Thermal runaway detection method, system and equipment for battery pack and medium

Abstract

The invention discloses a thermal runaway detection method, system and device for a battery pack and a medium, and belongs to the technical field of battery safety detection.The method comprises the steps that charging and discharging current signals of a to-be-detected battery pack are obtained to extract high-frequency ripples, the current fluctuation rate and the harmonic distortion rate, and then a preliminary thermal runaway detection result of the to-be-detected battery pack is obtained; removing a direct current component in the charging and discharging current signal to carry out continuous wavelet transform on the obtained direct current-removed signal to obtain wavelet coefficients corresponding to the direct current-removed signal under different wavelet scales; performing wavelet reconstruction on the DC-removed signal according to the wavelet coefficient and the wavelet scale to obtain a reconstructed signal; and obtaining a plurality of power spectrums corresponding to the reconstructed signal according to the wavelet scale, and obtaining a real-time thermal runaway detection result of the battery pack to be detected according to the plurality of power spectrums. Therefore, by implementing the method, the technical problem that thermal runaway detection of the battery pack cannot be timely and accurately performed in the prior art can be solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery safety detection, and in particular to a thermal runaway detection method, system, device and medium for a battery pack. BACKGROUND

[0002] With the increasing degree of coordination between new energy power generation and power grid dispatching, electrochemical energy storage power stations play a key role in peak load shifting and emergency power supply. Such energy storage facilities usually contain large-scale battery packs. However, the characteristics of thermal runaway of battery packs and the abuse of charging and discharging processes pose risks to the use of battery packs and hinder the application of battery packs in large-scale energy storage power stations.

[0003] In the initial stage of thermal runaway, immediately stopping charging is extremely effective in curbing the progress of thermal runaway. In this process, how to timely and accurately identify the initial stage of thermal runaway becomes an important technical means to ensure battery safety. The prior art detects thermal runaway by measuring voltage fluctuations of the battery pack or using gas detection technology. However, before the safety problem of the battery pack appears, the voltage of the battery will not fluctuate significantly, so relying solely on voltage as a warning signal for thermal runaway will result in significant time delay. In the gas detection technology, carbon monoxide and hydrocarbons detected are confirmed as indicators of thermal abuse or overcharging, which cannot timely identify thermal runaway of the battery pack. In addition to the above detection technologies, the prior art also detects thermal runaway of the battery pack based on a single parameter such as temperature. However, these methods have certain limitations. For example, temperature sensors have a response delay and cannot timely detect thermal runaway. SUMMARY

[0004] The present application provides a thermal runaway detection method, system, device and medium for a battery pack, which can solve the technical problem that the prior art cannot timely and accurately detect thermal runaway of the battery pack.

[0005] In a first aspect, the present application discloses a thermal runaway detection method for a battery pack, characterized in that it comprises: obtaining a charging and discharging current signal of a battery pack to be tested, and extracting high-frequency ripples, current fluctuation rate and harmonic distortion rate of the charging and discharging current signal; obtaining a preliminary thermal runaway detection result of the battery pack to be tested according to the high-frequency ripples, the current fluctuation rate and the harmonic distortion rate; removing a direct current component in the charging and discharging current signal according to the preliminary thermal runaway detection result to obtain a direct current removed signal; performing continuous wavelet transform on the direct current removed signal to obtain wavelet coefficients corresponding to the direct current removed signal at different wavelet scales; wavelet reconstruction is performed on the direct current removing signal according to the wavelet coefficients and the wavelet scales, to obtain a reconstructed signal; a plurality of power spectrums corresponding to the reconstructed signal are obtained according to the wavelet scales, to obtain a real-time thermal runaway detection result of the battery pack according to the plurality of power spectrums.

[0006] The application discloses a thermal runaway detection method for a battery pack, which adopts a multi-stage and multi-parameter fusion progressive thermal runaway detection structure to timely and accurately detect the thermal runaway of the battery pack. The charging and discharging current signals of the battery pack are selected to comprehensively detect the battery pack, and then the high-frequency components such as high-frequency ripples, slope edge distortion rates and the like in the current signals, which can directly reflect the thermal runaway of the battery pack, are extracted to detect the thermal runaway, so that the timeliness of detection is improved. The thermal runaway is detected by extracting multiple parameters such as high-frequency ripples, current fluctuation rates and harmonic distortion rates, so that false positives caused by single signal interference can be effectively avoided, and the accuracy of detection is improved. After preliminary detection, continuous wavelet transformation and power spectrum analysis are performed on the direct current removing signal, the energy change characteristics of the signal in a specific frequency band are accurately extracted, then the power spectrums of the reconstructed signal under different scales are analyzed, the subtle but trend-oriented characteristic changes in the thermal runaway process can be captured, and the thermal runaway of the battery pack can be accurately identified, so that the accuracy of thermal runaway detection of the battery pack is improved.

[0007] As a preferred example, the charging and discharging current signals of the battery pack are obtained, and the high-frequency ripples, current fluctuation rates and harmonic distortion rates of the charging and discharging current signals are extracted, including: The charging current signals, discharging current signals and reverse current signals at the charging and discharging switching time of the battery pack are obtained, and the charging current signals, discharging current signals and reverse current signals are taken as the charging and discharging current signals of the battery pack; The total current signals of the total input end of the battery pack and the branch current signals of the branch nodes of the battery pack are extracted from the charging and discharging current signals, and the branch nodes are arranged in the charging loop of the battery pack; The signal difference value of the total current signals and the branch current signals is obtained, and the signal difference value is filtered to obtain the high-frequency ripples corresponding to the branch nodes; The ratio of the signal difference value to the total current signals is obtained, and the ratio is determined as the current fluctuation rate of the battery pack; The total current signals and the branch current signals are respectively subjected to Fourier transformation to obtain the total harmonic distortion rate corresponding to the total current signals and the branch harmonic distortion rate corresponding to the branch current signals; Obtaining a difference value of the total harmonic distortion rate and the branch harmonic distortion rate to determine the harmonic distortion rate of the battery pack to be tested according to the difference value.

