Battery thermal runaway early warning method and device, computing device, and storage medium

Abstract

The application discloses a battery thermal runaway early warning method and device, a computing device and a storage medium, and belongs to the technical field of batteries. The method comprises the following steps: acquiring charging data in a charging process of a to-be-identified battery, and determining the internal resistance value of each battery monomer at each sampling time point according to the charging data; determining the internal resistance range, internal resistance standard deviation and coefficient of variation of the to-be-identified battery according to the internal resistance value of each battery monomer at each sampling time point; inputting the internal resistance range, internal resistance standard deviation and coefficient of variation into a pre-trained early warning model to determine the thermal runaway risk of the to-be-identified battery. The above method simplifies the identification of the thermal runaway risk into a classification problem based on characteristic indexes, realizes the prediction of the thermal runaway risk of the to-be-identified battery, can improve the accuracy of the thermal runaway early warning, reduces the risk of vehicle thermal runaway, and further reduces personal and property losses.

Description

Technical Field

[0001] This invention relates to the field of battery technology, and in particular to a method, device, computing equipment, and storage medium for early warning of battery thermal runaway. Background Technology

[0002] In recent years, several thermal runaway incidents have occurred in new energy vehicles. When a vehicle experiences thermal runaway, the heat inside the battery cannot dissipate and accumulates, causing the internal temperature to rise. After a large amount of oxygen is released from the positive electrode, it will burn violently in a very short time, seriously threatening the lives and property of the people.

[0003] Current technologies typically employ single voltage and temperature offset measurements, while simultaneously monitoring combustible gas concentration and pressure in real time. This involves determining if the combustible gas concentration and pressure have reached threshold values, assigning different fault bits to different faults, and summing these fault bits to obtain the total fault bit. These parameters are then used for thermal runaway early warning. However, a battery is a hysteretic system; its terminal voltage fluctuates continuously during use. Therefore, it is difficult to determine battery faults using fixed threshold values, resulting in low accuracy for this early warning method. Summary of the Invention

[0004] To address this issue, the present invention proposes a battery thermal runaway early warning method, as well as a battery thermal runaway early warning device, a computing device, and a computer-readable storage medium, aiming to at least partially solve the technical problem of low early warning accuracy in related technologies.

[0005] To achieve the above objectives, a first aspect of the present invention provides a battery thermal runaway early warning method, the method comprising:

[0006] Acquire charging data during the charging process of the battery to be identified, and determine the internal resistance value of each battery cell at each sampling time point based on the charging data;

[0007] Based on the internal resistance value of each battery cell at each sampling time point, determine the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified.

[0008] The internal resistance range, internal resistance standard deviation, and coefficient of variation are input into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified.

[0009] According to one embodiment of the present invention, the charging data includes the temperature value of each temperature detection point at each sampling time point, the sampling current value corresponding to each sampling time point, the sampling voltage value of each battery cell at each sampling time point, and the remaining battery capacity.

[0010] Based on the charging data, determine the internal resistance value of each battery cell at each sampling time point, including:

[0011] Based on the temperature values ​​of each temperature detection point at each sampling time point, determine the corresponding battery temperature value at each sampling time point;

[0012] Based on the battery temperature value at each sampling time point and the remaining battery charge of each battery cell at each sampling time point, determine the open circuit voltage value of each battery cell at each sampling time point.

[0013] Based on the sampling voltage and open-circuit voltage values ​​and the sampling current values ​​of each battery cell at each sampling time point, the internal resistance value of each battery cell at each sampling time point is determined.

[0014] According to one embodiment of the present invention, the internal resistance value of each battery cell at each sampling time point is determined based on the sampling voltage value and open-circuit voltage value of each battery cell at each sampling time point, and the sampling current value at each sampling time point, including:

[0015] Based on the sampling voltage and open-circuit voltage values ​​of the target battery cell at the target sampling time point, the sampling voltage and open-circuit voltage values ​​of the target battery cell at the previous sampling time point, the sampling current values ​​at the target sampling time point and the previous sampling time point, the first parameter and the second parameter at the target sampling time point, and the preset parameters, the internal resistance value of the target battery cell at the target sampling time point is determined. Here, the target battery cell is any battery cell of the battery to be identified, and the target sampling time point is any sampling time point.

[0016] According to one embodiment of the present invention, the internal resistance value of the target battery cell at the target sampling time point is determined by the following formula:

[0017] φ(k)=[U(k-1)-OCV(k-1)-I(k) I(k-1)] T

[0018] Y(k)=U(k)-OCV(k)

[0019] E(k)=Y(k)-θ(k-1) T φ(k)

[0020]

[0021] Wherein, U(k) represents the sampling voltage value of the target battery cell at the target sampling time point, OCV(k) represents the open-circuit voltage value of the target battery cell at the target sampling time point, U(k-1) represents the sampling voltage value of the target battery cell at the previous sampling time point, OCV(k-1) represents the open-circuit voltage value of the target battery cell at the previous sampling time point, I(k) represents the sampling current value at the target sampling time point, I(k-1) represents the sampling current value at the previous sampling time point, θ(k-1) represents the first parameter at the target sampling time point, P(k-1) represents the second parameter at the target sampling time point, β represents the preset parameter, K(k)E(k) represents the internal resistance value of the target battery cell at the target sampling time point, φ(k) represents the third parameter at the target sampling time point, K(k) represents the fourth parameter at the target sampling time point, and E(k) represents the fifth parameter at the target sampling time point.

[0022] According to an embodiment of the present invention, the first parameter and the second parameter corresponding to the target sampling time point are determined by the following formula:

[0023] θ(k-1)=θ(k-2)+K(k-1)E(k-1)

[0024]

[0025] Where θ(k-1) represents the first parameter corresponding to the target sampling time point, P(k-1) represents the second parameter corresponding to the target sampling time point, θ(k-2) represents the first parameter corresponding to the previous sampling time point, P(k-2) represents the second parameter corresponding to the previous sampling time point, φ(k-1) represents the third parameter corresponding to the previous sampling time point, K(k-1) represents the fourth parameter corresponding to the previous sampling time point, E(k-1) represents the fifth parameter corresponding to the previous sampling time point, and K(k-1)E(k-1) represents the internal resistance value of the target battery cell at the previous sampling time point.

[0026] According to one embodiment of the present invention, the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified are determined based on the internal resistance value of each battery cell at various sampling time points, including:

[0027] The charging process of the battery to be identified is divided into multiple charging time periods, and multiple sampling time points are determined for each charging time period.

