Battery heat shrink tube and cell matching degree detection method and related equipment

Abstract

The invention relates to the technical field of battery manufacturing, and discloses a battery heat shrink tube and battery cell matching degree detection method and related equipment. The method comprises the following steps: collecting multi-period electrostatic evolution data generated when a target battery cell is coated with a target heat shrink tube through standardized electrostatic detection parameters; electrostatic time sequence characteristic parameters are obtained through thermal shrinkage process time domain segmentation and electrostatic accumulation-discharge period identification, and a frequency domain characteristic parameter set is obtained through electrostatic time-space distribution characteristic extraction, spatial distribution calculation, frequency domain transformation and thermal shrinkage tube-cell interface friction frequency spectrum characteristic calculation; generating a comprehensive feature vector through multi-dimensional feature fusion and weight optimization calculation; and performing matching mode classification and membership function mapping based on the comprehensive feature vector to obtain matching degree evaluation data, and generating a matching degree detection result of the heat shrink tube and the battery cell through quantitative calculation and confidence calculation. And the detection accuracy of the matching degree of the battery heat shrink tube and the battery cell and the multi-dimensional electrostatic feature fusion efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery manufacturing, and in particular to a method for detecting the degree of matching between a battery heat shrink tube and a battery cell, and related equipment. Background Art

[0002] In the field of battery manufacturing technology, testing the matching of heat shrink tubing and battery cells is a crucial step in battery assembly and quality control. Real-time monitoring of electrostatic properties and diagnostics of matching are the most critical and challenging stages of the testing process. Accurately detecting the matching status of heat shrink tubing and battery cells is crucial for battery manufacturers, R&D institutions, and testing centers, as it directly impacts the safety assessment, performance optimization, and overall service life of battery products. Therefore, an effective matching test method to monitor the physical properties of heat shrink tubing during the wrapping process is crucial to ensuring the assembly quality and safety of battery products.

[0003] Currently, mechanical measurement techniques are used to analyze the dimensional parameters of heat shrink tubing, establish geometric matching models to evaluate the coating state, or attempt to apply image recognition technology to the appearance inspection process to better evaluate the coating quality and defect types. However, these methods still face challenges in integrating multidimensional physical parameters, processing nonlinear deformation characteristics, and adapting to dynamic heat shrinkage processes. In addition, these detection methods often ignore some unique physical phenomena in the heat shrinkage process, such as static electricity accumulation, friction effects, and interfacial stress distribution, which may have a significant impact on matching judgment and quality assessment. In other words, the existing heat shrink tubing and battery cell matching detection methods ignore the multidimensional electrostatic characteristics and the unique physical phenomena of the heat shrinkage coating process, resulting in low accuracy of the final matching test results. Summary of the Invention

[0004] The main purpose of the present invention is to solve the problem that the existing heat shrink tubing and battery cell matching detection method ignores the multi-dimensional electrostatic characteristics and the unique physical phenomena of the heat shrink coating process, resulting in low accuracy of the final matching detection result.

[0005] The first aspect of the present invention provides a method for detecting the matching degree between a battery heat shrink tube and a battery cell. The method comprises: based on preset standardized electrostatic detection parameters, collecting multi-period electrostatic evolution data generated when a target heat shrink tube is wrapped around a target battery cell; performing heat shrink process time domain segmentation and electrostatic accumulation-discharge cycle identification on the multi-period electrostatic evolution data to obtain electrostatic time series characteristic parameters; and based on preset multi-point positions around the battery cell, performing electrostatic spatiotemporal distribution feature extraction and spatial distribution calculation on the multi-period electrostatic evolution data to obtain electrostatic fingerprint map data. The electrostatic fingerprint map data is subjected to frequency domain transformation and spectral feature calculation of the heat shrink tube-battery cell interface friction frequency to obtain a frequency domain feature parameter set; based on the preset battery cell and heat shrink tube specification parameters, the electrostatic timing feature parameters and the frequency domain feature parameter set are subjected to feature fusion and weight optimization calculation of the heat shrink tube and battery cell matching degree to obtain a comprehensive feature vector, and the comprehensive feature vector is subjected to matching mode classification and membership function mapping of the heat shrink tube and battery cell to obtain matching evaluation data; the matching evaluation data is subjected to quantitative calculation and confidence calculation to generate a matching detection result.

