Battery pack detection method, detection device, electronic equipment and storage medium

Abstract

This invention relates to the field of energy storage technology, and provides a method, device, electronic equipment, and storage medium for detecting battery packs. The invention includes: pre-pressurizing the battery pack under test and acquiring multi-channel acoustic signals from the pressurized battery pack using a microphone array; performing noise reduction processing on the multi-channel acoustic signals to obtain target acoustic signals; determining the signal characteristics of each detection location on the battery pack under test based on the target acoustic signals, and matching the signal characteristics of each detection location with a pre-stored leakage feature library to obtain the leakage situation corresponding to each detection location; and visually presenting the leakage situation corresponding to each detection location based on a visible light image of the battery pack under test. This addresses the shortcomings of existing battery pack airtight leakage detection methods in identifying minute leaks, spatially locating leak points, and providing intuitive traceability of detection results.

Description

Technical Field

[0001] This invention relates to the field of energy storage technology, and in particular to a method, device, electronic equipment, and storage medium for testing battery packs. Background Technology

[0002] With the rapid development of new energy vehicles and energy storage, the production capacity and installed capacity of power battery packs and energy storage battery packs continue to grow. Battery pack airtightness is one of the key indicators affecting the lifespan and operational safety of battery systems. Airtightness leaks inside the battery pack can cause internal components to become damp and their insulation performance to deteriorate, leading to safety risks. Therefore, testing, locating, and generating traceable test results for battery pack airtightness is a routine procedure in the factory inspection and after-sales troubleshooting of battery packs.

[0003] The battery pack airtight leakage detection methods in related technologies still have shortcomings in terms of identifying minute leaks, locating leak points in space, and providing intuitive traceability of detection results. Summary of the Invention

[0004] This invention provides a method, device, electronic equipment, and storage medium for detecting battery packs, which helps to solve the problems of insufficient detection methods for airtight leaks in battery packs in terms of identifying minute leaks, locating leak points in space, and providing intuitive traceability of detection results.

[0005] This invention provides a method for detecting a battery pack, comprising: pre-pressurizing the battery pack under test and acquiring multi-channel acoustic signals from the pressurized battery pack using a microphone array; performing noise reduction processing on the multi-channel acoustic signals to obtain target acoustic signals; determining the signal characteristics of each detection location on the battery pack under test based on the target acoustic signals, and matching the signal characteristics of each detection location with a pre-stored leakage feature library to obtain the leakage situation corresponding to each detection location; and visually presenting the leakage situation corresponding to each detection location based on a visible light image of the battery pack under test.

[0006] Another aspect of this invention provides a battery pack testing device, comprising: a pressurization module for pressurizing and preprocessing the battery pack under test; an acoustic wave acquisition module including a microphone array for acquiring multi-channel acoustic wave signals from the pressurized battery pack under test; a noise reduction module for performing noise reduction processing on the multi-channel acoustic wave signals to obtain a target acoustic wave signal; an identification module for determining the signal characteristics of each detection location on the battery pack under test based on the target acoustic wave signal, and matching the signal characteristics of each detection location with a pre-stored leakage feature library to obtain the leakage status corresponding to each detection location; and an output module for visually outputting the leakage status corresponding to each detection location based on a visible light image of the battery pack under test.

[0007] Another aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the steps of the battery pack detection method as described in any of the preceding claims.

[0008] In another aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the battery pack detection method as described in any of the preceding claims.

[0009] The technical solution provided by the embodiments of the present invention has at least the following advantages: By pressurizing and pre-processing the battery pack under test, the gas escaping from the leaking part is excited by the internal and external pressure difference to generate high-frequency sound waves that can be acoustically picked up. These high-frequency sound waves are picked up by a microphone array in a multi-channel manner, and after noise reduction processing to increase the proportion of the leaking sound waves relative to the workshop environment noise, they are matched with a pre-stored leak feature library. This allows even small leaks with small leakage amounts and weak acoustic features to be identified, which is beneficial to improving the ability to identify small leaks. Since the above matching is performed separately for each detection position on the battery pack under test, the identification results are organized according to the detection position and naturally contain the spatial distribution information of the leak point. Thus, while determining whether a leak exists, the specific spatial location of the leak point is determined, which is beneficial to realizing the spatial localization of the leak point. Furthermore, based on the visible light image of the battery pack under test, the leakage situation corresponding to each detection position is visualized, so that the acoustic identification results and the physical shape of the battery pack are established with an intuitive spatial correspondence. The detection results can be intuitively read by the inspection personnel and traced back to the specific location, which is beneficial to the intuitive traceability of the detection results. Attached Figure Description

[0010] One or more embodiments are illustrated by way of example with corresponding pictures in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Unless otherwise stated, the pictures in the accompanying drawings do not constitute a limitation on scale. In order to more clearly illustrate the technical solutions in the embodiments of this application or in the conventional technology, 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 based on these drawings without creative effort.

[0011] Figure 1 A flowchart illustrating a battery pack detection method according to an embodiment of the present invention; Figure 2 This is a structural block diagram of a battery pack detection device provided in one embodiment of the present invention; Figure 3 This is a structural block diagram of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of 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.

[0013] In the description of the embodiments of this invention, 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 invention, "multiple" means two or more, unless otherwise explicitly defined.

[0014] 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 the invention. 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.

[0015] In the description of the embodiments of this invention, the term "and / or" is merely a description of the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: the existence of A, the simultaneous existence of A and B, and the existence of B. In addition, the character " / " in this document generally indicates that the related objects before and after are in an "or" relationship.

[0016] In the description of the embodiments of the present invention, the technical terms "center", "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "axial", "radial", "circumferential", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of the present invention.

[0017] In the description of the embodiments of the present invention, unless otherwise explicitly 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 the present invention can be understood according to the specific circumstances.

[0018] In the description of embodiments of the present invention, the terms "about," "approximately," "roughly," or "about" for a value referring to a specific parameter include that the value is within acceptable tolerances for the specific parameter and will be understood by those skilled in the art. For example, "about" or "about" for a value may include additional values ​​that are in the range of 90.0% to 110.0% of the value, such as 95.0% to 105.0%, 97.5% to 102.5%, 99.0% to 101.0%, 99.5% to 100.5%, or 99.9% to 100.1%.

[0019] The term "electrical connection" as used in this invention refers to the connection relationship between two components that enables the transmission of electrical signals or electrical energy through a conductive medium, including but not limited to direct electrical contact, connection through wires, connection through printed circuits, or connection through electrical connectors.

[0020] The “microphone array” referred to in this invention refers to an acoustic sensor set consisting of two or more microphone units arranged in a predetermined spatial geometric relationship, with each microphone unit outputting the corresponding channel’s acoustic wave signal under the same time reference.