[0028] The internal resistance value of each battery cell at the last sampling time point in each charging period is determined as the internal resistance value of each battery cell in each charging period.

[0029] Based on the internal resistance value of each battery cell within the same charging time period, determine the internal resistance range corresponding to each charging time period.

[0030] The internal resistance range of the battery to be identified is determined based on the internal resistance range corresponding to each charging period.

[0031] The standard deviation of internal resistance for each charging period is determined based on the internal resistance value of each battery cell in each charging period and the number of battery cells in each charging period.

[0032] The standard deviation of the internal resistance of the battery to be identified is determined based on the standard deviation of the internal resistance corresponding to each charging period.

[0033] The coefficient of variation for each charging time period is determined based on the standard deviation of the internal resistance for each charging time period and the average internal resistance for each charging time period.

[0034] The coefficient of variation of the battery to be identified is determined based on the coefficient of variation corresponding to each charging period.

[0035] According to one embodiment of the present invention, the early warning model is trained in the following manner:

[0036] Obtain the training dataset, which includes multiple sets of training charging data and the actual thermal runaway risk of the training battery corresponding to each set of training charging data;

[0037] The training dataset is input into the decision tree model to determine the predicted thermal runaway risk of the training battery corresponding to each set of training charging data.

[0038] The loss value is determined based on the predicted thermal runaway risk and the actual thermal runaway risk of the training battery;

[0039] The decision tree model is trained based on the loss value until the training stops, resulting in a well-trained early warning model.

[0040] According to one embodiment of the present invention, the method further includes:

[0041] The trained early warning model is pruned to obtain an updated early warning model.

[0042] Accordingly, the internal resistance range, internal resistance standard deviation, and coefficient of variation are input into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified, including:

[0043] The internal resistance range, internal resistance standard deviation, and coefficient of variation are input into the updated early warning model to determine the thermal runaway risk of the battery to be identified.

[0044] To achieve the above objectives, a second aspect of the present invention provides a battery thermal runaway early warning device, the device comprising:

[0045] The first determining module is configured to acquire charging data during the charging process of the battery to be identified, and determine the internal resistance value of each battery cell at each sampling time point based on the charging data.

[0046] The second determining module is configured to determine the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified based on the internal resistance value of each battery cell at each sampling time point.

[0047] The third determination module is configured to input the internal resistance range, internal resistance standard deviation, and coefficient of variation into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified.

[0048] To achieve the above objectives, a third aspect of the present invention provides a computing device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the battery thermal runaway early warning method as described in any of the first aspects above.

[0049] To achieve the above objectives, a third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the battery thermal runaway early warning method described in any of the first aspects.

[0050] The battery thermal runaway early warning method provided in this invention acquires charging data during the charging process of the battery to be identified, and determines the internal resistance value of each battery cell at each sampling time point based on the charging data; based on the internal resistance value of each battery cell at each sampling time point, it determines the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified; and inputs the internal resistance range, internal resistance standard deviation, and coefficient of variation into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified. The above method determines the internal resistance value of the battery cells in the battery to be identified based on charging data, processes the internal resistance value to obtain the internal resistance range, internal resistance standard deviation, and coefficient of variation, and inputs these three parameters as features into a pre-trained early warning model. This simplifies the identification of thermal runaway risk into a classification problem based on feature indicators, enabling the prediction of the thermal runaway risk of the battery to be identified. This improves the accuracy of thermal runaway early warning, reduces the risk of vehicle thermal runaway, and thus reduces personal injury and property damage.

[0051] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0052] Figure 1 This is a flowchart of a battery thermal runaway early warning method provided by an embodiment of the present invention;

[0053] Figure 2This is a schematic diagram of a first-order RC equivalent circuit model provided by an embodiment of the present invention;

[0054] Figure 3 This is a flowchart of another battery thermal runaway early warning method provided according to an embodiment of the present invention;

[0055] Figure 4 This is a schematic diagram of the structure of a battery thermal runaway early warning device according to an embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of the structure of a computing device provided according to an embodiment of the present invention. Detailed Implementation

[0057] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0058] The following description, with reference to the accompanying drawings, outlines the battery thermal runaway early warning method, apparatus, computing device, and storage medium proposed in embodiments of the present invention.

[0059] Figure 1 This is a flowchart of a battery thermal runaway early warning method according to an embodiment of the present invention, which may include the following steps.

[0060] Step 101: Obtain charging data during the charging process of the battery to be identified, and determine the internal resistance value of each battery cell at each sampling time point based on the charging data.

[0061] In this embodiment of the invention, the battery to be identified can be a battery in a vehicle. The battery thermal runaway early warning of this solution can identify whether the vehicle has a risk of thermal runaway, so that measures can be taken in advance when there is a risk to reduce personal and property losses.

[0062] As an example, the battery to be identified may include multiple individual battery cells, and the charging data may include the temperature values ​​at each temperature detection point at each sampling time point, the sampling current values ​​corresponding to each sampling time point, the sampling voltage values ​​of each individual battery cell at each sampling time point, and the remaining battery capacity. The temperature values ​​at each temperature detection point can be the temperatures collected by temperature sensors at various temperature detection points inside the vehicle.

[0063] In practice, the relationship between the remaining charge (SOC) and open circuit voltage (OCV) of each battery cell at different temperatures can be pre-detected, and SOC-OCV curves at different temperatures can be generated. As an example, the operating voltage and operating temperature range of each battery cell can be determined according to the cell product specifications, such as an operating voltage range of 2.5V to 4.2V and an operating temperature range of -20℃ to 55℃. Assuming eight temperature points are selected—-20℃, -10℃, 0℃, 15℃, 25℃, 35℃, 45℃, and 55℃—SOC-OCV curves can be calibrated at different temperatures. Furthermore, to ensure the accuracy of subsequent early warning algorithm calculations, a total of 21 SOC points (0%, 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, and 100%) can be selected to measure the open-circuit voltage.

[0064] In some embodiments, determining the internal resistance value of each battery cell at each sampling time point based on charging data may include: determining the battery temperature value corresponding to each sampling time point based on the temperature values ​​of each temperature detection point at each sampling time point; determining the open-circuit voltage value of each battery cell at each sampling time point based on the battery temperature value corresponding to each sampling time point and the remaining battery charge of each battery cell at each sampling time point; and determining the internal resistance value of each battery cell at each sampling time point based on the sampling voltage value, open-circuit voltage value, and sampling current value of each battery cell at each sampling time point.