[0006] Optionally, in a first implementation of the first aspect of the present invention, the multi-period electrostatic evolution data is subjected to time-domain segmentation of the heat shrink process and identification of the electrostatic accumulation-discharge cycle to obtain electrostatic time series characteristic parameters, including: based on the shrinkage temperature characteristics of the target heat shrink tube, the multi-period electrostatic evolution data is subjected to time-domain segmentation of multiple heat shrink process stages to obtain segmented electrostatic data, and the segmented electrostatic data is subjected to filtering, denoising and multiple data preprocessing to obtain a preprocessed electrostatic signal sequence; the preprocessed electrostatic signal sequence is subjected to electrostatic accumulation rate calculation and The peak value of electrostatic accumulation on the surface of the battery cell is identified to obtain the accumulation key characteristic points, and the preprocessed electrostatic signal sequence is subjected to periodic pattern recognition and periodic consistency calculation to obtain periodic characteristic parameters. Based on the accumulation key characteristic points, the preprocessed electrostatic signal sequence is fitted with a piecewise exponential function to obtain accumulation-discharge kinetic parameters; the accumulation-discharge kinetic parameters and the periodic characteristic parameters are time-series aligned and parameter normalized to obtain standardized characteristic data, and the standardized characteristic data is subjected to time-feature matrix construction and parameter encoding to obtain electrostatic time series characteristic parameters.

[0007] Optionally, in a second implementation of the first aspect of the present invention, based on the accumulated key feature points, the pre-processed electrostatic signal sequence is fitted with a piecewise exponential function to obtain accumulation-discharge kinetic parameters, including: determining the accumulation peak value in the accumulation kinetic parameters and the accumulation segment boundary and the discharge segment boundary corresponding to the peak moment, and based on the accumulation segment boundary and the discharge segment boundary, extracting data points and eliminating outliers for the accumulation segment and the discharge segment in the pre-processed electrostatic signal sequence to obtain a segmented data set; performing exponential function fitting on the accumulation segment data in the segmented data set and determining the accumulation time constant to generate an accumulation fitting result, and based on the accumulation fitting result, performing segmented fitting on the segmented data set. The discharge segment data in the embodiment of the present invention are used to determine the discharge segment starting position and the discharge time constant, and generate a discharge fitting result; a time constant ratio is calculated for the accumulation fitting result and the discharge fitting result to obtain a time constant ratio, and multiple complete accumulation-discharge cycles in the preprocessed electrostatic signal sequence are identified based on the time constant ratio; a time constant variation coefficient and a reproducibility index are calculated for the accumulation time constant and the discharge time constant corresponding to each of the accumulation-discharge cycles to obtain a time constant reproducibility feature, and a dynamic parameter vector is constructed and numerically normalized for the accumulation fitting result, the discharge fitting result, the time constant ratio and the time constant reproducibility feature to obtain an accumulation-discharge dynamic parameter.

[0008] Optionally, in a third implementation of the first aspect of the present invention, the electrostatic spatiotemporal distribution feature extraction and spatial distribution calculation are performed on the multi-period electrostatic evolution data based on the preset multi-point circumferential positions of the battery cell to obtain electrostatic fingerprint map data, including: based on the preset multi-point circumferential positions of the battery cell, the angle-static charge density mapping and spatial interpolation of the sensor array are performed on the multi-period electrostatic evolution data to obtain a spatially continuous electrostatic distribution, and spatial statistical calculation and distribution correlation calculation are performed on the spatially continuous electrostatic distribution to obtain spatial distribution statistical features; based on the spatial distribution statistical features, spatial gradient calculation and gradient mutation detection are performed on the corresponding adjacent angular positions in the spatially continuous electrostatic distribution to obtain spatial gradient feature parameters, and based on a preset local charge density threshold, local extreme value detection and outlier identification are performed on the spatial gradient feature parameters to obtain island marking data; spatial aggregation calculation is performed on the island marking data to obtain spatial abnormal distribution features, and multi-dimensional feature matrix construction and typical spatial distribution pattern identification are performed on the spatial distribution statistical features, the spatial gradient feature parameters and the spatial abnormal distribution features to obtain electrostatic fingerprint map data.

