Battery leakage detection method and detection system

Abstract

The application discloses a battery leakage detection method and a detection system, and belongs to the technical field of batteries. The battery leakage detection method comprises the following steps: step 1, acquiring impedance information of the battery in real time, generating an ELS curve of the battery, and extracting leakage information of the battery based on the ELS curve; step 2, preprocessing the leakage information to decompose leakage characteristics in different frequency states; step 3, pre-training a prediction model; and step 4, inputting the leakage characteristics into the prediction model to generate a leakage probability of the battery. In the technical scheme provided in the application, when detecting whether the battery leaks, the impedance information of the battery is mainly collected. The impedance information of the battery can be directly acquired by measuring voltage and current, and the impedance information can be acquired in a simple manner and with high precision. Moreover, whether the battery leaks can be determined without obvious liquid leakage of the battery, and the safety of the battery in use can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of batteries, in particular to a battery leakage detection method and system. BACKGROUND

[0002] Lithium batteries have high voltage, high energy density and long charge-discharge life, and are an excellent electrical energy storage body. They are widely used in many fields. Due to the high energy density of lithium batteries, safety has always been an important issue in the development of lithium batteries. There is a large amount of electrolyte in lithium batteries, and when the electrolyte leaks, it is easy to cause a fire, and a single lithium battery can quickly ignite other lithium batteries, causing the lithium battery stack to explode.

[0003] In order to improve the safety performance of lithium batteries, when using lithium batteries, corresponding sensors are installed to monitor whether the electrolyte of the lithium battery leaks. However, this monitoring method has low timeliness. The liquid sensor is used to determine whether the electrolyte leaks, and a large amount of electrolyte needs to be spilled into the effective monitoring position to trigger an alarm, so the early warning capability is quite limited. When the electrolyte leak is detected, the battery has already leaked significantly and has the condition of fire.

[0004] In addition, a gas detector can also be used to detect the products generated after the electrolyte leaks, such as carbon dioxide, carbon monoxide, hydrogen, etc., to reflect whether the electrolyte has leaked. Unlike using a liquid sensor to detect whether the electrolyte has leaked, using a gas detector does not require the electrolyte to spill into the effective detection position, but still requires the amount of electrolyte leakage to reach a certain magnitude.

[0005] In summary, the above two ways of electrolyte leakage detection are limited by the amount of electrolyte leakage, and cannot timely issue an early warning when the electrolyte leaks slightly. SUMMARY

[0006] The summary portion of the present application is used to introduce the concept in a brief form, which will be described in detail in the specific embodiments section. The summary portion of the present application is not intended to identify key or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0007] As a first aspect of the present application, in order to solve the technical problems mentioned in the background section, some embodiments of the present application provide a battery leakage detection method, comprising the following steps:

[0008] Step 1: Real-time acquisition of the impedance information of the battery, generation of the EIS curve of the battery, and extraction of the leakage information of the battery based on the EIS curve;

[0009] Step 2: preprocessing the leakage information to decompose leakage features in different frequency states from the leakage information;

[0010] Step 3: pre-training a prediction model;

[0011] Step 4: inputting the leakage features into the prediction model to generate the leakage probability of the battery.

[0012] In the technical scheme provided in the application, when detecting whether the battery leaks, the impedance information of the battery is collected. The impedance information of the battery can be directly obtained by measuring voltage and current, and the impedance information is relatively simple to obtain and has high accuracy. After the battery leaks, the impedance information changes due to the loss of electrolyte. Compared with a normally aged battery, the impedance information of the battery with slow loss of electrolyte is different. Based on this, it is determined whether the battery has a leakage phenomenon. In this way, without requiring the battery to have obvious leakage, it can be determined whether the battery leaks, thereby improving the safety performance of the battery.