[0008] The above scheme can accurately locate the abnormal source by comparing and analyzing the total current signal and the branch current signal. The above scheme can amplify the local fault characteristics by calculating the signal difference value and the ratio value and comparing the total harmonic distortion rate with the branch harmonic distortion rate, so that the preliminary detection result is more targeted, the local hidden danger in the battery pack can be found earlier, and the timeliness of detection is improved.

[0009] As a preferred example, the preliminary thermal runaway detection result of the battery pack to be tested according to the high-frequency ripple, the current fluctuation rate and the harmonic distortion rate comprises: obtaining a first comparison result of the high-frequency ripple and a preset ripple threshold value; obtaining a second comparison result of the current fluctuation rate and a preset fluctuation rate threshold value; obtaining a third comparison result of the harmonic distortion rate and a preset distortion rate threshold value; determining the preliminary thermal runaway detection result of the battery pack to be tested according to the first comparison result, the second comparison result and the third comparison result.

[0010] The above scheme constructs a multi-condition triggering decision mechanism by comparing the high-frequency ripple, the current fluctuation rate and the harmonic distortion rate with the corresponding preset threshold value respectively, and determining according to the comparison result, which greatly enhances the accuracy of thermal runaway detection.

[0011] As a preferred example, the removing of the direct current component in the charging and discharging current signal according to the preliminary thermal runaway detection result to obtain a direct current removed signal comprises: When it is determined according to the preliminary thermal runaway detection result that the battery pack to be tested has a thermal runaway risk, performing a first derivative operation on the charging and discharging current signal to obtain a current change rate signal; performing multi-layer decomposition wavelet transform on the current change rate signal according to a preset wavelet function to obtain a wavelet coefficient data set with different frequency subbands; zero processing the wavelet coefficients corresponding to the direct current component in the wavelet coefficient data set, and performing wavelet reconstruction using the processed wavelet coefficient data set to obtain a direct current removed current change rate signal; performing integral operation on the direct current removed current change rate signal to obtain a direct current removed current signal corresponding to the charging and discharging current signal.

[0012] The above scheme employs a method combining first-order derivative operations with wavelet decomposition, zeroing, reconstruction, and integration to provide a high-precision signal preprocessing technique. This provides high-precision data for subsequent thermal runaway detection, thereby improving detection accuracy. Specifically, the first-order derivative highlights the rapidly changing parts of the signal, wavelet decomposition accurately separates the slowly changing DC component and zeros it, and finally, integration restores the signal shape, thus removing the DC component more thoroughly and cleanly while preserving AC details containing fault information to the greatest extent possible. This provides a more accurate input signal for subsequent wavelet transform and feature extraction, thereby improving detection accuracy.

[0013] As a preferred example, the step of performing continuous wavelet transform on the de-DC signal to obtain the wavelet coefficients of the de-DC signal at different wavelet scales includes: The DC-DC current signal is subjected to continuous wavelet transform based on preset wavelet transform basis functions and multiple preset wavelet scales to obtain wavelet coefficient datasets corresponding to the DC-DC current signal under different frequency band information; wherein, the frequency band information is determined according to the wavelet scales. Each original wavelet coefficient in the wavelet coefficient dataset is thresholded so that the original wavelet coefficient that has passed the thresholding is used as the wavelet coefficient corresponding to the DC-DC signal.

[0014] The above scheme, after obtaining the original wavelet coefficient dataset, adds a thresholding step, which filters out coefficients with small amplitudes that may be caused by background noise, retaining only those key wavelet coefficients with significant amplitudes that are more likely to be generated by real fault characteristics. By highlighting the frequency band information related to the fault, subsequent reconstruction and power spectrum analysis can focus more on effective information, improving the accuracy of detection.

[0015] As a preferred example, the step of performing wavelet reconstruction on the DC-de-switched signal based on the wavelet coefficients and the wavelet scale to obtain the reconstructed signal includes: Based on the wavelet coefficients corresponding to each wavelet scale, the DC-de-switched signal is reconstructed using a preset wavelet reconstruction algorithm to obtain an initial reconstructed signal. The initial reconstructed signal is subjected to amplitude calibration and smoothing filtering to obtain the reconstructed signal.

[0016] The above scheme performs amplitude calibration and smoothing filtering after obtaining the initial reconstructed signal. Amplitude calibration ensures that the magnitude of the reconstructed signal matches the original signal, facilitating subsequent quantitative analysis. Smoothing filtering suppresses high-frequency glitches or noise that may be introduced during the reconstruction process. These two post-processing steps together guarantee the quality and smoothness of the final reconstructed signal, making the power spectrum calculated based on the reconstructed signal more stable and reliable, thereby improving the accuracy of detection.