[0065] In practice, for each sampling time point, the average or median of the temperature values ​​at each temperature detection point at that sampling time point can be taken as the battery temperature value corresponding to that sampling time point. Then, based on the battery temperature value and the remaining charge of each battery cell, the open-circuit voltage value of that battery cell is determined from the SOC-OCV curve corresponding to that battery cell. Next, based on the sampling voltage and open-circuit voltage values ​​and the sampling current values ​​of each battery cell at each sampling time point, the internal resistance value of each battery cell at each sampling time point is determined using Recursive Least Squares with Forgetting Factor (FFRLS). Furthermore, the calculation of the internal resistance value of each battery cell at the current sampling time point depends on the calculation at the previous sampling time point.

[0066] As an example, the calculation process for the battery temperature values ​​at each sampling time point includes:

[0067] T_avg(n) = median(T1(n),T2(n),...,Tm(n)), or,

[0068] T_avg(n)=(T1(n)+T2(n)+...+Tm(n)) / m.

[0069] Where T_avg(n) represents the battery temperature value corresponding to the nth sampling time point, median() is the median calculation formula, Tm(n) represents the temperature value of the mth temperature detection point at the nth sampling time point, n represents the number of sampling time points, and m represents the number of temperature detection points, i.e. the number of temperature sensors.

[0070] As an example, SOC-OCV curves for each battery cell at different temperatures are pre-generated. For any given battery cell, the curve for that cell at the specified battery temperature can be determined first. Then, based on the SOC (sustainable charge) of that cell (i.e., the remaining battery capacity), the corresponding OCV (open-circuit voltage) value can be determined. For instance, if the remaining battery capacity is not the same as the sampled remaining battery capacity when the SOC-OCV curve was generated, a linear interpolation method is used to calculate the open-circuit voltage value corresponding to that remaining battery capacity.

[0071] In this embodiment of the invention, the open-circuit voltage of a battery cell is calculated using linear interpolation based on the remaining battery capacity, temperature, and SOC-OCV curves calibrated at different temperatures. This method is simple, time-saving, and improves efficiency.

[0072] According to one embodiment of the present invention, determining the internal resistance value of each battery cell at each sampling time point based on the sampling voltage value and open-circuit voltage value of each battery cell at each sampling time point, and the sampling current value at each sampling time point, may include:

[0073] Based on the sampling voltage and open-circuit voltage values ​​of the target battery cell at the target sampling time point, the sampling voltage and open-circuit voltage values ​​of the target battery cell at the previous sampling time point, the sampling current values ​​at the target sampling time point and the previous sampling time point, the first parameter and the second parameter at the target sampling time point, and the preset parameters, the internal resistance value of the target battery cell at the target sampling time point is determined. Here, the target battery cell is any battery cell of the battery to be identified, and the target sampling time point is any sampling time point.

[0074] In this embodiment of the invention, when calculating the internal resistance of a single battery cell at the target sampling time point, it is necessary to use not only the sampling voltage, open-circuit voltage, and sampling current values ​​of the battery cell at the current sampling time point, as well as some other parameters, but also the sampling voltage, open-circuit voltage, and sampling current values ​​of the battery cell at the previous sampling time point. This allows for dynamic determination of the battery cell's internal resistance, resulting in more accurate results.

[0075] In practical implementation, a single battery cell can be represented by a first-order RC equivalent circuit model, with the voltage and current reference directions as follows: Figure 2 As shown, Figure 2 This is a schematic diagram of a first-order RC equivalent circuit model provided by an embodiment of the present invention. Based on this equivalent circuit model, Kirchhoff's voltage law and Kirchhoff's current law are applied, letting... Where τ=R1C1, we can obtain the following formula (1):

[0076] U 1,k =U 1,k-1 α k-1 +R 1,k-1 I k-1 (1-α k-1 )

[0077] U k =OCV k -I k R 0,k -U 1,k (1)

[0078] Among them, I k U represents the sampled current value at time k (the kth sampling time point). k OCV represents the sampled voltage value of a single battery cell at time k, i.e., the voltage between the positive and negative terminals of the battery cell. k U represents the open-circuit voltage of a single battery cell at time k. 1,k Let R be the voltage of the RC parallel network at time k. 0,k and R 1,k-1 These are the internal resistance values ​​of a single cell at time k and time k-1, respectively.

[0079] Based on the above formula, we can obtain in:

[0080]

[0081] In some embodiments, the internal resistance of the target battery cell at the target sampling time point is determined by the following formula (2):

[0082] φ(k)=[U(k-1)-OCV(k-1)-I(k) I(k-1)] T

[0083] Y(k)=U(k)-OCV(k)

[0084] E(k)=Y(k)-θ(k-1) T φ(k)

[0085]

[0086] Wherein, U(k) represents the sampling voltage value of the target battery cell at the target sampling time point, OCV(k) represents the open-circuit voltage value of the target battery cell at the target sampling time point, U(k-1) represents the sampling voltage value of the target battery cell at the previous sampling time point, OCV(k-1) represents the open-circuit voltage value of the target battery cell at the previous sampling time point, I(k) represents the sampling current value at the target sampling time point, I(k-1) represents the sampling current value at the previous sampling time point, θ(k-1) represents the first parameter at the target sampling time point, P(k-1) represents the second parameter at the target sampling time point, β represents the preset parameter, K(k)E(k) represents the internal resistance value of the target battery cell at the target sampling time point, φ(k) represents the third parameter at the target sampling time point, K(k) represents the fourth parameter at the target sampling time point, and E(k) represents the fifth parameter at the target sampling time point.

[0087] Taking the target sampling time point as the first sampling time point and calculating the internal resistance of the target battery cell at the target sampling time point as an example, the first parameter θ(0) and the second parameter P(0) corresponding to the first sampling time point can be set empirically, and the preset parameter β can also be set empirically, and the preset parameter remains unchanged. U(0), OCV(0) and I(0) are all 0, U(1) is the sampling voltage value of the target battery cell at the first sampling time point, OCV(1) is the open-circuit voltage value of the target battery cell at the first sampling time point, and I(1) is the sampling current value at the first sampling time point. According to the above formula, the third parameter φ(1) corresponding to the target sampling time point can be calculated as [0 - I(1) 0]. T The fourth parameter Y(1)=U(1)-OCV(1) and the fifth parameter E(1)=Y(1)-θ(0) T If φ(1), then the internal resistance value of the battery cell at the first sampling time point is K(1)E(1).