[0009] Optionally, in a fourth implementation of the first aspect of the present invention, the frequency domain transformation of the electrostatic fingerprint spectrum data and the spectrum characteristic calculation of the heat shrink tube-battery core interface friction frequency are performed to obtain a set of frequency domain characteristic parameters, including: setting a time-frequency analysis window and frame windowing in the heat shrink process stage of the electrostatic fingerprint spectrum data to obtain a windowed time-frequency data matrix, and performing time-frequency transformation and amplitude-phase separation of the heat shrink tube-battery core interface friction signal on the windowed time-frequency data matrix to obtain a time-frequency amplitude-phase distribution; performing static analysis of the time-frequency amplitude-phase distribution on the battery surface The instantaneous frequency and mean variance of the electrical change are calculated to obtain the instantaneous frequency characteristic parameters, and the power spectrum density estimation and frequency band energy distribution calculation of the time-frequency amplitude and phase distribution in the heat shrinkage process stage are performed to obtain the frequency band energy distribution characteristics, and the multi-dimensional modulation depth of the heat shrink tube-battery core interface is calculated for the time-frequency amplitude and phase distribution to obtain the multi-dimensional modulation characteristic parameters; the characteristic vector combination and characteristic screening of the instantaneous frequency characteristic parameters, the frequency band energy distribution characteristics and the multi-dimensional modulation characteristic parameters are performed for heat shrink tube-battery core matching to obtain a frequency domain characteristic parameter set.

[0010] Optionally, in the fifth implementation method of the first aspect of the present invention, based on the preset battery cell and heat shrink tube specification parameters, the electrostatic timing feature parameters and the frequency domain feature parameter set are subjected to feature fusion and weight optimization calculation of the heat shrink tube and battery cell matching degree to obtain a comprehensive feature vector, including: feature dimension alignment and time synchronization of the electrostatic timing feature parameters and the frequency domain feature parameter set to obtain the original feature data set, and feature correlation calculation and principal component dimensionality reduction of the original feature data set to obtain the reduced dimension feature space; weight coefficient optimization search of the electrostatic timing feature and frequency domain feature in the reduced dimension feature space to obtain a multi-objective optimization weight matrix, and based on the preset battery cell and heat shrink tube specification parameters, weight compensation calculation is performed on the multi-objective optimization weight matrix to obtain compensated feature weights; based on the compensated feature weights, weighted fusion calculation and matching pattern clustering recognition of heat shrink tube and battery cell matching are performed on the reduced dimension feature space to obtain a comprehensive feature vector.

[0011] Optionally, in a sixth implementation method of the first aspect of the present invention, the matching evaluation data is quantitatively calculated and the confidence calculated to generate a matching detection result, including: quantitatively calculating and grading the matching level of the heat shrink tube and the battery cell on the matching evaluation data to obtain a quantitative matching score; based on the quantitative matching score, performing confidence calculation and stability evaluation on the matching evaluation data to obtain a confidence evaluation index; and comprehensively evaluating the matching performance of the heat shrink tube and the battery cell on the quantitative matching score and the confidence evaluation index to generate a matching detection result.