[0021] The "detection location" referred to in this invention refers to a predetermined spatial location point or area on the outer surface of the battery pack to be tested or corresponding to its sealing structure, including but not limited to locations where there may be weak points in airtightness, such as shell welds, sealant seams, terminal interfaces, liquid injection ports, and explosion-proof valves.

[0022] The “signal characteristics” referred to in this invention are characteristic quantities that can be used to distinguish between leaking and non-leaking signals after mathematical transformation or statistical analysis of acoustic signals, including but not limited to time-domain statistics, frequency-domain energy distribution, and time-frequency joint distribution.

[0023] The "leakage feature library" referred to in this invention refers to a set of leakage acoustic signal features that are pre-collected, labeled and stored, corresponding to different shell materials, different leakage levels and different leakage locations.

[0024] With the rapid development of new energy vehicles and energy storage, the production capacity and installed capacity of power battery packs and energy storage battery packs continue to grow. Battery pack airtightness is one of the key indicators affecting the lifespan and operational safety of battery systems. Airtightness leaks inside the battery pack can cause internal components to become damp, reduce insulation performance, and subsequently lead to safety risks such as thermal runaway. Therefore, detecting battery pack airtightness, locating leaks, and generating traceable test results are routine procedures in battery pack factory inspection and after-sales troubleshooting.

[0025] Battery pack airtightness leak detection methods in related technologies, such as differential pressure method combined with manual application of soapy water, and helium mass spectrometry leak detection combined with manual suction gun scanning, still have shortcomings in terms of identifying minute leaks, spatially locating leak points, and intuitively tracing detection results. After disassembling and analyzing the failure cases of related technologies, it was found that the shortcomings of these technologies did not stem from the limits of detection sensitivity itself, but rather from the following three hidden root causes: First, the physical media (soapy water, helium) used in the detection process are strongly coupled with the detection environment. The environmental adaptability of the media and the handling of residues are difficult to guarantee under the production line cycle, resulting in the consistency of the detection results being affected by fluctuations in operating conditions. Second, most related technologies separate the qualitative judgment of "whether there is a leak" and the spatial location of "where the leak is" into two sequential steps. First, a comprehensive judgment is made based on differential pressure or helium concentration, and then a local scan is performed manually. This sequential structure causes spatial information to be lost in the judgment stage, and the location stage must reconstruct it based on the lost spatial information, which limits both efficiency and accuracy. Third, the output results of related technologies are mainly numerical readings or manual markings, which lack a direct spatial correspondence with the actual physical morphology of the battery pack, making it difficult to form traceable and verifiable electronic archives of the detection results.

[0026] This invention identifies the root causes mentioned above and proposes a detection scheme based on multi-channel acoustic acquisition and visible light superposition, which simultaneously improves three aspects: environmental coupling of the physical medium, serial loss in discrimination and positioning, and lack of correspondence between the output and the physical morphology.

[0027] Example 1 Figure 1 This is a flowchart illustrating a battery pack detection method according to an embodiment of the present invention. See also... Figure 1 The method includes steps S100 to S400.

[0028] Step S100: Pre-process the battery pack under test by pressurizing it, and collect multi-channel acoustic signals from the pressurized battery pack using a microphone array.

[0029] Pressurization pretreatment refers to applying a stable gas pressure higher than ambient pressure to the inside of the battery pack under test before the test begins. This causes the gas inside the battery pack to escape through any possible leaks, and the pressure difference between the inside and outside of these leaks generates high-frequency sound waves that can be picked up by acoustic sensors. In one example, the pressurizing medium is a dry inert gas or dry air, and the dew point of the pressurizing medium is lower than the minimum operating temperature of the environment in which the battery pack is located, to prevent water vapor in the medium from condensing during the pressurization process.

[0030] Optionally, the pressurizing medium can be one or more of dry nitrogen, dry air, dry argon, or dry helium, and the purity of the pressurizing medium can be from 95% to 99.999%, for example, 95%, 98%, 99%, 99.9%, or 99.999%. Using a dry inert gas or dry air as the pressurizing medium helps to prevent moisture from entering the battery pack and corroding the cells and metal interfaces, and its chemical properties are stable and it is not easy to react with the internal materials of the battery pack. This invention does not limit the specific type of pressurizing medium.

[0031] Optionally, the pressurizing medium can be from 0.05 MPa to 0.5 MPa, for example, 0.05 MPa, 0.1 MPa, 0.2 MPa, 0.3 MPa, or 0.5 MPa. The specific value of the pressurizing pressure only needs to meet the allowable pressure of the battery pack casing and the seals under test, and this invention does not limit it.

[0032] Optionally, the pressurization pretreatment also includes a pressure holding step after the pressurization pressure reaches the target value. The pressure holding time can be from 3 seconds to 60 seconds, for example, 3 seconds, 5 seconds, 10 seconds, 15 seconds, 20 seconds, 30 seconds or 60 seconds. The pressure holding time is used to allow the gas escape process in the leakage channel to enter a stable state, so as to avoid directly collecting sound waves during the transient stage of pressure establishment and introducing non-stationary transient components.

[0033] Multichannel acoustic signals refer to a collection of independent acoustic signal channels output by each microphone unit in a microphone array at the same time reference. In one example, the microphone array includes two or more microphone units, which are arranged in space according to a predetermined geometric relationship.

[0034] Optionally, the number of microphone units can be 4, 8, 16, 24, 32, 48, or 64, and the spatial arrangement of the microphone array can be any one of linear array, planar array, circular array, spherical array, or random array. A larger number of microphone units is beneficial for improving spatial resolution. Different arrangements adapt to the geometric spatial distribution and pickup direction requirements of different detection stations. This invention does not limit the specific number or arrangement of microphone units.

[0035] Optionally, the microphone unit is a microelectromechanical system (MEMS) microphone, and the sampling rate of the microphone unit can be from 16kHz to 192kHz, for example, 16kHz, 48kHz, 96kHz, or 192kHz; the frequency response range of the microphone unit can be from 2kHz to 100kHz, and the frequency band of interest for the microphone array to acquire sound wave signals can be from 3kHz to 60kHz, for example, 3kHz, 5kHz, 10kHz, 20kHz, 30kHz, 50kHz, or 60kHz. The specific value of the frequency band of interest is adapted to the leakage characteristic frequency band corresponding to the sealing structure of the battery pack under test, and this invention does not limit it.

[0036] In one example, the visible light image is acquired based on a visible light sensor, and the visible light sensor and microphone array are synchronously started acquiring data based on the same trigger signal. The same trigger signal refers to a level signal or pulse signal emitted by the main control unit of the detection device or an external timing source, which can be simultaneously responded to by the visible light sensor and microphone array. Examples include a hardware second pulse signal, an external trigger signal, or a network time synchronization signal based on a precision time protocol. In one example, the shutter synchronization error of the same trigger signal is no greater than 10ms, for example, 0.1ms, 1ms, or 10ms. By using the same trigger signal to start acquisition under the same time reference for the visible light sensor and microphone array, the battery pack morphology recorded in the visible light image and the acoustic events corresponding to the multi-channel acoustic signals are at the same moment. Therefore, when the acoustic recognition results are subsequently superimposed onto the visible light image, they correspond to the state of the battery pack under test at the same moment, avoiding timing misalignment of the image and acoustic waves due to inconsistencies in the acquisition times (e.g., the battery pack still undergoing minor movements during the production line cycle or vibrations at the detection front end).

