Lithium battery internal short circuit diagnosis method and system based on impedance spectrum analysis

Abstract

The application discloses a lithium battery internal short circuit diagnosis method and system based on impedance spectrum analysis, and relates to the technical field of lithium battery short circuit diagnosis. The method comprises the following steps: constructing a standard impedance spectrum library based on the identification information of a target battery; performing electrochemical impedance spectrum testing on the target battery to obtain a monitoring impedance spectrum; mapping the monitoring impedance spectrum to the standard impedance spectrum library to identify the spectrum interval of the monitoring impedance spectrum; calculating the offset of the upper-limit impedance spectrum and the lower-limit impedance spectrum; obtaining a virtual impedance spectrum; fusing the virtual impedance spectrum and the monitoring impedance spectrum, extracting an impedance feature set, and performing short circuit identification to obtain a short circuit diagnosis result. The application solves the technical problem that the lithium battery internal short circuit diagnosis in the prior art is affected by temperature and SOC and is difficult to accurately determine, and achieves the technical effects of accurately identifying the lithium battery internal short circuit and improving the accuracy and reliability of the lithium battery internal short circuit diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery short circuit diagnosis, and in particular to a lithium battery internal short circuit diagnosis method and system based on impedance spectroscopy analysis. Background Art

[0002] Lithium batteries are widely used in modern electronic devices and electric vehicles, and their safety and reliability are of paramount importance. Internal short circuits are a serious potential malfunction during the operation of lithium batteries, which may cause problems such as battery performance degradation, overheating, and even fire and explosion. In existing technologies, temperature and SOC (state of charge) both have a significant impact on the impedance spectrum of lithium batteries. During the battery charging and discharging process, the temperature of the battery cells is difficult to accurately measure and define, which makes it complicated to diagnose internal short circuits in lithium batteries based on impedance spectrum analysis. At the same time, traditional diagnostic methods simply rely on setting fixed thresholds to determine short circuits. Due to the differences in the ohmic resistance of each battery in the battery pack, false alarms or missed alarms are prone to occur, resulting in the inability to accurately identify internal short circuits in the battery, making it difficult to meet the demand for efficient and accurate diagnosis of lithium batteries.

[0003] The existing technology has a technical problem that the diagnosis of internal short circuit of lithium batteries is affected by temperature and SOC, making it difficult to accurately judge. Summary of the Invention

[0004] The present application provides a lithium battery internal short circuit diagnosis method and system based on impedance spectroscopy analysis, which is used to solve the technical problem in the prior art that the internal short circuit diagnosis of lithium batteries is affected by temperature and SOC and is difficult to accurately judge.

[0005] In view of the above problems, the present application provides a lithium battery internal short circuit diagnosis method and system based on impedance spectroscopy analysis.

[0006] In a first aspect of the present application, a method for diagnosing an internal short circuit of a lithium battery based on impedance spectroscopy analysis is provided, the method comprising:

[0007] construct a standard impedance spectrum library based on the identification information of the target battery, wherein the standard impedance spectrum library comprises standard impedance spectra under each temperature-SOC combination; perform electrochemical impedance spectrum testing on the target battery in combination with the standard impedance spectrum library to obtain a monitoring impedance spectrum; map the monitoring impedance spectrum to the standard impedance spectrum library, identify a spectrum interval of the monitoring impedance spectrum, and correspondingly obtain an upper-limit impedance spectrum and a lower-limit impedance spectrum; calculate a shift amount of the monitoring impedance spectrum relative to the upper-limit impedance spectrum and the lower-limit impedance spectrum, and locate a relative spectrum position according to the shift amount; calculate and obtain a virtual impedance spectrum based on an interpolation method in combination with the relative spectrum position, the upper-limit impedance spectrum, and the lower-limit impedance spectrum; fuse the virtual impedance spectrum and the monitoring impedance spectrum, extract an impedance feature set, and perform short-circuit discrimination according to the impedance feature set to obtain a short-circuit diagnosis result, wherein the short-circuit diagnosis result comprises a short-circuit discrimination result and a short-circuit level label.

[0008] In a second aspect of the present application, a lithium battery internal short-circuit diagnosis system based on impedance spectrum analysis is provided, and the system comprises:

[0009] a standard impedance spectrum library construction module configured to construct a standard impedance spectrum library based on the identification information of the target battery, wherein the standard impedance spectrum library comprises standard impedance spectra under each temperature-SOC combination; a monitoring impedance spectrum acquisition module configured to perform electrochemical impedance spectrum testing on the target battery in combination with the standard impedance spectrum library to obtain a monitoring impedance spectrum; a monitoring impedance spectrum mapping module configured to map the monitoring impedance spectrum to the standard impedance spectrum library, identify a spectrum interval of the monitoring impedance spectrum, and correspondingly obtain an upper-limit impedance spectrum and a lower-limit impedance spectrum; a shift amount calculation module configured to calculate a shift amount of the monitoring impedance spectrum relative to the upper-limit impedance spectrum and the lower-limit impedance spectrum, and locate a relative spectrum position according to the shift amount; a virtual impedance spectrum acquisition module configured to calculate and obtain a virtual impedance spectrum based on an interpolation method in combination with the relative spectrum position, the upper-limit impedance spectrum, and the lower-limit impedance spectrum; and a short-circuit diagnosis result acquisition module configured to fuse the virtual impedance spectrum and the monitoring impedance spectrum, extract an impedance feature set, and perform short-circuit discrimination according to the impedance feature set to obtain a short-circuit diagnosis result, wherein the short-circuit diagnosis result comprises a short-circuit discrimination result and a short-circuit level label.

[0010] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0011] Construct a standard impedance spectrum library based on the identification information of the target battery; perform electrochemical impedance spectrum testing on the target battery to obtain a monitoring impedance spectrum; map the monitoring impedance spectrum to the standard impedance spectrum library, identify the spectrum interval of the monitoring impedance spectrum, and correspondingly obtain an upper limit impedance spectrum and a lower limit impedance spectrum; calculate the offset of the monitoring impedance spectrum relative to the upper limit impedance spectrum and the lower limit impedance spectrum, and locate the relative spectrum position according to the offset; based on an interpolation method, combine the relative spectrum position, the upper limit impedance spectrum and the lower limit impedance spectrum to calculate and obtain a virtual impedance spectrum; fuse the virtual impedance spectrum and the monitoring impedance spectrum, extract an impedance feature set, and perform short circuit discrimination according to the impedance feature set to obtain a short circuit diagnosis result. The technical effect of realizing accurate discrimination of internal short circuit of a lithium battery and improving the accuracy and reliability of internal short circuit diagnosis of the lithium battery is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0013] Figure 1 A flowchart of a lithium battery internal short circuit diagnosis method based on impedance spectrum analysis is provided for the embodiments of the present application.

[0014] Figure 2 A structure diagram of a lithium battery internal short circuit diagnosis system based on impedance spectrum analysis is provided for the embodiments of the present application.

[0015] Explanation of reference signs: standard impedance spectrum library construction module 10, monitoring impedance spectrum acquisition module 20, monitoring impedance spectrum mapping module 30, offset calculation module 40, virtual impedance spectrum acquisition module 50, and short circuit diagnosis result acquisition module 60. DETAILED DESCRIPTION

[0016] The present application provides a lithium battery internal short circuit diagnosis method and system based on impedance spectrum analysis, which is used to solve the technical problem that the existing lithium battery internal short circuit diagnosis is affected by temperature and SOC, and it is difficult to accurately determine.

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0018] Embodiment one, as shown in Figure 1 The application provides a lithium battery internal short circuit diagnosis method based on impedance spectrum analysis, which comprises the following steps:

[0019] Step S100: Based on the identification information of the target battery, a standard impedance spectrum library is constructed, wherein the standard impedance spectrum library comprises standard impedance spectra under different temperature-SOC combinations.