[0012] The second aspect of the present invention provides a device for detecting the degree of matching between a battery heat shrink tube and a battery cell. The device comprises: a data acquisition module for collecting, based on preset standardized electrostatic detection parameters, multi-period electrostatic evolution data generated when a target heat shrink tube is wrapped around a target battery cell; a feature extraction module for performing time domain segmentation of the heat shrink process and electrostatic accumulation-discharge cycle identification on the multi-period electrostatic evolution data to obtain electrostatic time series characteristic parameters, and performing electrostatic spatiotemporal distribution feature extraction and spatial distribution calculation on the multi-period electrostatic evolution data based on preset multi-point positions around the battery cell to obtain electrostatic fingerprint map data. The electrostatic fingerprint map data is subjected to frequency domain transformation and spectral feature calculation of the heat shrink tube-battery cell interface friction frequency to obtain a frequency domain feature parameter set; a feature fusion module is used to perform feature fusion and weight optimization calculation of the matching degree between the heat shrink tube and the battery cell on the electrostatic timing feature parameters and the frequency domain feature parameter set based on preset battery cell and heat shrink tube specification parameters to obtain a comprehensive feature vector, and the comprehensive feature vector is subjected to matching mode classification and membership function mapping between the heat shrink tube and the battery cell to obtain matching evaluation data; a result generation module is used to perform quantitative calculation and confidence calculation on the matching evaluation data to generate a matching detection result.

[0013] A third aspect of the present invention provides a device for detecting the degree of matching between a battery heat shrink tube and a battery cell, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory to enable the device for detecting the degree of matching between a battery heat shrink tube and a battery cell to perform each step of the above-mentioned method for detecting the degree of matching between a battery heat shrink tube and a battery cell.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is run on a computer, it enables the computer to execute each step of the above-mentioned method for detecting the matching degree between the battery heat shrink tube and the battery cell.

[0015] The above-mentioned battery heat shrink tubing and battery cell matching detection method and related equipment. In an embodiment of the present invention, the electrostatic parameters of the target heat shrink tubing wrapped around the target battery cell are collected in multiple time periods and feature decoupled to obtain basic electrostatic features, and then these features are segmented in the time domain and reconstructed in spatial distribution to obtain electrostatic time series features and electrostatic fingerprint maps; then, a set of frequency domain feature parameters is obtained through frequency domain transformation and spectral feature calculation, and multi-layer feature fusion and weight optimization are performed in combination with the battery cell and heat shrink tubing specification parameters to obtain a comprehensive feature vector and map the matching evaluation data; finally, fuzzy logic reasoning and multi-level decision-making are performed based on the matching evaluation data to output the matching detection results. Through hierarchical electrostatic data processing and feature analysis, the problem of accurate identification of the matching degree between heat shrink tubing and battery cells was solved. In particular, in terms of static electricity accumulation-discharge cycle, interface friction characteristics and spatial distribution uniformity, the dynamic characteristics and matching characteristics of the heat shrink wrapping process were fully considered, effectively improving the detection accuracy. In addition, multi-dimensional electrostatic feature fusion and pattern classification strategies were adopted to realize the correlation analysis between time domain, spatial domain and frequency domain features, and enhance the reliability of matching degree assessment. In addition, through electrostatic fingerprint map construction and membership function mapping, the matching pattern and confidence level were accurately identified, thereby realizing efficient and accurate detection of the matching degree between heat shrink tubing and battery cells as a whole.

[0016] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the present invention. The objectives and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a schematic diagram of a first embodiment of a method for detecting the degree of matching between a battery heat shrink tube and a battery cell according to an embodiment of the present invention; Figure 2 A schematic diagram of an embodiment of a device for detecting the degree of matching between a battery heat shrink tube and a battery cell according to an embodiment of the present invention; Figure 3 Schematic diagram of an embodiment of a device for detecting the degree of matching between a battery heat shrink tube and a battery cell in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0020] The terms "including," "having," and any variations thereof, as used in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to the process, method, product, or device.

[0021] To facilitate understanding of this embodiment, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the method for detecting the matching degree between a battery heat shrink tube and a battery cell in the embodiment of the present invention includes: 101. Based on the preset standardized electrostatic detection parameters, collect the multi-period electrostatic evolution data generated when the target heat shrink tubing is wrapped around the target battery cell; The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0022] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0023] In this embodiment, before collecting multi-time period electrostatic evolution data, it also includes: performing zero point calibration, sensitivity calibration and linearity calibration on a preset electrostatic sensor array by using a preset standard electrostatic power source to obtain response characteristic parameters of each electrostatic sensor, and performing inter-sensor deviation calculation and correction coefficient determination on each response characteristic parameter to obtain a sensor calibration parameter table; based on a pre-established multi-point grounding network and an insulating isolation device to form an electrostatic potential reference system, the detection environment is temperature and humidity controlled and electromagnetic interference shielded to obtain environmental control parameters, and based on the environmental control parameters, the sensor calibration parameter table is subjected to environmental compensation and inter-channel consistency correction to obtain corrected calibration data; background noise measurement and signal quality evaluation are performed on the corrected calibration data to obtain standardized electrostatic detection parameters, and then based on the standardized electrostatic detection parameters, a variety of electrostatic characteristic parameter data generated by each electrostatic sensor channel when the target heat shrink tube covers the target battery cell is collected to obtain multi-time period electrostatic evolution data.