[0017] As a preferred example, the step of obtaining the power spectrum of the reconstructed signal at different frequencies according to the wavelet scale, and obtaining the real-time thermal runaway detection result of the battery pack under test according to the power spectrum, includes: The power spectrum of the reconstructed signal at each wavelet scale is obtained according to the preset power spectral density estimation algorithm. The power spectra are sorted according to the wavelet scale to obtain a multi-scale power spectrum; Time-domain statistical analysis was performed on the multi-scale power spectrum to obtain the mean change rate of the multi-scale power spectrum and the variance of the power spectrum at each scale. The real-time thermal runaway detection results of the battery pack under test are obtained based on the mean change rate of the multi-scale power spectrum, the variance of the power spectrum at each scale, and the preset mean change rate threshold and the preset variance threshold.

[0018] The above-described scheme, by monitoring the rate of change of the power spectrum mean, can keenly detect the trend of signal energy entering a stable upward channel, thus enabling earlier warning. Simultaneously, monitoring the variance of the power spectrum at various scales can determine whether fault characteristics appear synchronously and consistently across different frequency bands, further confirming the certainty of fault occurrence. Combining trend judgment and consistency verification ensures that the final real-time thermal runaway detection results are both fast and reliable.

[0019] In a second aspect, the present invention discloses a thermal runaway detection system for battery packs, including a feature extraction module, a preliminary detection module, a DC removal module, a wavelet transform module, a wavelet reconstruction module, and a real-time detection module. The feature extraction module is used to acquire the charging and discharging current signal of the battery pack under test, and extract the high-frequency ripple, current fluctuation rate and harmonic distortion rate of the charging and discharging current signal. The preliminary detection module is used to obtain the preliminary thermal runaway detection results of the battery pack under test based on the high-frequency ripple, the current fluctuation rate, and the harmonic distortion rate. The DC removal module is used to remove the DC component from the charge / discharge current signal based on the preliminary thermal runaway detection results, to obtain a DC removal signal. The wavelet transform module is used to perform continuous wavelet transform on the de-DC signal to obtain the wavelet coefficients of the de-DC signal at different wavelet scales. The wavelet reconstruction module is used to perform wavelet reconstruction on the DC-de-DC signal according to the wavelet coefficients and the wavelet scale to obtain the reconstructed signal. The real-time detection module is used to obtain multiple power spectra corresponding to the reconstructed signal according to the wavelet scale, so as to obtain the real-time thermal runaway detection result of the battery pack under test according to the multiple power spectra.

[0020] This invention discloses a thermal runaway detection system for battery packs. It employs a multi-stage, multi-parameter fusion-based progressive thermal runaway detection structure to detect thermal runaway in a timely and accurate manner. Specifically, it comprehensively detects the battery pack by selecting its charging and discharging current signals. Then, it extracts high-frequency components from the current signals, such as high-frequency ripple and bevel distortion rate, which directly reflect thermal runaway, for thermal runaway detection, thus improving detection timeliness. By extracting multiple parameters such as high-frequency ripple, current fluctuation rate, and harmonic distortion rate for thermal runaway detection, false alarms caused by single-signal interference can be effectively avoided, improving detection accuracy. After preliminary detection, continuous wavelet transform and power spectrum analysis are performed on the de-DC signal to accurately extract the energy change characteristics of the signal in a specific frequency band. Then, by analyzing and reconstructing the power spectrum of the signal at different scales, subtle but trend-like characteristic changes in the thermal runaway process can be captured, thereby accurately identifying thermal runaway in the battery pack and improving the accuracy of thermal runaway detection.

[0021] Thirdly, the present invention discloses a terminal device including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a thermal runaway detection method for a battery pack as described in the first aspect.

[0022] Fourthly, the present invention discloses a computer-readable storage medium comprising: a stored computer program, wherein, when the computer program is executed, the device in which the computer-readable storage medium is located is controlled to perform a thermal runaway detection method for a battery pack as described in the first aspect. Attached Figure Description

[0023] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a schematic flowchart of a method for detecting thermal runaway in a battery pack, as disclosed in an embodiment of the present invention. Figure 2 This is a schematic diagram of the original waveform of the battery pack under test in charging condition 1, as disclosed in an embodiment of the present invention; Figure 3 This is a schematic diagram of the original waveform of the battery pack under test in charging condition 2, as disclosed in an embodiment of the present invention; Figure 4 This is a schematic diagram of a thermal runaway detection system for battery packs disclosed in an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0027] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0028] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0029] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character "" in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0030] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0031] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0032] See Figure 1 To address the technical problem that existing technologies cannot perform timely and accurate thermal runaway detection of battery packs, an embodiment of the present invention provides a method for thermal runaway detection of battery packs, comprising: Step 101: Obtain the charging and discharging current signal of the battery pack under test, and extract the high-frequency ripple, current fluctuation rate and harmonic distortion rate of the charging and discharging current signal.

[0033] In this embodiment, the main steps are as follows: acquiring the charging current signal, discharging current signal, and reverse current signal during charge / discharge switching of the battery pack under test, and using the charging current signal, discharging current signal, and reverse current signal as the charging / discharging current signals of the battery pack under test; extracting the total current signal of the total input terminal of the battery pack under test and the branch current signal of the branch node of the battery pack under test from the charging / discharging current signals; wherein the branch node is deployed in the charging circuit of the battery pack under test; acquiring the signal difference between the total current signal and the branch current signal, and filtering the signal difference to obtain the high-frequency ripple corresponding to the branch node; acquiring the ratio of the signal difference to the total current signal, and determining the ratio as the current fluctuation rate of the battery pack under test; performing Fourier transform on the total current signal and the branch current signal respectively to obtain the total harmonic distortion rate corresponding to the total current signal and the branch harmonic distortion rate corresponding to the branch current signal; acquiring the difference between the total harmonic distortion rate and the branch harmonic distortion rate, and determining the harmonic distortion rate of the battery pack under test based on the difference.

