Battery thermal runaway early warning diagnosis method based on acoustic energy integration

Abstract

The invention discloses an acoustic diagnosis method for early warning of thermal runaway of a power battery. Relates to the field of battery energy storage safety monitoring. According to the method, an attached electronic stethoscope is arranged on the outer side of a battery to serve as a single acoustic sensor, and original sound signals in the battery running process are collected; carrying out direct current removal, band-pass filtering and normalization preprocessing on the sound signal; selecting a background noise segment in an initial or static stable working condition, calculating a background energy threshold and setting an energy upper limit; instantaneous energy is calculated by adopting a sliding window, the instantaneous energy is added to an acoustic energy integral AEI only when window energy is met, and early warning is triggered at the moment; wherein the integral alarm threshold value is obtained in a self-adaptive mode according to the robust statistic of the background section and is used for avoiding accidental strong impact false triggering. The method is simple to deploy and low in cost, can realize early warning in advance when weak events such as swelling, micro gas permeation and initial side reaction occur, and has the capability of resisting environmental noise interference and good generalization.

Description

Technical Field

[0001] This invention relates to the field of battery energy storage safety monitoring, and in particular to an early warning and diagnosis method for battery thermal runaway based on acoustic energy integration. Background Technology

[0002] In recent years, with the widespread application of electric vehicles, energy storage power stations, and portable electronic devices, the installed capacity of lithium-ion batteries has continued to expand, and their safety has become increasingly important. Lithium-ion batteries may experience thermal runaway under conditions such as overcharging, over-discharging, short circuits, mechanical shock, or excessively high ambient temperatures. Thermal runaway is a catastrophic failure process caused by the rapid accumulation of heat inside the battery and its self-accelerating reaction. It is often accompanied by intense heat release, fire, or even explosion, which not only damages equipment but also poses a serious threat to human life and the surrounding environment.

[0003] Currently, early warning of battery thermal runaway mainly relies on methods such as temperature monitoring, voltage / current anomaly detection, and gas sensing. However, these methods typically only trigger alarms after significant internal reactions have occurred, temperatures have risen markedly, or large amounts of flammable gases have been released. The warning time is relatively delayed, leaving limited time for the system to take emergency measures.

[0004] Research has found that before a battery enters the early stages of thermal runaway, internal side reactions and structural changes trigger weak but distinctive acoustic signals. These signals differ significantly from those under normal operating conditions in terms of frequency domain and energy integral characteristics, and they appear earlier than traditional monitoring indicators such as temperature rise and gas release. Therefore, detection methods based on acoustic signal analysis can provide early warnings in the very early stages of thermal runaway, buying valuable response time for the battery management system (BMS) and safety measures, thereby significantly reducing the risk of accidents. Summary of the Invention

[0005] This invention provides an early warning and diagnosis method for battery thermal runaway based on acoustic energy integration to solve the problems of delayed warning, untimely response, and inability to effectively identify anomalies in the early stage of thermal runaway in existing battery thermal runaway monitoring methods.

[0006] In view of the above problems, the present invention provides a method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration, comprising: S1: The sound sensor is attached and installed as a single acoustic sensor on the outside of the battery to collect the raw sound signal; S2: Preprocess the audio signal by removing DC, bandpass filtering and amplitude normalization; S3: Select a background noise segment under initial or stable operating conditions, calculate its energy sequence using a sliding window, and determine the background energy threshold based on the energy sequence. And set an energy limit. ; S4: Calculate the instantaneous energy of the preprocessed signal according to the window length L and step size S. ; S5: When satisfied At that time, Accumulated acoustic energy integral (AEI); S6: When The system outputs an early warning signal for battery thermal runaway, in which... This is the threshold for the integrated alarm.

