0016] The safety warning method for lithium-ion battery packs based on enhanced nuclear Mahalanobis distance provided by this invention introduces a Gaussian kernel function to capture nonlinearity, thereby obtaining a target relation characterizing the enhanced nuclear Mahalanobis distance.

[0017] In one alternative implementation, the control limits are determined based on a second historical monitoring dataset and the target relation, including:

[0018] Based on the second historical monitoring dataset and the target relation, the second enhanced core Markov distance of the lithium-ion battery pack that has not experienced thermal runaway is calculated; based on the second enhanced core Markov distance, the control limit is obtained after processing with the central limit theorem.

[0019] The safety early warning method for lithium-ion battery packs based on enhanced nuclear Marvin distance provided by this invention can generate control limits for safety early warning by combining historical monitoring data of lithium-ion battery packs that have not experienced thermal runaway with target relationships characterizing enhanced nuclear Marvin distance, thus providing support for safety early warning of subsequent lithium-ion battery packs to be warned.

[0020] In one optional implementation, a safety warning is issued for the lithium-ion battery pack to be warned based on the first enhanced nuclear Martens distance and the control limit, resulting in a safety warning result for the lithium-ion battery pack to be warned, including:

[0021] Based on the first enhanced core Marsh distance and control limit, it is determined whether the lithium-ion battery pack to be warned has malfunctioned; if the lithium-ion battery pack to be warned has not malfunctioned, it is determined that the lithium-ion battery pack to be warned is in a safe state; if the lithium-ion battery pack to be warned has malfunctioned, it is determined that the lithium-ion battery pack to be warned is in an unsafe state, and a safety warning message is issued.

[0022] The present invention provides a lithium-ion battery pack safety early warning method based on enhanced nuclear Marvin distance. By using the first enhanced nuclear Marvin distance of the lithium-ion battery pack to be warned and the control limit for safety early warning, it can be determined whether the lithium-ion battery pack to be warned has failed. Furthermore, by combining the judgment result, the safety status of the lithium-ion battery pack to be warned can be determined, and thus a safety early warning result can be obtained.

[0023] In a second aspect, the present invention provides a lithium-ion battery pack early warning device based on enhanced nuclear Martens distance, the device comprising:

[0024] The system comprises the following modules: an acquisition module for acquiring a first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and a second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway; a processing module for processing the first historical monitoring dataset using data augmentation and nonlinear transformation methods to obtain a target relation characterizing the enhanced core Mahalanobis distance; a determination module for determining the control limit based on the second historical monitoring dataset and the target relation; an acquisition and calculation module for acquiring the offline monitoring dataset of the lithium-ion battery pack to be warned and calculating the first enhanced core Mahalanobis distance of the lithium-ion battery pack to be warned based on the offline monitoring dataset and the target relation; and a safety warning module for issuing a safety warning for the lithium-ion battery pack to be warned based on the first enhanced core Mahalanobis distance and the control limit, and obtaining the safety warning result for the lithium-ion battery pack to be warned.

[0025] Thirdly, the present invention provides a computer device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the lithium-ion battery pack safety warning method based on enhanced nuclear Mahalanobis distance described in the first aspect or any corresponding embodiment.

[0026] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the lithium-ion battery pack safety warning method based on enhanced nuclear Mahalanobis distance according to the first aspect or any corresponding embodiment described above.

[0027] Fifthly, the present invention provides a computer program product, including computer instructions for causing a computer to execute the lithium-ion battery pack safety early warning method based on enhanced nuclear Mahalanobis distance according to the first aspect or any corresponding embodiment described above. Attached Figure Description

[0028] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific 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 from these drawings without creative effort.

[0029] Figure 1 This is a flowchart illustrating a lithium-ion battery pack safety early warning method based on enhanced nuclear Martens distance according to an embodiment of the present invention.

[0030] Figure 2 This is a flowchart illustrating another lithium-ion battery pack safety early warning method based on enhanced nuclear Martens distance according to an embodiment of the present invention;

[0031] Figure 3 This is a flowchart illustrating another lithium-ion battery pack safety early warning method based on enhanced nuclear Martens distance according to an embodiment of the present invention;

[0032] Figure 4 This is a flowchart illustrating a multi-feature early warning method for lithium-ion batteries based on enhanced nuclear Markov distance according to an embodiment of the present invention.

[0033] Figure 5 This is a schematic diagram of the safety early warning principle based on enhanced nuclear Mahalanobis distance according to an embodiment of the present invention;

[0034] Figure 6 This is a structural block diagram of a lithium-ion battery pack safety early warning device based on enhanced nuclear Martens distance according to an embodiment of the present invention;

[0035] Figure 7 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] This invention provides a safety warning method for lithium-ion battery packs based on enhanced nuclear Marvin distance. By determining the target relation characterizing the enhanced nuclear Marvin distance and the control limit for safety warning, the method achieves the effect of safety warning for the nonlinear process of multi-feature fusion of lithium-ion battery packs.