[0020] Specifically, starting from the identification information carried by the target battery, this information can uniquely determine the battery, ensuring that the standard impedance spectrum library constructed subsequently corresponds to a specific battery. Then, the key parameter range required for constructing the standard impedance spectrum library is determined, i.e., the appropriate temperature range and SOC (State of Charge) range are determined. According to the orthogonal experimental design principle, a series of representative temperature-SOC combination point sets are planned within the selected temperature and SOC ranges. For each combination point, the target battery is subjected to multiple electrochemical impedance spectrum tests, which can obtain impedance data of the battery under different states. Subsequently, the multiple test results under the same combination point are fused to remove errors and disturbances that may exist in the test process, thereby obtaining more accurate and reliable standard impedance spectrum data. Finally, the fused standard impedance spectrum data are classified and stored according to different temperature-SOC combinations, and a standard impedance spectrum library containing standard impedance spectra under different temperature-SOC combinations is constructed. This standard impedance spectrum library covers the normal impedance characteristics of the target battery under various temperatures and SOC states, and provides a key reference for subsequent battery state monitoring and internal short circuit diagnosis, which is an important basis for the smooth development of the entire diagnosis process.

[0021] Step S200: Based on the standard impedance spectrum library, the target battery is subjected to electrochemical impedance spectrum test to obtain a monitoring impedance spectrum.

[0022] Specifically, by means of the standard impedance spectrum library that has been constructed, a comprehensive and accurate impedance spectrum test is carried out on the target battery to obtain the monitoring impedance spectrum. First, statistical analysis method is used to deeply analyze the standard impedance spectrum library, and the confidence frequency interval is selected from a large number of historical data. The determination of this interval can provide a reliable reference range for subsequent tests. Then, according to the confidence frequency interval, an electrochemical test device is used to apply an alternating excitation signal to the target battery electrode, and the signal frequency covers a range from low frequency to high frequency, thereby simulating the running state of the battery under different working conditions. At the same time of applying the excitation signal, high-precision measuring instruments are used to synchronously collect the voltage response data of the target battery, which contains the electrochemical information inside the battery. Subsequently, according to the collected voltage response data and the known excitation signal parameters, the complex impedance data of the battery are calculated, which can more comprehensively reflect the impedance characteristics inside the battery. Finally, the calculated complex impedance data are arranged and analyzed, and the monitoring impedance spectrum is established according to certain rules and models, which intuitively presents the current impedance state of the target battery, and provides a key basis for subsequent judgment of whether the battery has internal short circuit and other abnormal conditions.

[0023] Step S300: mapping the monitoring impedance spectrum to the standard impedance spectrum library, identifying the spectrum interval of the monitoring impedance spectrum, and correspondingly obtaining the upper limit impedance spectrum and the lower limit impedance spectrum.

[0024] Specifically, the SOC marker of the monitoring impedance spectrum is first extracted, which records the current state of charge information of the battery. Then, taking the SOC marker as a clue, a comprehensive search is carried out in the standard impedance spectrum library to select the data corresponding to the SOC marker and construct a standard impedance spectrum sub-library. Then, a special impedance spectrum coordinate system is constructed, which can accurately reflect the characteristics of the impedance spectrum in different dimensions. The constructed standard impedance spectrum sub-library is mapped into this coordinate system, and multiple different spectrum intervals are divided according to the data distribution. Then, the monitoring impedance spectrum is also mapped into this impedance spectrum coordinate system, and the specific spectrum interval to which it belongs is determined by comparing its position in the coordinate system. Once the spectrum interval is determined, the upper limit part of the boundary data in this interval is extracted as the upper limit impedance spectrum, and the lower limit part is extracted as the lower limit impedance spectrum. This step provides a key data basis for subsequent in-depth analysis of the difference between the monitoring impedance spectrum and the standard impedance spectrum, and calculation of the virtual impedance spectrum, which is crucial for accurately judging whether the lithium battery has internal short circuit.

[0025] Step S400: calculating the offset of the monitoring impedance spectrum relative to the upper limit impedance spectrum and the lower limit impedance spectrum, and positioning the relative spectrum position according to the offset.

[0026] Specifically, the difference between the monitoring impedance spectrum and the standard impedance spectrum is analyzed in depth to more accurately locate the battery state. First, representative key frequency points in the impedance spectrum are selected as a comparison benchmark. These key frequency points are determined based on research on the electrochemical properties of the battery and a large amount of experimental data, and they can sensitively reflect changes in the internal state of the battery. After determining the benchmark points, the monitoring impedance spectrum is compared with the upper and lower limit impedance spectra at each key frequency point, and the offset difference value of each key point is calculated. These differences reflect the degree of deviation of the monitoring impedance spectrum from the upper and lower boundaries of the standard impedance spectrum at different key frequencies. Then, according to the importance of each key frequency point to the overall state of the battery, each key point difference is given a corresponding weight, and the overall offset is calculated by weighting. The upper offset of the monitoring impedance spectrum relative to the upper limit impedance spectrum and the lower offset relative to the lower limit impedance spectrum are obtained. Finally, the calculated upper and lower offsets are normalized to convert them to values between 0 and 1. Through these two normalized offsets, the relative position of the monitoring impedance spectrum within the corresponding spectrum interval can be accurately determined, and its specific state within the standard impedance spectrum range can be understood. This provides an accurate basis for subsequent calculation of the virtual impedance spectrum and further improves the accuracy of lithium battery internal short circuit diagnosis.

[0027] Step S500: Based on the interpolation method, the relative spectrum position, the upper limit impedance spectrum, and the lower limit impedance spectrum are combined to calculate and obtain the virtual impedance spectrum.

[0028] Specifically, since the actual temperature of the on-site battery is difficult to accurately determine, it is difficult to directly obtain a standard impedance spectrum that perfectly matches the actual state. Therefore, a specific method is needed to construct a more practical reference spectrum, i.e., the virtual impedance spectrum. First, the relative spectrum position information of the monitoring impedance spectrum relative to the upper and lower limit impedance spectra is used, combined with the data of the upper and lower limit impedance spectra. Then, a suitable interpolation algorithm is selected, such as linear interpolation. Taking linear interpolation as an example, according to the relative spectrum position, the impedance values of each frequency point within the interval formed by the upper and lower limit impedance spectra are calculated according to the linear proportional relationship. Assuming that the relative spectrum position is x, the impedance value of the upper limit impedance spectrum at a certain frequency point is Z upper , and the impedance value of the lower limit impedance spectrum at the frequency point is Z lower , then the virtual impedance value Z virtual of the frequency point is calculated by linear interpolation as Z lower = (1-x) Z upperThe virtual impedance spectrum can be used as a more accurate reference spectrum to effectively correct deviations caused by temperature or SOC estimation errors and other factors, and provide a reference basis for subsequent extraction of reliable impedance features from the fusion of the virtual impedance spectrum and the monitored impedance spectrum, and then accurate short circuit discrimination.

[0029] Step S600: fuse the virtual impedance spectrum and the monitored impedance spectrum, extract an impedance feature set, and perform short circuit discrimination according to the impedance feature set to obtain a short circuit diagnosis result, wherein the short circuit diagnosis result includes a short circuit discrimination result and a short circuit level label.