[0088] Taking the example of calculating the internal resistance of a target battery cell at the target sampling time point when the target sampling time point is not the first sampling time point, the first parameter corresponding to the target sampling time point can be determined based on the first parameter corresponding to the previous sampling time point and the internal resistance of the battery cell at the previous sampling time point. The second parameter corresponding to the target sampling time point can be determined based on the preset parameter, the second parameter corresponding to the previous sampling time point, the third parameter, and the fourth parameter.

[0089] According to an embodiment of the present invention, the first parameter and the second parameter corresponding to the target sampling time point are determined by the following formula (3):

[0090] θ(k-1)=θ(k-2)+K(k-1)E(k-1)

[0091]

[0092] Where θ(k-1) represents the first parameter corresponding to the target sampling time point, P(k-1) represents the second parameter corresponding to the target sampling time point, θ(k-2) represents the first parameter corresponding to the previous sampling time point, P(k-2) represents the second parameter corresponding to the previous sampling time point, φ(k-1) represents the third parameter corresponding to the previous sampling time point, K(k-1) represents the fourth parameter corresponding to the previous sampling time point, E(k-1) represents the fifth parameter corresponding to the previous sampling time point, and K(k-1)E(k-1) represents the internal resistance value of the target battery cell at the previous sampling time point.

[0093] In specific implementation, taking the target sampling time point as the second sampling time point as an example, the first parameter θ(1) and the second parameter P(1) corresponding to the second sampling time point can be determined according to the above formula (3). Thus, θ(1), P(1), Y(1), E(1) and

[0094] As an example, determining the internal resistance of the target battery cell at the second sampling time point includes: first, determining the battery temperature value T_avg(2) corresponding to the second sampling time point based on the average or median of the temperature values ​​at each temperature detection point at the second sampling time point; then, based on the remaining battery charge SOC(2) and battery temperature value T_avg(2) of the target battery cell, finding the open circuit voltage value OCV(2) of the target battery cell in the pre-calibrated SOC-OCV curve table of the target battery cell at different temperatures; and then, using the following formula (4) to calculate the fourth parameter K(2) and the fifth parameter E(2) corresponding to the second sampling time point, thereby obtaining the internal resistance of the target battery cell at the second sampling time point.

[0095] φ(2)=[U(1)-OCV(1)-I(2) I(1)] T

[0096] Y(2)=U(2)-OCV(2)

[0097] E(2)=Y(2)-θ(1) T φ(2)

[0098]

[0099] θ(2)=θ(1)+K(2)E(2)

[0100]

[0101] Wherein, U(1) is the sampled voltage value of the target battery cell at the first sampling time point, OCV(1) is the open-circuit voltage value of the target battery cell at the first sampling time point, I(1) is the sampled current value at the first sampling time point, U(2) is the sampled voltage value of the target battery cell at the second sampling time point, OCV(2) is the open-circuit voltage value of the target battery cell at the second sampling time point, and I(2) is the sampled current value at the second sampling time point. θ(1) is the third parameter corresponding to the second sampling time point, P(1) is the first parameter corresponding to the second sampling time point, K(2) and E(2) are the fourth and fifth parameters corresponding to the second sampling time point, respectively, and the product of K(2) and E(2) is the internal resistance value of the target battery cell at the second sampling time point. Then, according to θ(1), K(2), E(2) and P(1) can be used to calculate θ(2) and P(2), which are the first and second parameters corresponding to the next sampling time point, and are used to calculate the internal resistance value corresponding to the next sampling time point.

[0102] By combining the above formulas, the internal resistance value of each battery cell in the battery to be charged at each sampling time point can be calculated, that is, the internal resistance value of all series-connected battery cells in the vehicle.

[0103] In this embodiment of the invention, based on the first-order RC equivalent circuit model and Kirchhoff's voltage and current laws, the recursive least squares method with a forgetting factor is used to calculate the internal resistance value of each battery cell at each sampling time point, which simplifies the internal resistance calculation process and improves the internal resistance calculation efficiency.

[0104] Step 102: Determine the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified based on the internal resistance value of each battery cell at each sampling time point.

[0105] In this embodiment of the invention, the charging process of the battery to be identified can be divided into multiple charging time periods. Each charging time period can include multiple sampling time points. For a target battery cell, the internal resistance value corresponding to the last sampling time point in each charging time period can be determined as the internal resistance value of the target battery cell in that charging time period. Then, based on the internal resistance value of each battery cell in each charging time period, the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified are determined.

[0106] The coefficient of variation, also known as the coefficient of dispersion, is the ratio of the standard deviation to the mean. Furthermore, the charging time period can be obtained by dividing the charging process into averages according to a preset duration, which can be set empirically; alternatively, the charging time period can also be obtained by dividing the charging process according to the battery's charging characteristics. This embodiment of the invention does not limit this approach.

[0107] In practice, the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified can be determined first. Then, the internal resistance range of the battery to be identified can be determined based on the internal resistance range of multiple charging time periods, the internal resistance standard deviation of the battery to be identified can be determined based on the internal resistance standard deviation of multiple charging time periods, and the coefficient of variation of the battery to be identified can be determined based on the coefficient of variation of multiple charging time periods.

[0108] In some embodiments, determining the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified based on the internal resistance value of each battery cell at each sampling time point may include: dividing the charging process of the battery to be identified into multiple charging time periods and determining multiple sampling time points corresponding to each charging time period; determining the internal resistance value of each battery cell at the last sampling time point within each charging time period as the internal resistance value of each battery cell within each charging time period; determining the internal resistance range corresponding to each charging time period based on the internal resistance value of each battery cell within the same charging time period; determining the internal resistance range of the battery to be identified based on the internal resistance range corresponding to each charging time period; determining the internal resistance standard deviation corresponding to each charging time period based on the internal resistance value of each battery cell within each charging time period and the number of battery cells within each charging time period; determining the internal resistance standard deviation of the battery to be identified based on the internal resistance standard deviation corresponding to each charging time period; determining the coefficient of variation corresponding to each charging time period based on the internal resistance standard deviation corresponding to each charging time period and the average internal resistance corresponding to each charging time period; and determining the coefficient of variation of the battery to be identified based on the coefficient of variation corresponding to each charging time period.

[0109] In practice, the charging process is divided into multiple charging time periods. Each charging time period includes multiple sampling time points. After determining the internal resistance value of each battery cell at each sampling time point, for the target charging time period, the internal resistance value of each battery cell at the last sampling time point of the target charging time period can be determined as the internal resistance value of each battery cell in the target charging time period.

