A battery internal short circuit detection method based on sliding time window

Abstract

The application relates to a battery internal short circuit detection method based on a sliding time window, which comprises the following steps: S1, collecting voltage and time data of each single battery of a power battery; S2, defining a sliding time window; S3, based on the voltage and time data of each single battery of the battery, respectively calculating the voltage Hausdorff distance of each single battery in each sliding time window and carrying out variational mode decomposition signal processing; and S4, determining whether the power battery has an internal short circuit and positioning the single battery where the internal short circuit occurs based on the result of the variational mode decomposition signal processing. Compared with the prior art, the application has the advantages of accurate short circuit battery detection and the like.

Description

Technical Field

[0001] This invention relates to the field of energy storage battery technology, and in particular to a method for detecting internal short circuits in batteries based on a sliding time window. Background Technology

[0002] Lithium-ion batteries are widely used in electric vehicles due to their high energy density, long cycle life, and environmental friendliness. However, lithium batteries have potential safety issues, the most common of which is thermal runaway, which seriously threatens the personal and property safety of consumers. Research shows that more than half of thermal runaway accidents are related to internal short circuits within the battery. Early internal short circuits are characterized by inconspicuous electrothermal features and slow development. If universally applicable detection methods can be used to identify internal short circuit characteristics in their early stages, it can be prevented from continuing to evolve into thermal runaway or even an explosion.

[0003] Currently, researchers have proposed various methods for internal short-circuit detection. For example, methods based on abnormal self-discharge and remaining charge capacity have the problem that identifying abnormal self-discharge takes a long time, and the remaining charge capacity can lead to significant errors due to battery balancing. Furthermore, model-based methods can transform the internal short-circuit detection problem into a parameter and state estimation problem, but updating model parameters and maintaining detection accuracy under complex operating conditions is quite challenging. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art by providing an accurate short-circuit detection method for batteries based on a sliding time window.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A method for detecting short circuits within a battery based on a sliding time window, comprising the following steps:

[0007] S1. Collect voltage and time data of each individual cell in the power battery;

[0008] S2. Define the sliding time window;

[0009] S3. Based on the voltage and time data of each individual cell, calculate the voltage Hausdorff distance of each individual cell in each sliding time window and perform variational mode decomposition signal processing.

[0010] S4. Based on the results of variational mode decomposition signal processing, determine whether the power battery has an internal short circuit and locate the individual battery cell that has an internal short circuit.

[0011] Furthermore, the specific steps of S3 are as follows:

[0012] S31. Based on the voltage and time data of each individual cell in the battery, calculate the voltage Hausdorff distance of each individual cell within each sliding time window.

[0013] S32. Perform variational mode decomposition signal processing on the voltage Hausdorff distance obtained in S31.

[0014] The specific calculation process for the voltage Hausdorff distance is as follows:

[0015] Calculate the median voltage of the individual cell corresponding to the l-th sampling point within the i-th sliding time window, and record all the median voltages of the individual cells in the voltage median set. Calculate the voltage of the r-th individual cell at the l-th sampling point within the i-th sliding time window, and record all the voltages in the individual cell voltage set. Calculate the Hausdorff distance between the individual cell voltage set and the voltage median set.

[0016] Furthermore, the expression for Hausdorff distance is:

[0017]

[0018] Wherein, h(U) i,r U i,m ) for U i,r To U i,m The maximum value of the nearest distance between the trajectories, h(U) i,m U i,r ) for U i,m To U i,r The maximum value of the nearest distance between the trajectories, U i,r For the collection of individual cell voltages, U i,m For the set of voltage medians, Let be the Hausdorff distance of the r-th individual cell within the i-th sliding time window.

[0019] Furthermore, U i,r To U i,m The maximum value of the nearest distance between the trajectories and U i,m To U i,r The expressions for the maximum distance between the nearest points between the trajectories are as follows:

[0020]

[0021]

[0022] Among them, u l,r Let u be the voltage of the r-th single cell at the l-th sampling point within the i-th sliding time window. l,m This represents the median cell voltage corresponding to the l-th sampling point within the i-th sliding time window.

[0023] Furthermore, in S32, the specific steps of variational mode decomposition signal processing are as follows:

[0024] Using the voltage Hausdorff distance of the r-th cell as the original signal, the original signal is decomposed into k components. These k components are the result of variational mode decomposition signal processing. The decomposition needs to satisfy constraint variation, and the expression of the constraint variation is as follows:

[0025]

[0026]

[0027] Where k is the number of patterns to be decomposed, {v k}、{ω k} represents the k-th component and its corresponding center frequency after decomposition, δ(t) is the Dirac function, * is the convolution operator, and f is the original signal.