[0038] According to an embodiment of the present invention, a method for safety warning of lithium-ion battery packs based on enhanced nuclear Martens distance is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0039] This embodiment provides a safety early warning method for lithium-ion battery packs based on enhanced nuclear Markov distance. Figure 1 This is a flowchart of a lithium-ion battery pack safety warning method based on enhanced nuclear Martens distance according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps:

[0040] Step S101: Obtain the first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway.

[0041] Specifically, thermal runaway refers to a phenomenon where a battery experiences a violent chain reaction in a short period of time due to overheating, overcharging, internal short circuits, collisions, or other reasons, resulting in a rapid increase in battery temperature and potentially causing an explosion or fire.

[0042] Furthermore, the monitoring data for lithium-ion battery packs can include data from the charging and discharging processes of the lithium-ion battery packs.

[0043] Step S102: Based on the first historical monitoring dataset, the target relation representing the augmented kernel Mahalanobis distance is obtained through data augmentation and nonlinear transformation methods.

[0044] Data augmentation is a technique commonly used in machine learning and deep learning. It generates additional training samples by applying various transformations or conversions.

[0045] Furthermore, nonlinear transformation refers to an image transformation method performed on a reference image according to a specific functional relationship, and there is no linear relationship between the generated target image and the reference image. That is, the transformation performed on an input variable is not a linear function.

[0046] Specifically, by processing the first historical monitoring dataset using data augmentation methods, the dynamic characteristics of lithium-ion battery packs can be better reflected, while preserving the correlation between each feature and the historical monitoring data.

[0047] Furthermore, Mahalanobis distance is not suitable for handling nonlinear problems. Therefore, by using nonlinear transformation methods, we can better describe the nonlinear relationships between monitoring data during the historical charge and discharge process of lithium-ion batteries and obtain the target relationship that characterizes the enhanced core Mahalanobis distance.

[0048] Step S103: Determine the control limits based on the second historical monitoring dataset and the target relation.

[0049] Specifically, by combining historical monitoring data of lithium-ion battery packs that have not experienced thermal runaway with the target relationship for enhancing the nuclear Marvin distance, control limits for safety warnings can be automatically generated.

[0050] Step S104: Obtain the offline monitoring dataset of the lithium-ion battery pack to be warned, and calculate the first enhanced nuclear Mahalanobis distance of the lithium-ion battery pack to be warned based on the offline monitoring dataset and the target relation.

[0051] Specifically, by taking the offline monitoring data of the lithium-ion battery pack to be warned as input and combining it with the target relation, the enhanced nuclear Marvin distance of the lithium-ion battery pack to be warned can be calculated.

[0052] Step S105: Based on the first enhanced core Marsh distance and control limit, a safety warning is issued for the lithium-ion battery pack to be warned, and the safety warning result of the lithium-ion battery pack to be warned is obtained.

[0053] Specifically, based on the known first enhanced core Marvin distance of the lithium-ion battery pack to be warned, and combined with the generated control limit for safety warning, a nonlinear process of multi-feature fusion of lithium-ion battery packs can be realized for safety warning.

[0054] The lithium-ion battery pack safety early warning method based on enhanced core Markov distance provided in this embodiment first acquires historical monitoring data of lithium-ion battery packs that have experienced thermal runaway and those that have not. Then, it processes the historical monitoring data of the thermal runaway lithium-ion battery pack using data augmentation and nonlinear transformation methods to obtain a target relation characterizing the enhanced core Markov distance. Data augmentation better reflects the dynamic characteristics of the lithium-ion battery pack, preserving the correlation between each feature and historical data. Simultaneously, the nonlinear transformation method better describes the nonlinear relationships between monitoring data during the historical charge and discharge processes of the lithium-ion battery. Furthermore, by combining the historical monitoring data of the non-thermal runaway lithium-ion battery pack with the target relation characterizing the enhanced core Markov distance, a control limit for safety early warning can be generated. Finally, by combining the enhanced core Markov distance of the lithium-ion battery pack to be warned, a safety early warning can be achieved. Therefore, by implementing this invention, a safety early warning for the nonlinear process of multi-feature fusion in lithium-ion battery packs is realized, and the safety early warning process does not require the establishment of a complex model.

[0055] This embodiment provides a safety early warning method for lithium-ion battery packs based on enhanced nuclear Markov distance. Figure 2This is a flowchart of a lithium-ion battery pack safety warning method based on enhanced nuclear Martens distance according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps:

[0056] Step S201: Obtain the first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway.

[0057] Specifically, step S201 includes:

[0058] Step S2011: Obtain the third historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway.

[0059] Specifically, monitoring data of lithium-ion battery packs that have experienced overheating and runaway were selected as the research object. The monitoring data was then filtered to select data such as time, alarm status, minimum cell voltage and cell number, maximum probe temperature and probe number, etc., to form a third historical monitoring dataset.

[0060] Step S2012: Obtain the fourth historical monitoring dataset of the lithium-ion battery pack that has experienced thermal runaway but has not yet experienced thermal runaway.

[0061] Specifically, normal monitoring data of lithium-ion battery packs that have experienced thermal runaway but have not experienced thermal runaway are selected to form the fourth historical monitoring dataset.