[0013] Step 1 includes the following steps:

[0014] Step 11: real-time monitoring of the SOC value of the battery;

[0015] Step 12: sending an excitation current signal i(t) to the battery when reaching a preset SOC measurement node, i = Isin(wt+Q i ), wherein I is a preset current value, t is time, w represents an angular frequency, and Q i is a current parameter;

[0016] Measuring the voltage across the battery to obtain a voltage response signal u(t); u(t) = Usin(wt+Q u ), Q u is a voltage parameter;

[0017] Calculating an impedance diagram according to i(t) and u(t);

[0018] Z = Z re +jZ im ;

[0019] wherein Z re is the real part of the impedance, Z im is the imaginary part of the impedance, and j is an imaginary unit;

[0020] Step 13: generating an impedance diagram of the battery with Z re as the horizontal coordinate and Z im as the vertical coordinate, and taking the impedance diagrams at different SOC values as leakage information.

[0021] In the judgment of whether the battery exists leakage by using impedance characteristics, the accuracy of impedance measurement needs to be ensured, and the accuracy of impedance measurement is related to multiple factors, in order to ensure the detection effect, the low precision data needs to be screened out, and the technical scheme provided in the application is as follows:

[0022] In step 13, the impedance diagram at each SOC value is verified, and if the validity of the impedance diagram is low, the impedance diagram at the SOC value is measured again.

[0023] The application can verify the validity of the impedance diagram, so that some inaccurate impedance information can be removed and the detection accuracy can be increased.

[0024] Further, step 13 includes the following steps:

[0025] Step 131: obtaining the impedance diagram to be verified, decomposing the impedance diagram into Z re (w) and Z im (w); im (w) is the imaginary part of the impedance when the angular frequency is w, Z re (w) is the real part of the impedance when the angular frequency is w;

[0026] Step 132: calculating the theoretical Z im `(w) based on Z re `(w);

[0027]

[0028] Calculating the theoretical Z re `(w) based on Z im `(w);

[0029]

[0030] Step 133: based on the theoretical Z im `(w) and Z im (w) error and Z re (w) and Z re `(w) error, the detection result of the impedance diagram is screened, and the impedance diagram under different frequencies is obtained.

[0031] In the technical scheme provided in the application, the internal relationship between the imaginary part and the real part of the impedance, that is, the causal response between the impedance spectrum and the current signal, and the specific relationship between the imaginary part and the real part in the impedance spectrum are used, so that whether the measured impedance is accurate can be verified, and the accuracy of the impedance information collection is increased.

[0032] The impedance information mainly reflects the capacitance characteristics of the battery, i.e. the change of impedance at different frequencies, and reflects the impedance characteristics of the battery at high frequencies. In practice, the change state of the electrochemical reaction inside the battery cannot be reflected, so the characteristics of the electrochemical change of the battery after leakage cannot be reflected.

[0033] The impedance information of the battery changes during the aging process, although the degree of change of the impedance information when the electrolyte of the battery leaks is inconsistent, there is a high similarity, and the pure impedance information cannot accurately reflect the dynamics of the battery, especially when the impedance information at different frequencies is superimposed together, the dynamics of the battery is lost when analyzing the impedance information.

[0034] Step 2 includes the following steps:

[0035] Step 21: Obtain the leakage information of the battery, and for each impedance graph in the leakage information, extract the original features of the impedance graph;

[0036] The original features include the initial resistance R1, the arc vertex resistance R2, the inflection point resistance R3, and the low-frequency fitting straight line f(re);

[0037] The initial resistance R1 is the point where the imaginary part starts to rise in the impedance graph, the arc vertex resistance R2 is the point where the imaginary part starts to drop in the impedance graph, the inflection point resistance R3 is the inflection point where the imaginary part starts to rise for the second time in the impedance graph, and the low-frequency fitting straight line f(re) is the fitting straight line of the subsequent part of the inflection point resistance in the impedance graph.