[0007] The technical solution provided in this application has at least the following technical effects or advantages: This invention utilizes energy integration analysis of acoustic signals to identify early signs of thermal runaway by monitoring changes in the acoustic signals of the battery system before it occurs. This method can make an early diagnosis by capturing subtle acoustic changes within the battery, even before significant changes in temperature and gas concentration are apparent, thus solving the problem of untimely warnings caused by the lag in the response of temperature and gas sensors in existing technologies.

[0008] Specifically, this limitation of being unable to identify thermal runaway in its early stages improves the accuracy and reliability of diagnosis. Furthermore, by integrating into the battery management system (BMS) in real time, it can quickly respond and activate corresponding safety protection measures, effectively preventing major safety incidents such as fires or explosions caused by thermal runaway. Attached Figure Description

[0009] Figure 1 This is a flowchart of an early warning and diagnosis method for battery thermal runaway based on acoustic energy integration, provided by an embodiment of the present invention.

[0010] Figure 2 This is a diagram showing the warning effect of a 314Ah lithium iron phosphate battery under 0.5C overcharge condition, provided in the first embodiment of the present invention.

[0011] Figure 3 These are the first set of time-domain and time-frequency domain diagrams of a 314Ah lithium iron phosphate battery under overcharge conditions provided in this embodiment of the invention. Detailed Implementation

[0012] The above technical solutions will now be described in detail with reference to the accompanying drawings and specific embodiments to provide a better understanding of them. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments used only to explain the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Furthermore, it should be noted that, for ease of description, only the parts related to the present invention are shown in the drawings, not all of them.

[0013] Please see Figure 1 This paper provides an early warning and diagnostic method for battery thermal runaway based on acoustic energy integration, including: S1: The sound sensor is attached and installed as a single acoustic sensor on the outside of the battery to collect the raw sound signal; S2: Preprocess the audio signal by removing DC, bandpass filtering and amplitude normalization; S3: Select a background noise segment under initial or stable operating conditions, calculate its energy sequence using a sliding window, and determine the background energy threshold based on the energy sequence. And set an energy limit. ; S4: Calculate the instantaneous energy of the preprocessed signal according to the window length L and step size S. ; S5: When satisfied At that time, Accumulated acoustic energy integral (AEI); S6: When The system outputs an early warning signal for battery thermal runaway, in which... This is the threshold for the integrated alarm.

[0014] The steps described in step S3 The median method is used to determine the background energy sequence, that is, the median of the background energy sequence is taken as the median. And the energy upper limit Determined by one of the following methods: (a) ,in ; (b) Take the upper limit of 30% to 80% in the normalized energy range.

[0015] The instantaneous energy is calculated using the following formula: ,in The signal is the preprocessed discrete signal, and N is the number of sampling points corresponding to the window length.

[0016] The window length L is 0.05 to 0.20 s, and the step size S is 0.02 to 0.10 s.

[0017] Duration of the background noise segment The time interval is 30 to 120 seconds, and the background fragments are updated according to the set period to adapt to environmental changes.

[0018] The integral alarm threshold It adaptively adjusts according to battery casing temperature, state of charge (SOC), or load intensity.

[0019] After the warning is triggered, the AEI is reset or decreased at a preset rate, and the original acoustic segments and threshold parameters of several windows before and after the trigger are recorded for source tracing.

[0020] After the warning is output in step S6, the system will work in conjunction with the battery management system (BMS) to execute control commands to reduce power, stop charging, or enter a safe mode.

[0021] This embodiment conducts two independent tests on 314Ah and 280Ah lithium iron phosphate batteries under 0.5C overcharge conditions to verify the early warning method for thermal runaway based on acoustic energy integration. It achieves a stable warning about 6 minutes before the thermal runaway valve opens, without relying on temperature or gas sensors throughout the process.