[0024] In practical applications, a standard electrostatic source is first used to perform a triple calibration of the electrostatic sensor array (i.e., zero-point calibration, sensitivity calibration, and linearity calibration). The electrostatic sensor array refers to multiple electrostatic sensors arranged at equal angular intervals along the circumference of the target cell, with each sensor spaced 30 degrees apart. A total of 12 sensors provide 360-degree coverage. Zero-point calibration involves recording the background reading of each sensor in the absence of an electrostatic source to establish a zero reference point. Sensitivity calibration involves stimulating each sensor individually with a standard electrostatic source of known charge, recording the ratio of the output response value to the input charge. Linearity calibration tests the linearity of the sensor response at different charge levels. For example, applying standard charges of 1pC, 5pC, and 10pC to sensor No. 1 yields response voltages of 0.98mV, 4.95mV, and 9.92mV, respectively. The calculated sensitivity of the sensor is 0.99mV / pC, which results in response characteristic parameters including zero-point offset, sensitivity coefficient, and linearity error. Then, the inter-sensor deviation calculation and correction coefficient determination are performed on each response characteristic parameter. The inter-sensor deviation calculation is to compare the response differences of each sensor under the same excitation, and the correction coefficient determination is to calculate the numerical correction required to eliminate the deviation; thus, a sensor calibration parameter table is obtained that records the zero point value, sensitivity coefficient, linear error and correction coefficient of each sensor.

[0025] An electrostatic potential reference system is constructed based on a multi-point grounding network and an insulation isolation device. The detection environment is temperature- and humidity-controlled, as well as shielded from electromagnetic interference (EMI). The multi-point grounding network involves placing multiple grounding electrodes around the detection area to form a stable zero-potential reference surface. The insulation isolation device places an insulating support platform between the battery cell and the grounding network to prevent direct grounding of the battery cell from affecting the electrostatic measurement. Temperature and humidity control is achieved by maintaining the detection environment at a constant temperature of 23±1°C and relative humidity of 45±3%RH using constant temperature and humidity equipment. EMI shielding involves activating a Faraday cage shielding system (a metal mesh shielding enclosure installed around the sensor array to block external 50Hz power frequency interference and high-frequency electromagnetic signals) and monitoring the ambient electromagnetic noise level to below -60dBm. This generates environmental control parameters, including real-time temperature, humidity, and electromagnetic noise intensity. Based on these environmental control parameters, environmental compensation and inter-channel consistency correction are applied to the sensor calibration parameter table. Environmental compensation involves adjusting sensor calibration parameters based on temperature and humidity changes. Temperature compensation corrects for sensor zero drift using the temperature coefficient, while humidity compensation corrects for electrostatic leakage using the humidity coefficient. Inter-channel consistency calibration includes gain consistency and phase synchronization. Gain consistency calibration adjusts the amplification of each sensor channel to maintain consistency, while phase synchronization eliminates differences in signal acquisition time between channels. For example, when the ambient temperature changes from 23°C to 25°C, the zero point of sensor 1 drifts from 0.05mV to 0.12mV. A temperature coefficient of -0.035mV / °C is applied for compensation. This results in post-calibration data, which records the final calibration parameters after environmental compensation and consistency correction.