[0062] Step S2013: Preprocess the third historical monitoring dataset based on the fourth historical monitoring dataset to obtain the first historical monitoring dataset of the lithium-ion battery pack that has experienced thermal runaway.

[0063] The preprocessing includes deduplication and omission correction, interpolation smoothing, and removal of outliers.

[0064] Specifically, preprocessing can improve the accuracy of the first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway.

[0065] Step S202: Based on the first historical monitoring dataset, the target relation representing the augmented kernel Mahalanobis distance is obtained through data augmentation and nonlinear transformation methods.

[0066] Specifically, step S202 includes:

[0067] Step S2021: Select the feature input dataset from the first historical monitoring dataset.

[0068] Specifically, based on the study of thermal runaway mechanism (a single cell that experiences thermal runaway will produce a pressure drop and a temperature rise), data such as the minimum single cell voltage and the maximum probe temperature can be selected from the first historical monitoring dataset as feature input data.

[0069] Step S2022: Establish a multidimensional feature matrix based on the feature input dataset.

[0070] Specifically, the multidimensional feature matrix is ​​shown in the following relation (1):

[0071] [x1,x2,…,x k ,…,x N ]∈R N×D (1)

[0072] In the formula: D represents the dimension of the variable; N represents the length of the sampling points; x k Let represent the D-dimensional feature vector of the k-th sampling point.

[0073] Step S2023: Based on the multidimensional feature matrix, the target augmentation matrix is ​​established through data augmentation methods.

[0074] Specifically, considering the autocorrelation of the data, data augmentation techniques are used to process the multidimensional feature matrix and establish the target augmentation matrix.

[0075] First, the enhancement parameter L is defined, where L is a non-negative integer. A larger value of L results in better enhancement; however, a larger L value also increases the number of dimensions involved in the calculation, leading to slower computation. Therefore, the selection and determination of the enhancement parameter L can be based on actual circumstances, and this embodiment of the invention does not impose specific limitations on it.

[0076] Secondly, the kth sampling point x k The enhancement matrix is ​​shown in the following relation (2):

[0077] x k (L)=[x(k),x(k-1),…,x(kl)]∈R 1×(L+1)D (2)

[0078] Where, when the sampling time k is less than the enhancement parameter L, the enhancement matrix X a The form is shown in the following relation (3):

[0079]

[0080] Furthermore, when the sampling time k is greater than the enhancement parameter L, the enhancement matrix X b The form is shown in the following relation (4):

[0081]

[0082] Finally, the final target enhancement matrix X can be determined. L The following relation (5) is shown:

[0083]

[0084] Step S2024: The target enhancement matrix is ​​processed by a nonlinear transformation method to obtain the target relation that characterizes the Mahalanobis distance of the enhancement kernel.

[0085] Specifically, since Mahalanobis distance is not suitable for handling nonlinear problems, a Gaussian kernel function is introduced to capture nonlinearity. Through nonlinear transformation, the enhancement vector is mapped to a high-dimensional feature space to obtain the enhancement kernel feature matrix, from which the target relation for calculating the Mahalanobis distance of the enhancement kernel can be obtained.

[0086] In some optional implementations, step S2024 above includes:

[0087] Step a1: The target enhancement matrix is ​​processed by a nonlinear transformation method to obtain the enhancement kernel feature matrix.

[0088] Step a2: Based on the enhanced kernel feature matrix, determine the initial relation that characterizes the enhanced kernel Mahalanobis distance.

[0089] Step a3: Based on the initial relation and the enhanced kernel feature matrix, the target relation representing the enhanced kernel Mahalanobis distance is obtained after processing with the Gaussian kernel function.

[0090] First, through nonlinear transformation φ k =F(x) k (L)), the enhancement matrix is ​​mapped to a high-dimensional feature space, the dimension of which is represented as h, and the mapping matrix, i.e. the enhancement kernel feature matrix, is expressed as the following relation (6):

[0091] Φ=[φ1,φ2,…,φ k ,…,φ N ]∈R h×N (6)

[0092] Secondly, the initial relation characterizing the enhanced nuclear Markovian distance can be determined, as shown in the following relation (7):

[0093]

[0094] In the formula: i represents the i-th sampling point, j represents the j-th sampling point, (φ i ,φ j )∈Φ;Σ φ The covariance matrix of Φ is represented by the following relation (8):

[0095]

[0096] in:

[0097]

[0098] Specifically, the above relation (9) satisfies Where H represents a symmetric idempotent matrix, as shown in the following relation (10):

[0099]

[0100] In the formula: I N ∈R N×N , represents an N×N identity matrix; e N ∈R N×1 , represents an N×1 unit vector.

[0101] Furthermore, μ φ The mean of Φ is represented by the following relationship (11):

[0102]

[0103] Then, the Gaussian kernel function is selected to calculate the dot product of the mapped vectors, as shown in the following relation (12):

[0104]

[0105] Furthermore, calculate the central kernel matrix. First, the central kernel matrix It can be initially represented as the following relation (13):

[0106]

[0107] Furthermore, using the singular value decomposition (SVD) method to perform full-rank decomposition, the following relation (14) is obtained:

[0108]

[0109] In the formula: U∈R h×h V∈R N×N Both represent orthogonal matrices; S∈R h×N ,Depend on It consists of singular values.