[0028] Furthermore, when solving the constrained variational problem, a Lagrange multiplicative operator is introduced. The constrained variational problem is transformed into an unconstrained optimization problem through an augmented Lagrange function. This unconstrained optimization problem is then solved using an alternating direction multiplier to obtain k components.

[0029] Furthermore, the expression for the augmented Lagrange function is:

[0030]

[0031] Where L is the augmented Lagrange function, k is the number of patterns to be decomposed, and {v k}、{ω k} represents the k-th component and its corresponding center frequency after decomposition, α is the quadratic penalty factor, δ(t) is the Dirac function, * is the convolution operator, f is the original signal, and λ is the Lagrange multiplicative operator.

[0032] Furthermore, the specific steps of S4 are as follows:

[0033] S41. Select one or more IMF components from the k components and set a threshold.

[0034] S42. Determine whether the IMF component of the voltage Hausdorff distance of the r-th cell within the i-th sliding time window is greater than the threshold. If yes, determine that the r-th cell is short-circuited within the sliding time window. If no, determine that the r-th cell is not short-circuited within the sliding time window. After the i-th sliding time window is determined, continue to determine the IMF components of other sliding time windows until all windows are determined, and obtain the final detection result.

[0035] Furthermore, the threshold is 0.03.

[0036] Furthermore, in S2, the expression for the sliding time window is defined as follows:

[0037] T i =[((i-1)n+1)Δt,i·n·Δt]

[0038] Where n represents n sampling points, and Δt is the sampling time interval.

[0039] Compared with the prior art, the present invention has the following advantages:

[0040] When an internal short circuit occurs in a single battery cell at a certain moment, self-discharge occurs inside the cell, and the voltage of the short-circuited cell becomes increasingly lower compared to the voltages of other normal cells. Therefore, this invention calculates the Hausdorff distance of the battery voltage within a sliding time window and performs variational mode decomposition signal processing. If a short circuit occurs in a single battery cell, the Hausdorff distance of the battery voltage within the sliding time window can reflect the change in voltage between the short-circuited cell and other normal cells. This effectively amplifies the early electrochemical characteristics of internally short-circuited cells, improves the accuracy of internally short-circuited cell detection, and is simple and easy to implement. Attached Figure Description

[0041] Figure 1 This is a flowchart of the present invention;

[0042] Figure 2 The voltage-time curves of each individual cell in this invention are shown.

[0043] Figure 3 The voltage-Hausdorff distance-time curves of each individual cell in this invention are shown.

[0044] Figure 4 The voltage Hausdorff distance component (IMF1-time curve) of each individual cell in this invention is shown below.

[0045] Figure 5 The voltage Hausdorff distance component (IMF2)-time curves of each individual cell in this invention are shown. Detailed Implementation

[0046] 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, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0047] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0048] Example 1

[0049] This invention proposes a battery internal short-circuit detection method based on a sliding time window, the flowchart of which is as follows: Figure 1 As shown. The specific steps of the method include:

[0050] S1. Collect voltage and time data of each individual cell in the power battery;

[0051] S2. Define the sliding time window;

[0052] S3. Based on the voltage and time data of each individual cell, calculate the voltage Hausdorff distance of each individual cell in each sliding time window and perform variational mode decomposition signal processing.

[0053] S4. Based on the results of variational mode decomposition signal processing, determine whether the power battery has an internal short circuit and locate the individual battery cell that has an internal short circuit.

[0054] Step S3 includes:

[0055] S31. Calculate the voltage Hausdorff distance of each individual cell in each sliding time window.

[0056] S32. Perform variational mode decomposition signal processing on the voltage Hausdorff distance of each individual cell.