[0030] Specifically, the virtual impedance spectrum and the monitored impedance spectrum are fused, which requires first analyzing a large number of sample short circuit data to determine a key feature index set, and labeling the sensitivity of each feature index. Based on the key feature index set, feature values are extracted from the virtual and monitored impedance spectra respectively to generate first and second impedance feature sets, the difference between the two is calculated to form an impedance feature set for short circuit discrimination. Then, short circuit discrimination is performed, a battery historical impedance feature sequence is obtained, combined with the current impedance feature set, a trend analysis method is used to calculate the impedance feature change trend to obtain a trend development rate. A preset short circuit discrimination constraint is determined by statistical analysis method, and the trend development rate is compared with the constraint. If the trend development rate is greater than or equal to the constraint, it is determined that the battery has a short circuit, and the short circuit level label is the ratio of the trend development rate to the short circuit discrimination constraint. In addition, a short circuit diagnosis model based on machine learning is used. The model has a virtual impedance spectrum correction channel and a difference analysis mapping channel. The target battery cycle log is obtained from the battery management end, the virtual impedance spectrum and the cycle log are input into the virtual impedance spectrum correction channel, and the reference impedance spectrum is obtained by optimization. The monitored impedance spectrum and the reference impedance spectrum are input into the difference analysis mapping channel to obtain a reference impedance feature set, which is compared with the previous impedance feature set to obtain a short circuit discrimination result. If it is determined that there is a short circuit, the reference impedance feature set is mapped to a short circuit level label, and finally a comprehensive and accurate short circuit diagnosis result is obtained, which provides strong support for the safety management of lithium batteries.

[0031] In one possible implementation, step S100 further includes:

[0032] Step S110: determine a temperature range and a SOC range, and define a temperature-SOC combination point set according to the orthogonal experimental design principle.

[0033] Step S120: perform a preset number of electrochemical impedance spectrum tests on the target battery at each combination point, and fuse the test results to obtain standard impedance spectrum data.

[0034] Step S130: The obtained standard impedance spectrum data is classified and stored according to the temperature-SOC combination, and a standard impedance spectrum library is constructed.

[0035] Specifically, the actual application scenarios and performance characteristics of the lithium battery are comprehensively considered to determine the required temperature range and SOC (state of charge) range of the research. For example, if the lithium battery is mainly applied to an electric vehicle, considering the operation of the vehicle at different ambient temperatures and the change of SOC during the charging and discharging of the battery, the temperature range can be set to -30°C to 50°C, and the SOC range can be set to 5% to 95%. After the range is determined, the orthogonal experimental design principle is used, which can obtain more comprehensive and representative data with fewer experimental times. Through the orthogonal table, a series of typical temperature values and SOC values are selected within the given temperature and SOC range for combination. For example, -30°C, 0°C, 25°C, and 50°C are selected as temperature points from the temperature range, and 5%, 30%, 60%, and 95% are selected as SOC points from the SOC range. These temperature points and SOC points are combined with each other to form a set of temperature-SOC combination points, which cover a variety of states that the battery may encounter, laying a foundation for subsequent acquisition of comprehensive and valuable impedance spectrum data.

[0036] The target battery is subjected to a preset number of electrochemical impedance spectrum tests at each combination point, and standard impedance spectrum data is obtained by fusing the test results. For each determined temperature-SOC combination point, the target battery is subjected to a preset number (e.g., 5 times) of electrochemical impedance spectrum tests using an electrochemical test device. During each test, an AC excitation signal of different frequency is applied to the battery, the voltage response of the battery is measured, and the corresponding impedance data is obtained. Since there may be some errors in the test process, to obtain more accurate and reliable data, the test results at the same combination point are fused. The average method can be used to integrate the data of multiple tests to obtain standard impedance spectrum data that can more accurately reflect the real impedance characteristics of the battery at the combination point.

[0037] The obtained standard impedance spectrum data is classified and stored according to the temperature-SOC combination, and a standard impedance spectrum library is constructed. The standard impedance spectrum data obtained by fusion processing at each combination point is classified and arranged according to its corresponding temperature-SOC combination, a database is established, temperature and SOC are used as indexes, the standard impedance spectrum data of each combination point is accurately stored in the corresponding position, and a complete standard impedance spectrum library is constructed. This standard impedance spectrum library contains the standard impedance spectrum of the target battery under different temperature and SOC combinations, which provides an important reference benchmark for subsequent real-time monitoring and accurate diagnosis of internal short circuit conditions of the target battery.

[0038] In one possible implementation, step S200 further includes:

[0039] Step S210: Based on a statistical analysis method, statistical analysis is performed on the standard impedance spectrum library to determine a confidence frequency interval.

[0040] Step S220: According to the confidence frequency interval, an alternating excitation signal covering low frequency to high frequency is applied to the target battery electrode, and the voltage response of the target battery is measured to obtain complex impedance data of the battery.

[0041] Step S230: According to the complex impedance data, the monitoring impedance spectrum is established.

[0042] Specifically, since the standard impedance spectrum library stores a large amount of impedance spectrum data at different temperature-SOC combinations, in order to more accurately and reliably test the target battery subsequently, statistical analysis method is used to process these data to determine the confidence frequency interval. First, the impedance data at each frequency is extracted from the standard impedance spectrum library, and statistical quantities such as mean, variance, and frequency distribution of the data are calculated by using statistical principles. Taking the calculation of the mean value as an example, the average level of the impedance at the frequency is obtained by summing all the impedance data at the same frequency and dividing by the number of data. Then, according to the pre-set confidence level, such as the commonly used 95% confidence level, combined with the distribution of the data, the normal distribution or other appropriate distribution model is used to determine the confidence frequency interval. If the data approximately obeys the normal distribution, the frequency boundary value at a certain confidence level can be calculated according to the characteristics of the normal distribution, and a frequency range is determined, which is the confidence frequency interval. The data in this interval can better represent the impedance characteristics of the battery under normal state, providing a scientific and reasonable frequency reference range for subsequent application of alternating excitation signal to the target battery, which helps to improve the accuracy of state monitoring of the target battery.

[0043] According to the determined confidence frequency interval, an alternating excitation signal covering low frequency to high frequency is applied to the target battery electrode, and the voltage response of the target battery is measured to obtain complex impedance data of the battery. Using electrochemical test equipment, an alternating excitation signal is applied to the target battery electrode according to the established confidence frequency interval, and the signal frequency gradually changes from low frequency to high frequency, such as from 0.01 Hz to 10 kHz. At the same time of applying the alternating excitation signal, a high-precision voltage measuring instrument is used to measure the voltage response of the target battery synchronously. Since the battery will exhibit different electrochemical characteristics under different frequency alternating excitation, its voltage response will also be different. According to Ohm's law and complex operation rules, the measured voltage response data is combined with the known alternating excitation signal parameters to calculate the complex impedance data of the battery at different frequencies. These complex impedance data contain comprehensive information such as battery resistance, capacitance and inductance, reflecting the electrochemical state of the battery.

[0044] According to the obtained complex impedance data, a monitoring impedance spectrum is established. The complex impedance data calculated at different frequencies is sorted and analyzed, and arranged in order from small to large frequency. With frequency as the horizontal coordinate, the real part and the imaginary part of the complex impedance as the vertical coordinate, the data points are plotted in the coordinate system, and a curve is constructed through curve fitting, which is a curve that can accurately reflect the current state of the target battery. The monitoring impedance spectrum intuitively shows the change of the impedance characteristics of the target battery under the current test conditions with frequency, which provides key data basis for subsequent judgment of whether the battery has internal short circuit and other abnormal conditions.

[0045] In one possible implementation, step S300 further includes:

[0046] Step S310: Extracting the SOC marker of the monitoring impedance spectrum.

[0047] Step S320: Based on the SOC marker, traversing the standard impedance spectrum library to construct a standard impedance spectrum sub-library.

[0048] Step S330: Constructing an impedance spectrum coordinate system, and mapping the standard impedance spectrum sub-library to the impedance spectrum coordinate system to obtain a plurality of spectrum intervals.