[0022] In the experiment, an electronic stethoscope was selected as the acoustic sensor, which was attached to the center of one side of the battery casing with tape to ensure uniform acquisition of acoustic signals from the large-capacity battery. The data acquisition unit captured characteristic acoustic signals such as SEI film decomposition and electrolyte micro-permeation during overcharging in real time at a sampling rate of 44.1kHz and a precision of 16 bits. The signals were preprocessed by first-order high-pass filtering (cutoff 0.1Hz) to eliminate DC drift, 100Hz–10kHz FIR filtering to extract characteristic frequency bands, and amplitude normalization (scaled to the [-1,1] interval) to unify the dimensions. To distinguish between effective signals and background noise, a stable segment of 300s before the battery was operating normally was selected. The instantaneous energy sequence was calculated with a 0.1s window (4410 sampling points) and a step size of 0.05s. The background threshold was calibrated using the median absolute deviation (MAD) algorithm. (Corresponding to the red dashed line in the "Instantaneous Energy Changes Over Time" graph, covering the 95% confidence interval), and set... (Blue dashed line) Filter housing collisions and other occasional strong interference. For the preprocessed signal, the instantaneous energy (the sum of squares of the normalized signal within the window) is calculated using the same window and step size, only when... The acoustic energy integral (AEI) is added over time to ensure that low energy fluctuations under normal operating conditions do not trigger invalid accumulation. The warning threshold is calibrated experimentally. (Corresponding to the dotted line in the "Cumulative AEI Changes Over Time" graph), the first test triggered an alarm at 363.7 seconds and the second at 473.6 seconds, both approximately 6 minutes ahead of the valve opening time; immediately after the alarm, the charging circuit was cut off, and the original acoustic data and threshold parameters for 1 second before and after the alarm were stored. To adapt to environmental noise drift, stable operating conditions (current fluctuation <5%, temperature change <2℃) were verified every 30 minutes, and a new 5-second segment was extracted to recalculate the MAD update. This ensures dynamic adaptability of the threshold. From the waveform characteristics, the instantaneous energy map... and The method accurately characterizes noise boundaries and interference thresholds. The orange trend line of the cumulative AEI curve clearly shows the energy accumulation pattern in the early stage of thermal runaway. The purple dashed line marks the early warning time, verifying the stable early warning performance of the method in two tests. This highlights the technological innovation of not requiring additional sensors and fills the early warning gap in the overcharge scenario of large-capacity lithium iron phosphate batteries.

[0023] Please see the appendix Figure 2 It displays the early warning effect diagram of the first set of tests. In the "Instantaneous Energy Changes Over Time" graph above, the dashed line... and dashed lines An effective range for energy integration was precisely constructed, effectively filtering out background noise and occasional strong interference. In the "Cumulative AEI vs. Time" graph below, the solid line shows a clear, non-linear increasing trend in the AEI value before the warning, indicating that the system is capturing the energy accumulation caused by continuous internal side reactions. Ultimately, the AEI curve exceeded the warning threshold indicated by the dashed line at 363.7 seconds (marked by the purple dashed line). The warning was successfully triggered. This occurred approximately 6 minutes before the actual battery valve opened, demonstrating the lead time and effectiveness of this method.

[0024] Please see the appendix Figure 3 It reveals the intrinsic mechanism of early warning from the perspective of the physical characteristics of the signal. (Appendix) Figure 3 The spectrum below shows that as overcharging progresses, a significant increase in energy over time occurs within the characteristic frequency band of 100Hz to 10kHz, manifested as multiple vertical bright bands. These characteristic signals correspond to the weak acoustic events generated during the initial side reactions inside the battery, such as SEI film decomposition and trace gas production. The method of this invention effectively integrates the energy of these weak but persistent characteristic signals, thereby capturing early signs of thermal runaway before significant changes occur in macroscopic physical quantities (such as temperature and voltage). The second set of tests also exhibits a highly similar pattern, verifying the stability and repeatability of the method.