[0026] The calibration data is then subjected to noise floor measurement and signal quality assessment. The noise floor measurement involves continuously sampling the background signal of each sensor channel for 30 seconds in an unloaded state without heat shrink tubing or battery cells. The RMS value of the signal is then calculated as the noise baseline. The signal quality assessment verifies the signal-to-noise ratio and stability of each sensor channel. The SNR is required to be greater than 40dB, and the stability requirement is that the standard deviation of continuous measurements be less than 0.1% of the full scale. This evaluation yields standardized ESD detection parameters, including calibrated sensor sensitivity, zero offset, noise level, and dynamic range. During the heat shrink tubing wrapping process, each sensor channel simultaneously collects electrostatic characteristic parameter data (including but not limited to electrostatic charge density, electrostatic field strength, and discharge current). Electrostatic charge density is the amount of charge per unit area and is calculated by dividing the sensor's induced voltage by the sensor area and dielectric constant. Electrostatic field strength is the strength of the electric field in space and is calculated by dividing the potential difference between adjacent sensors by the distance. The discharge current is the current flowing through the discharge path and is measured by monitoring the ground loop current. Ultimately, multi-period static evolution data is collected (referring to the time series of static characteristic parameter changes during the entire process of heat shrink tubing, from initial heating to complete shrinkage). For example, when wrapping a battery cell with heat shrink tubing, the preheating phase occurs from 0 to 30 seconds, during which the static charge density slowly rises from 0.1 μC / m² to 0.3 μC / m². The rapid shrinkage phase occurs from 30 to 90 seconds, during which the static charge density rapidly rises to a peak of 1.2 μC / m² before beginning to decline. This step establishes a precise static detection benchmark, enabling real-time and accurate measurement of static characteristics during the heat shrink wrapping process.