[0049] Step S340: Mapping the monitoring impedance spectrum to the impedance spectrum coordinate system, extracting the upper limit of the interval as the upper limit impedance spectrum and the lower limit of the interval as the lower limit impedance spectrum according to the spectrum interval to which it belongs.

[0050] Specifically, since the monitoring impedance spectrum is a reflection of the current state of the target battery, and the SOC (state of charge) marker contains the current power information of the battery, which is crucial for subsequent analysis. First, according to the data storage format of the monitoring impedance spectrum and the relevant technical specifications, the storage location or encoding method of the SOC marker in the data is determined. If the monitoring impedance spectrum is stored in a specific file format, such as CSV format, there is a clear column title indicating the column where the SOC data is located; if it is a binary data format, the SOC marker data segment needs to be located according to the pre-set parsing rules. After finding the SOC marker data, the data reading algorithm is used to accurately extract the stored value. If the SOC marker is stored in a certain encoding form, decoding operation is also needed to convert it into an intuitive state of charge value that can be used for subsequent processing. This extracted SOC marker represents the state of charge of the target battery when the monitoring impedance spectrum is obtained, and is the key basis for subsequent construction of the standard impedance spectrum sub-library and further analysis of the relationship between the monitoring impedance spectrum and the standard impedance spectrum. It can ensure that the subsequent analysis focuses on the standard data matching the current state of charge of the battery, and improves the accuracy and pertinence of the analysis.

[0051] Based on the extracted SOC label, a standard impedance spectrum sub-library is constructed by traversing the standard impedance spectrum library. The standard impedance spectrum library contains a large number of standard impedance spectrum data under different temperature-SOC combinations. With the SOC label as the index, the corresponding data is searched one by one in the standard impedance spectrum library. All standard impedance spectrum data meeting the SOC label are filtered out and re-integrated to form a special standard impedance spectrum sub-library, providing a more targeted data set for subsequent accurate analysis.

[0052] An impedance spectrum coordinate system is constructed, and the standard impedance spectrum sub-library is mapped to the coordinate system to obtain a plurality of spectrum intervals. In order to more intuitively and effectively analyze the impedance spectrum data, a suitable impedance spectrum coordinate system is constructed. The coordinate axes of the coordinate system are set according to the key characteristic parameters of the impedance spectrum, for example, the frequency is taken as the horizontal axis, and the real and imaginary parts of the impedance are taken as the vertical axes. All data in the constructed standard impedance spectrum sub-library are mapped one by one into this coordinate system according to their characteristics in the frequency, impedance real part and imaginary part dimensions. According to the distribution of the data in the coordinate system, the clustering analysis method is used to divide the entire data region into a plurality of different spectrum intervals. These spectrum intervals are an effective way to classify the standard impedance spectrum sub-library data, which helps to more carefully analyze the position and characteristics of the monitoring impedance spectrum within the standard spectrum range.

[0053] The monitoring impedance spectrum is mapped to the impedance spectrum coordinate system, and the upper limit of the interval is extracted as the upper limit impedance spectrum and the lower limit of the interval is extracted as the lower limit impedance spectrum according to the spectrum interval to which it belongs. The monitoring impedance spectrum obtained previously is mapped into the impedance spectrum coordinate system constructed according to the same rule. By comparing the position of the monitoring impedance spectrum in the coordinate system with the range of each spectrum interval, the specific spectrum interval to which it belongs is determined. After finding the spectrum interval, the upper limit part is extracted from the boundary data of the interval, which constitutes the upper limit impedance spectrum; the lower limit part is extracted to form the lower limit impedance spectrum. The upper limit impedance spectrum and the lower limit impedance spectrum respectively represent the two boundary states of the spectrum interval where the monitoring impedance spectrum is located, providing a key reference for subsequent accurate calculation of the shift amount of the monitoring impedance spectrum and judgment of the battery state, which helps to more accurately diagnose whether the lithium battery has internal short circuit and other problems.

[0054] In one possible implementation, step S400 further includes:

[0055] Step S410: selecting representative key frequency points in the impedance spectrum as comparison reference.

[0056] Step S420: comparing the impedance values of the monitoring impedance spectrum, the upper limit impedance spectrum and the lower limit impedance spectrum at each key frequency point, and calculating the shift difference value of each key point.

[0057] Step S430: weighting the difference values of each key point, calculating the overall shift amount, and outputting the upper shift amount and the lower shift amount.

[0058] Step S440: Normalize the upper and lower offsets to determine the relative position of the monitoring impedance spectrum in the corresponding spectrum interval.

[0059] Specifically, representative key frequency points in the impedance spectrum are selected as the comparison reference. Since the impedance spectrum contains numerous frequency point data, not all frequency points have the same sensitivity to the analysis of the internal state change of the battery. Through in-depth research on the electrochemical characteristics of lithium batteries, combined with a large amount of experimental data and theoretical analysis, those frequency points that can sensitively reflect the change of the internal state of the battery are selected as key frequency points. For example, in the impedance spectrum of a lithium battery, some frequency points in the low frequency band are more sensitive to the charge transfer process in the battery, and some frequency points in the high frequency band are closely related to the ohmic resistance characteristics of the battery. Selecting these representative key frequency points lays the foundation for subsequent accurate comparison of the differences between different impedance spectra.

[0060] The impedance values of the monitoring impedance spectrum, the upper limit impedance spectrum and the lower limit impedance spectrum at each key frequency point are compared, and the offset difference value of each key point is calculated. The impedance values of the monitoring impedance spectrum at each selected key frequency point are compared with the impedance values of the upper limit impedance spectrum and the lower limit impedance spectrum at the same key frequency point, respectively. For each key frequency point, the impedance difference between the monitoring impedance spectrum and the upper limit impedance spectrum, and the impedance difference between the monitoring impedance spectrum and the lower limit impedance spectrum are calculated by subtraction operation. These differences can intuitively reflect the deviation of the monitoring impedance spectrum at each key frequency point compared with the upper and lower boundaries of the standard impedance spectrum, providing detailed data support for subsequent evaluation of the overall offset of the monitoring impedance spectrum.

[0061] The differences of each key point are weighted, the overall offset is calculated, and the upper and lower offsets are output. Considering that different key frequency points have different effects on the overall state of the battery, each key point difference is given a corresponding weight. The determination of the weight is based on the understanding of the physical and chemical processes inside the battery and the analysis of a large amount of experimental data. The difference of each key frequency point is multiplied by the corresponding weight, and all the product results are added to obtain the overall offset of the monitoring impedance spectrum relative to the upper limit impedance spectrum, i.e. the upper offset, and the overall offset relative to the lower limit impedance spectrum, i.e. the lower offset. These two offsets comprehensively reflect the deviation of the monitoring impedance spectrum at all key frequency points relative to the upper and lower boundaries of the standard impedance spectrum, and can more comprehensively measure the difference between the monitoring impedance spectrum and the standard impedance spectrum.

[0062] The normalized upper and lower offsets determine the relative position of the monitoring impedance spectrum in the corresponding spectrum interval. Since the numerical size of the upper and lower offsets can be affected by factors such as measurement units and data magnitudes, in order to accurately compare them on a unified scale, the two offsets need to be normalized. Through the normalization algorithm, the upper and lower offsets are converted to values between 0 and 1. According to the normalized upper and lower offsets, the relative position of the monitoring impedance spectrum in its corresponding spectrum interval can be determined. For example, if the normalized upper offset is 0.3 and the lower offset is 0.2, it indicates that the monitoring impedance spectrum is closer to the lower limit impedance spectrum and is relatively close to the lower limit in the spectrum interval. The determination of this relative position is of great significance for further analyzing the state of the battery and judging whether there is an internal short circuit or other abnormal conditions, and can provide key basis for health assessment and fault diagnosis of lithium batteries.