[0110] Furthermore, since S is a full-rank row, its Moore-Penrose pseudoinverse matrix is ​​given by the following relation (15):

[0111] S + =S T (SS T )-1 (15)

[0112] Therefore, the following relation (16) can be obtained:

[0113] (S T S) + =S + (S + ) T =S T (SS T ) -2 S (16)

[0114] Furthermore, the central kernel matrix can be calculated. and its Moore-Penrose pseudoinverse matrix The following relationships are shown respectively: (17) and (18):

[0115]

[0116]

[0117] Finally, the covariance matrix shown in relation (8) is rewritten as shown in relation (19):

[0118]

[0119] Furthermore, the inverse of the rewritten covariance matrix is ​​calculated, as shown in the following relation (20):

[0120] Σ φ -1 =(N-1)U(SS) T ) -1 U T (20)

[0121] Furthermore, combining the above relation (14), we can obtain the following relation (21):

[0122]

[0123] Furthermore, based on the above relation (16), we can obtain the following relation (22):

[0124] S[(S T S) + ] 2 S T =(SS) T ) -1 (twenty two)

[0125] Furthermore, combining the above relations (20) and (21), we can obtain the inverse matrix Σ of the rewritten covariance matrix.φ -1 The following relation (23) is shown:

[0126]

[0127] Furthermore, by rewriting the initial relation of the enhanced kernel Mahalanobis distance shown in relation (7), the target relation for calculating the enhanced kernel Mahalanobis distance can be obtained, as shown in the following relation (24):

[0128]

[0129] In the formula:

[0130] Step S203: Determine the control limits based on the second historical monitoring dataset and the target relationship. For details, please refer to [link to relevant documentation]. Figure 1 Step S103 of the illustrated embodiment will not be described again here.

[0131] Step S204: Obtain the offline monitoring dataset of the lithium-ion battery pack to be warned, and calculate the first enhanced core Mahalanobis distance of the lithium-ion battery pack to be warned based on the offline monitoring dataset and the target relation. For details, please refer to [link to relevant documentation]. Figure 1 Step S104 of the illustrated embodiment will not be described again here.

[0132] Step S205: Based on the first enhanced core Marsh distance and control limit, a safety warning is issued for the lithium-ion battery pack to be warned, yielding the safety warning result for the lithium-ion battery pack to be warned. For details, please refer to [link to relevant documentation]. Figure 1 Step S105 of the illustrated embodiment will not be described again here.

[0133] This embodiment provides a lithium-ion battery pack safety early warning method based on enhanced nuclear Mahalanobis distance. By preprocessing the monitoring data of a lithium-ion battery pack that has experienced thermal runaway before thermal runaway, the accuracy of the monitoring data for thermal runaway lithium-ion battery packs is improved. By using multiple historical monitoring data points of the lithium-ion battery pack as feature inputs, the robustness and reliability of subsequent safety early warning results can be improved. Furthermore, data augmentation methods can better reflect the dynamic characteristics of the lithium-ion battery pack, preserving the correlation between each feature and historical data. Simultaneously, the introduction of a Gaussian kernel function can capture nonlinearity, and combined with nonlinear transformation methods, it can better describe the nonlinear relationships between monitoring data during the historical charge and discharge processes of the lithium-ion battery. Furthermore, by using historical monitoring data of lithium-ion battery packs that have not experienced thermal runaway and combining it with the target relationship characterizing the enhanced nuclear Mahalanobis distance, a control limit for safety early warning can be generated. Finally, by combining the enhanced nuclear Mahalanobis distance of the lithium-ion battery pack to be warned, a safety early warning can be achieved for the lithium-ion battery pack to be warned. Therefore, by implementing this invention, a safety early warning for the nonlinear process of multi-feature fusion in lithium-ion battery packs is realized, and the safety early warning process does not require the establishment of a complex model.

[0134] This embodiment provides a safety early warning method for lithium-ion battery packs based on enhanced nuclear Markov distance. Figure 3 This is a flowchart of a lithium-ion battery pack safety warning method based on enhanced nuclear Martens distance according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps:

[0135] Step S301: Obtain the first historical monitoring dataset of lithium-ion battery packs that experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that did not experience thermal runaway. For details, please refer to [link to relevant documentation]. Figure 2 Step S201 of the illustrated embodiment will not be described again here.

[0136] Step S302: Based on the first historical monitoring dataset, the target relation representing the augmented kernel Mahalanobis distance is obtained through data augmentation and nonlinear transformation methods. For details, please refer to [link to relevant documentation]. Figure 2 Step S202 of the illustrated embodiment will not be described again here.

[0137] Step S303: Determine the control limits based on the second historical monitoring dataset and the target relationship.

[0138] Specifically, step S303 includes:

[0139] Step S3031: Based on the second historical monitoring dataset and the target relation, calculate the second enhanced nuclear Marvin distance of the lithium-ion battery pack that has not experienced thermal runaway.