[0037] In a specific example, the visible light sensor and the microphone array are mounted coaxially and coplanarly on the detection front end. The field of view center of the visible light sensor and the geometric center of the microphone array coincide mechanically. The deviation of the field of view center from the normal direction of the microphone array is no greater than ±5°, such as ±0.5°, ±1°, ±2° or ±5°. This ensures that the image center of the visible light image and the acoustic positioning center of the microphone array are aligned on the same physical reference point, reducing parallax deviation in subsequent coordinate transformations.

[0038] In one embodiment of the invention, the pressurization preprocessing may further include a step of low-frequency small-amplitude modulation of the pressurization pressure. In a specific example, during the pressure holding phase, the pressurization pump superimposes a low-frequency small-amplitude periodic signal onto the steady-state pressurization pressure. The modulation frequency can be from 0.2Hz to 10Hz, for example, 0.2Hz, 0.5Hz, 1Hz, 3Hz, 5Hz, or 10Hz; the modulation amplitude can be from 0.5% to 10% of the steady-state pressurization pressure, for example, 0.5%, 1%, 3%, 5%, or 10%. This low-frequency small-amplitude modulation causes the escaping airflow at the leak point to fluctuate in intensity at the same beat as the modulation frequency. The power of the leak sound wave fluctuates with the modulation pressure at the same frequency and exhibits a first-order harmonic component, while the ambient noise in the workshop is unrelated to the aforementioned modulation beat. Subsequently, in the noise reduction and feature matching stages, the modulation frequency is used as a reference phase to perform correlation demodulation on the acquired sound wave signal, which can further extract the synchronous component related to the leak intensity from the ambient noise, improving the detectability of small leaks under low signal-to-noise ratio conditions.

[0039] Step S200: Perform noise reduction processing on the multi-channel acoustic wave signal to obtain the target acoustic wave signal.

[0040] Noise reduction processing refers to the process of suppressing environmental noise components, electrical noise components, and non-stationary interference components introduced by the measurement system itself in a multi-channel acoustic signal that are unrelated to leakage, so that the output target acoustic signal has a relatively higher signal proportion in the frequency band related to leakage. In one example, the noise reduction processing suppresses the multi-channel acoustic signal in at least one of the frequency domain, time domain, or time-frequency domain to obtain a target acoustic signal that is consistent with the multi-channel acoustic signal in terms of the number of channels and time reference.

[0041] In one example, before noise reduction of the multi-channel acoustic signal, a bandpass filter can be used to filter out the non-leakage characteristic frequency bands of the multi-channel acoustic signal. A bandpass filter is a filter that only allows frequency components within a preset passband to pass through, attenuating frequency components outside the passband. Bandpass filters can be implemented as hardware circuits, such as active bandpass filter circuits or analog filter banks, or as digital filters, such as finite impulse response (FIR) filters or infinite impulse response (IR) filters. In a specific example, the lower passband limit of the bandpass filter can be 1kHz to 5kHz, for example, 1kHz, 2kHz, 3kHz, or 5kHz; the upper passband limit can be 40kHz to 100kHz, for example, 40kHz, 50kHz, 60kHz, 80kHz, or 100kHz; the specific range of the passband is adapted to the leakage characteristic frequency band corresponding to the material of the battery pack casing under test.

[0042] Optionally, for battery packs with aluminum alloy casings, the passband of the bandpass filter can be selected in the mid-to-high frequency range, such as 5kHz to 60kHz; for battery packs with engineering plastic or composite material casings, the passband of the bandpass filter can be selected in the mid-frequency range, such as 2kHz to 30kHz. By filtering out non-leakage characteristic frequency bands (such as low-frequency structural noise below 1kHz and electrical interference above 100kHz) before entering the noise reduction process, the noise components that need to be processed in subsequent noise reduction processes can be reduced, and the adverse effects of strong out-of-band interference on the numerical stability of the noise reduction algorithm can be avoided.

[0043] Step S300: For each detection location on the battery pack under test, determine the signal characteristics of each detection location based on the target acoustic signal, and match the signal characteristics of each detection location with the pre-stored leakage feature library to obtain the leakage situation corresponding to each detection location.

[0044] Each detection location refers to a finite number of spatial sampling points or areas obtained by spatially discretizing the outer surface of the battery pack under test. Each detection location can be pre-generated based on the geometric model of the battery pack under test before pressurization preprocessing. In one example, each detection location covers weak sealing points of the battery pack under test, such as shell welds, sealant seams, terminal interfaces, liquid injection ports, explosion-proof valves, and shell through-holes. The spatial interval between adjacent detection locations can be from 1mm to 50mm, for example, 1mm, 2mm, 5mm, 10mm, 20mm, or 50mm.

[0045] The signal features matched with the leakage feature library must include at least one of the following: time-domain features, frequency-domain features, and time-frequency features. Time-domain features refer to the statistical or waveform characteristics of the acoustic signal along the time axis. These include, but are not limited to, kurtosis, zero-crossing rate, peak factor, attenuation coefficient, short-time energy distribution, and autocorrelation function value. Frequency-domain features refer to the spectral distribution characteristics obtained after transforming the acoustic signal to the frequency domain using Fourier transform or similar methods. These include, but are not limited to, the energy proportion of the frequency band of interest, spectral centroid, spectral flatness, spectral peak position, and peak spacing distribution. Time-frequency features refer to the characteristics obtained through joint time-frequency transforms such as short-time Fourier transform or wavelet transform, which characterize the energy changes of the acoustic signal with time and frequency. These include, but are not limited to, wavelet scale distribution, wavelet coefficient modulus maxima, the evolution trajectory of the short-time spectral envelope, and time-frequency ridge features. Using at least one of time-domain features, frequency-domain features, and time-frequency features as matching input, one or more features can be flexibly selected according to the casing material and leakage level of the battery pack under test. When using only a single-dimensional feature, a discrimination criterion for leakage sound waves can be formed in that dimension, with low computational load, making it suitable for detection stations with limited computing power. When using a combination of two or three features as matching input, it can cover multiple discrimination cues of leakage sound waves in the time dimension (vibration impact and attenuation process), frequency dimension (energy distribution of characteristic frequency bands), and time-frequency joint dimension (non-stationary time-frequency structure). Compared with single-dimensional feature matching, it is beneficial to maintain a high recognition accuracy under different casing materials and different leakage levels.