[0038] Step 22: For the impedance graph at different frequencies for each SOC value, extract the frequency feature Z(w);

[0039]

[0040] Where w is the angular frequency, t is the relaxation time, j is the imaginary unit, R1 is the initial resistance, g(τ) is the relaxation time distribution function, and lg is the logarithm with base 10;

[0041] Step 23: Extract the peak points V1, V2, V3, V4 from Z(w);

[0042] Step 24: V1, V2, V3, V4, initial resistance R1, arc vertex resistance R2, inflection point resistance R3, and low-frequency fitting straight line f(re) are taken as leakage features.

[0043] In the technical solution provided by the present application, in order to allow the leakage characteristics to accurately reflect the electrochemical characteristics inside the battery, the impedance information in the frequency domain is converted into a relaxation time distribution in the time domain. This conversion process is based on the assumption that the electrochemical model is a series connection of an ohmic resistance and a polarization impedance, and the polarization impedance is expressed as a continuous series RC parallel circuit. A single ideal polarization process is generally expressed by a parallel resistor and capacitor, corresponding to (relaxation time) τ = RC. In practice, the capacitor C is generally represented by a constant phase angle element, which is represented by a relaxation time distribution function g(τ) centered on the characteristic time constant τ in the relaxation time distribution diagram. This modeling analysis can effectively separate the overlapping processes of the battery dynamics process in the frequency spectrum through time domain conversion and the solution of the relaxation time distribution function, thereby more accurately reflecting the characteristics of the battery dynamics process.

[0044] During the use of batteries, multiple batteries are generally combined together to be used as a battery pack, so in practice, it is necessary to ensure the accuracy of battery prediction. To this end, in practice, in order to ensure safety, it is preferred to increase the prediction threshold to avoid omissions. When the neural network model predicts whether the battery is leaking, the internal feature transmission process is basically not visible, so when it is used, it will lead to the inability to trust the model's detection results too accurately. To this end, this application provides the following technical solutions:

[0045] Furthermore, step 3 includes the following steps:

[0046] Step 31: Obtain a set of training samples P in advance, each training sample P includes a leakage feature X and a label Y, and the label y includes leakage and non-leakage; construct the original problem;

[0047] y i (∈x i +b)>1, or y i (∈x i +b)<-1; i=1, 2…n, n is an integer, n>1, i represents the sample index, ∈ is the normal vector of the hyperplane, b is the intercept of the hyperplane, x i is the leakage feature of the i-th training sample, y i is the set of corresponding category labels;

[0048] Step 32: Introduce the Lagrangian function:

[0049] i and j represent the index of the training sample, α, α i , α j They represent Lagrange multipliers, L represents Lagrange function, and K represents kernel function;

[0050] Step 33: Solve the dual problem and find the optimal Lagrange multiplier αi ;

[0051]

[0052] max α W(α) represents the value of maximizing the objective function W(α) with respect to the Lagrange multiplier α i W(α) is the objective function;

[0053] satisfy the following constraint conditions:

[0054]

[0055] K(x i , x j ) is a kernel function for calculating the inner product between samples, and C represents a penalty function;

[0056]

[0057] u i is a slack variable for indicating the degree of each sample allowed to violate the interval constraint;

[0058] Step 34: continuously adjust the Lagrange multiplier α i until the verification accuracy reaches the maximum value.

[0059] In the technical solution provided in the present application, the SVM model is used to design the prediction model. Compared with the neural network model, the decision boundary of the SVM model is easier to understand and explain, so it is easier to adjust the key factors of the SVM model decision in practice, and the SVM model decision mode can be adjusted in the required direction to eliminate the missed detection problem as much as possible. And compared with other prediction models, the decision boundary of the SVM model is less affected by abnormal points and noise points, and the model is more stable.