[0140] Specifically, the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway is used as the input training dataset, and the length N of the training dataset is determined.

[0141] Further, the target enhancement matrix of the second historical monitoring dataset is obtained according to step S2023.

[0142] Furthermore, the idempotent matrix H is calculated using the above relation (10).

[0143] Furthermore, the kernel matrix K and the central kernel matrix are calculated according to the description in step S2024. and its Moore-Penrose pseudoinverse matrix

[0144] Furthermore, calculate the parameters in the above relation (24).

[0145] Furthermore, based on the above relationship (24), the second enhanced core Marvin distance D of the lithium-ion battery pack that has not experienced thermal runaway can be calculated. KM (φ i ,φ j ), where φ i φ represents the enhanced kernel eigenvalue at the i-th sampling time. j The mean μ of the second historical monitoring dataset is represented. φ .

[0146] Step S3032: Based on the second enhanced core Markov distance, the control limit is obtained after processing with the central limit theorem.

[0147] The central limit theorem represents a class of theorems in probability theory that discuss the asymptotic approximation of the distribution of a sequence of random variables to a normal distribution.

[0148] Specifically, assuming that Φ follows a multinomial normal distribution, the control limit can be determined by the central limit theorem. Where α represents the significance level; h represents the dimension of Φ.

[0149] Step S304: Obtain the offline monitoring dataset of the lithium-ion battery pack to be warned, and calculate the first enhanced core Mahalanobis distance of the lithium-ion battery pack to be warned based on the offline monitoring dataset and the target relation.

[0150] Specifically, the offline monitoring dataset of the lithium-ion battery packs to be warned is used as the input training dataset, and the length N of the training dataset is determined.

[0151] Further, the target augmentation matrix of the offline monitoring dataset is obtained according to step S2023.

[0152] Furthermore, by substituting the Y calculated in step S3031 into the above relation (24), the first enhanced core Marsh distance D of the lithium-ion battery pack to be warned can be calculated. KM (φ i ,φ j ), where φ i φ represents the enhanced kernel eigenvalue at the i-th sampling time. j The mean μ of the offline monitoring dataset is represented. φ .

[0153] Step S305: Based on the first enhanced core Marsh distance and control limit, a safety warning is issued for the lithium-ion battery pack to be warned, and the safety warning result of the lithium-ion battery pack to be warned is obtained.

[0154] Specifically, step S305 includes:

[0155] Step S3051: Based on the first enhanced core Marshall distance and control limit, determine whether the lithium-ion battery pack to be warned has malfunctioned.

[0156] Specifically, by comparing the first enhanced kernel Mahalanobis distance D KM (φ i ,φ j and control limits It can determine whether the lithium-ion battery pack under warning has malfunctioned.

[0157] Among them, if This indicates that a fault has occurred in the lithium-ion battery pack awaiting warning; if This indicates that the lithium-ion battery pack awaiting warning has not malfunctioned.

[0158] Step S3052: If the lithium-ion battery pack to be warned has not malfunctioned, it is determined that the lithium-ion battery pack to be warned is in a safe state.

[0159] Specifically, if the lithium-ion battery pack to be warned has not malfunctioned, it means that the lithium-ion battery pack to be warned is currently in a safe state and does not need to be warned.

[0160] Step S3053: When the lithium-ion battery pack to be warned malfunctions, it is determined that the lithium-ion battery pack to be warned is in an unsafe state, and a safety warning message is issued.

[0161] Specifically, if a lithium-ion battery pack under warning malfunctions, it indicates that the lithium-ion battery pack under warning is currently in an unsafe state, and a corresponding safety warning message needs to be issued.

[0162] The lithium-ion battery pack safety early warning method based on enhanced core Markov distance provided in this embodiment first acquires historical monitoring data of lithium-ion battery packs that have experienced thermal runaway and those that have not. Then, it processes the historical monitoring data of the thermal runaway lithium-ion battery pack using data augmentation and nonlinear transformation methods to obtain a target relation characterizing the enhanced core Markov distance. Data augmentation better reflects the dynamic characteristics of the lithium-ion battery pack, preserving the correlation between each feature and historical data. Simultaneously, the nonlinear transformation method better describes the nonlinear relationships between monitoring data during the historical charge and discharge processes of the lithium-ion battery. Furthermore, using the historical monitoring data of the lithium-ion battery pack that has not experienced thermal runaway and the target relation characterizing the enhanced core Markov distance, a control limit for safety early warning can be generated. Finally, by using the first enhanced core Markov distance of the lithium-ion battery pack to be warned and the control limit for safety early warning, it can be determined whether the lithium-ion battery pack to be warned has malfunctioned. Furthermore, combining the judgment result, the safety status of the lithium-ion battery pack to be warned can be determined, thus obtaining the safety early warning result.

[0163] In one example, a multi-feature early warning method for lithium-ion batteries based on enhanced nuclear Markov distance is provided, such as... Figure 4 As shown, it includes the following steps:

[0164] S1: Data preprocessing: Obtain monitoring data of lithium-ion battery packs that have experienced thermal runaway, select initial normal data as training data, select pre-accident data as test data, and preprocess the selected data.