[0046] In a specific example, for each detection location, the target acoustic signal is extracted within the time window corresponding to that detection location, and at least one of the time-domain features, frequency-domain features, and time-frequency features is extracted as the signal feature of that detection location. This signal feature is then input into the feature matching module. Based on a pre-stored leakage feature library, the feature matching module calculates the similarity between the signal feature and the features of each sample in the leakage feature library. The specific method for calculating the similarity can be any combination of one or more of the following: cosine similarity, Euclidean distance, Mahalanobis distance, correlation coefficient, and discriminant function based on a neural network. When the similarity exceeds a preset threshold, the leakage situation corresponding to that detection location is determined to be present, and the leakage level corresponding to that detection location is further determined based on the leakage level label of the matched sample.

[0047] Optionally, the leakage level can be divided into four levels: micro-leakage, small leakage, medium leakage, and large leakage. The above level division is only an example to facilitate the connection of test results with the quality control grading of the production line. This invention does not limit the specific level division of leakage and the corresponding leakage range.

[0048] The pre-stored leakage feature library is a set of feature samples pre-built and stored in the detection device or external database. In a specific example, each sample in the leakage feature library includes at least the following fields: shell material label (e.g., aluminum alloy, stainless steel, engineering plastic, composite material), leakage location label (e.g., weld, sealant joint, pole interface), leakage level label, corresponding time-domain feature vector, frequency-domain feature vector, and time-frequency feature vector; the sample sources of the leakage feature library include, but are not limited to, laboratory calibration samples, on-site labeling samples from the production line, and historical after-sales investigation samples.

[0049] Step S400: Based on the visible light image of the battery pack under test, visualize the leakage situation corresponding to each detection location.

[0050] A visible light image is a two-dimensional image covering the detection area of ​​the battery pack under test, acquired by a visible light sensor at the same time reference as the multi-channel acoustic signal. In one example, visualization refers to establishing a spatial correspondence between the leakage conditions at each detection location and the corresponding pixel positions in the visible light image, and then overlaying layers using at least one visual representation of color, brightness, transparency, or shape to obtain a fused image that can be directly read by the inspector.

[0051] Optionally, the output format of the visualization can be any one of a static image, a dynamic video stream, or an interactive 3D point cloud. The above output format can be flexibly selected according to the type of viewing terminal used by the inspector and the requirements of the viewing scenario; this invention does not impose any limitations on this.

[0052] In one embodiment of the present invention, the visualized results can also be output in conjunction with a test report. The test report includes at least the identification information of the battery pack under test, the test timestamp, the leakage status at each test location, the leakage level, and the test image fused with the visible light image.

[0053] Optionally, the test report can be exported in formats such as PDF (Portable Document Format), JSON (JavaScript Object Notation), and XML (Extensible Markup Language), and uploaded to the production execution system, enterprise resource planning system, or after-sales management system via wired interface or wireless network. The specific format and upload channel of the report can be adapted to the interface specifications of the downstream receiving system, and this invention does not limit this.

[0054] Example 2 The difference from Embodiment 1 lies in the further explanation of the method for acquiring the signal features of the detection position in step S300. Step S300, which determines the signal features of each detection position based on the target acoustic signal, essentially involves spatially focusing the multi-channel acoustic signal acquired by the microphone array onto the spatial direction corresponding to the detection position and acquiring the signal features corresponding one-to-one with that position from the focusing result. This spatial focusing can be achieved in various ways, such as aligning the channel components based on propagation delay, adaptive beamforming, or subspace spatial spectrum estimation. This embodiment uses the method of aligning the channel components based on propagation delay as an example for explanation.

[0055] In one example, for each detection location on the battery pack under test, the propagation delay from the detection location to each microphone unit is calculated based on the positional relationship between the detection location and each microphone unit in the microphone array; based on the propagation delay, the channel components of the target acoustic signal are aligned, and the signal characteristics of the detection location are obtained based on the aligned channel components.

[0056] The positional relationship between the detection position and each microphone unit refers to the relative positional relationship between the spatial coordinates of the detection position and the spatial coordinates of the geometric center of each microphone unit in the same spatial coordinate system. In a specific example, a detection coordinate system is established with the geometric center of the microphone array as the origin. After the battery pack to be tested is placed on the detection station, the three-dimensional spatial coordinates of each detection position in the detection coordinate system are calculated based on the relative position of the geometric model of the battery pack to be tested and the positioning fixture. For each detection position, the Euclidean distance from the detection position to the geometric center of each microphone unit is calculated, and the Euclidean distance is divided by the speed of sound propagation in the detection environment to obtain the propagation time delay from the detection position to each microphone unit.

[0057] Optionally, the propagation speed of sound waves in the detection environment can be dynamically corrected based on the temperature, humidity, and atmospheric pressure parameters of the detection environment. The range of the propagation speed can be from 320 m / s to 360 m / s, for example, 330 m / s, 340 m / s, or 350 m / s. Dynamically correcting the sound speed based on environmental parameters helps ensure the accuracy of propagation delay calculation and makes the peak energy of the aligned signal after focusing more stable. This invention does not limit the specific algorithm for sound speed correction.

[0058] Aligning the channel components of a target acoustic signal based on propagation delay refers to shifting each channel component along the time axis by an amount corresponding to its propagation delay, so that acoustic components from the same detection position exhibit a time-consistent waveform structure across all channels. In a specific example, after shifting each channel component according to its propagation delay, the shifted waveforms of all channels are weighted and summed or subjected to consistency weighting to obtain an aligned signal spatially focused with the detection position. This aligned signal is then used as the source signal for that detection position, and at least one of the time-domain features, frequency-domain features, and time-frequency features is extracted from the aligned signal as the signal feature for that detection position.

[0059] By calculating the propagation delay based on the relationship between the detection position and the microphone array position, and then aligning each channel component with the propagation delay before extracting signal features, compared to the method of directly extracting features from the original multi-channel signal without time delay alignment, this embodiment forms spatial focus on the detection position at the array pickup level: the sound waves from the detection position remain time-consistent and coherently superimposed after alignment, and the interfering sound waves from other spatial positions are relatively weakened due to the time delay mismatch and present an incoherent structure after alignment. This is beneficial to further improve the distinguishability of the leakage sound wave relative to the ambient noise under low signal-to-noise ratio conditions, and makes the signal features of each detection position correspond one-to-one with that position in space.

[0060] In one embodiment of the present invention, an effective detection area is determined based on the geometric model of the battery pack under test; the effective detection area is the spatial area occupied by the sealing trajectory of the battery pack under test; for detection positions located within the effective detection area, the propagation delay from the detection position to each microphone unit is calculated based on the positional relationship between the detection position and each microphone unit in the microphone array; based on the propagation delay, the channel components of the target acoustic signal are aligned, and the signal characteristics of the detection position are obtained based on the aligned channel components.