[0060] When training the SVM model, the Lagrange multiplier needs to be continuously adjusted, and each adjustment of the Lagrange multiplier requires multiple training. Therefore, when training the SVM model, the training difficulty is very large, and because the number of features introduced in the present application is too large, the training difficulty is large. Therefore, in the present application, the following technical solution is adopted:

[0061] Further, step 34 includes the following steps:

[0062] Step 341: taking all value ranges of the Lagrange multiplier as a search space d;

[0063] Set the particle swarm size N;

[0064] Initialize the particle position G h ; h represents the index of the particle;

[0065] Set the maximum number of iterations T;

[0066] At the tth iteration, the position vector of the particle is

[0067] Step 342: continuously iterate, for each iteration:

[0068] Calculate the convergence factor Coefficient vector Swing coefficient vector

[0069]

[0070] For each particle G h , calculate the fitness value f(G h ), sort all current particles according to the fitness value to obtain the optimal particle L1, the suboptimal particle L2 and the third particle L3;

[0071] Select the position of the current optimal solution as the target position, and calculate the distance vector between the target position and each particle

[0072]

[0073] Update the particle position:

[0074]

[0075] Wherein is the coefficient vector, t is the current iteration number, represents the current target position, represents the current particle position, represents the distance between the particle and the target position;

[0076] Step 343: after updating for several times:

[0077] For each particle, respectively calculate the distance vector between it and the current optimal particle L1, the suboptimal particle L2 and the third particle L3

[0078] Update the temporary position of the particle according to the distance coefficient;

[0079]

[0080] Update the temporary position of the particle according to the distance and coefficient vector

[0081]

[0082] wherein, is the distance between the optimal particle L1, the sub-optimal particle L2 and the third particle L3 and other particles, is the position of other particles, respectively represent the current position of the optimal particle L1, the sub-optimal particle L2 and the third particle L3, and is the corresponding coefficient vector, represents the position of the current particle;

[0083]

[0084] Step 343: the iteration ends when the maximum iteration number T is reached, or the target function value is lower than the preset value.

[0085] In the scheme, the positions of the optimal particle L1, the sub-optimal particle L2 and the third particle L3 are introduced as target positions, and when optimizing in the entire problem dimension space d, the remaining examples are guided to constantly approach the target positions, so in the entire optimization process, the global perspective can be taken into account, the balance between the local and the global during optimization is avoided, and the precision and convergence ability of the model are increased.

[0086] When the particle swarm is used to find the optimal scheme, there is a certain luck factor in iterating to the best scheme. Because the iteration is mainly determined by the coefficient vector , and the coefficient vector is determined by the convergence factor . The convergence factor changes linearly when converging. Therefore, the entire optimization process is actually a linear process, which may result in low convergence ability of the model during optimization. Therefore, the present application provides the following technical scheme:

[0087] Further,

[0088] In the scheme, the convergence factor is modified from linear decrease to nonlinear decrease (exponential decrease), which can increase the diversity of the convergence factor change. Therefore, the convergence speed of the convergence factor is slow in the early iteration, and the global optimal solution is used to search at this time. With the iteration, the convergence factor decreases faster, and the local optimal solution can be better found at this time. Therefore, in the scheme, the convergence factor has sufficient performance when searching for the local optimal solution and the global optimal solution.

[0089] As a second aspect of the present application, the present application provides a battery leakage detection system comprising:

[0090] An information collection device is used to collect impedance information of each battery.

[0091] A prediction device predicts a leakage probability of the battery according to the battery leakage detection method. BRIEF DESCRIPTION OF DRAWINGS

[0092] The accompanying drawings, which form a part of this application, are included to provide a further understanding of the application and are incorporated in and constitute a part of this application. The illustrations, together with the description, serve to explain the application, but are not intended to limit the application.

[0093] In addition, throughout the drawings, same or similar reference numerals are used to denote same or similar elements. It should be understood that the drawings are schematic and elements and features are not necessarily to scale.

[0094] In the drawings:

[0095] Figure 1 A flow chart of the battery leakage detection method.

[0096] Figure 2 An impedance spectrum diagram.

[0097] Figure 3 A peak position diagram of Z(w).

[0098] Figure 4 A In linear and exponential decrease. DETAILED DESCRIPTION

[0099] Embodiments of the present application will be described in more detail with reference to the drawings. Although certain embodiments of the present application are shown in the drawings, it is understood that the present application can be embodied in various forms and should not be construed as being limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present application will be thoroughly and completely understood. It should be understood that the drawings and embodiments of the present application are for illustrative purposes only and are not intended to limit the scope of the present application.