[0027] 102. Perform time-domain segmentation of the heat shrink process and identification of the electrostatic accumulation-discharge cycle on the multi-period electrostatic evolution data to obtain electrostatic time series characteristic parameters. Based on the preset multi-point positions of the battery cell circumference, perform electrostatic spatiotemporal distribution feature extraction and spatial distribution calculation on the multi-period electrostatic evolution data to obtain electrostatic fingerprint spectrum data. Perform frequency domain transformation on the electrostatic fingerprint spectrum data and calculate the spectrum characteristics of the heat shrink tube-battery cell interface friction frequency to obtain a set of frequency domain characteristic parameters. In this embodiment, based on the shrinkage temperature characteristics of the target heat shrinkable tube, the multi-period electrostatic evolution data is segmented in the time domain of multiple heat shrinkage process stages to obtain segmented electrostatic data, and the segmented electrostatic data is filtered, denoised, and multiple data preprocessed to obtain a preprocessed electrostatic signal sequence; the electrostatic accumulation rate of the target heat shrinkable tube when the target battery cell is wrapped is calculated for the preprocessed electrostatic signal sequence, and the electrostatic accumulation peak on the surface of the battery cell is identified to obtain accumulation key feature points, and the preprocessed electrostatic signal sequence is subjected to periodic pattern recognition and periodic consistency calculation to obtain periodic characteristic parameters, and based on the accumulation key feature points, the preprocessed electrostatic signal sequence is fitted with a piecewise exponential function to obtain accumulation-discharge kinetic parameters (i.e., the accumulation peak value in the accumulation kinetic parameters and the accumulation segment boundary and the discharge segment boundary corresponding to the peak moment are determined, and based on the accumulation segment boundary and the discharge segment boundary, the accumulation segment and the discharge segment in the preprocessed electrostatic signal sequence are extracted and outliers are eliminated to obtain a segmented data set; the accumulation segment data in the segmented data set is exponentially fit. Function fitting and accumulation time constant determination generate accumulation fitting results, and based on the accumulation fitting results, determine the discharge segment starting position and discharge time constant of the discharge segment data in the segmented data set to generate discharge fitting results; calculate the time constant ratio of the accumulation fitting results and the discharge fitting results to obtain the time constant ratio, and identify multiple complete accumulation-discharge cycles in the preprocessed electrostatic signal sequence based on the time constant ratio; calculate the time constant variation coefficient and reproducibility index of the accumulation time constant and the discharge time constant corresponding to each accumulation-discharge cycle to obtain the time constant reproducibility feature, and construct and numerically normalize the dynamic parameter vector of the accumulation fitting results, the discharge fitting results, the time constant ratio and the time constant reproducibility feature to obtain the accumulation-discharge dynamic parameters); perform time series alignment and parameter normalization on the accumulation-discharge dynamic parameters and the periodic characteristic parameters to obtain standardized characteristic data, and construct a time-feature matrix and parameter encode the standardized characteristic data to obtain the electrostatic time series characteristic parameters. Based on the preset multi-point positions around the battery cell, the electrostatic evolution data of multiple time periods are mapped by angle-static charge density of the sensor array and spatial interpolation is performed to obtain the spatially continuous electrostatic distribution, and the spatially continuous electrostatic distribution is subjected to spatial statistical calculation and distribution correlation calculation to obtain the spatial distribution statistical characteristics; based on the spatial distribution statistical characteristics, the corresponding adjacent angular positions in the spatially continuous electrostatic distribution are subjected to spatial gradient calculation and gradient mutation detection to obtain the spatial gradient characteristic parameters, and based on the preset local charge density threshold, the spatial gradient characteristic parameters are subjected to local extreme value detection and outlier identification to obtain the island marking data; the island marking data is subjected to spatial aggregation calculation to obtain the spatial abnormal distribution characteristics, and the spatial distribution statistical characteristics, spatial gradient characteristic parameters and spatial abnormal distribution characteristics are subjected to multi-dimensional feature matrix construction and typical spatial distribution pattern identification to obtain the electrostatic fingerprint map data.The electrostatic fingerprint map data is subjected to time-frequency analysis window setting and frame windowing in the heat shrinking process stage to obtain a windowed time-frequency data matrix, and the windowed time-frequency data matrix is ​​subjected to time-frequency transformation and amplitude-phase separation of the friction signal of the heat shrink tube-battery cell interface to obtain a time-frequency amplitude-phase distribution; the time-frequency amplitude-phase distribution is subjected to instantaneous frequency calculation and mean-variance calculation of the electrostatic change on the cell surface to obtain instantaneous frequency characteristic parameters, and the time-frequency amplitude-phase distribution is subjected to power spectral density estimation and frequency band energy distribution calculation in the heat shrinking process stage to obtain frequency band energy distribution characteristics, and the time-frequency amplitude-phase distribution is subjected to multi-dimensional modulation depth calculation of the heat shrink tube-battery cell interface to obtain multi-dimensional modulation characteristic parameters; the instantaneous frequency characteristic parameters, frequency band energy distribution characteristics and multi-dimensional modulation characteristic parameters are subjected to characteristic vector combination and feature screening for heat shrink tube-battery cell matching to obtain a set of frequency domain characteristic parameters.

[0028] In practical applications, firstly, based on the shrinkage temperature characteristics of the target heat shrinkable tube (referring to the shrinkage behavior of the heat shrinkable tube at different temperatures, for example, PVC heat shrinkable tube begins to soften at 60°C, shrinks rapidly at 80°C, and shrinks completely at 120°C), the multi-period electrostatic evolution data is subjected to time-domain segmentation processing of various heat shrinkage process stages (including preheating stage (0-30 seconds, temperature rises from room temperature to 60°C), rapid shrinkage stage (30-90 seconds, temperature rises from 60°C to 120°C), and stabilization stage (90-180 seconds, temperature maintained at 120°C)). That is, according to the real-time temperature value monitored by the temperature sensor, the continuous electrostatic evolution data is time-sliced ​​according to the process stage to obtain segmented electrostatic data (referring to the time of each process stage). The electrostatic charge density, electrostatic field strength and discharge current time series data corresponding to each process stage are collected); and the segmented electrostatic data are filtered, denoised and preprocessed with multiple data. The filtering and denoising adopts a Butterworth low-pass filter with a cutoff frequency set to 50Hz to eliminate high-frequency noise interference. The multiple data preprocessing includes outlier detection, missing value interpolation and data smoothing. The outlier detection adopts the 3 times standard deviation criterion to eliminate data points that deviate too much from the mean. The missing value interpolation adopts the linear interpolation method to fill the missing data. The data smoothing adopts a 5-point moving average filter to obtain the preprocessed electrostatic signal sequence. It can eliminate the noise interference and abnormal fluctuations in the original data and provide a stable and reliable electrostatic signal foundation.