[0110] As an example, the range refers to the difference between the maximum and minimum values ​​of a set of data, expressed by the formula R = max(x) - min(x), where x represents the data of the object under study, which in this scheme is the internal resistance value of a single battery cell. Within any target charging time period, the internal resistance values ​​corresponding to multiple battery cells are compared, and the difference between the maximum and minimum internal resistance values ​​is determined as the internal resistance range corresponding to that target charging time period. This process is repeated to determine the internal resistance range for each charging time period. Averaging the internal resistance ranges across multiple charging time periods yields the internal resistance range of the battery to be identified.

[0111] For example, it can be done through formula R i =max(R) 0,i )-min(R 0,i ) Calculate the internal resistance range corresponding to each charging time period, where max(R 0,i ) represents the maximum internal resistance value during the i-th charging time period, min(R 0,i R represents the minimum internal resistance value during the i-th charging time period. i This represents the range of internal resistance during the i-th charging time period. (Using the formula...) Determine the internal resistance range of the battery to be identified, where k represents the number of charging time periods during the charging process.

[0112] As an example, standard deviation can reflect the dispersion of a dataset. Based on the internal resistance value of each battery cell and the number of battery cells within the target charging time period, the average internal resistance of multiple battery cells within the target charging time period is first determined. Then, based on the average internal resistance, the internal resistance value of each battery cell, and the number of battery cells, the standard deviation of the internal resistance corresponding to the target charging time period is determined. This process can be repeated to determine the standard deviation of the internal resistance for each charging time period. By averaging the standard deviations of the internal resistance for multiple charging time periods, the standard deviation of the internal resistance of the battery to be identified can be obtained.

[0113] For example, it can be done through the formula Determine the standard deviation of the internal resistance corresponding to the i-th charging time period, where s i R represents the standard deviation of the internal resistance corresponding to the i-th charging time period. 0,j Let represent the internal resistance value of the j-th battery cell during the i-th charging time period, and l represent the number of battery cells during the i-th charging time period. This represents the average value of all internal resistance values ​​during the i-th charging time period. (Using the formula...) Determine the standard deviation of the internal resistance of the battery to be identified, where k represents the number of charging time periods during the charging process.

[0114] As an example, the coefficient of variation, also known as the coefficient of dispersion, is the ratio of the standard deviation to the mean. The ratio of the standard deviation to the mean of the internal resistance corresponding to the target charging time period is determined as the coefficient of variation for that target charging time period. This process can be repeated to determine the coefficient of variation for each charging time period. The average of the coefficients of variation for multiple charging time periods is then taken as the coefficient of variation for the battery to be identified.

[0115] For example, it can be done through the formula Among them, c v,i Let s represent the coefficient of variation for the i-th charging time period. i This represents the standard deviation of the internal resistance corresponding to the i-th charging time period. This represents the average value of all internal resistance values ​​during the i-th charging time period. (Using the formula...) Determine the coefficient of variation of the battery to be identified, where k represents the number of charging time periods during the charging process.

[0116] In this embodiment of the invention, based on the internal resistance value of each battery cell during each charging period, three variables—internal resistance range, internal resistance standard deviation, and coefficient of variation—are used to characterize the internal resistance characteristics. This can accurately characterize the characteristics of the battery to be identified, providing an accurate data foundation for subsequent identification of the thermal runaway risk of the battery based on these three variables.

[0117] Step 103: Input the internal resistance range, internal resistance standard deviation, and coefficient of variation into the pre-trained early warning model to determine the thermal runaway risk of the battery to be identified.

[0118] In this embodiment of the invention, the early warning model can be obtained by pre-training a decision tree model, and the output of the early warning model includes 0 and 1, where 0 indicates that the battery to be identified has no risk of thermal runaway, that is, the vehicle is normal, and 1 indicates that the battery to be identified has a risk of thermal runaway, that is, the vehicle is abnormal.

[0119] In this embodiment of the invention, the range, standard deviation, and coefficient of variation of internal resistance are selected as vehicle feature data. A machine learning decision tree algorithm is applied to perform batch risk identification of vehicles, simplifying vehicle risk identification into a classification problem based on feature indicators, thus improving the efficiency and accuracy of thermal runaway risk identification. Furthermore, based on the early warning model, batch risk identification of vehicles is performed; vehicles with thermal runaway risk are output as 1, and vehicles without thermal runaway risk are output as 0. A vehicle early warning list is periodically generated, and corresponding maintenance is carried out to avoid thermal runaway accidents that could cause personal injury and property damage.

[0120] According to one embodiment of the present invention, the early warning model can be trained in the following manner:

[0121] Obtain a training dataset, which includes multiple sets of training charging data and the actual thermal runaway risk of the training battery corresponding to each set of training charging data; input the training dataset into a decision tree model to determine the predicted thermal runaway risk of the training battery corresponding to each set of training charging data; determine the loss value based on the predicted thermal runaway risk and the actual thermal runaway risk of the training battery; train the decision tree model based on the loss value until the training stopping condition is met to obtain a trained early warning model.

[0122] In practice, the training charging data includes charging data from normal vehicles and charging data from vehicles at risk of thermal runaway. The actual thermal runaway label includes 1 and 0, where 1 indicates a normal vehicle with no thermal runaway risk, and 0 indicates an abnormal vehicle with thermal runaway risk. After obtaining the training dataset, the training charging data can be processed according to the steps described above to determine the internal resistance value of each individual cell in the training batteries. Based on the internal resistance value, the internal resistance range, internal resistance standard deviation, and coefficient of variation are then determined to obtain the feature data of each training battery. This feature data is input into the decision tree model for processing, outputting the predicted thermal runaway risk for each training battery. The loss value is then determined based on the predicted and actual thermal runaway risks of each training battery, and the iteration count is incremented by 1. If the loss value is greater than the loss threshold and the iteration count is less than the number of iterations threshold, the decision tree model continues to be trained based on the training dataset until the training stops when the loss value is less than the loss threshold or the number of iterations is greater than the number of iterations threshold. At this point, training stops, and a well-trained warning model is obtained.

[0123] In some embodiments, a test dataset can also be constructed, which includes multiple sets of test charging data and the actual thermal runaway risk of the test battery corresponding to each set of test charging data. After the early warning model is trained, the early warning model can be tested according to the test dataset to improve the early warning accuracy of the trained early warning model.

[0124] As an example, various early warning models can be generated by training different decision tree models, and these models can be evaluated based on a test dataset. Evaluation metrics include, but are not limited to, precision and recall. The optimal early warning model can be selected from these models based on the evaluation results. This optimal model has the ability to determine whether a vehicle is at risk of thermal runaway and meets the pre-specified early warning metrics.