[0061] The applicant discovered that among the discretely divided detection locations on the surface of the battery pack under test, only a portion were actually located at weak sealing points such as shell welds, sealant seams, terminal interfaces, liquid injection ports, and explosion-proof valves. The remaining detection locations were distributed on flat surfaces of the shell, decorative covers, and non-pressure-bearing areas. Performing propagation delay calculations and alignment operations on these non-sealed detection locations presents two potential risks: First, propagation delay alignment is essentially a coherent superposition of sound wave components from a given spatial direction. When there is no actual leakage source in that spatial direction, the alignment operation consumes computing power but produces no effective output. Second, strong sound waves scattered by metal clamps, robotic arm joints, and reflective obstructions around the detection station, after being picked up by the array, may coincide with the geometric direction corresponding to some detection locations in non-sealed areas. In the alignment operation, these waves coherently superimpose into high-energy pseudo-peaks unrelated to the actual leakage source, thus interfering with subsequent matching and discrimination with the leakage feature database.

[0062] The geometric model of the battery pack under test refers to the geometric data that digitally describes the shape, sealing structure, and key interface contours of the battery pack. The specific form of the geometric model can be a three-dimensional solid model, a surface mesh model, or a set of feature points derived from two-dimensional engineering drawings. This invention does not limit the specific form of the geometric model. The sealing trajectory refers to the geometric line segments or regions in the geometric model that correspond to the sealing structure of the battery pack. The sealing trajectory includes, but is not limited to, the direction of the shell weld seam, the direction of the sealing glue seam, the contour of the terminal interface, the contour of the injection port, and the edge of the explosion-proof valve. The spatial area occupied by the sealing trajectory refers to the strip-shaped or annular space on the surface of the battery pack under test, which is jointly covered by the sealing trajectory according to its own geometric width and allowable process deviations.

[0063] In a specific example, the process of generating the effective detection area includes: mapping the geometric model of the battery pack under test, along with its sealing trajectory, onto a detection coordinate system with the geometric center of the microphone array as the origin; extending the bandwidth corresponding to the allowable process deviation along both sides of each sealing trajectory according to its own geometric width to obtain a strip or annular space corresponding to the sealing trajectory; merging the strip or annular spaces corresponding to all sealing trajectories to obtain the effective detection area. The bandwidth value is adapted to the sealing structure morphology and the directional resolution capability of the microphone array at the working distance: for narrow linear sealing structures such as welds and sealant seams, the bandwidth is chosen so that the effective detection area just covers the allowable sealing deviation of the process and the directional resolution limit of the array at that working distance; for local annular sealing structures such as terminal interfaces, injection ports, and explosion-proof valves, the bandwidth is extended along the interface contour to both the inner and outer sides.

[0064] In the propagation delay alignment process, for detection positions located within the effective detection area, the propagation delay from the detection position to each microphone unit is calculated according to the propagation delay calculation method given in the aforementioned embodiment. Based on the propagation delay, the channel components of the target acoustic signal are aligned, and the signal characteristics of the detection position are obtained based on the aligned channel components. For detection positions located outside the effective detection area, the propagation delay calculation, channel component alignment, and signal characteristic acquisition process are not entered, nor are they involved in the subsequent matching and discrimination with the leakage feature library.

[0065] By determining the effective detection area based on the geometric model and constraining the propagation delay alignment range with the effective detection area, this embodiment achieves improvements at both the computational power distribution and sound field pickup levels compared to performing propagation delay alignment on all detection positions. At the computational power distribution level, the propagation delay calculation and the alignment operation of each channel component are converged to a limited number of detection positions related to the sealing trajectory. The cumulative time consumed by the alignment operation decreases accordingly with the reduction in the number of effective detection positions, ensuring that the increase in array size and detection position density does not come at the cost of a year-on-year increase in alignment time. At the sound field pickup level, since the alignment operation only occurs with... The array unfolds along the geometric direction corresponding to the actual sealing trajectory. Even if strong reflection interference sources from outside the field of view of the battery pack under test (such as metal shields above the detection station or robotic arm joints of adjacent stations) are picked up by the array, they will not be included in the alignment calculation because their corresponding geometric directions on the array do not fall within the effective detection area. This avoids spatial aliasing and the formation of false peaks when such interference sources are aligned with the detection position of a non-sealed area during propagation delay. As a result, the energy structure of the aligned signal more purely reflects the real acoustic events from within the effective detection area, which is beneficial to reducing the probability of misjudgment in the subsequent matching process with the leakage feature library.

[0066] In one embodiment of the present invention, the alignment signals obtained after alignment of each detection position based on propagation delay can be further evaluated based on the phase coherence between multiple channels and the timing and attenuation relationship of the waveforms to eliminate false reflection sources. A false reflection source refers to a false sound source formed on the array after the sound wave from the real leak point is reflected from the surface of the battery pack casing, the metal structure of the detection station, or a surrounding robotic arm, etc., creating a spatial orientation inconsistent with the real leak point. In a specific example, for the alignment signals of two adjacent candidate detection positions, the cross-spectral phase coherence between the two channels is calculated, with the coherence value ranging from 0 to 1. Simultaneously, the relationship between the onset time and peak energy of the alignment signals at the two candidate detection positions is examined: the sound wave from the real leak point arrives at the array via a single direct path, has high phase coherence, starts oscillating earlier, and its energy varies with spatial distance according to the attenuation law of near-field sound waves; while the sound wave from the false reflection source arrives at the array via a longer path, has lower phase coherence, a relatively delayed onset time, and its energy attenuation deviates from the attenuation law of near-field sound waves. Of the two candidate detection locations, those with low phase coherence, delayed oscillation timing, and energy attenuation deviating from the near-field acoustic wave attenuation pattern are identified as false reflection sources and eliminated, leaving only the other candidate detection location as the true leak point. This avoids misidentifying multipath false peaks introduced by the high-reflectivity interface of the casing as true leak points, improving the uniqueness and reliability of leak point spatial localization.

[0067] Example 3 The difference from Embodiment 1 lies in the further explanation of the specific method of noise reduction processing in step S200. Step S200 involves noise reduction processing of the multi-channel acoustic signal to obtain the target acoustic signal. Essentially, this process suppresses noise components unrelated to leakage in the multi-channel acoustic signal and increases the signal proportion in the leakage-related frequency bands. This noise reduction processing can be achieved in various ways, such as matched filtering based on environmental noise feature templates, multi-channel adaptive differential noise reduction, spectral subtraction or Wiener filtering, and wavelet threshold noise reduction. This embodiment uses a combination of matched filtering based on environmental noise feature templates and multi-channel adaptive differential noise reduction as an example for explanation.

[0068] In one example, noise reduction processing is performed on a multi-channel acoustic signal to obtain a target acoustic signal, including: matching and filtering the multi-channel acoustic signal based on a pre-stored environmental noise feature template to obtain a first-level suppressed signal; and adaptive differential noise reduction is performed on the first-level suppressed signal to obtain the target acoustic signal.