[0063] In one possible implementation, step S500 further includes:

[0064] Step S510: Calculate the coefficient of variation of the virtual impedance spectrum and the monitoring impedance spectrum.

[0065] Step S520: If the coefficient of variation meets the preset threshold, output the virtual impedance spectrum.

[0066] Step S530: If the coefficient of variation does not meet the preset threshold, perform iterative updating of the virtual impedance spectrum based on the over-limit direction of the coefficient of variation, and recalculate the corresponding coefficient of variation until the preset threshold is met.

[0067] Specifically, the coefficient of variation of the virtual impedance spectrum and the monitoring impedance spectrum is calculated. The coefficient of variation is a relative index that measures the degree of data dispersion, which is used here to evaluate the difference between the virtual impedance spectrum and the monitoring impedance spectrum. The impedance value data sequences of the virtual impedance spectrum and the monitoring impedance spectrum at each frequency point are obtained respectively. According to the calculation formula of the coefficient of variation, the standard deviations of the two groups of data are calculated first, and then divided by their respective means to obtain the coefficient of variation of the virtual impedance spectrum and the monitoring impedance spectrum. This coefficient of variation reflects the relative dispersion degree between the two, and the smaller the value, the higher the similarity between the virtual impedance spectrum and the monitoring impedance spectrum, and the more the virtual impedance spectrum can represent the actual state of the target battery.

[0068] If the coefficient of variation meets the preset threshold, the virtual impedance spectrum is output. The preset threshold is determined based on a large number of experiments and data analysis, and is used to judge whether the virtual impedance spectrum is accurate enough. When the calculated coefficient of variation is less than or equal to the preset threshold, it indicates that the difference between the virtual impedance spectrum and the monitoring impedance spectrum is within an acceptable range, and the virtual impedance spectrum can well reflect the actual situation of the target battery. At this time, the virtual impedance spectrum can be directly output as a reliable basis for subsequent feature extraction and short circuit discrimination.

[0069] If the coefficient of variation does not meet the preset threshold, the virtual impedance spectrum is iteratively updated based on the over-limit direction of the coefficient of variation, and the corresponding coefficient of variation is recalculated until the preset threshold is met. If the coefficient of variation is greater than the preset threshold, it means that the difference between the virtual impedance spectrum and the monitoring impedance spectrum is large, and the virtual impedance spectrum needs to be optimized. According to the direction of the over-limit coefficient of variation, that is, whether the virtual impedance spectrum is too large or too small relative to the monitoring impedance spectrum, the corresponding adjustment strategy is adopted. For example, if the virtual impedance spectrum is too large, the parameters for calculating the virtual impedance spectrum by interpolation need to be fine-tuned to make it closer to the monitoring impedance spectrum. After adjustment, the virtual impedance spectrum is recalculated, and the coefficient of variation between the new virtual impedance spectrum and the monitoring impedance spectrum is calculated again. This iterative updating and coefficient of variation calculation process is repeated until the coefficient of variation meets the preset threshold. In this way, the final virtual impedance spectrum obtained can accurately reflect the state of the target battery, providing strong support for subsequent accurate diagnosis of internal short circuit of lithium batteries.

[0070] In one possible implementation manner, step S600 further includes:

[0071] Step S610: determining a set of key feature indicators by analyzing the sample short circuit data, wherein each feature indicator included in the set of key feature indicators is marked with a sensitivity.

[0072] Step S620: extracting feature values from the virtual impedance spectrum and the monitoring impedance spectrum based on the set of key feature indicators, respectively, to obtain a first impedance feature set and a second impedance feature set.

[0073] Step S630: calculating impedance feature difference information of the first impedance feature set and the second impedance feature set, and outputting the impedance feature set.

[0074] Specifically, in order to accurately diagnose the internal short circuit condition of lithium batteries, it is necessary to start from a large amount of sample short circuit data. These sample short circuit data cover various conditions of lithium batteries under different operating conditions and different short circuit degrees. Through the use of data mining, statistical analysis methods and electrochemical professional knowledge, these rich data are comprehensively and deeply analyzed. From a large number of possible characteristic indicators, those closely related to the short circuit condition of lithium batteries and capable of significantly reflecting the short circuit characteristics are selected to determine the key characteristic indicator set. For example, the impedance change trend in a specific frequency interval, the impedance real and imaginary part values at some key frequency points, etc. may be included. At the same time, in order to reflect the importance and sensitivity difference of each characteristic indicator in reflecting the short circuit problem, each characteristic indicator in the key characteristic indicator set is marked with sensitivity. The determination of sensitivity is based on the observation of the response of the indicator to the change in short circuit degree in a large number of sample data. If a certain indicator changes significantly when the short circuit degree changes slightly, it means that it is more sensitive to the short circuit condition and has a higher sensitivity; on the contrary, if the indicator does not change significantly or lags, its sensitivity is lower. After marking the sensitivity, these key characteristic indicators can be more targeted in the subsequent diagnosis process, improving the accuracy and reliability of short circuit diagnosis.

[0075] With the determined key characteristic indicator set as a guide, the feature values of the virtual impedance spectrum and the monitoring impedance spectrum are extracted respectively to obtain the first impedance feature set and the second impedance feature set. For each indicator in the key characteristic indicator set, find the corresponding feature value on the virtual impedance spectrum and the monitoring impedance spectrum. For example, if the key characteristic indicator set contains the impedance real part at frequency f1, extract the impedance real part value at frequency f1 from the virtual impedance spectrum and the monitoring impedance spectrum. In this way, the feature values corresponding to all indicators in the key characteristic indicator set are extracted in turn to form the first impedance feature set (from the virtual impedance spectrum) and the second impedance feature set (from the monitoring impedance spectrum). These two feature sets respectively represent the battery impedance feature information based on virtual and actual monitoring.

[0076] The first impedance feature set and the second impedance feature set are one-to-one corresponding according to the same feature index sequence. For each pair of corresponding feature values, the difference between them is quantified. For example, the absolute value of the difference between the two is calculated to intuitively reflect the degree of deviation in the numerical value; or the ratio of the two is calculated to measure the relative change. If a certain feature index is the impedance real part at a specific frequency, the feature values of the impedance real part at this frequency in the first and second impedance feature sets are extracted respectively, and the absolute value of the difference between the two is calculated as the difference information of the feature index. For all corresponding feature values, such calculation is performed, and a series of difference information obtained is integrated according to the feature index sequence to form a new impedance feature set. This new impedance feature set highlights the difference between the virtual state and the actual monitoring state of the battery impedance feature, can more effectively reflect the actual condition of the battery, provides core data support for subsequent short circuit discrimination based on the impedance feature set, and helps to improve the accuracy of lithium battery internal short circuit diagnosis.

[0077] In a possible implementation manner, the step S600 further includes:

[0078] Step S640: Obtain a historical impedance feature sequence, and combine the impedance feature set and the historical impedance feature sequence to calculate a change trend of the impedance feature by using a trend analysis method to obtain a trend development rate.

[0079] Step S650: Determine whether the trend development rate meets a preset short circuit discrimination constraint, wherein the short circuit discrimination constraint is obtained based on a statistical analysis method.

[0080] Step S660: If the trend development rate is greater than or equal to the short circuit discrimination constraint, the short circuit discrimination result is that there is a short circuit, and the short circuit level mark is equal to the ratio of the trend development rate to the short circuit discrimination constraint.