[0100] It should also be noted that only parts related to the present application are shown in the drawings for the convenience of description. The embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0101] The present application will be described in detail with reference to the drawings and embodiments.

[0102] Referring to Figure 1 A battery leakage detection method, comprising the following steps:

[0103] Step 1: Real-time acquisition of impedance information of the battery, generation of EIS curve of the battery, and extraction of leakage information of the battery based on the EIS curve.

[0104] Step 11: Real-time monitoring the SOC value of the battery;

[0105] The SOC value of the battery can be calculated by detecting the battery inflow in the battery, and the detection of SOC is the prior art which will not be discussed here.

[0106] Step 12: When reaching the preset SOC measurement node, send the excitation current signal i(t) to the battery, i = Isin(wt+Q i ), wherein I is a preset current value, t is time, w represents angular frequency, and Q i is a current parameter;

[0107] The voltage response signal u(t) is obtained by measuring the voltage across the battery; u(t) = Usin(wt+Q u ), Q u is a voltage parameter;

[0108] The impedance Z is calculated according to i(t) and u(t);

[0109] Z = Z re +jZ im ;

[0110] Wherein Z re is the real part of the impedance, Z im is the imaginary part of the impedance, and j is the imaginary unit.

[0111] Step 13: Taking Z re as the horizontal coordinate and Z im as the vertical coordinate to generate the impedance diagram of the battery, and taking the impedance diagrams under different SOC values as leakage information. As shown in Figure 2 .

[0112] In this scheme, the measurement of the impedance is related to the frequency of the current. Specifically, when measuring the impedance diagram, not one impedance information under a certain SOC value is obtained, but multiple impedance information under multiple SOC values.

[0113] First, the battery is pre-set with different SOC values, such as 10%, 30%, 50%, 60%, and 90%, during a charging or discharging process, and when the battery reaches these SOC values, impedance information acquisition will be performed. And at each time of acquiring impedance information, the excitation current signal i(t) generates multiple angular frequencies, so the final obtained impedance information will include the SOC under the frequency of different excitation current signals i(t) under multiple SOC values.

[0114] Step 13 includes the following steps:

[0115] Step 131: Obtain the impedance diagram to be verified, and decompose the impedance diagram into Z re (w) and Z im(w), Z im (w) is the imaginary part of the impedance at the angular frequency w, Z re (w) is the real part of the impedance at the angular frequency w;

[0116] Step 132: based on Z im (w) to obtain theoretical Z re (w) is calculated;

[0117]

[0118] Based on Z re (w) to obtain theoretical Z im (w) is calculated;

[0119]

[0120] Step 133: based on theoretical Z im (w) and Z im (w) error and Z re (w) and Z re (w) error screening impedance map detection results, obtain the impedance map under different frequencies.

[0121] Step 133 is mainly to screen the impedance map, mainly using the specific relationship between the real part and the imaginary part in the impedance map. If it meets the relationship, it means that the impedance calculation is accurate, otherwise it is not accurate.

[0122] Step 2: preprocessing the leakage information to decompose the leakage characteristics under different frequency states.

[0123] There are many curves in the leakage information, and the battery records multiple SOC value nodes for each charging. In addition, the impedance is also measured by multiple current frequencies for each node, so there are multiple information curves. Such data is very complex, and the application provides the following scheme:

[0124] Step 2 includes the following steps:

[0125] Step 21: obtain the leakage information of the battery, and extract the original features of the impedance map from each impedance map in the leakage information;

[0126] The original features include the starting resistance R1, the arc vertex resistance R2, the inflection point resistance R3, and the low-frequency fitting straight line f(re);

[0127] Among them, the starting resistance R1 is the point where the imaginary part of the impedance diagram starts to rise, the arc vertex resistance R2 is the point where the imaginary part of the impedance diagram starts to decrease, the inflection point resistance R3 is the inflection point where the imaginary part of the impedance diagram starts to rise for the second time, and the low-frequency fitting straight line f(re) is the fitting straight line of the subsequent part of the inflection point resistance in the impedance diagram; f(re) is a fitting straight line, and its fixed item information has been determined by the inflection point, and its slope is mainly recorded during analysis.