[0025] In one embodiment, to further improve the robustness of the early warning model and its adaptability to different operating conditions, the early warning threshold is... It is not fixed, but rather adaptively adjusted based on battery state parameters obtained in real time from the battery management system (BMS). This adaptive adjustment mechanism can be specifically implemented as follows: The processor unit has a built-in preset adjustment module that dynamically adjusts the baseline threshold based on the battery casing temperature (T) and state of charge (SOC). One specific implementation method is to use a multi-dimensional lookup table. A two-dimensional lookup table is established through offline experiments to store the adjustment coefficient K(T,SOC) for different temperature and SOC ranges. For example, when the battery is in a high-risk condition with high temperature (e.g., T>45°C) and high SOC (e.g., SOC>80%), the K value is less than 1 (e.g., 0.8), thereby reducing the warning threshold and improving warning sensitivity; while in a low-risk condition with low temperature and low SOC, the K value can be greater than 1 (e.g., 1.2) to avoid false alarms caused by environmental noise.

[0026] By introducing this adaptive adjustment mechanism, the present invention can intelligently balance the timeliness and anti-interference of early warning while ensuring safety, thus greatly improving the reliability of the early warning system.

[0027] This detailed embodiment elaborates on the internal operating logic of each major functional module, aiming to provide a detailed basis and explanation for those skilled in the art to understand and implement it. It should be emphasized that the above description constitutes a specific, preferred embodiment, but the concept of the present invention is not limited thereto. Any equivalent transformations, modifications, or improvements based on the core spirit of the present invention, without departing from the technical principles and scope disclosed in this specification, should be considered to fall within the scope of protection claimed by the present invention, as long as they achieve the same or similar technical effects.

Claims

1. A method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration, characterized in that, include: S1: The sound sensor is attached and installed as a single acoustic sensor on the outside of the battery to collect the raw sound signal; S2: Preprocess the audio signal by removing DC, bandpass filtering and amplitude normalization; S3: Select a background noise segment under initial or stable operating conditions, calculate its energy sequence using a sliding window, and determine the background energy threshold based on the energy sequence. And set an energy limit. ; S4: Calculate the instantaneous energy of the preprocessed signal according to the window length L and step size S. ; S5: When satisfied At that time, Accumulated acoustic energy integral (AEI); S6: When The system outputs an early warning signal for battery thermal runaway, in which... This is the threshold for the integrated alarm.

2. The method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration according to claim 1, characterized in that, The steps described in step S3 The median method is used to determine the background energy sequence, that is, the median of the background energy sequence is taken as the median. And the energy upper limit Determined by one of the following methods: (a) ,in ; (b) Take the upper limit of 30% to 80% in the normalized energy range.

3. The method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration according to claim 1, characterized in that, The instantaneous energy is calculated using the following formula: ,in The signal is the preprocessed discrete signal, and N is the number of sampling points corresponding to the window length.

4. The method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration according to claim 1, characterized in that, The window length L is 0.05 to 0.20 s, and the step size S is 0.02 to 0.10 s.

5. The method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration according to claim 1, characterized in that, Duration of the background noise segment The time interval is 30 to 120 seconds, and the background fragments are updated according to the set period to adapt to environmental changes.

6. The method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration according to claim 1, characterized in that, The integral alarm threshold It adaptively adjusts according to battery casing temperature, state of charge (SOC), or load intensity.

7. The method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration according to claim 1, characterized in that, After the warning is triggered, the AEI is reset or decreased at a preset rate, and the original acoustic segments and threshold parameters of several windows before and after the trigger are recorded for source tracing.

8. The method for early warning and diagnosis of battery thermal runaway based on acoustic energy integration according to claim 1, characterized in that, After the warning is output in step S6, the system will work in conjunction with the battery management system (BMS) to execute control commands to reduce power, stop charging, or enter a safe mode.

9. A battery acoustic early warning device, characterized in that, include: Acoustic sensors, signal acquisition and processing units, memory and communication interfaces; The processing unit is configured to perform the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, the program implementing the method of any one of claims 1 to 8 when executed on a processor.

Images