[0081] Specifically, in order to more accurately determine whether an internal short circuit of a lithium battery occurs, a historical impedance characteristic sequence of the battery needs to be obtained, which is a record set of impedance characteristics of the battery at different time points in the past use process of the battery, and contains impedance change information of the battery under various working conditions. Then, the current obtained impedance characteristic set is combined with the historical impedance characteristic sequence. Because in actual application, there are differences in ohmic resistance of each battery in the battery pack, and the internal short circuit will increase the ohmic resistance of the battery, but relying on setting a fixed threshold to determine the short circuit is not accurate. For the battery with relatively high ohmic resistance, false alarm may occur when it is not short-circuited according to the conventional threshold setting; and for the battery with relatively low ohmic resistance, the same threshold may cause missed alarm when short-circuited. Therefore, the trend analysis method is used to further process these data. Common trend analysis methods such as linear regression are used to fit the data points of the historical impedance characteristic sequence and the current impedance characteristic set, and find a straight line that best reflects the trend of the data. The slope of the straight line is the trend development rate. The trend development rate directly reflects the speed and direction of the change of the battery impedance characteristic with time. By analyzing this change trend instead of simply relying on a fixed threshold, the internal short circuit of the battery during the cycle use process can be more reliably determined, and key basis is provided for subsequent short circuit diagnosis.

[0082] It is determined whether the trend development rate meets a preset short circuit discrimination constraint, wherein the short circuit discrimination constraint is obtained based on a statistical analysis method. Based on a large amount of historical data and experimental results of the battery, a statistical analysis method such as hypothesis testing, probability distribution analysis, etc. is used to determine a reasonable short circuit discrimination constraint value. This constraint value is a measure standard for distinguishing the boundary between the normal operation of the battery and the occurrence of internal short circuit. The calculated trend development rate is compared with the preset short circuit discrimination constraint to determine whether the trend development rate exceeds the normal range.

[0083] If the trend development rate is greater than or equal to the short circuit discrimination constraint, it indicates that the state of the battery is abnormal, the short circuit discrimination result is that there is a short circuit, and the short circuit level mark is equal to the ratio of the trend development rate to the short circuit discrimination constraint. This means that the larger the trend development rate is, the higher the degree of exceeding the short circuit discrimination constraint is, and the more serious the short circuit of the battery is, and the higher the short circuit level is. For example, if the trend development rate is twice the short circuit discrimination constraint, the short circuit level mark is 2; if the trend development rate is only slightly greater than the short circuit discrimination constraint, the short circuit level mark is close to 1, indicating that the short circuit degree is relatively light. In this way, not only can the internal short circuit of the battery be accurately determined, but also the severity of the short circuit can be quantitatively marked, which provides key information for battery maintenance and management, and helps to take corresponding measures in time to ensure the safe and stable operation of the lithium battery.

[0084] In one possible implementation manner, step S600 further includes:

[0085] Step S670: Establish a short circuit diagnosis model based on machine learning, wherein the short circuit diagnosis model includes a virtual impedance spectrum correction channel and a difference analysis mapping channel. Interact with the battery management end to obtain the cycle log of the target battery. Take the virtual impedance spectrum and the cycle log as input, correct the impedance spectrum through the virtual impedance spectrum correction channel, obtain the reference impedance spectrum, and perform difference analysis on the monitoring impedance spectrum and the reference impedance spectrum through the difference analysis mapping channel. Obtain the reference impedance feature set, compare the reference impedance feature set with the impedance feature set, and obtain the short circuit discrimination result. If the short circuit discrimination result is that there is a short circuit, map the reference impedance feature set to the short circuit level label.

[0086] Specifically, to more accurately diagnose the internal short circuit condition of the lithium battery, a short circuit diagnosis model based on machine learning needs to be established. This model has two key channels: a virtual impedance spectrum correction channel and a difference analysis mapping channel. The role of the virtual impedance spectrum correction channel is to optimize the virtual impedance spectrum calculated in the early stage. It learns the complex relationship between the virtual impedance spectrum and the actual state of the battery through machine learning algorithms, and then corrects the virtual impedance spectrum, so that the virtual impedance spectrum can more accurately reflect the current true state of the battery. For example, a neural network algorithm can be used, with the virtual impedance spectrum data as input, to train the network to learn the characteristics of the virtual impedance spectrum under different working conditions, and thus output a more accurate virtual impedance spectrum. The difference analysis mapping channel focuses on comparative analysis, taking the monitoring impedance spectrum and the corrected reference impedance spectrum (output by the virtual impedance spectrum correction channel) as input, using a neural network algorithm to deeply analyze the differences between the two, and extracting a reference impedance feature set therefrom. By comparing the reference impedance feature set with the impedance feature set obtained by fusing the virtual and monitoring impedance spectra, this channel can determine whether the battery has a short circuit, providing key decision support for the entire short circuit diagnosis.

[0087] The battery management end, as the core control unit of the battery system, stores detailed running data of the target battery, including the cycle log which records key information of the battery during multiple charge and discharge cycles. To obtain these data, the diagnosis system establishes a communication connection with the battery management end, using a specific communication protocol such as Controller Area Network (CAN) protocol or other protocols suitable for battery management systems to ensure the accuracy and stability of data transmission. By sending a request instruction, the diagnosis system asks the battery management end for the cycle log of the target battery. After receiving the request, the battery management end retrieves and extracts the corresponding cycle log data from its internal storage unit, which includes rich information such as the start and end time of each charge and discharge, the charge and discharge current size, the voltage change, the battery temperature, and the SOC (State of Charge) change history.

[0088] After acquiring the virtual impedance spectrum and cycle log, they are transmitted as input data to the virtual impedance spectrum correction channel. This channel uses a pre-trained internal machine learning model, such as a neural network or other algorithmic model suitable for processing such data, to adjust the input virtual impedance spectrum. Based on information such as the number of battery charge and discharge cycles recorded in the cycle log, the current and voltage changes during each charge and discharge, and temperature conditions, combined with the characteristics of the virtual impedance spectrum itself, the model optimizes and corrects the virtual impedance spectrum, ultimately outputting a reference impedance spectrum that better reflects the battery's current actual state.

[0089] The difference analysis mapping channel begins operation, taking the monitored impedance spectrum and the reference impedance spectrum as input. Using a neural network algorithm, the channel performs a detailed difference analysis on the two impedance spectra, extracting from the differences between the two a reference impedance feature set that represents the characteristics of the battery's current state. This reference impedance feature set contains information about the difference between the battery's actual monitored state and its corrected ideal state. Subsequently, the reference impedance feature set is compared with the impedance feature set previously obtained by fusing the virtual impedance spectrum and the monitored impedance spectrum. By comparing the degree of similarity and feature differences between the two, the preset discrimination rules are applied to obtain a short circuit discrimination result to determine whether the battery has an internal short circuit.

[0090] After determining that the battery has a short circuit through the previous analysis process, it is necessary to further clarify the severity of the short circuit. The reference impedance feature set is obtained after the difference analysis mapping channel performs a difference analysis on the monitoring impedance spectrum and the reference impedance spectrum. It contains key information that reflects the difference between the current state of the battery and the normal state. According to the pre-set mapping rules, the values, change trends and other information of each characteristic indicator in the reference impedance feature set are converted. For example, if the impedance value at certain specific frequencies in the reference impedance feature set deviates more from the normal range, or the change trend of certain characteristic indicators is more obvious, according to the established rules, the short circuit level mark value obtained by mapping will be larger, indicating that the short circuit situation is more serious; conversely, the smaller the short circuit level mark value, the relatively lighter the short circuit is. In this way, the abstract reference impedance feature set is converted into an intuitive short circuit level mark, providing a clear and quantitative basis for subsequent decisions on battery maintenance, replacement, etc.