[0034] Specifically, in this first embodiment, to form a monitoring network covering the entire power flow path for comprehensive and timely detection of thermal runaway in the battery pack under test, power quality parameter acquisition points for collecting charging and discharging current signals can be set at the total input terminal, total output terminal, and multiple preset branch nodes of the battery pack under test. These branch nodes are respectively deployed in the charging circuit (e.g., between the total input terminal and the charging device), the discharging circuit (e.g., between the total output terminal and the load), and near the charge / discharge switching switch of the battery pack under test. These branch nodes, in conjunction with the acquisition points at the total input and total output terminals of the battery pack under test, acquire charging current, discharging current, and reverse current information during charge / discharge switching, forming a monitoring network covering the entire power flow path.

[0035] In some embodiments of this example, to improve the accuracy of the charging and discharging current signal acquisition, an 18-bit resolution AD converter can be used at the power quality parameter acquisition point to acquire the charging and discharging current signal. Preferably, the charging and discharging current signal acquisition device, including the AD converter, deployed at the power quality parameter acquisition point should have high precision and a high sampling frequency, capable of accurately recording changes in the charging and discharging current signal in real time. Furthermore, the sampling frequency of the acquisition device should be greater than or equal to 1kHz and comply with the EN IEC61010 standard, while its insulation performance should reach 1500V DC CATⅡ to ensure measurement safety. The AD converter can capture μA-level current fluctuations and connects signal acquisition with subsequent digital processing through analog-to-digital conversion, converting analog signals into digital signals.

[0036] Specifically, in some embodiments of this example, after acquiring the charging and discharging current signal through the acquisition device, high-frequency ripple, current fluctuation rate, and harmonic distortion rate can be extracted from the charging and discharging current signal for subsequent preliminary detection. Specifically, the current signals from the total input terminal and the branch node are acquired and fused in the following manner to obtain the high-frequency ripple, current fluctuation rate, and harmonic distortion rate corresponding to the battery pack under test.

[0037] Specifically, calculate the current signal at the total input terminal. Current signal of branch node signal difference The signal difference ΔI is filtered (e.g., using a 1 kHz-8 kHz bandpass filter) to obtain the high-frequency ripple corresponding to the branch node. Then, the current signal from the total input terminal is used... Calculate the current fluctuation rate based on the baseline. It is used to identify reverse current fluctuations during charge / discharge switching; while the total current signal... and branch current signal By performing Fourier transforms on each, the 3rd and 5th harmonic components can be extracted. , Then, the harmonic distortion rate corresponding to the branch current signal or the total current signal is calculated respectively. Furthermore, based on the difference between the total harmonic distortion rate corresponding to the total current signal and the branch harmonic distortion rate corresponding to the branch current signal, combined with... The difference is used to locate the source of harmonics.

[0038] In this embodiment, the above steps, by comparing and analyzing the total current signal and the branch current signal, can accurately locate the source of the anomaly. The above scheme, by calculating the signal difference and ratio and comparing the total harmonic distortion rate with the branch harmonic distortion rate, can amplify this local fault characteristic, making the preliminary detection results more targeted, enabling earlier detection of localized hidden dangers within the battery pack, and improving the timeliness of detection.

[0039] Step 102: Obtain preliminary thermal runaway detection results of the battery pack under test based on the high-frequency ripple, the current fluctuation rate, and the harmonic distortion rate.

[0040] In this embodiment, the main steps are: obtaining a first comparison result of the high-frequency ripple and a preset ripple threshold; obtaining a second comparison result of the current fluctuation rate and a preset fluctuation rate threshold; obtaining a third comparison result of the harmonic distortion rate and a preset distortion rate threshold; and determining the preliminary thermal runaway detection result of the battery pack under test based on the first comparison result, the second comparison result, and the third comparison result.

[0041] Specifically, in some embodiments of this example, the resistance inside the battery pack fluctuates abnormally before thermal runaway, leading to increased high-frequency ripple, volatility, and harmonic distortion of the battery pack's current. These parameters directly reflect early battery pack failures. Therefore, in this embodiment, to promptly identify these early failures, the real-time acquired high-frequency ripple, current volatility, and harmonic distortion can be compared with their corresponding thresholds to obtain timely and accurate thermal runaway detection results based on the comparison results.

[0042] In this embodiment, the above steps involve comparing the three parameters—high-frequency ripple, current fluctuation rate, and harmonic distortion rate—with their respective preset thresholds, and then making a comprehensive judgment based on the comparison results. This constructs a multi-condition triggered decision-making mechanism, which greatly enhances the accuracy of thermal runaway detection.

[0043] Step 103: Remove the DC component from the charge / discharge current signal based on the preliminary thermal runaway detection results to obtain the de-DC signal.

[0044] In this embodiment, the step mainly includes: when it is determined that the battery pack under test has a risk of thermal runaway based on the preliminary thermal runaway detection result, performing a first derivative operation on the charge / discharge current signal to obtain a current conversion rate signal; performing a multi-level decomposition wavelet transform on the current conversion rate signal according to a preset wavelet function to obtain a wavelet coefficient dataset with different frequency sub-bands; setting the wavelet coefficients corresponding to the DC component in the wavelet coefficient dataset to zero, and using the processed wavelet coefficient dataset for wavelet reconstruction to obtain a DC-free current conversion rate signal; and performing an integral operation on the DC-free current conversion rate signal to obtain the DC-free current signal corresponding to the charge / discharge current signal.