[0165] S2: Data Augmentation: Select the minimum voltage and maximum temperature as input features, establish a feature matrix, consider the autocorrelation of process data, and use data augmentation techniques to establish an augmentation matrix;

[0166] S3: Enhanced Kernel Mahalanobis Distance: Since Mahalanobis distance is not suitable for handling nonlinear problems, an appropriate kernel function is introduced and selected to capture nonlinearity. Through nonlinear transformation, the enhancement vector is mapped to a high-dimensional feature space to obtain the enhancement kernel feature matrix, and then the enhanced kernel Mahalanobis distance is calculated.

[0167] S4: Offline Training: Calculate the augmented kernel Mahalanobis distance and related parameters for normal-state training data. Assuming the variables in the feature matrix follow a multinormal distribution, the control limit is obtained using the central limit theorem as the warning threshold.

[0168] S5: Online Early Warning: Load relevant parameters and alarm thresholds obtained from offline training, calculate the test data to enhance the nuclear Mahalanobis distance, and realize online safety early warning for lithium-ion battery packs.

[0169] Furthermore, step S1 specifically includes the following steps:

[0170] S11: Select monitoring data of lithium-ion battery packs that have experienced overheating runaway as the research object, and filter the data to select data such as time, alarm status, minimum cell voltage and cell number, maximum probe temperature and probe number;

[0171] S12: Select the initial normal data as the training dataset and the data before the accident as the test dataset. Perform data preprocessing on the selected data, such as deduplication and filling, interpolation smoothing, and removal of outliers.

[0172] Furthermore, step S2 specifically includes the following steps:

[0173] S21: For the dataset preprocessed in step S1, select the minimum value of the single-unit voltage, the maximum value of the probe temperature, etc. as feature inputs and establish a multi-dimensional feature matrix, as shown in the above relation (1).

[0174] S22: Determine the enhancement parameter L, where L is a non-negative integer.

[0175] S23: The kth sampling point x k The enhancement matrix is ​​shown in the above relation (2).

[0176] S24: When the sampling time k is less than the enhancement parameter L, the enhancement matrix X a The form is shown in the above relation (3).

[0177] S25: When the sampling time k is greater than the enhancement parameter L, the enhancement matrix X n The form is shown in the above relation (4).

[0178] S26: The final enhancement matrix X L The form is shown in the above relation (5).

[0179] Furthermore, step S3 specifically includes the following steps:

[0180] S31: Through nonlinear transformation φ k =F(x) k (L)), the enhancement matrix is ​​mapped to a high-dimensional feature space, the dimension of which is represented as h, and the mapping matrix is ​​represented as the above relation (6).

[0181] S32: Determine the enhanced nucleus Marvin distance, referring to the above relationships (7) to (11).

[0182] S33: Select the Gaussian kernel function to calculate the dot product of the mapped vectors, as shown in the above relation (12).

[0183] S34: Central Kernel Matrix Refer to the above relations (13) to (18).

[0184] S35: Rewrite the covariance matrix, referring to the above relations (19) to (23).

[0185] S36: Rewrite the enhanced kernel Mahalanobis distance as shown in the above relation (24).

[0186] Furthermore, step S4 specifically includes the following steps:

[0187] S41: Select the training dataset as input, determine the length N of the training dataset, and obtain the augmentation matrix of the training dataset according to step S2;

[0188] S42: Calculate the matrix

[0189] S43: Calculate the kernel matrix K according to relation (12);

[0190] S44: Calculate the central kernel matrix according to step S34. and its Moore-Penrose pseudoinverse matrix

[0191] S45: Calculation

[0192] S46: Calculate the enhanced nuclear Mahalanobis distance D according to relation (24) KM (φ i ,φ j ), where φ i φ represents the enhanced kernel eigenvalue at the i-th sampling time. j This indicates that the mean μ of the selected training dataset is used. φ .

[0193] S47: Assuming Φ follows a multinomial normal distribution, the control limit can be determined by the central limit theorem. Where α represents the significance level; h represents the dimension of Φ.

[0194] Furthermore, step S5 specifically includes the following steps:

[0195] S51: Load parameters Y from offline training.

[0196] S52: Select the test dataset as input, determine the length N of the test dataset, and obtain the enhancement matrix of the test dataset according to step S2;

[0197] S53: Substitute the Υ obtained from offline training into relation (24) to calculate the enhanced kernel Mahalanobis distance, D KM (φ i ,φ j ), where φ iφ represents the enhanced kernel eigenvalue at the i-th sampling time. j This indicates that the mean μ of the selected test dataset is used. φ .

[0198] S54: Determine if a fault has occurred. Then a warning message will be issued.

[0199] In summary, the safety early warning principle based on enhanced nuclear Mahalanobis distance is as follows: Figure 5 As shown.