[0069] An environmental noise feature template refers to a set of noise features pre-collected and extracted from known and relatively stable noise sources in the environment where the testing station is located. In a specific example, the construction of the environmental noise feature template includes the following process: under no-load conditions when the battery pack to be tested is not put into the testing station, multiple sound wave acquisitions are performed on the environment where the testing station is located; frequency domain energy distribution, time domain statistics, and time-frequency ridge features corresponding to the operation of production line equipment, workshop ventilation, periodic movements of robotic arms, and lighting electrical noise are extracted from the no-load sound wave signals; the above features are classified and stored according to the noise source category to form the environmental noise feature template.

[0070] Optionally, the update frequency of the environmental noise characteristic template can be based on shifts, workdays, or equipment maintenance cycles. The specific selection of the update frequency should ensure the consistency between the template characteristics and the actual noise background at the workstation. A higher update frequency is beneficial for tracking the gradual trend of noise background changes, but this invention does not impose any limitations on this.

[0071] Matching filtering based on environmental noise feature templates involves calculating the similarity between each channel component of the acquired multi-channel acoustic signal and each template feature in the environmental noise feature template. When the similarity exceeds a preset threshold, the signal component in the multi-channel acoustic signal corresponding to that template feature is attenuated or filtered out, resulting in a first-level suppression signal that has suppressed the noise components within the template. In a specific example, the preset threshold for matching filtering can be between 0.7 and 0.99, such as 0.7, 0.8, 0.85, 0.9, or 0.99. The specific value of the threshold is adapted to the stability of the noise background of the production line.

[0072] Adaptive differential denoising refers to adaptively adjusting the weights of the differential combinations between the components of each channel within the same time window. This allows common-mode components (i.e., stable noise highly correlated among multiple channels) to cancel each other out, while leaking sound waves, due to their spatially dependent propagation delay differences on the array, are not completely common-mode and are thus preserved in the differential combination. In a specific example, the first-stage suppression signal is weighted and differentially combined among multiple adjacent channels. The weights are obtained by minimizing the residual noise power in the current frame using an adaptive algorithm. The adaptive algorithm can employ any combination of one or more of the following: least mean square algorithm, normalized least mean square algorithm, and recursive least square algorithm. After adaptive differential denoising, residual non-stationary interference components (such as the sound of people walking or the intermittent movement of a robotic arm) are further suppressed, ultimately resulting in a target sound wave signal with a significantly improved signal-to-noise ratio compared to the original multi-channel sound wave signal.

[0073] By combining the first-level suppression based on environmental noise feature templates with the second-level suppression based on multi-channel differential, the noise reduction process covers both stable and non-stationary noise interference. The first-level suppression is based on template priors, which suppresses the overall stable noise that can be modeled in the workstation environment, and it can suppress noise within the template without relying on spatial information. The second-level suppression is based on multi-channel spatial differential, which further strips away non-stationary noise not covered by the template, and it has a natural suppression selectivity for real leakage sound waves due to their propagation delay related to spatial location, thus avoiding the suppression of real leakage sound waves along with noise.

[0074] In one embodiment of the present invention, the low-frequency small-amplitude pressure modulation applied in the pressure preprocessing stage of Embodiment 1 can be combined with the adaptive differential noise reduction in this embodiment. In a specific example, the modulation frequency of the low-frequency small-amplitude pressure modulation is used as a reference phase input to the weight update logic of the adaptive differential noise reduction, so that the weight update is biased towards retaining signal components that fluctuate synchronously with the modulation frequency and suppressing noise components that are unrelated to the modulation frequency. Thus, the pressure modulation excitation and the two-stage noise reduction form a positive synergy in the same signal chain: the modulation stage first enhances the distinguishability of the leaked sound wave relative to the ambient noise at the source; the two-stage noise reduction further amplifies this distinguishability at the channel layer and the differential layer.

[0075] Example 4 The difference from Embodiment 1 lies in the further explanation of the specific method for visualizing the leakage situation corresponding to each detection location in step S400. Step S400, which visualizes the leakage situation corresponding to each detection location based on the visible light image, essentially involves establishing the correspondence between the physical spatial coordinates of each detection location and the pixel positions of the visible light image, and then superimposing the leakage situation corresponding to each detection location onto the visible light image. The mapping from detection location to pixel position can be achieved in various ways, such as based on a pre-calibrated coordinate transformation matrix, based on a camera projection model, or based on feature point registration. This embodiment uses the method based on a pre-calibrated coordinate transformation matrix as an example for explanation.

[0076] In one example, based on the visible light image of the battery pack under test, the leakage situation corresponding to each detection location is visualized, including: mapping each detection location to the corresponding pixel position in the visible light image through a pre-calibrated coordinate transformation matrix; and rendering a heat map at the corresponding pixel position in the visible light image based on the leakage situation corresponding to each detection location to obtain the fused visualization image.

[0077] The pre-calibrated coordinate transformation matrix refers to the transformation matrix determined in advance through the calibration process before the detection device leaves the factory or is used in the field. This matrix maps the physical spatial coordinates of each detection position to the pixel coordinates of a visible light image. In a specific example, the calibration process of the coordinate transformation matrix includes the following steps: placing a calibration board with known geometric dimensions and known feature point positions on the detection station; acquiring a visible light image of the calibration board using a visible light sensor; and acquiring corresponding sound wave signals at known sound source excitation positions on the calibration board using a microphone array; solving the coordinate transformation matrix based on the correspondence between the physical spatial coordinates of each feature point on the calibration board and their corresponding pixel coordinates in the visible light image; and storing the obtained coordinate transformation matrix in the detection device as a fixed parameter for mapping each detection position to a pixel position in subsequent detection processes.

[0078] Optionally, the coordinate transformation matrix can be represented as a 3×3 homography matrix, a 3×4 projection matrix, or a 4×4 homogeneous transformation matrix. The matrix form can be flexibly selected based on the field of view coverage and computational resources of the detection device; this invention does not limit the specific representation of the coordinate transformation matrix.

[0079] Rendering a heatmap of pixel locations on a visible light image refers to overlaying the pixel locations corresponding to each detection location in the visible light image with a combination of color and transparency, based on the leakage status (including whether there is leakage and the leakage level) at each detection location. In a specific example, the process of rendering the heatmap includes: normalizing the leakage level or probability corresponding to each detection location to a value range of 0 to 1; mapping the normalized value to the corresponding RGB (Red, Green, Blue) color value according to a preset color mapping rule, where the color mapping rule can be a warm / cool transition mapping (e.g., blue-cyan-green-yellow-red), with higher normalized values ​​corresponding to more reddish colors; mapping the normalized value to the corresponding transparency according to a preset transparency mapping rule, with higher normalized values ​​corresponding to lower transparency, and a normalized value of 0 corresponding to complete transparency (not obscuring the visible light image); and applying the mapped RGB color values ​​and transparency to the corresponding pixel locations in the visible light image to obtain the fused visualization image.