[0045] Specifically, in some embodiments of this example, the first derivative of the charging / discharging current signal is subjected to wavelet transform, and then the coefficients corresponding to the DC component in the wavelet-transformed signal are set to zero before reconstruction. Integration then yields the de-DC signal. Preferably, a preset wavelet function, such as the db4 wavelet, can be used to perform multi-level decomposition on the signal after the first derivative, such as a 3-level decomposition, to obtain a wavelet coefficient dataset with different frequency sub-bands. Then, the wavelet coefficients corresponding to the DC component in the wavelet coefficient dataset are set to zero for subsequent integration and reconstruction.

[0046] In this embodiment, the above steps employ a method combining first-order derivative operations with wavelet decomposition, zeroing, reconstruction, and integration. This provides a high-precision signal preprocessing technique, thereby providing high-precision data for subsequent thermal runaway detection and improving detection accuracy. Specifically, the first-order derivative highlights the rapidly changing parts of the signal, wavelet decomposition accurately separates the slowly changing DC component and zeros it, and finally, integration restores the signal shape, enabling a more thorough and clean removal of the DC component while preserving AC details containing fault information to the maximum extent. This provides a more accurate input signal for subsequent wavelet transform and feature extraction, thus improving detection accuracy.

[0047] Step 104: Perform continuous wavelet transform on the de-DC signal to obtain the wavelet coefficients of the de-DC signal at different wavelet scales.

[0048] In this embodiment, the step mainly includes performing continuous wavelet transform on the de-DC current signal according to a preset wavelet transform basis function and multiple preset wavelet scales to obtain wavelet coefficient datasets corresponding to the de-DC current signal under different frequency band information; wherein, the frequency band information is determined according to the wavelet scale; and each original wavelet coefficient in the wavelet coefficient dataset is thresholded to use the original wavelet coefficients that have passed the thresholding as the wavelet coefficients corresponding to the de-DC signal.

[0049] Specifically, in some embodiments of this example, the deDC signal obtained after removing the DC component undergoes a continuous wavelet transform, such as using the Morlet wavelet. The deDC signal is processed according to multiple preset wavelet scales (e.g., multiple wavelet scales from 1 to 8) using a preset continuous wavelet transform function (CWT) to obtain wavelet coefficient datasets in different frequency bands. The frequency bands correspond to the wavelet scales (e.g., multiple wavelet scales from 1 to 8 correspond to multiple frequency bands from 1 kHz to 8 kHz). It should be noted that the wavelet transform function used is the Morlet wavelet. At that time, the center frequency of the Morlet wavelet was approximately 0.8 Hz, and the scaling parameter α ranged from 0.5 to 8 (corresponding to the analysis frequency). Ts is the sampling period, covering 1.25 kHz to 10 kHz. The time-domain signal is mapped to the time-frequency domain by CWT to extract fault features at different frequencies (such as early-stage features of battery micro-short circuits corresponding to 3 kHz to 8 kHz).

[0050] Specifically, the process of obtaining the wavelet coefficients is as follows: Wherein, t is a time variable, representing the time-domain dimension of the charging and discharging current signal, describing the dynamic process of current change with time t, such as... This represents the time series within the collection period; j is the imaginary unit, where j is the imaginary unit of complex number operations (satisfying...). The function α is used to construct complex-valued wavelet functions, enabling the wavelet to simultaneously possess amplitude and phase information, facilitating the analysis of instantaneous signal characteristics in the time-frequency domain. α is a scaling parameter; in this embodiment, α is used to analyze the evolution of high-frequency ripples from 1 kHz to 10 kHz. Different α values ​​correspond to different fault characteristic frequency bands, such as a scaling parameter α ranging from 0.5 to 8 (corresponding to the analysis frequency). (covering 1.25 kHz to 10 kHz); the The sampling period is the fundamental time parameter for signal acquisition and determines the frequency resolution accuracy.

[0051] Furthermore, after obtaining multiple wavelet coefficients, each wavelet coefficient undergoes thresholding, meaning the original wavelet coefficients that pass the thresholding process are used as the wavelet coefficients corresponding to the de-DC signal. It should be noted that the threshold can be set to a number of times the standard deviation σ of the de-DC signal, such as 1.5 times the standard deviation σ, to filter the wavelet coefficients using the threshold.

[0052] Let the DC signal be f(t), and calculate its standard deviation σ to quantify the signal dispersion: Where N is the number of signal sampling points (determined by the sampling frequency and duration, such as N=10000 if the sampling frequency is 1kHz and the sampling time is 10s). The average value of the DC signal; Let be the value of the i-th DC signal.

[0053] In this embodiment, after obtaining the original wavelet coefficient dataset, a thresholding step is added to the above steps. This step filters out coefficients with small amplitudes that may be caused by background noise, retaining only those key wavelet coefficients with significant amplitudes that are more likely to be generated by real fault characteristics. By highlighting the frequency band information related to the fault, subsequent reconstruction and power spectrum analysis can focus more on effective information, improving the accuracy of detection.

[0054] Step 105: Perform wavelet reconstruction on the DC-de-switched signal according to the wavelet coefficients and the wavelet scale to obtain the reconstructed signal.

[0055] In this embodiment, the step mainly includes: performing wavelet reconstruction on the DC-de-DC signal according to the wavelet coefficients corresponding to each wavelet scale, and obtaining an initial reconstructed signal; performing amplitude calibration and smoothing filtering on the initial reconstructed signal to obtain a reconstructed signal.