[0200] The multi-feature early warning method for lithium-ion batteries based on enhanced nuclear Markov distance provided in this example has the following beneficial effects:

[0201] 1. By using multiple variables from lithium-ion battery pack charge and discharge data as feature inputs, the early warning results are made more robust and reliable;

[0202] 2. Data augmentation methods are used to process the original feature matrix, which can better reflect the dynamic characteristics of the battery and retain the correlation between each feature and historical data;

[0203] 3. A kernel method was introduced, which can better describe the nonlinear relationships between data during the charging and discharging process of lithium-ion batteries;

[0204] 4. It can automatically generate warning thresholds based on normal data.

[0205] In summary, the multi-feature early warning method for lithium-ion batteries based on enhanced nuclear Mahalanobis distance provided in this example takes multiple battery pack signals as input, uses kernel functions to process the input signals to characterize the nonlinear process of the battery, and realizes safety early warning of the nonlinear process of multi-feature fusion of lithium-ion battery packs.

[0206] This embodiment also provides a lithium-ion battery pack early warning device based on enhanced nuclear Markov distance. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0207] This embodiment provides a lithium-ion battery pack early warning device based on enhanced nuclear Martens distance, such as... Figure 6 As shown, it includes:

[0208] The acquisition module 601 is used to acquire the first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway.

[0209] The processing module 602 is used to obtain the target relation representing the augmented kernel Mahalanobis distance based on the first historical monitoring dataset through data augmentation and nonlinear transformation methods.

[0210] The determination module 603 is used to determine the control limits based on the second historical monitoring dataset and the target relationship.

[0211] The acquisition and calculation module 604 is used to acquire the offline monitoring dataset of the lithium-ion battery pack to be warned, and calculate the first enhanced nuclear Mahalanobis distance of the lithium-ion battery pack to be warned based on the offline monitoring dataset and the target relation.

[0212] The safety warning module 605 is used to provide safety warnings for the lithium-ion battery pack under warning based on the first enhanced core Marsh distance and control limit, and to obtain the safety warning results for the lithium-ion battery pack under warning.

[0213] In some alternative implementations, the acquisition module 601 includes:

[0214] The first acquisition submodule is used to acquire the third historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway.

[0215] The second acquisition submodule is used to acquire the fourth historical monitoring dataset of the lithium-ion battery pack that has experienced thermal runaway but has not yet experienced thermal runaway.

[0216] The processing submodule is used to preprocess the third historical monitoring dataset based on the fourth historical monitoring dataset to obtain the first historical monitoring dataset of the lithium-ion battery pack that has experienced thermal runaway.

[0217] In some alternative implementations, the processing module 602 includes:

[0218] The selection submodule is used to select the feature input dataset from the first historical monitoring dataset.

[0219] Create a submodule for building a multidimensional feature matrix based on the feature input dataset.

[0220] The processing and creation submodule is used to create a target augmentation matrix based on a multidimensional feature matrix and processed by data augmentation methods.

[0221] The first processing submodule is used to process the target enhancement matrix through a nonlinear transformation method to obtain the target relation that characterizes the enhancement kernel Mahalanobis distance.

[0222] In some alternative implementations, the first processing submodule includes:

[0223] The first processing unit is used to process the target enhancement matrix through a nonlinear transformation method to obtain the enhancement kernel feature matrix.

[0224] The unit is defined to determine the initial relation characterizing the Mahalanobis distance of the enhanced kernel based on the enhanced kernel feature matrix.

[0225] The second processing unit is used to obtain the target relation representing the enhanced kernel Mahalanobis distance based on the initial relation and the enhanced kernel feature matrix, after processing with a Gaussian kernel function.

[0226] In some alternative implementations, the determining module 603 includes:

[0227] The calculation submodule is used to calculate the second enhanced nuclear Mahalanobis distance of lithium-ion battery packs that have not experienced thermal runaway, based on the second historical monitoring dataset and the target relation.

[0228] The second processing submodule is used to obtain the control limit based on the second enhanced kernel Markov distance and the central limit theorem.

[0229] In some alternative implementations, the safety warning module 605 includes:

[0230] The judgment submodule is used to determine whether the lithium-ion battery pack to be warned has malfunctioned based on the first enhanced core Marshall distance and control limit.

[0231] The first determination submodule is used to determine that the lithium-ion battery pack to be warned is in a safe state when no fault has occurred.

[0232] The second determination submodule is used to determine that the lithium-ion battery pack to be warned is in an unsafe state when a fault occurs, and to issue a safety warning message.

[0233] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0234] The lithium-ion battery pack safety warning device based on enhanced nuclear Mahalanobis distance in this embodiment is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0235] This invention also provides a computer device having the above-described features. Figure 6 The image shows a lithium-ion battery pack safety warning device based on enhanced nuclear Marsh distance.

[0236] Please see Figure 7 ,Figure 7 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 7 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 7 Take a processor 10 as an example.