[0091] The second embodiment is based on the same inventive concept as the lithium battery internal short circuit diagnosis method based on impedance spectrum analysis in the above embodiment. Figure 2 As shown, the present application provides a lithium battery internal short circuit diagnosis system based on impedance spectroscopy analysis. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0092] A standard impedance spectrum library construction module 10 is configured to construct a standard impedance spectrum library based on the identified identity information of the target battery, wherein the standard impedance spectrum library comprises standard impedance spectra under each temperature-SOC combination.

[0093] A monitoring impedance spectrum acquisition module 20 is configured to perform an electrochemical impedance spectrum test on the target battery in combination with the standard impedance spectrum library, and acquire a monitoring impedance spectrum.

[0094] A monitoring impedance spectrum mapping module 30 is configured to map the monitoring impedance spectrum to the standard impedance spectrum library, identify the spectrum interval of the monitoring impedance spectrum, and correspondingly acquire an upper-limit impedance spectrum and a lower-limit impedance spectrum.

[0095] An offset calculation module 40 is configured to calculate the offset of the monitoring impedance spectrum relative to the upper-limit impedance spectrum and the lower-limit impedance spectrum, and locate a relative spectrum position according to the offset.

[0096] A virtual impedance spectrum acquisition module 50 is configured to calculate and acquire a virtual impedance spectrum based on an interpolation method, in combination with the relative spectrum position, the upper-limit impedance spectrum and the lower-limit impedance spectrum.

[0097] A short-circuit diagnosis result acquisition module 60 is configured to fuse the virtual impedance spectrum and the monitoring impedance spectrum, extract an impedance feature set, and perform short-circuit identification according to the impedance feature set to acquire a short-circuit diagnosis result, wherein the short-circuit diagnosis result comprises a short-circuit identification result and a short-circuit level label.

[0098] Further, the system is further configured to implement the following functions:

[0099] A temperature range and an SOC range are determined, and a set of temperature-SOC combination points is defined according to the principle of orthogonal experiment design; the target battery is subjected to a preset number of electrochemical impedance spectrum tests at each combination point, and standard impedance spectrum data is acquired by fusing the test results; and the acquired standard impedance spectrum data is classified and stored according to temperature-SOC combinations to construct a standard impedance spectrum library.

[0100] Further, the system is further configured to implement the following functions:

[0101] A confidence frequency interval is determined by performing statistical analysis in the standard impedance spectrum library based on a statistical analysis method; an alternating current excitation signal covering low frequencies to high frequencies is applied to the electrode of the target battery according to the confidence frequency interval, and the voltage response of the target battery is measured to acquire complex impedance data of the battery; and the monitoring impedance spectrum is established according to the complex impedance data.

[0102] Further, the system is further configured to implement the following functions:

[0103] extracting an SOC marker of the monitoring impedance spectrum; based on the SOC marker, traversing the standard impedance spectrum library to construct a standard impedance spectrum sub-library; constructing an impedance spectrum coordinate system and mapping the standard impedance spectrum sub-library to the impedance spectrum coordinate system to obtain a plurality of spectrum intervals; mapping the monitoring impedance spectrum to the impedance spectrum coordinate system, extracting an upper limit of the interval as the upper limit impedance spectrum and an lower limit of the interval as the lower limit impedance spectrum according to the spectrum interval to which the monitoring impedance spectrum belongs.

[0104] Further, the system is also used to implement the following functions:

[0105] selecting a representative key frequency point in the impedance spectrum as a comparison reference; comparing the impedance values of the monitoring impedance spectrum, the upper limit impedance spectrum and the lower limit impedance spectrum at each key frequency point to calculate the deviation value of each key point; weighting the deviation values of each key point to calculate the overall deviation, and outputting the upper deviation and the lower deviation; normalizing the upper deviation and the lower deviation to determine the relative position of the monitoring impedance spectrum in the spectrum interval to which it belongs.

[0106] Further, the system is also used to implement the following functions:

[0107] calculating the coefficient of variation of the virtual impedance spectrum and the monitoring impedance spectrum; if the coefficient of variation meets the preset threshold, outputting the virtual impedance spectrum; if the coefficient of variation does not meet the preset threshold, performing iterative updating of the virtual impedance spectrum based on the over-limit direction of the coefficient of variation, and recalculating the corresponding coefficient of variation until the preset threshold is met.

[0108] Further, the system is also used to implement the following functions:

[0109] determining a set of key feature indicators by analyzing sample short-circuit data, wherein each feature indicator in the set of key feature indicators is marked with a sensitivity; based on the set of key feature indicators, extracting feature values from the virtual impedance spectrum and the monitoring impedance spectrum respectively to obtain a first impedance feature set and a second impedance feature set; calculating impedance feature difference information of the first impedance feature set and the second impedance feature set, and outputting the impedance feature set.

[0110] Further, the system is also used to implement the following functions:

[0111] obtaining a historical impedance feature sequence, and combining the impedance feature set and the historical impedance feature sequence to calculate a change trend of the impedance feature using a trend analysis method to obtain a trend development rate; determining whether the trend development rate meets a preset short-circuit discrimination constraint, wherein the short-circuit discrimination constraint is obtained based on a statistical analysis method; if the trend development rate is greater than or equal to the short-circuit discrimination constraint, the short-circuit discrimination result is that there is a short circuit, and the short-circuit level marker is equal to the ratio of the trend development rate to the short-circuit discrimination constraint.

[0112] Further, the system is also used to realize the following functions:

[0113] establish a short circuit diagnosis model based on machine learning, wherein the short circuit diagnosis model includes a virtual impedance spectrum correction channel and a difference analysis mapping channel; an interactive battery management end acquires a cycle log of a target battery; the virtual impedance spectrum and the cycle log are taken as inputs, impedance spectrum correction is performed through the virtual impedance spectrum correction channel to acquire a reference impedance spectrum; the difference analysis mapping channel takes the monitoring impedance spectrum and the reference impedance spectrum as inputs to perform difference analysis, acquires a reference impedance feature set, and compares the reference impedance feature set and the impedance feature set to acquire the short circuit discrimination result; if the short circuit discrimination result is that there is a short circuit, the reference impedance feature set is mapped to the short circuit level label.