[0056] Specifically, in some embodiments of this example, based on the wavelet coefficients extracted by the CWT, the waverec function is used to perform wavelet reconstruction on the DC-depleted signal through the decomposition structure (scales 1 to 8) of the Morlet wavelet to obtain the reconstructed signal.

[0057] In this embodiment, after obtaining the initial reconstructed signal, the above steps perform amplitude calibration and smoothing filtering. Amplitude calibration ensures that the magnitude of the reconstructed signal matches the original signal, facilitating subsequent quantitative analysis. Smoothing filtering suppresses high-frequency glitches or noise that may be introduced during the reconstruction process. These two post-processing steps together guarantee the quality and smoothness of the final reconstructed signal, making the power spectrum calculated based on the reconstructed signal more stable and reliable, thereby improving the accuracy of detection.

[0058] Step 106: Obtain multiple power spectra corresponding to the reconstructed signal according to the wavelet scale, so as to obtain the real-time thermal runaway detection results of the battery pack under test based on the multiple power spectra.

[0059] In this embodiment, the step mainly includes: obtaining the power spectrum of the reconstructed signal at each wavelet scale according to a preset power spectral density estimation algorithm; sorting the multiple power spectra according to the wavelet scale to obtain a multi-scale power spectrum; performing time-domain statistical analysis on the multi-scale power spectrum to obtain the mean change rate of the multi-scale power spectrum and the variance of the power spectrum at each scale; and obtaining the real-time thermal runaway detection result of the battery pack under test based on the mean change rate of the multi-scale power spectrum, the variance of the power spectrum at each scale, a preset mean change rate threshold, and a preset variance threshold.

[0060] Specifically, in some embodiments of this example, based on the reconstructed signal, the Welch method is used to calculate the power spectrum of the reconstructed signal at each wavelet scale (corresponding to analysis frequencies of 1 kHz to 8 kHz) for each wavelet scale corresponding to 1 to 8. When the mean change rate of the power spectrum at multiple scales is less than a preset change rate threshold, such as less than 3% / min, or the variance of the power spectrum at each wavelet scale is less than or equal to a preset variance threshold, it is determined that the power spectrum tends to be stable. At this time, a thermal runaway warning is triggered, and a thermal runaway recovery operation is performed on the battery pack under test according to the warning.

[0061] In this first embodiment, the reconstructed signal is processed by Fast Fourier Transform (FFT) to obtain the power spectral density after averaging. Stability criterion: when the mean change rate of the multi-scale power spectrum within three consecutive sampling periods is <3% / min and the variance is ≤5%, the system is considered to have entered the risk accumulation stage. The early warning model employs an improved deep belief network, trained with 5000 normal samples (0.5C–2C charging / discharging) and 800 fault samples (simulating micro-short circuits), outputting three levels of early warning corresponding to the risk probability, to execute different thermal runaway recovery operations based on different early warnings. A yellow alert is triggered when the risk probability is between 30% and 50%, at which point a local audible and visual alarm is activated, and data is uploaded to the monitoring platform. An orange alert is triggered when the risk probability is between 50% and 80%, at which point the liquid cooling flow rate of the battery pack under test is increased by 50%, and the charging and discharging current of the battery pack under test is limited to 60% of the rated value. A red alert is triggered when the risk probability is ≥80%, at which point the main circuit of the battery pack under test is cut off, the inert gas fire extinguishing device is activated, and the EMS system is linked to transfer the load.

[0062] In this embodiment, the above steps, by monitoring the rate of change of the power spectrum mean, can keenly capture the trend of signal energy entering a stable upward channel, thereby achieving earlier warning. Simultaneously, monitoring the variance of the power spectrum at each scale can determine whether fault characteristics appear synchronously and consistently across different frequency bands, further confirming the certainty of fault occurrence. Combining trend judgment and consistency verification ensures that the final real-time thermal runaway detection results are both fast and reliable.

[0063] In this first embodiment, using as follows Figure 1 The monitoring method shown is used to test battery packs with different charging conditions in order to verify, based on the test results, that... Figure 1 The accuracy and timeliness of the thermal runaway detection method for battery packs are shown. Specifically, the original waveform of the battery under charging condition 1 is shown below. Figure 2 As shown, from Figure 2 It can be seen that this battery has been under poor load conditions for a long time, and the discharge process is severely affected by interference from equipment such as motors, and the charging conditions are also poor. Furthermore, its maximum peak current is 10.21A, the minimum is 6.31A, and the average current during charging is 8.15A; the original waveform of the battery under charging condition 2 is shown below. Figure 3 As shown. From Figure 3 It can be seen that this battery has been used in energy storage power stations for a long time, with high-quality PCS charging and discharging equipment and favorable operating environment conditions. Its maximum peak current is 7.28A, the minimum is 6.47A, and the average current during charging is 8.15A. The charging waveforms under different power quality conditions are then compared to illustrate the implementation of different solutions.

[0064] Through such Figure 1The specific results after processing by the method shown are as follows: Charging condition 1 (harsh operating conditions): After DC removal, the signal passes through CWT ( Features were extracted from 3 kHz to 5 kHz. The reconstructed signal power spectrum showed a sustained stable band at scale 3 (mean change rate 2.1% / min), with the model outputting a risk probability of 68%, triggering an orange alert. In charging scenario 2 (normal operating condition), the power spectrum showed no stable band, with a mean change rate of 12.5% / min, and the model outputting a risk probability of 11%, with no alert. Validation shows that... Figure 1 The method shown can identify thermal runaway risks 15-20 minutes in advance, which is 8-10 minutes earlier than traditional temperature warnings. In summary, the case studies illustrate how different charge / discharge qualities can be addressed under two charging scenarios. Figure 1 The method shown can effectively distinguish between dangerous trends and general disturbances, enabling accurate and timely detection of thermal runaway in the battery pack under test, and providing strong protection for the safe operation of the battery pack.