[0237] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0238] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0239] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0240] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0241] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0242] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0243] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0244] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A safety early warning method for lithium-ion battery packs based on enhanced nuclear Markov distance, characterized in that, The method includes: Obtain the first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway. Based on the first historical monitoring dataset, the target relation representing the augmented kernel Mahalanobis distance is obtained through data augmentation and nonlinear transformation methods. Based on the second historical monitoring dataset and the target relation, the control limits are determined; Obtain the offline monitoring dataset of the lithium-ion battery pack to be warned, and calculate the first enhanced nuclear Mahalanobis distance of the lithium-ion battery pack to be warned based on the offline monitoring dataset and the target relation. Based on the first enhanced nuclear Martens distance and the control limit, a safety warning is issued for the lithium-ion battery pack to be warned, and the safety warning result of the lithium-ion battery pack to be warned is obtained.

2. The method according to claim 1, characterized in that, Obtain the first historical monitoring dataset of lithium-ion battery packs that experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that did not experience thermal runaway, including: The third historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway are obtained. The third historical monitoring dataset is obtained by filtering the monitoring data of lithium-ion battery packs that have experienced thermal runaway, including time, alarm status, minimum value of single cell voltage and its single cell number, maximum value of probe temperature and its probe number. Obtain the fourth historical monitoring dataset of a lithium-ion battery pack that has experienced thermal runaway but has not yet experienced thermal runaway; The third historical monitoring dataset is preprocessed based on the fourth historical monitoring dataset to obtain the first historical monitoring dataset of the lithium-ion battery pack that has experienced thermal runaway. The first historical monitoring dataset includes data of the lithium-ion battery pack that has experienced thermal runaway during the charging and discharging process.

3. The method according to claim 1, characterized in that, Based on the first historical monitoring dataset, after processing with data augmentation and nonlinear transformation methods, the target relation representing the augmented kernel Mahalanobis distance is obtained, including: Select a feature input dataset from the first historical monitoring dataset; A multidimensional feature matrix is ​​established based on the aforementioned feature input dataset; Based on the multidimensional feature matrix, and after processing by the data augmentation method, a target augmentation matrix is ​​established. The target enhancement matrix is ​​processed by the nonlinear transformation method to obtain the target relation characterizing the Mahalanobis distance of the enhancement kernel.

4. The method according to claim 3, characterized in that, The target enhancement matrix is ​​processed by the nonlinear transformation method to obtain the target relation characterizing the Mahalanobis distance of the enhancement kernel, including: The target enhancement matrix is ​​processed by the nonlinear transformation method to obtain the enhancement kernel feature matrix; Based on the enhanced kernel feature matrix, an initial relational expression characterizing the Mahalanobis distance of the enhanced kernel is determined; Based on the initial relation and the enhanced kernel feature matrix, the target relation representing the enhanced kernel Mahalanobis distance is obtained after processing with a Gaussian kernel function.

5. The method according to claim 1, characterized in that, Based on the second historical monitoring dataset and the target relation, control limits are determined, including: Based on the second historical monitoring dataset and the target relation, the second enhanced nuclear Marvin distance of the lithium-ion battery pack that has not experienced thermal runaway is calculated; Based on the second enhanced core Markov distance, the control limit is obtained after processing with the central limit theorem.

6. The method according to claim 1, characterized in that, Based on the first enhanced nuclear Marvin distance and the control limit, a safety warning is issued for the lithium-ion battery pack to be warned, and the safety warning result for the lithium-ion battery pack to be warned is obtained, including: Based on the first enhanced nuclear Marshall distance and the control limit, it is determined whether the lithium-ion battery pack to be warned has malfunctioned. If the lithium-ion battery pack to be warned does not malfunction, it is determined that the lithium-ion battery pack to be warned is in a safe state; When the lithium-ion battery pack to be warned malfunctions, it is determined that the lithium-ion battery pack to be warned is in an unsafe state, and a safety warning message is issued.

7. A lithium-ion battery pack early warning device based on enhanced nuclear Martens distance, characterized in that, The device includes: The acquisition module is used to acquire the first historical monitoring dataset of lithium-ion battery packs that have experienced thermal runaway and the second historical monitoring dataset of lithium-ion battery packs that have not experienced thermal runaway. The processing module is used to process the first historical monitoring dataset using data augmentation and nonlinear transformation methods to obtain the target relation representing the augmented kernel Mahalanobis distance. The determination module is used to determine the control limits based on the second historical monitoring dataset and the target relation. The acquisition and calculation module is used to acquire the offline monitoring dataset of the lithium-ion battery pack to be warned, and calculate the first enhanced nuclear Mahalanobis distance of the lithium-ion battery pack to be warned based on the offline monitoring dataset and the target relation. The safety warning module is used to provide a safety warning for the lithium-ion battery pack to be warned based on the first enhanced nuclear Marshall distance and the control limit, and to obtain the safety warning result of the lithium-ion battery pack to be warned.

8. A computer device, characterized in that, include: A memory and a processor are interconnected, the memory stores computer instructions, and the processor executes the computer instructions to perform the lithium-ion battery pack safety warning method based on enhanced nuclear Mahalanobis distance as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the lithium-ion battery pack safety warning method based on enhanced nuclear Mahalanobis distance as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes computer instructions for causing a computer to execute the lithium-ion battery pack safety warning method based on enhanced nuclear Mahalanobis distance as described in any one of claims 1 to 6.

Images