[0114] It should be noted that the above sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the above describes a specific embodiment of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0115] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0116] The specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that fall within the scope of the present application should be considered. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A lithium battery internal short circuit diagnosis method based on impedance spectroscopy analysis, characterized by, The method comprises: Based on the identification information of the target battery, a standard impedance spectrum library is constructed, wherein the standard impedance spectrum library comprises standard impedance spectra under each temperature-SOC combination; In combination with the standard impedance spectrum library, an electrochemical impedance spectrum test is performed on the target battery to obtain a monitoring impedance spectrum; The monitoring impedance spectrum is mapped to the standard impedance spectrum library to identify the spectrum interval of the monitoring impedance spectrum, and the upper limit impedance spectrum and the lower limit impedance spectrum are correspondingly obtained, wherein the upper limit impedance spectrum is extracted from the upper limit part of the boundary data of the specific spectrum interval to which the monitoring impedance spectrum belongs, and the lower limit impedance spectrum is extracted from the lower limit part of the boundary data of the specific spectrum interval to which the monitoring impedance spectrum belongs, and the upper limit impedance spectrum and the lower limit impedance spectrum respectively represent two boundary states of the spectrum interval where the monitoring impedance spectrum is located; The offset of the monitoring impedance spectrum relative to the upper limit impedance spectrum and the lower limit impedance spectrum is calculated, and the relative spectrum position is located according to the offset; Based on an interpolation method, the relative spectrum position, the upper limit impedance spectrum and the lower limit impedance spectrum are combined to calculate and obtain a virtual impedance spectrum, which is a more accurate reference spectrum and effectively corrects the deviation caused by temperature or SOC estimation error factors; The virtual impedance spectrum and the monitoring impedance spectrum are fused to extract an impedance feature set, and short circuit discrimination is performed according to the impedance feature set to obtain a short circuit diagnosis result, wherein the short circuit diagnosis result comprises a short circuit discrimination result and a short circuit level label; Based on an interpolation method, the relative spectrum position, the upper limit impedance spectrum and the lower limit impedance spectrum are combined to calculate and obtain a virtual impedance spectrum, which is a more accurate reference spectrum and effectively corrects the deviation caused by temperature or SOC estimation error factors; The variation coefficient of the virtual impedance spectrum and the monitoring impedance spectrum is calculated; If the variation coefficient meets the preset threshold, the virtual impedance spectrum is output; If the variation coefficient does not meet the preset threshold, the virtual impedance spectrum is iteratively updated based on the over-limit direction of the variation coefficient, and the corresponding variation coefficient is recalculated until the preset threshold is met; The virtual impedance spectrum and the monitoring impedance spectrum are fused to extract an impedance feature set, and short circuit discrimination is performed according to the impedance feature set to obtain a short circuit diagnosis result, wherein the short circuit diagnosis result comprises a short circuit discrimination result and a short circuit level label; Based on the key feature index set, feature values of the virtual impedance spectrum and the monitoring impedance spectrum are extracted respectively to obtain a first impedance feature set and a second impedance feature set; The impedance feature difference information of the first impedance feature set and the second impedance feature set is calculated, and the impedance feature set is output; According to the impedance feature set, short circuit discrimination is performed to obtain a short circuit diagnosis result, comprising: A historical impedance feature sequence is obtained, and the impedance feature set and the historical impedance feature sequence are combined to calculate the change trend of the impedance feature by using a trend analysis method to obtain a trend development rate; It is judged whether the trend development rate meets a preset short circuit discrimination constraint, wherein the short circuit discrimination constraint is obtained based on a statistical analysis method; If the trend development rate is greater than or equal to the short circuit discrimination constraint, the short circuit discrimination result is that there is a short circuit, and the short circuit level label is equal to the ratio of the trend development rate to the short circuit discrimination constraint. ​ 2. The impedance spectroscopy-based lithium battery internal short circuit diagnosis method of claim 1, wherein, Based on the identification information of the target battery, a standard impedance spectrum library is constructed, wherein the standard impedance spectrum library includes standard impedance spectra under each temperature-SOC combination, including: Determine the temperature range and SOC range, and define the temperature-SOC combination point set according to the orthogonal experimental design principle; Perform a preset number of electrochemical impedance spectrum tests on the target battery at each combination point, and fuse the test results to obtain standard impedance spectrum data; The obtained standard impedance spectrum data is classified and stored according to the temperature-SOC combination to construct a standard impedance spectrum library.

3. The impedance spectroscopy-based lithium battery internal short circuit diagnosis method of claim 2, wherein, In combination with the standard impedance spectrum library, the target battery is subjected to electrochemical impedance spectrum test to obtain a monitoring impedance spectrum, including: Based on statistical analysis method, statistical analysis is carried out in the standard impedance spectrum library to determine the confidence frequency interval; According to the confidence frequency interval, an alternating current excitation signal covering low frequency to high frequency is applied to the target battery electrode, and the voltage response of the target battery is measured to obtain the complex impedance data of the battery; According to the complex impedance data, the monitoring impedance spectrum is established.

4. The impedance spectroscopy-based lithium battery internal short circuit diagnosis method of claim 3, wherein, Map the monitoring impedance spectrum to the standard impedance spectrum library to determine the spectrum interval of the monitoring impedance spectrum, and correspondingly obtain the upper and lower limit impedance spectra, including: Extract the SOC marker of the monitoring impedance spectrum; Based on the SOC marker, traverse the standard impedance spectrum library to construct a standard impedance spectrum sub-library; Construct an impedance spectrum coordinate system and map the standard impedance spectrum sub-library to the impedance spectrum coordinate system to obtain a plurality of spectrum intervals; Map the monitoring impedance spectrum to the impedance spectrum coordinate system, and according to the corresponding spectrum interval, extract the upper limit as the upper limit impedance spectrum and the lower limit as the lower limit impedance spectrum.

5. The impedance spectroscopy-based lithium battery internal short circuit diagnosis method of claim 4, wherein, Calculate the offset of the monitoring impedance spectrum relative to the upper and lower limit impedance spectra, and locate the relative spectrum position according to the offset, including: Select representative key frequency points in the impedance spectrum as comparison reference; Compare the impedance values of the monitoring impedance spectrum, the upper limit impedance spectrum and the lower limit impedance spectrum at each key frequency point to calculate the offset difference value of each key point; Weight the difference values of each key point to calculate the overall offset, and output the upper and lower offset values; Normalize the upper and lower offset values to determine the relative position of the monitoring impedance spectrum in the corresponding spectrum interval.

6. The impedance spectroscopy-based lithium battery internal short diagnosis method of claim 1, wherein, The method further includes: Establish a short circuit diagnosis model based on machine learning, wherein the short circuit diagnosis model includes a virtual impedance spectrum correction channel and a difference analysis mapping channel; Interact with the battery management terminal to obtain the cycle log of the target battery; Input the virtual impedance spectrum and the cycle log into the virtual impedance spectrum correction channel to correct the impedance spectrum and obtain a reference impedance spectrum; The difference analysis mapping channel takes the monitoring impedance spectrum and the reference impedance spectrum as input to perform difference analysis and obtain a reference impedance feature set, and compares the reference impedance feature set with the impedance feature set to obtain the short circuit judgment result; If the short circuit judgment result is that there is a short circuit, map the reference impedance feature set to the short circuit level marker.

7. A lithium battery internal short circuit diagnosis system based on impedance spectroscopy analysis, characterized by, The system is used to execute any one of the methods of claims 1-6, including: a standard impedance spectrum library construction module, configured to construct a standard impedance spectrum library based on the identified identity information of the target battery, wherein the standard impedance spectrum library comprises standard impedance spectra under each temperature-SOC combination; a monitoring impedance spectrum acquisition module, configured to perform an electrochemical impedance spectrum test on the target battery in combination with the standard impedance spectrum library, and acquire a monitoring impedance spectrum; a monitoring impedance spectrum mapping module, configured to map the monitoring impedance spectrum to the standard impedance spectrum library, identify a spectrum interval of the monitoring impedance spectrum, and correspondingly acquire an upper-limit impedance spectrum and a lower-limit impedance spectrum, wherein the upper-limit impedance spectrum is extracted from a boundary data of a specific spectrum interval to which the monitoring impedance spectrum belongs, and the lower-limit impedance spectrum is extracted from a boundary data of a specific spectrum interval to which the monitoring impedance spectrum belongs, the upper-limit impedance spectrum and the lower-limit impedance spectrum respectively representing two boundary states of the spectrum interval to which the monitoring impedance spectrum belongs; an offset calculation module, configured to calculate an offset of the monitoring impedance spectrum relative to the upper-limit impedance spectrum and the lower-limit impedance spectrum, and locate a relative spectrum position according to the offset; a virtual impedance spectrum acquisition module, configured to calculate and acquire a virtual impedance spectrum based on an interpolation method, in combination with the relative spectrum position, the upper-limit impedance spectrum and the lower-limit impedance spectrum, the virtual impedance spectrum being a more accurate reference spectrum that effectively corrects deviations caused by temperature or SOC estimation error factors; a short-circuit diagnosis result acquisition module, configured to fuse the virtual impedance spectrum and the monitoring impedance spectrum, extract an impedance feature set, and perform short-circuit identification according to the impedance feature set, and acquire a short-circuit diagnosis result, wherein the short-circuit diagnosis result comprises a short-circuit identification result and a short-circuit level label.

Images