[0125] Furthermore, in order to improve the generalization ability of the early warning model, the number of leaf nodes in the decision model can be reduced by model compression. This not only reduces the steps for identifying thermal runaway risks and improves early warning efficiency, but also avoids the decision tree model from overfitting to the training set data.

[0126] Therefore, after training the early warning model, model pruning can be performed to obtain an updated early warning model. Correspondingly, inputting the internal resistance range, internal resistance standard deviation, and coefficient of variation into the pre-trained early warning model to determine the thermal runaway risk of the battery to be identified can specifically include: inputting the internal resistance range, internal resistance standard deviation, and coefficient of variation into the updated early warning model to determine the thermal runaway risk of the battery to be identified.

[0127] In practice, model pruning can be performed before training the decision tree model, followed by training the pruned model to obtain a trained early warning model. Alternatively, model pruning can be performed after training the early warning model to obtain an updated model. Therefore, when identifying the thermal runaway risk of a battery, the internal resistance range, internal resistance standard deviation, and coefficient of variation constructed based on the battery's charging data can be input into the updated early warning model to determine the thermal runaway risk of the battery, thereby determining whether the vehicle is abnormal and enabling appropriate measures to be taken.

[0128] The battery thermal runaway early warning method provided in this embodiment of the invention acquires charging data during the charging process of the battery to be identified, and determines the internal resistance value of each battery cell in each charging time period based on the charging data; determines the internal resistance range, internal resistance standard deviation and coefficient of variation of the battery to be identified based on the internal resistance value of each battery cell in each charging time period; and inputs the internal resistance range, internal resistance standard deviation and coefficient of variation into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified. The above method is based on battery SOC, temperature, and SOC-OCV curves calibrated at different temperatures. It uses linear interpolation to calculate the open-circuit voltage, which is simple and time-saving. Furthermore, based on the first-order RC equivalent circuit model and Kirchhoff's voltage and current laws, it uses recursive least squares method with forgetting factor to calculate the internal resistance value of each battery cell during each charging period, simplifying the internal resistance calculation process. In addition, the internal resistance range, internal resistance standard deviation, and coefficient of variation are selected as vehicle feature data. The machine learning decision tree algorithm is applied to identify the thermal runaway risk of vehicles in batches, simplifying the thermal runaway risk identification into a classification problem based on feature indicators. This enables the prediction of the thermal runaway risk of the battery to be identified, which can improve the accuracy of thermal runaway warning, reduce the risk of vehicle thermal runaway, and thus reduce personal and property losses.

[0129] Figure 3 This is a flowchart of another battery thermal runaway early warning method provided according to an embodiment of the present invention, which may include the following steps.

[0130] Step 301: Obtain charging data of the vehicle during the charging process. The charging data includes the temperature value of each temperature detection point at each sampling time point, the sampling current value corresponding to each sampling time point, the sampling voltage value of each battery cell at each sampling time point, and the remaining battery power.

[0131] Step 302: Determine the battery temperature value corresponding to each sampling time point based on the temperature values ​​of each temperature detection point at each sampling time point.

[0132] Step 303: Based on the battery temperature value at each sampling time point and the remaining battery charge of each battery cell at each sampling time point, determine the open circuit voltage value of each battery cell at each sampling time point.

[0133] Step 304: Based on the sampling voltage and open-circuit voltage values ​​of the target battery cell at the target sampling time point, the sampling voltage and open-circuit voltage values ​​of the target battery cell at the previous sampling time point, the sampling current values ​​at the target sampling time point and the previous sampling time point, the first parameter and the second parameter at the target sampling time point, and the preset parameters, determine the internal resistance value of the target battery cell at the target sampling time point. Similarly, the internal resistance value of each battery cell at each sampling time point can be determined.

[0134] Step 305: Divide the charging process into multiple charging time periods, and determine the internal resistance value of each battery cell at the last sampling time point in each charging time period as the internal resistance value of each battery cell in each charging time period.

[0135] Step 306: Determine the internal resistance range, internal resistance standard deviation, and coefficient of variation based on the internal resistance value of each battery cell during each charging period.

[0136] Step 307: Construct training and testing datasets. Train various decision tree models using the training dataset to obtain various early warning models, and test various early warning models using the testing dataset to select the optimal early warning model.

[0137] Step 308: Perform model pruning on the optimal early warning model to obtain the updated early warning model.

[0138] Step 309: Input the internal resistance range, internal resistance standard deviation, and coefficient of variation into the updated early warning model to determine the risk of thermal runaway of the vehicle.

[0139] It should be noted that any content not described in detail in this embodiment can be found in the relevant records of the above embodiments, and will not be repeated here.

[0140] The battery thermal runaway early warning method provided in this invention is based on battery SOC, temperature, and SOC-OCV curves calibrated at different temperatures. It uses linear interpolation to calculate the open-circuit voltage, which is simple and time-saving. Furthermore, based on the first-order RC equivalent circuit model and Kirchhoff's voltage and current laws, it uses recursive least squares method with forgetting factor to calculate the internal resistance value of each battery cell during each charging period, simplifying the internal resistance calculation process. In addition, it selects the internal resistance range, internal resistance standard deviation, and coefficient of variation as vehicle feature data and applies machine learning decision tree algorithm to identify the thermal runaway risk of vehicles in batches. This simplifies the thermal runaway risk identification into a classification problem based on feature indicators, realizing the prediction of the thermal runaway risk of the battery to be identified. This can improve the accuracy of thermal runaway early warning, reduce the risk of vehicle thermal runaway, and thus reduce personal and property losses.

[0141] Figure 4 This is a schematic diagram of a battery thermal runaway early warning device according to an embodiment of the present invention. The device may include:

[0142] The first determining module 401 is configured to acquire charging data during the charging process of the battery to be identified, and determine the internal resistance value of each battery cell at each sampling time point based on the charging data.

[0143] The second determining module 402 is configured to determine the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified based on the internal resistance value of each battery cell at each sampling time point.

[0144] The third determining module 403 is configured to input the internal resistance range, internal resistance standard deviation and coefficient of variation into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified.

[0145] According to one embodiment of the present invention, the charging data includes the temperature value of each temperature detection point at each sampling time point, the sampling current value corresponding to each sampling time point, the sampling voltage value of each battery cell at each sampling time point, and the remaining battery capacity.