[0128] Step 22: For each impedance graph at different frequencies under each SOC value, extract the frequency feature Z(w);

[0129]

[0130] Where w is the angular frequency, t is the relaxation time, j is the imaginary unit, R1 is the starting resistance, g(τ) is the relaxation time distribution function, and lg is the logarithm to base 10;

[0131] Step 23: Extract peak points V1, V2, V3, and V4 from Z(w).

[0132] like Figure 3 As shown, there are 4 specific peak points in Z(w). The positions and peak values ​​of the peak points are different in different Z(w), but there are 4 peak points in each Z(w).

[0133] Step 24: Take V1, V2, V3, V4, the starting resistance R1, the arc top resistance R2, the inflection point resistance R3 and the low-frequency fitting straight line f(re) as leakage characteristics.

[0134] In this scheme, the features of the impedance information are extracted, and the impedance information is frequency-converted. The impedance change characteristics at different frequencies are compared, and the changes of the impedance information at different levels are further extracted.

[0135] Step 3: Pre-train the prediction model.

[0136] When predicting whether there is leakage, this solution uses the SVM model for prediction. Specifically:

[0137] Step 31: Obtain a set of training samples P in advance, each training sample P includes a leakage feature X and a label Y, and the label y includes leakage and non-leakage; construct the original problem;

[0138] y i (∈x i +b)>1, or y i (∈x i +b)<-1; i=1, 2…n, n is an integer, n>1, i represents the sample index, ∈ is the normal vector of the hyperplane, b is the intercept of the hyperplane, x iis the leakage feature of the i-th training sample, y i is the set of corresponding category labels;

[0139] Step 32: Introduce the Lagrangian function:

[0140] i and j represent the index of the training sample, α, α i , α j They represent Lagrange multipliers, L represents Lagrange function, and K represents kernel function;

[0141] Step 33: Solve the dual problem and find the optimal Lagrange multiplier α i ;

[0142]

[0143] max α W(α) represents the maximization of the objective function W(α) with respect to the Lagrange multiplier α i The value of , W(α) is the objective function;

[0144] The following constraints are met:

[0145]

[0146] K(x i , x j ) is the kernel function, which is used to calculate the inner product between samples, and C represents the penalty function;

[0147]

[0148] u i is a slack variable that indicates the extent to which each sample is allowed to violate the interval constraint;

[0149] Step 34: Continuously adjust the Lagrange multiplier α i Until the verification accuracy reaches the maximum value.

[0150] Step 34 includes the following steps:

[0151] Step 341: taking the entire range of Lagrange multipliers as the search space d;

[0152] Set the particle swarm size N;

[0153] Initialize particle position G h ; h represents the index of the particle;

[0154] Set the maximum number of iterations T;

[0155] At the tth iteration, the position vector of the particle is

[0156] Step 342: continuously iterate, for each iteration:

[0157] Convergence factor is calculated Coefficient vector Swing coefficient vector

[0158]

[0159] For each particle G h , the fitness value f(G h ) is calculated, and all current particles are sorted according to the fitness value to obtain the optimal particle L1, the suboptimal particle L2 and the third particle L3;

[0160] The position of the current optimal solution is selected as the target position, and the distance vector between the target position and each particle is calculated

[0161]

[0162] Update the particle position:

[0163]

[0164] Wherein is the coefficient vector, t is the current iteration number, represents the current target position, represents the current particle position, represents the distance between the particle and the target position;

[0165] Step 343: after updating for several times:

[0166] For each particle, the distance vector between it and the current optimal particle L1, the suboptimal particle L2 and the third particle L3 is calculated respectively

[0167] According to the distance coefficient, the temporary position of the particle is updated;

[0168]

[0169] According to the distance and coefficient vector, the temporary position of the particle is updated

[0170]

[0171] Wherein, is the distance between the optimal particle L1, the suboptimal particle L2 and the third particle L3 and other particles, the position of other particles, respectively represent the current position of the optimal particle L1, the suboptimal particle L2 and the third particle L3, and are corresponding coefficient vectors, represent the position of the current particle;

[0172]

[0173] Step 343: the iteration ends when the maximum number of iterations T is reached, or the value of the objective function is lower than the preset value.