[0080] Optionally, for detection locations where the normalized value exceeds a preset threshold (e.g., 0.7, 0.8, 0.9), circular, cross, or rectangular markers can be superimposed on the fused image, and the leakage level and normalized value can be labeled in text form near the markers to facilitate quick identification by operators.

[0081] In one embodiment of the present invention, when the visible light sensor is not coaxially and coplanarly mounted with the microphone array but is independently mounted at the inspection station in the form of an external industrial camera, the visualization presentation may further include a parallax compensation step. Parallax compensation refers to the process of correcting the geometric deviation between the pixel position in the visible light image and the acoustic positioning result caused by the spatial offset of the external industrial camera relative to the microphone array. In a specific example, the parallax compensation step includes: pre-measuring the physical offset (including three-axis translation and three-axis rotation components) between the optical center of the external industrial camera and the geometric center of the microphone array, and storing it as a parallax parameter; for each detection position, constructing a dynamic parallax compensation matrix based on its physical spatial coordinates in the detection coordinate system and the parallax parameter, the compensation matrix dynamically changing with the distance between the detection position and the array; multiplying the preliminary pixel position obtained by mapping each detection position through a coordinate transformation matrix with the parallax compensation matrix to obtain the final pixel position after parallax compensation; and rendering a heatmap at the final pixel position. Therefore, in application scenarios using external industrial cameras, the parallax deviation caused by the non-coaxial and coplanar nature of the external industrial camera relative to the microphone array is dynamically corrected, ensuring that the spatial correspondence between the rendered position of the thermal map and the actual leakage location of the battery pack under test remains accurate.

[0082] Example 5 Figure 2 This is a structural block diagram of a battery pack detection device according to an embodiment of the present invention. See also... Figure 2 The detection device includes a pressurization module, an acoustic wave acquisition module, a noise reduction module, an identification module, and an output module.

[0083] The battery pack testing device provided in this embodiment is a device embodiment corresponding to the above-described method embodiments (Embodiments 1 to 4), and can implement each step of the battery pack testing method provided in the above method embodiments. Each functional module in the device corresponds one-to-one with each step in the method embodiments. The parameter selection, optional extensions, and technical effect descriptions of each step in the method embodiments also apply to the corresponding modules in the device of this embodiment, and will not be repeated.

[0084] The pressurization module is configured to pre-pressurize the battery pack under test. Hardware-wise, the pressurization module may include an air source, a pressurization pump, a pressure regulating valve, a pressure sensor, and a quick-connect fitting that seals with the battery pack under test. Logically, the pressurization module receives pressurization commands from the main control unit, completes the pre-pressurization process according to the preset pressurization pressure and holding time, and outputs a pressurization completion signal to the main control unit after the target pressure is reached.

[0085] Optionally, the pressurizing medium of the pressurizing module can be dry nitrogen, dry air, or dry argon, the pressurizing pressure can be from 0.05 MPa to 0.5 MPa, and the pressure holding time can be from 3 seconds to 60 seconds. Specific parameters of the pressurizing module can be adapted to the relevant descriptions of pressurization pretreatment in the foregoing method embodiments; this invention does not limit these parameters.

[0086] The acoustic wave acquisition module includes a microphone array configured to acquire multi-channel acoustic wave signals from a pressurized battery pack under test. The acoustic wave acquisition module hardware may include a microphone array, an analog-to-digital converter, signal conditioning circuitry, an electromagnetic interference-resistant housing, and a trigger interface linked to a visible light sensor. The microphone units in the microphone array are arranged spatially according to a predetermined geometric relationship and output acoustic wave signals from their respective channels at the same time reference.

[0087] Optionally, the microphone array can have 4 to 64 microphone units, the sampling rate of the microphone units can be 16 kHz to 192 kHz, and the frequency response range of the microphone units can be 2 kHz to 100 kHz. Specific parameters of the microphone array can be adapted to the relevant descriptions of multi-channel acoustic signal acquisition in the foregoing method embodiments; this invention does not limit these parameters.

[0088] The noise reduction module is configured to perform noise reduction processing on multi-channel acoustic signals to obtain the target acoustic signal. The noise reduction module can be implemented in hardware using a processor and a memory: the memory stores the program instructions corresponding to the noise reduction algorithm and pre-stored environmental noise feature templates; the processor retrieves the program instructions and environmental noise feature templates from the memory, and sequentially performs at least one of the following operations on the multi-channel acoustic signals: bandpass filtering, matched filtering based on environmental noise feature templates, and adaptive differential noise reduction, outputting the target acoustic signal to the recognition module.

[0089] The identification module is configured to determine the signal characteristics of each detection location on the battery pack under test based on the target acoustic signal, and then match the signal characteristics of each detection location with a pre-stored leakage feature library to obtain the leakage status corresponding to each detection location. The identification module can be implemented in hardware using a processor and memory: the memory stores the pre-stored leakage feature library, feature extraction algorithm, and program instructions corresponding to the feature matching algorithm; the processor calls the program instructions from the memory, sequentially performs channel alignment, feature extraction, and feature matching on the target acoustic signal, and outputs the leakage status corresponding to each detection location to the output module.

[0090] The output module is configured to visualize the leakage status at each detection location based on a visible light image of the battery pack under test. The output module hardware may include a visible light sensor, a display screen, a data interface, and a wireless communication unit. The visible light sensor and microphone array synchronously initiate data acquisition based on the same trigger signal. The display screen is configured to show the fused visualized image and detection report. The data interface and wireless communication unit are configured to export the fused visualized image and detection report to an external system.

[0091] Example 6 like Figure 3 An embodiment of the present invention also provides an electronic device. The electronic device includes a memory 202 and a processor 201. The memory 202 stores a computer program, which, when executed by the processor 201, implements the steps of the battery pack detection method as described in any of the above embodiments.

[0092] The electronic device provided in this embodiment is a hardware carrier embodiment corresponding to the above method embodiment. It implements the battery pack detection method provided in the above method embodiment by loading and running a computer program.

[0093] In one example, the processor can be any one or more combinations of a central processing unit, a graphics processing unit, a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), a digital signal processor (DSP), or a neural network processor. The memory can include any one or more combinations of read-only memory (ROM), random access memory (RAM), electrically erasable programmable read-only memory (EPROM), flash memory, solid-state drive (SSD), and hard disk drive (HDD). The electronic device may also include input / output interfaces, network communication interfaces, and buses, with the input / output interfaces and network communication interfaces interconnected with the processor and memory via the bus.

[0094] In one example, the electronic device can take the form of any of the following: an industrial control computer, an embedded industrial control motherboard, a portable testing terminal, a server, or a cloud computing node.

[0095] Example 7 One embodiment of the present invention also provides a computer-readable storage medium. This computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the battery pack detection method as described in any of the above embodiments.