[0057] The voltage Hausdorff distance is obtained as follows:

[0058] First, calculate the median cell voltage u corresponding to the l-th sampling point within the i-th sliding time window. l,m (l = 1, 2, 3, ..., n) and recorded in set U i,m In the above, the voltage u of the r-th single cell at the l-th sampling time point within the i-th sliding time window is... l,r Recorded in set U i,r Then, calculate the voltage U of the r-th single cell within the i-th sliding time window. i,r With the median voltage U i,m Hausdorf Distance

[0059] The voltage Hausdorff distance is determined by the following formula:

[0060]

[0061]

[0062]

[0063] In the formula, h(U) i,r U i,m U is the voltage U of the r-th individual cell within the i-th sliding time window. i,r to the median voltage U i,m The maximum distance between the nearest points of the trajectories, similarly, h(U) i,m U i,r ) represents the median voltage U within the i-th sliding time window. i,m The voltage U up to the r-th cell i,r The maximum value of the nearest distance between the trajectories.

[0064] Variational mode decomposition is achieved through the following formula:

[0065] Assuming the original signal f is decomposed into k components, the corresponding constrained variational expression is as follows:

[0066]

[0067]

[0068] Where k is the number of patterns to be decomposed, {v k}、{ω k Let} represent the k-th component and its corresponding center frequency after decomposition, δ(t) be the Dirac function, and * be the convolution operator. Introducing the Lagrange multiplicative operator λ, and using the augmented Lagrange function, the above equation-constrained optimization problem is transformed into an unconstrained optimization problem, yielding the augmented Lagrange expression as follows:

[0069]

[0070] Here, α is the quadratic penalty factor. This constrained problem can be solved using an alternating direction multiplier, with the following results:

[0071]

[0072]

[0073]

[0074] Where γ is the noise margin. and They are v q Fourier transforms of f(t), f(t), and λ(t).

[0075] Here, the original signal f is the set of Hausdorff distances for the voltage of the r-th individual cell.

[0076] The specific steps for S4 are as follows:

[0077] S41. Select one or more IMF components from the k components and set a threshold.

[0078] S42. Determine whether the IMF component of the voltage Hausdorff distance of the r-th cell within the i-th sliding time window is greater than the threshold. If yes, determine that the r-th cell is short-circuited within the sliding time window. If no, determine that the r-th cell is not short-circuited within the sliding time window. After the i-th sliding time window is determined, continue to determine the IMF components of other sliding time windows until all windows are determined, and obtain the final detection result.

[0079] The following example will be used to verify this:

[0080] The voltage-time curves of the module consisting of four individual cells in this embodiment over 15 cycles are as follows: Figure 2 As shown, starting from the 6th cycle, an external 1000Ω resistor is connected to simulate an internal short circuit in the battery.

[0081] S2, Define the sliding time window.

[0082] In this example, the experimental sampling time interval is 0.1s, and the sliding time window contains 5000 sampling points, meaning the sliding time window length is 500s.

[0083] S3. Calculate the voltage Hausdorff distance of each individual cell in each sliding time window and perform variational mode decomposition signal processing.

[0084] S31. Calculate the voltage Hausdorff distance of each individual cell within each sliding time window.

[0085] Because the individual cells in a power battery pack are connected in series, and the current flowing through each cell is the same, the voltage of each cell should be consistent for a healthy battery. When a cell experiences an internal short circuit at a certain moment, self-discharge occurs inside the battery, causing the voltage of that internally short-circuited cell to be lower than that of other healthy cells at the same moment. As time increases, this leakage accumulates, causing the voltage of the internally short-circuited cell to become increasingly lower than that of other normal cells. The more severe the short circuit, the more pronounced this phenomenon. Therefore, within the i-th sliding time window, if the r-th cell experiences an internal short circuit, the distance between its voltage change trajectory and the median voltage change trajectory within that window is much larger than the distance between the voltage change trajectories of other normal cells and the median voltage change trajectory. The Hausdorff distance effectively reflects the distance change between the two trajectories.

[0086] The voltage-time curves of each individual cell in this embodiment are as follows: Figure 2 As shown, the voltage-Hausdorff distance-time curves of each individual cell are as follows: Figure 3 As shown. It should be noted that: Figure 2 The voltage-time curves of the four cells basically overlap, with slight deviations.

[0087] S32. Perform variational mode decomposition signal processing on the voltage Hausdorff distance of each individual cell.

[0088] In this embodiment, the number of signal components is selected as k=2, and the voltage Hausdorff distance component IMF1-time curve of each individual cell is as follows: Figure 4 As shown, the component IMF2-time curve is as follows: Figure 5 As shown;

[0089] S4. Based on the signal processing results, determine whether the power battery has an internal short circuit and locate the individual battery cell that has an internal short circuit.