[0065] like Figure 4 As shown, based on the above method embodiments, this embodiment also provides corresponding device embodiments; this embodiment provides a thermal runaway detection system for battery packs, including a feature extraction module 201, a preliminary detection module 202, a DC removal module 203, a wavelet transform module 204, a wavelet reconstruction module 205, and a real-time detection module 206.

[0066] The feature extraction module 201 is used to acquire the charging and discharging current signal of the battery pack under test, and extract the high-frequency ripple, current fluctuation rate and harmonic distortion rate of the charging and discharging current signal.

[0067] The preliminary detection module 202 is used to obtain the preliminary thermal runaway detection results of the battery pack under test based on the high-frequency ripple, the current fluctuation rate and the harmonic distortion rate.

[0068] The DC removal module 203 is used to remove the DC component from the charge / discharge current signal based on the preliminary thermal runaway detection results to obtain a DC removal signal.

[0069] The wavelet transform module 204 is used to perform continuous wavelet transform on the de-DC signal to obtain the wavelet coefficients of the de-DC signal at different wavelet scales.

[0070] The wavelet reconstruction module 205 is used to perform wavelet reconstruction on the DC-de-switched signal according to the wavelet coefficients and the wavelet scale to obtain the reconstructed signal.

[0071] The real-time detection module 206 is used to obtain multiple power spectra corresponding to the reconstructed signal according to the wavelet scale, so as to obtain the real-time thermal runaway detection result of the battery pack under test according to the multiple power spectra.

[0072] It is understood that the above-described device embodiment corresponds to the method embodiment, and can implement the thermal runaway detection method for battery packs provided by the method embodiment.

[0073] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided in this embodiment, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0074] Based on the above embodiment of a method for detecting thermal runaway in a battery pack, this embodiment also provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a method for detecting thermal runaway in a battery pack according to the method embodiment.

[0075] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete this embodiment. The one or more module units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0076] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0077] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0078] Based on the above-described method embodiments, this embodiment also provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the thermal runaway detection method for a battery pack described in the method embodiments.

[0079] The modules / units integrated in the aforementioned device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0080] This embodiment provides a method, system, device, and medium for detecting thermal runaway in battery packs. By removing the DC component and eliminating DC baseline interference from the battery's rated charging and discharging current, it highlights the μA-level current fluctuation characteristics caused by the early stage of thermal runaway (micro-short circuit / internal resistance anomaly), solving the problem of weak fault signals being easily submerged under strong DC background. Wavelet transform (CWT) maps the time-domain current signal to a two-dimensional time-frequency space, and analyzes the transient evolution law of high-frequency ripple from 1 kHz to 10 kHz through the scale parameter (α), accurately capturing the frequency component distortion in the early stage of thermal runaway. Wavelet reconstruction, based on the screened fault feature components (such as α=2~6 corresponding to the 3 kHz~8 kHz frequency band), restores the enhanced current signal containing early fault information, providing a clean input for subsequent power spectrum analysis. Through power spectrum detection, the frequency energy distribution of the reconstructed signal is quantified. When the mean fluctuation of the multi-scale power spectrum is less than 3% / min, it is determined that the risk of thermal runaway has entered the energy accumulation steady state stage, realizing "earlier electrical dimension early warning than temperature rise / gas production", breaking through the hysteresis limitation of traditional reliance on thermodynamic / chemical parameters. By learning from the above, we can detect the potential for thermal runaway within the battery as early as possible and take appropriate measures to prevent personal injury and damage to lithium-ion battery equipment caused by fires and explosions due to thermal runaway.

[0081] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for detecting thermal runaway in battery packs, characterized in that, include: Acquire the charging and discharging current signal of the battery pack under test, and extract the high-frequency ripple, current fluctuation rate and harmonic distortion rate of the charging and discharging current signal; The preliminary thermal runaway detection results of the battery pack under test are obtained based on the high-frequency ripple, the current fluctuation rate, and the harmonic distortion rate. Based on the preliminary thermal runaway detection results, the DC component in the charge / discharge current signal is removed to obtain the de-DC signal; Perform continuous wavelet transform on the de-DC signal to obtain the wavelet coefficients of the de-DC signal at different wavelet scales; The DC-de-signal is reconstructed using wavelet coefficients and wavelet scale to obtain the reconstructed signal. Multiple power spectra corresponding to the reconstructed signal are obtained based on the wavelet scale, so as to obtain the real-time thermal runaway detection result of the battery pack under test based on the multiple power spectra.

2. The method for detecting thermal runaway in a battery pack as described in claim 1, characterized in that, The process of acquiring the charging and discharging current signal of the battery pack under test and extracting the high-frequency ripple, current fluctuation rate, and harmonic distortion rate of the charging and discharging current signal includes: Acquire the charging current signal, discharging current signal, and reverse current signal during charge / discharge switching of the battery pack under test, and use the charging current signal, the discharging current signal, and the reverse current signal as the charging / discharging current signal of the battery pack under test; Extract the total current signal of the total input terminal of the battery pack under test and the branch current signal of the branch node of the battery pack under test from the charging and discharging current signal; wherein, the branch node is deployed in the charging circuit of the battery pack under test; The signal difference between the total current signal and the branch current signal is obtained, and the signal difference is filtered to obtain the high-frequency ripple corresponding to the branch node. The ratio of the signal difference to the total current signal is obtained, and the ratio is determined as the current fluctuation rate of the battery pack under test. Fourier transforms are performed on the total current signal and the branch current signal respectively to obtain the total harmonic distortion rate corresponding to the total current signal and the branch harmonic distortion rate corresponding to the branch current signal. The difference between the total harmonic distortion rate and the branch harmonic distortion rate is obtained, and the harmonic distortion rate of the battery pack under test is determined based on the difference.