[0096] The computer-readable storage medium provided in this embodiment is a carrier embodiment corresponding to the above method embodiment. The computer program carried thereon implements the battery pack detection method provided in the above method embodiment when executed by a processor.

[0097] In one example, the computer-readable storage medium can be any one of non-volatile or volatile storage media. Non-volatile storage media can include any one or more combinations of read-only memory, programmable read-only memory, electrically erasable programmable read-only memory, flash memory, magnetic storage media (e.g., hard disk, magnetic tape, magnetic disk), and optical storage media (e.g., optical disc, digital versatile optical disc, Blu-ray disc). Volatile storage media can include any one or more combinations of random access memory and cache memory.

[0098] The battery pack testing method, testing device, electronic equipment, and storage medium provided in this invention can be widely applied in application scenarios requiring efficient testing, precise positioning, and traceable recording of battery pack airtightness. These include, but are not limited to, full inspection processes on power battery module and battery pack production lines, after-sales fault diagnosis processes for power batteries, health status assessment processes before the secondary use of retired power batteries, factory inspection processes for energy storage battery packs, and on-site operation and maintenance inspection processes for energy storage power stations. The battery pack testing method, testing device, electronic equipment, and storage medium provided in this invention can detect the airtightness of energy storage battery packs used in long-term energy storage scenarios of 4 hours or more, such as 5-hour, 6-hour, and 8-hour energy storage scenarios. Long-term energy storage refers to the ability to continuously discharge at rated power for 4 hours or even longer, or to achieve large-scale, low-cost energy storage for several days or months.

[0099] The battery pack testing device provided in this invention can be used in conjunction with battery devices, energy storage devices, and electrical devices. Battery devices include one or more of battery modules, battery packs, and energy storage batteries. Energy storage devices include, but are not limited to, residential energy storage cabinets, commercial energy storage cabinets, energy storage containers, energy storage racks, energy storage power stations, energy storage battery packs, or portable energy storage systems. Energy storage devices may also include energy management systems (EMS), battery management systems (BMS), and power conversion systems (PCS). Electrical devices include, but are not limited to, mobile phones, tablets, laptops, electric toys, power tools, electric vehicles, electric cars, ships, spacecraft, etc.

[0100] The battery pack detection method, detection device, electronic device, and storage medium provided in the above embodiments can extend the serial operation mode of "differential pressure qualitative judgment + manual assisted positioning + manual recording and tracing" in related technologies to an integrated operation mode of "pressurized acquisition + multi-channel noise reduction + position-by-position feature matching + visible light superposition". This allows the three objectives of airtight leakage judgment, positioning, and tracing to be achieved simultaneously in the same signal chain. Compared with the implementation path of each objective relying on independent processes in related technologies, this is beneficial to balance detection efficiency, positioning accuracy, and traceability integrity in production line conditions with limited production cycle.

[0101] Those skilled in the art will understand that the above embodiments are specific examples of implementing the present invention, and in practical applications, various changes in form and detail can be made without departing from the spirit and scope of the present invention. Any person skilled in the art can make various alterations and modifications without departing from the spirit and scope of the present invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A method for detecting a battery pack, characterized in that, include: The battery pack under test is pre-pressurized, and multi-channel acoustic signals are acquired from the pressurized battery pack using a microphone array. The multi-channel acoustic signal is subjected to noise reduction processing to obtain the target acoustic signal; For each detection location on the battery pack under test, the signal characteristics of each detection location are determined based on the target acoustic signal, and the signal characteristics of each detection location are matched with a pre-stored leakage feature library to obtain the leakage situation corresponding to each detection location. Based on the visible light image of the battery pack under test, the leakage situation corresponding to each detection location is visualized.

2. The battery pack testing method according to claim 1, characterized in that, The step of determining the signal characteristics of each detection location on the battery pack under test based on the target acoustic signal includes: For each detection location, the propagation delay from the detection location to each microphone unit is calculated based on the positional relationship between the detection location and each microphone unit in the microphone array. Based on the propagation delay, the channel components of the target acoustic signal are aligned, and the signal characteristics of the detection position are obtained based on the aligned channel components.

3. The battery pack testing method according to claim 1, characterized in that, Before performing noise reduction processing on the multi-channel acoustic signal to obtain the target acoustic signal, the method further includes: The non-leaking characteristic frequency bands in the multi-channel acoustic signal are filtered out by a bandpass filter.

4. The method for detecting a battery pack according to claim 1 or 3, characterized in that, The noise reduction process for the multi-channel acoustic signal to obtain the target acoustic signal includes: Based on the pre-stored environmental noise feature template, the multi-channel acoustic signal is matched and filtered to obtain the first-level suppression signal; Adaptive differential noise reduction is performed on the first-level suppression signal to obtain the target acoustic signal.

5. The battery pack testing method according to claim 1, characterized in that, The visible light image of the battery pack under test is used to visualize the leakage situation at each detection location, including: By using a pre-calibrated coordinate transformation matrix, each detection position is mapped to the corresponding pixel position in the visible light image; Based on the leakage situation corresponding to each detection location, a heat map is rendered at the corresponding pixel position on the visible light image to obtain a fused visualization image.

6. The method for detecting a battery pack according to claim 1 or 2, characterized in that, The signal features matched with the leakage feature library include at least one of time-domain features, frequency-domain features, and time-frequency features.

7. The method for detecting a battery pack according to claim 1 or 5, characterized in that, The visible light image is acquired based on a visible light sensor; the visible light sensor and the microphone array are synchronously started acquiring the image based on the same trigger signal.

8. The battery pack testing method according to claim 2, characterized in that, The method further includes: Based on the geometric model of the battery pack under test, an effective detection area is determined; the effective detection area is the spatial area occupied by the sealing trajectory of the battery pack under test. For the detection positions located within the effective detection area only, the propagation delay from the detection position to each microphone unit is calculated based on the positional relationship between the detection position and each microphone unit in the microphone array; Based on the propagation delay, the channel components of the target acoustic signal are aligned, and the signal characteristics of the detection position are obtained based on the aligned channel components.

9. A battery pack testing device, characterized in that, include: The pressurization module is used to pre-pressurize the battery pack under test. The acoustic wave acquisition module includes a microphone array, which is used to acquire multi-channel acoustic wave signals from the pressurized battery pack under test; A noise reduction module is used to perform noise reduction processing on the multi-channel acoustic wave signal to obtain the target acoustic wave signal; The identification module is used to determine the signal characteristics of each detection location on the battery pack under test based on the target acoustic signal, and match the signal characteristics of each detection location with a pre-stored leakage feature library to obtain the leakage situation corresponding to each detection location. The output module is used to visualize and output the leakage status corresponding to each detection location based on the visible light image of the battery pack under test.

10. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program that, when executed by the processor, implements the steps of the battery pack detection method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the battery pack detection method as described in any one of claims 1 to 8.

Images