[0146] The first determining module 401 is further configured as follows:

[0147] Based on the temperature values ​​of each temperature detection point at each sampling time point, determine the battery temperature value corresponding to each sampling time point;

[0148] Based on the battery temperature value at each sampling time point and the remaining battery charge of each battery cell at each sampling time point, determine the open circuit voltage value of each battery cell at each sampling time point.

[0149] Based on the sampling voltage and open-circuit voltage values ​​and the sampling current values ​​of each battery cell at each sampling time point, the internal resistance value of each battery cell at each sampling time point is determined.

[0150] According to one embodiment of the present invention, the first determining module 401 is further configured to:

[0151] Based on the sampling voltage and open-circuit voltage values ​​of the target battery cell at the target sampling time point, the sampling voltage and open-circuit voltage values ​​of the target battery cell at the previous sampling time point, the sampling current values ​​at the target sampling time point and the previous sampling time point, the first parameter and the second parameter at the target sampling time point, and the preset parameters, the internal resistance value of the target battery cell at the target sampling time point is determined. Here, the target battery cell is any battery cell of the battery to be identified, and the target sampling time point is any sampling time point.

[0152] According to one embodiment of the present invention, the first determining module 401 is further configured to:

[0153] φ(k)=[U(k-1)-OCV(k-1)-I(k) I(k-1)] T

[0154] Y(k)=U(k)-OCV(k)

[0155] E(k)=Y(k)-θ(k-1) T φ(k)

[0156]

[0157] Wherein, U(k) represents the sampling voltage value of the target battery cell at the target sampling time point, OCV(k) represents the open-circuit voltage value of the target battery cell at the target sampling time point, U(k-1) represents the sampling voltage value of the target battery cell at the previous sampling time point, OCV(k-1) represents the open-circuit voltage value of the target battery cell at the previous sampling time point, I(k) represents the sampling current value at the target sampling time point, I(k-1) represents the sampling current value at the previous sampling time point, θ(k-1) represents the first parameter at the target sampling time point, P(k-1) represents the second parameter at the target sampling time point, β represents the preset parameter, K(k)E(k) represents the internal resistance value of the target battery cell at the target sampling time point, φ(k) represents the third parameter at the target sampling time point, K(k) represents the fourth parameter at the target sampling time point, and E(k) represents the fifth parameter at the target sampling time point.

[0158] According to one embodiment of the present invention, the first determining module 401 is further configured to:

[0159] θ(k-1)=θ(k-2)+K(k-1)E(k-1)

[0160]

[0161] Where θ(k-1) represents the first parameter corresponding to the target sampling time point, P(k-1) represents the second parameter corresponding to the target sampling time point, θ(k-2) represents the first parameter corresponding to the previous sampling time point, P(k-2) represents the second parameter corresponding to the previous sampling time point, φ(k-1) represents the third parameter corresponding to the previous sampling time point, K(k-1) represents the fourth parameter corresponding to the previous sampling time point, E(k-1) represents the fifth parameter corresponding to the previous sampling time point, and K(k-1)E(k-1) represents the internal resistance value of the target battery cell at the previous sampling time point.

[0162] According to one embodiment of the present invention, the second determining module 402 is further configured to:

[0163] The charging process of the battery to be identified is divided into multiple charging time periods, and multiple sampling time points are determined for each charging time period.

[0164] The internal resistance value of each battery cell at the last sampling time point in each charging period is determined as the internal resistance value of each battery cell in each charging period.

[0165] Based on the internal resistance value of each battery cell within the same charging time period, determine the internal resistance range corresponding to each charging time period.

[0166] The internal resistance range of the battery to be identified is determined based on the internal resistance range corresponding to each charging period.

[0167] The standard deviation of internal resistance for each charging period is determined based on the internal resistance value of each battery cell in each charging period and the number of battery cells in each charging period.

[0168] The standard deviation of the internal resistance of the battery to be identified is determined based on the standard deviation of the internal resistance corresponding to each charging period.

[0169] The coefficient of variation for each charging time period is determined based on the standard deviation of the internal resistance for each charging time period and the average internal resistance for each charging time period.

[0170] The coefficient of variation of the battery to be identified is determined based on the coefficient of variation corresponding to each charging period.

[0171] According to one embodiment of the present invention, the device further includes a training module configured to:

[0172] Obtain the training dataset, which includes multiple sets of training charging data and the actual thermal runaway risk of the training battery corresponding to each set of training charging data;

[0173] The training dataset is input into the decision tree model to determine the predicted thermal runaway risk of the training battery corresponding to each set of training charging data.

[0174] The loss value is determined based on the predicted thermal runaway risk and the actual thermal runaway risk of the training battery;

[0175] The decision tree model is trained based on the loss value until the training stops, resulting in a well-trained early warning model.

[0176] According to one embodiment of the present invention, the third determining module 403 is further configured to:

[0177] The trained early warning model is pruned to obtain an updated early warning model.

[0178] The internal resistance range, internal resistance standard deviation, and coefficient of variation are input into the updated early warning model to determine the thermal runaway risk of the battery to be identified.

[0179] The battery thermal runaway early warning method provided in this invention acquires charging data during the charging process of the battery to be identified, and determines the internal resistance value of each battery cell at each sampling time point based on the charging data. Based on the internal resistance value of each battery cell at each sampling time point, the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified are determined. The internal resistance range, internal resistance standard deviation, and coefficient of variation are input into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified. This method determines the internal resistance value of each battery cell in the battery to be identified based on charging data, processes the internal resistance value to obtain the internal resistance range, internal resistance standard deviation, and coefficient of variation, and inputs these three parameters as features into a pre-trained early warning model. This simplifies thermal runaway risk identification into a classification problem based on feature indicators, enabling the prediction of thermal runaway risk of the battery to be identified. This improves the accuracy of thermal runaway early warning, reduces the risk of vehicle thermal runaway, and thus reduces personal injury and property damage.

[0180] The above is a schematic scheme of a battery thermal runaway early warning device according to an embodiment of the present invention. It should be noted that the technical solution of this battery thermal runaway early warning device and the technical solution of the above-described battery thermal runaway early warning method belong to the same concept. For details not described in detail in the technical solution of the battery thermal runaway early warning device, please refer to the description of the technical solution of the above-described battery thermal runaway early warning method.

[0181] Figure 5This is a schematic diagram of a computing device according to an embodiment of the present invention. The computing device 500 includes: a memory 501, a processor 502, and a computer program stored in the memory 501 and executable on the processor 502. When the processor 502 executes the computer program, it implements a battery thermal runaway early warning method as proposed in any of the above embodiments.

[0182] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a battery thermal runaway early warning method as proposed in any of the above embodiments.