3. The method for detecting thermal runaway in a battery pack according to claim 2, characterized in that, The preliminary thermal runaway detection results of the battery pack under test obtained based on the high-frequency ripple, the current fluctuation rate, and the harmonic distortion rate include: Obtain the first comparison result between the high-frequency ripple and the preset ripple threshold; Obtain a second comparison result between the current volatility and a preset volatility threshold; Obtain a third comparison result between the harmonic distortion rate and a preset distortion rate threshold; Based on the first comparison result, the second comparison result, and the third comparison result, the preliminary thermal runaway detection result of the battery pack under test is determined.

4. The method for detecting thermal runaway in a battery pack according to claim 3, characterized in that, The step of removing the DC component from the charge / discharge current signal based on the preliminary thermal runaway detection results to obtain a de-DC signal includes: When it is determined that the battery pack under test has a risk of thermal runaway based on the preliminary thermal runaway detection results, the first derivative of the charging and discharging current signal is calculated to obtain the current conversion rate signal. The current conversion rate signal is subjected to multi-level decomposition wavelet transform according to a preset wavelet function to obtain a wavelet coefficient dataset with different frequency sub-bands. The wavelet coefficients corresponding to the DC component in the wavelet coefficient dataset are set to zero, and wavelet reconstruction is performed using the processed wavelet coefficient dataset to obtain the current change rate signal after DC removal. The de-DC current rate of change signal is integrated to obtain the de-DC current signal corresponding to the charging and discharging current signal.

5. The method for detecting thermal runaway in a battery pack according to claim 4, characterized in that, The step of performing continuous wavelet transform on the de-DC signal to obtain the wavelet coefficients of the de-DC signal at different wavelet scales includes: The DC-DC current signal is subjected to continuous wavelet transform based on preset wavelet transform basis functions and multiple preset wavelet scales to obtain wavelet coefficient datasets corresponding to the DC-DC current signal under different frequency band information; wherein, the frequency band information is determined according to the wavelet scales. Each original wavelet coefficient in the wavelet coefficient dataset is thresholded so that the original wavelet coefficient that has passed the thresholding is used as the wavelet coefficient corresponding to the DC-DC signal.

6. The method for detecting thermal runaway in a battery pack according to claim 5, characterized in that, The step of performing wavelet reconstruction on the DC-de-switched signal based on the wavelet coefficients and the wavelet scale to obtain the reconstructed signal includes: Based on the wavelet coefficients corresponding to each wavelet scale, the DC-de-switched signal is reconstructed using a preset wavelet reconstruction algorithm to obtain an initial reconstructed signal. The initial reconstructed signal is subjected to amplitude calibration and smoothing filtering to obtain the reconstructed signal.

7. The method for detecting thermal runaway in a battery pack according to claim 6, characterized in that, The step of obtaining the power spectrum of the reconstructed signal at different frequencies according to the wavelet scale, and obtaining the real-time thermal runaway detection result of the battery pack under test according to the power spectrum, includes: The power spectrum of the reconstructed signal at each wavelet scale is obtained according to the preset power spectral density estimation algorithm. The power spectra are sorted according to the wavelet scale to obtain a multi-scale power spectrum; Time-domain statistical analysis was performed on the multi-scale power spectrum to obtain the mean change rate of the multi-scale power spectrum and the variance of the power spectrum at each scale. The real-time thermal runaway detection results of the battery pack under test are obtained based on the mean change rate of the multi-scale power spectrum, the variance of the power spectrum at each scale, and the preset mean change rate threshold and the preset variance threshold.

8. A thermal runaway detection system for battery packs, characterized in that, It includes a feature extraction module, a preliminary detection module, a DC removal module, a wavelet transform module, a wavelet reconstruction module, and a real-time detection module; The feature extraction module is used to acquire the charging and discharging current signal of the battery pack under test, and extract the high-frequency ripple, current fluctuation rate and harmonic distortion rate of the charging and discharging current signal. The preliminary detection module is used to obtain the preliminary thermal runaway detection results of the battery pack under test based on the high-frequency ripple, the current fluctuation rate, and the harmonic distortion rate. The DC removal module is used to remove the DC component from the charge / discharge current signal based on the preliminary thermal runaway detection results, to obtain a DC removal signal. The wavelet transform module is used to perform continuous wavelet transform on the de-DC signal to obtain the wavelet coefficients of the de-DC signal at different wavelet scales. The wavelet reconstruction module is used to perform wavelet reconstruction on the DC-de-DC signal according to the wavelet coefficients and the wavelet scale to obtain the reconstructed signal. The real-time detection module is used to obtain multiple power spectra corresponding to the reconstructed signal according to the wavelet scale, so as to obtain the real-time thermal runaway detection result of the battery pack under test according to the multiple power spectra.

9. A terminal device, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements a thermal runaway detection method for a battery pack as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform a thermal runaway detection method for a battery pack as described in any one of claims 1-7.

Images