[0174] In some embodiments,

[0175] Specifically, in the foregoing fact mode, In the iteration, it belongs to linear change, while in the present scheme, As the iteration proceeds, the word number changes, and when the word number changes, The change rate is small, which can facilitate global consideration for optimization to avoid falling into a local optimal solution, while in the subsequent iteration, The change rate is large, which can better consider locally and increase the convergence rate. As Figure 4 shown.

[0176] Step 4: input the leakage feature into the prediction model to generate the leakage probability of the battery.

[0177] In step 3, the optimization process of the prediction model is provided, and in step 2, the preprocessing method of the leakage feature is provided, so in practice, only the leakage feature needs to be collected and input into the prediction model, and the leakage probability of the battery can be generated.

[0178] Embodiment 2: A battery leakage detection system includes an information collection device and a prediction device. The information collection device is used to collect impedance information of each battery. The prediction device predicts the leakage probability of the battery based on the foregoing battery leakage detection method.

[0179] The information collection device is mainly used to obtain the impedance information of the battery. Generally, in order to ensure that the impedance information can be obtained, a current generator and a voltage detector are needed. The current generator is used to generate an alternating current signal, and the voltage detector is used to detect the voltage across the battery. The prediction device is built-in with a prediction model, which adopts the method in embodiment 1 to detect the electrolyte leakage.

[0180] The above description is merely exemplary of some preferred embodiments and applications of the application. It is to be understood that the application is not limited to those specific embodiments disclosed, and that modifications and variations are possible in light of the above teachings and the principles of the application disclosed herein. For example, the above described features and applications can be combined in any combination. It is therefore contemplated to cover any and all modifications and variations of the specific embodiments disclosed above and to include them within the scope of the present application.

Claims

1. A method of detecting a battery leakage, the method comprising: The method comprises the following steps: Step 1: Real-time acquisition of impedance information of the battery, generation of an EIS curve of the battery, and extraction of leakage information of the battery based on the EIS curve; Step 2: Preprocessing of the leakage information to decompose leakage characteristics under different frequencies from the leakage information; Step 3: Pre-training of a prediction model; Step 4: Inputting the leakage characteristics into the prediction model to generate a leakage probability of the battery; Step 1 comprises the following steps: Step 11: Real-time monitoring of the SOC value of the battery; Step 12: send an excitation current signal i(t) to the battery when reaching a preset SOC measurement node, i = Isin(wt+Q i ), wherein I is a preset current value, t is time, w represents an angular frequency, and Q i is a current parameter; The voltage response signal u(t) is obtained by measuring the voltage across the battery; u(t) = Usin(wt + Q u ), Q u is the voltage parameter; Impedance Z is calculated according to i(t) and u(t); Z = Z re + jZ im ; wherein Z re is the real part of the impedance, Z im is the imaginary part of the impedance, and j is the imaginary unit; Step 13: generate the impedance graph of the battery with Z re as the horizontal coordinate and Z im as the vertical coordinate, and take the impedance graph at different SOC values as the leakage information; In step 13, the effectiveness of the impedance diagram under each SOC value is verified, and if the effectiveness of the impedance diagram is low, the impedance diagram under the SOC value is re-measured; Step 2 comprises the following steps: Step 21: Acquisition of the leakage information of the battery, and extraction of original characteristics of the impedance diagram from each impedance diagram in the leakage information; The original characteristics include a starting resistance R1, an arc vertex resistance R2, a turning point resistance R3, and a low-frequency fitting straight line f(re); The starting resistance R1 is a point where the imaginary part starts to rise in the impedance diagram, the arc vertex resistance R2 is a point where the imaginary part starts to fall in the impedance diagram, the turning point resistance R3 is a turning point where the imaginary part starts to rise for the second time in the impedance diagram, and the low-frequency fitting straight line f(re) is a fitting straight line of the subsequent part of the impedance diagram after the turning point resistance; Step 22: Extraction of frequency characteristics Z(w) from the impedance diagram under different frequencies for each SOC value; Step 23: Extraction of peak points V1, V2, V3, and V4 from Z(w); Step 24: Taking V1, V2, V3, V4, the starting resistance R1, the arc vertex resistance R2, the turning point resistance R3, and the low-frequency fitting straight line f(re) as the leakage characteristics.