[0183] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disks (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0184] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0185] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0186] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0187] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0188] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for early warning of battery thermal runaway, characterized in that, The method includes: Acquire charging data during the charging process of the battery to be identified, and determine the internal resistance value of each battery cell at each sampling time point based on the charging data; Based on the internal resistance value of each battery cell at each sampling time point, determine the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified; The internal resistance range, internal resistance standard deviation, and coefficient of variation are input into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified. The early warning model was trained in the following manner: Obtain a training dataset, wherein the training dataset includes multiple sets of training charging data and the actual thermal runaway risk of the training battery corresponding to each set of training charging data; The training dataset is input into the decision tree model to determine the predicted thermal runaway risk of the training battery corresponding to each set of training charging data. The loss value is determined based on the predicted thermal runaway risk and the actual thermal runaway risk of the training battery; The decision tree model is trained based on the loss value until the training stops, resulting in a well-trained early warning model.

2. The method according to claim 1, characterized in that, The charging data includes the temperature values ​​at each temperature detection point at each sampling time point, the sampling current values ​​corresponding to each sampling time point, the sampling voltage values ​​of each battery cell at each sampling time point, and the remaining battery power. The step of determining the internal resistance value of each battery cell at each sampling time point based on the charging data includes: Based on the temperature values ​​of each temperature detection point at each sampling time point, determine the battery temperature value corresponding to each sampling time point; Based on the battery temperature value at each sampling time point and the remaining battery charge of each battery cell at each sampling time point, determine the open circuit voltage value of each battery cell at each sampling time point. Based on the sampling voltage and open-circuit voltage values ​​and the sampling current values ​​of each battery cell at each sampling time point, the internal resistance value of each battery cell at each sampling time point is determined.

3. The method according to claim 2, characterized in that, The step of determining the internal resistance value of each battery cell at each sampling time point based on the sampling voltage value, open-circuit voltage value, and sampling current value at each sampling time point includes: Based on the sampling voltage and open-circuit voltage values ​​of the target battery cell at the target sampling time point, the sampling voltage and open-circuit voltage values ​​of the target battery cell at the previous sampling time point, the sampling current values ​​corresponding to the target sampling time point and the previous sampling time point, the first parameter and the second parameter corresponding to the target sampling time point, and preset parameters, the internal resistance value of the target battery cell at the target sampling time point is determined. The target battery cell is any battery cell of the battery to be identified, and the target sampling time point is any sampling time point.

4. The method according to claim 3, characterized in that, The internal resistance value of the target battery cell at the target sampling time point is determined by the following formula: in, This represents the sampled voltage value of the target battery cell at the target sampling time point. This represents the open-circuit voltage value of the target battery cell at the target sampling time point. This indicates the sampled voltage value of the target battery cell at the previous sampling time point. This represents the open-circuit voltage value of the target battery cell at the previous sampling time point. This represents the sampling current value corresponding to the target sampling time point. This represents the sampled current value corresponding to the previous sampling time point. This represents the first parameter corresponding to the target sampling time point. This represents the second parameter corresponding to the target sampling time point. Indicates preset parameters. This indicates the internal resistance value of the target battery cell at the target sampling time point. This represents the third parameter corresponding to the target sampling time point. This represents the fourth parameter corresponding to the target sampling time point. This represents the fifth parameter corresponding to the target sampling time point.

5. The method according to claim 4, characterized in that, The first and second parameters corresponding to the target sampling time point are determined by the following formula: in, This represents the first parameter corresponding to the target sampling time point. This represents the second parameter corresponding to the target sampling time point. This represents the first parameter corresponding to the previous sampling time point. This represents the second parameter corresponding to the previous sampling time point. This represents the third parameter corresponding to the previous sampling time point. This represents the fourth parameter corresponding to the previous sampling time point. This represents the fifth parameter corresponding to the previous sampling time point. This indicates the internal resistance value of the target battery cell at the previous sampling time point.

6. The method according to any one of claims 3-5, characterized in that, The step of determining the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified based on the internal resistance value of each individual battery cell at each sampling time point includes: The charging process of the battery to be identified is divided into multiple charging time periods, and multiple sampling time points are determined for each charging time period. The internal resistance value of each battery cell at the last sampling time point in each charging period is determined as the internal resistance value of each battery cell in each charging period. Based on the internal resistance value of each battery cell within the same charging time period, determine the internal resistance range corresponding to each charging time period. The internal resistance range of the battery to be identified is determined based on the internal resistance range corresponding to each charging time period. The standard deviation of internal resistance for each charging period is determined based on the internal resistance value of each battery cell in each charging period and the number of battery cells in each charging period. The standard deviation of the internal resistance of the battery to be identified is determined based on the standard deviation of the internal resistance corresponding to each charging time period. The coefficient of variation for each charging time period is determined based on the standard deviation of the internal resistance for each charging time period and the average internal resistance for each charging time period. The coefficient of variation of the battery to be identified is determined based on the coefficient of variation corresponding to each charging time period.

7. The method according to claim 1, characterized in that, The method further includes: The trained early warning model is pruned to obtain an updated early warning model. Accordingly, the internal resistance range, internal resistance standard deviation, and coefficient of variation are input into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified, including: The internal resistance range, internal resistance standard deviation, and coefficient of variation are input into the updated early warning model to determine the thermal runaway risk of the battery to be identified.

8. A battery thermal runaway early warning device, characterized in that, The device includes: The first determining module is configured to acquire charging data during the charging process of the battery to be identified, and determine the internal resistance value of each battery cell at each sampling time point based on the charging data. The second determining module is configured to determine the internal resistance range, internal resistance standard deviation, and coefficient of variation of the battery to be identified based on the internal resistance value of each battery cell at each sampling time point. The third determination module is configured to input the internal resistance range, internal resistance standard deviation and coefficient of variation into a pre-trained early warning model to determine the thermal runaway risk of the battery to be identified. The early warning model was trained in the following manner: Obtain a training dataset, wherein the training dataset includes multiple sets of training charging data and the actual thermal runaway risk of the training battery corresponding to each set of training charging data; The training dataset is input into the decision tree model to determine the predicted thermal runaway risk of the training battery corresponding to each set of training charging data. The loss value is determined based on the predicted thermal runaway risk and the actual thermal runaway risk of the training battery; The decision tree model is trained based on the loss value until the training stops, resulting in a well-trained early warning model.

9. A computing device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the battery thermal runaway early warning method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the battery thermal runaway early warning method as described in any one of claims 1-7.

Images