[0090] In this embodiment, the threshold for determining internal short circuits is set to τ = 0.03. Figure 5 As shown, it can be seen that at 137500s (≈38.2h), that is, the 275th sliding time window, the single cell 2 was diagnosed to have an internal short circuit.

[0091] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for detecting short circuits within a battery based on a sliding time window, characterized in that, The method includes the following steps: S1. Collect voltage and time data of each individual cell in the power battery; S2. Define the sliding time window; S3. Based on the voltage and time data of each individual cell, calculate the voltage Hausdorff distance of each individual cell in each sliding time window and perform variational mode decomposition signal processing. S4. Based on the results of variational mode decomposition signal processing, determine whether the power battery has an internal short circuit and locate the individual battery cells that have an internal short circuit. The specific steps for S3 are as follows: S31. Based on the voltage and time data of each individual cell in the battery, calculate the voltage Hausdorff distance of each individual cell within each sliding time window. S32. Perform variational mode decomposition signal processing on the voltage Hausdorff distance obtained in S31. The specific calculation process for the voltage Hausdorff distance is as follows: Calculate the first i The first sliding time window within the l The median voltage of each individual cell at the sampling point is recorded in the voltage median set. The calculation is then performed to determine the median voltage of the sampling point. i Within the first sliding time window r The single cell in the first l The voltage of each sampling point is recorded in the individual cell voltage set, and the Hausdorff distance between the individual cell voltage set and the voltage median set is calculated. The expression for Hausdorff distance is: in, For U i,r To U i,m The maximum value of the nearest distance between the trajectories. For U i,m To U i,r The maximum value of the nearest distance between the trajectories, U i,r For the collection of individual cell voltages, U i,m For the set of voltage medians, For the first i Within the first sliding time window r Hausdorf distance of a single cell; U i,r To U i,m The maximum value of the nearest distance between the trajectories and U i,m To U i,r The expressions for the maximum distance between the nearest points between the trajectories are as follows: in, u l,r For the first i Within the first sliding time window r The single cell in the first l The voltage at each sampling point u l,m For the first i The first sliding time window within the l The median voltage of a single cell corresponding to each sampling point.

2. The battery internal short-circuit detection method based on a sliding time window according to claim 1, characterized in that, In S32, the specific steps of variational mode decomposition signal processing are as follows: The first r The voltage Hausdorff distance of each individual cell is used as the original signal, and the original signal is decomposed into... k One portion, k Each component is the result of variational mode decomposition signal processing. During decomposition, constraint variations must be satisfied, and the expression for these constraint variations is: in, k The number of patterns to be decomposed, { v k }、{ ω k } is the first k Each component and its corresponding decomposed center frequency, δ ( t () represents the Dirac function, and * represents the convolution operator. f This is the original signal.

3. The battery internal short-circuit detection method based on a sliding time window according to claim 2, characterized in that, When solving the constrained variational problem, a Lagrange multiplicative operator is introduced. The constrained variational problem is transformed into an unconstrained optimization problem using an augmented Lagrange function. This unconstrained optimization problem is then solved using an alternating direction multiplier, yielding the desired result. k Each component.

4. The battery internal short-circuit detection method based on a sliding time window according to claim 3, characterized in that, The expression for the augmented Lagrange function is: in, L To augment the Lagrange function, k The number of patterns to be decomposed, { v k }、{ ω k } is the first k Each component and its corresponding decomposed center frequency, α It is a secondary penalty factor. δ ( t () represents the Dirac function, and * represents the convolution operator. f Let λ be the original signal and λ be the Lagrange multiplicative operator.

5. The battery internal short-circuit detection method based on a sliding time window according to claim 2, characterized in that, The specific steps for S4 are as follows: S41, in k Select one or more IMF components from the available components and set a threshold. S42, Determine the first i Within the first sliding time window r If the IMF component of the voltage Hausdorff distance of each individual cell is greater than the threshold, then determine the first... r If a single cell experiences a short circuit within the sliding time window, then determine the first... r No single cell experienced a short circuit within the sliding time window. i After the first sliding time window is judged, the IMF components of other sliding time windows are judged until all windows are judged, and the final detection result is obtained.

6. The battery internal short-circuit detection method based on a sliding time window according to claim 5, characterized in that, The threshold is 0.

03.

7. The battery internal short-circuit detection method based on a sliding time window according to claim 1, characterized in that, In S2, the expression for the sliding time window is defined as follows: in, n express n sampling points, Δ t This represents the sampling time interval.

Images