2. The battery liquid leakage detection method of claim 1, wherein: Step 3 comprises the following steps: Step 31: Pre-acquisition of a set of training samples P, each training sample P comprising leakage characteristics X and a label Y, the label y comprising leakage and no leakage; and construction of an original problem; y i (∈x i +b)>1, or y i (∈x i +b)<-1; i=1, 2…n, n is an integer, n>1, i represents the sample index, ∈ is the normal vector of the hyperplane, b is the intercept of the hyperplane, x i is the leakage feature of the i-th training sample, y i is the set of corresponding category labels; Step 32: Introduction of a Lagrange function: i and j represent indices of training samples, a, a i , a j represent Lagrange multipliers, respectively, L represents a Lagrangian function, and K represents a kernel function. Step 33: Solve the dual problem to find the optimal Lagrange multiplier a i ; max α W(a) denotes maximizing the objective function W(a) with respect to the Lagrange multiplier a i W(a) is the objective function; The following constraint conditions are met: 0 < a i ≤ C, i = 1...n, K(x i , x j ) is a kernel function for calculating inner products between samples, and C represents a penalty function. u i is a slack variable used to represent the extent to which each sample is allowed to violate the interval constraint; Step 34: Continuously adjust the Lagrange multiplier a i until the accuracy of the verification reaches a maximum value.

3. The battery liquid leakage detection method of claim 2, wherein: Step 34 comprises the following steps: Step 341: Taking all value ranges of the Lagrange multiplier as a search space d; The size of the particle swarm N is set; Initialize particle position G h ; h denotes the index of the particle; The maximum number of iterations T is set; At the tth iteration, the position vector of a particle is Step 342: Continuously iterating, for each iteration: Computing convergence factor Coefficient vector Wobble coefficient vector For each particle G h , a fitness value f(G h ) is calculated, all current particles are sorted according to the fitness value, and the optimal particle L1, the suboptimal particle L2 and the third particle L3 are obtained. selecting the position of the current best solution as a target position, calculating a distance vector between the target position and each particle Update the particle position: wherein is a coefficient vector, t is the current iteration number, denotes the current target position, denotes the current particle position, denotes the distance between the particle and the target position; Step 343: After being updated for several times: For each particle, the distance vector between it and the current best particle L1, the second best particle L2 and the third particle L3 is calculated respectively Update the temporary position of the particle according to the distance coefficient; updating the temporary position of the particle according to the distance and the coefficient vector wherein is the distance between the optimal particle L1, the suboptimal particle L2 and the third particle L3 and other particles, is the position of other particles, denote the current position of the optimal particle L1, the suboptimal particle L2 and the third particle L3, respectively, and is the corresponding coefficient vector, denotes the position of the current particle; Step 343: When the maximum number of iterations T is reached, the iteration ends, and the iteration ends when the value of the objective function is lower than the preset value.

4. The battery liquid leakage detection method of claim 3, wherein:

5. A battery liquid leakage detection system characterized by: The method comprises the following steps: An information collection device is used to collect impedance information of each battery; A prediction device is used to generate a leakage probability of the battery according to the battery leakage detection method in any one of claims 1-4.

Images