[0029] The electrostatic accumulation rate and electrostatic accumulation peak are calculated for the pre-processed electrostatic signal sequence, that is, the instantaneous accumulation rate is obtained by calculating the difference in electrostatic charge density between adjacent time points and dividing it by the time interval, and the peak position and value are determined by finding the local maximum point of the electrostatic charge density, and the accumulation key feature points (the time position of the accumulation starting point, the accumulation peak point and the discharge starting point and the corresponding electrostatic charge density value) are obtained; and the periodic pattern recognition and periodic consistency calculation are performed on the pre-processed electrostatic signal sequence, that is, by detecting the repeated accumulation-discharge pattern in the electrostatic signal, the start and end time points of each complete cycle are identified, and then the similarity of each accumulation-discharge cycle in time length and amplitude change is evaluated, and the coefficient of variation of the time difference and amplitude difference between adjacent cycles is calculated to obtain the periodic characteristic parameters (including the number of cycles, average cycle length, cycle length coefficient of variation and Periodic amplitude variation coefficient); then, based on the accumulation key characteristic points, a piecewise exponential function is fitted to obtain the accumulation-discharge kinetic parameters, that is, by determining the accumulation peak value and the accumulation segment boundary and the discharge segment boundary corresponding to the peak moment in the accumulation kinetic parameters, wherein the accumulation segment boundary corresponding to the accumulation peak value and the peak moment refers to the time interval from the static accumulation starting point to the peak point, and the discharge segment boundary refers to the time interval from the peak point to the next accumulation starting point; and based on the accumulation segment boundary and the discharge segment boundary, the accumulation segment and the discharge segment in the pre-processed electrostatic signal sequence are subjected to data point extraction and outlier removal, wherein data point extraction is to extract all data points within the corresponding time range of the accumulation segment and the discharge segment from the pre-processed electrostatic signal sequence, and outlier removal is to delete data points that deviate from the fitting curve by more than 2 times the standard deviation, and obtain a segmented data set (including all accumulation segment data sets and discharge segment data sets). Then, an exponential function is fitted to the accumulation segment data in the segmented data set. Since the accumulation process conforms to the law of exponential growth, ,in represents the electrostatic charge density at time t, A represents the accumulated saturation value, represents the accumulation time constant, t represents the time variable, and the accumulation time constant Reflects the speed of static electricity accumulation. The smaller the value, the faster the accumulation. This determines the generation of accumulation fitting results (record the saturation value A and time constant of each accumulation segment). Based on the accumulated fitting results, the exponential function is fitted to the discharge segment data in the segmented data set. Since the discharge process conforms to the exponential decay law , where B represents the discharge start value, Indicates the discharge time constant, and the discharge time constant Reflects the speed of electrostatic discharge. The smaller the value, the faster the discharge. This determines the discharge fitting result (record the starting value B and time constant of each discharge segment). ); and then calculate the time constant ratio of the accumulation fitting result and the discharge fitting result, that is, by dividing the accumulation time constant by Divide by the discharge time constant Get the ratio , this ratio reflects the relative speed relationship between the accumulation and discharge processes. The complete accumulation-discharge cycle is identified based on the time constant ratio. When the ratio R is in the range of 0.5-2.0, it is considered to be a complete accumulation-discharge cycle. Then, for the accumulation time constant and discharge time constant corresponding to each accumulation-discharge cycle, the standard deviation of the time constant corresponding to each accumulation-discharge cycle is divided by the average value to reflect the stability of the time constant, and the consistency of the time constant under the same conditions is evaluated. The relative standard deviation of the multiple measurement results is calculated to obtain the time constant reproducibility characteristics (including the accumulation time constant variation coefficient, the discharge time constant variation coefficient and the time constant ratio reproducibility); then, the accumulation fitting results, discharge fitting results, time constant ratio and time constant reproducibility characteristics are constructed and numerically normalized for the kinetic parameter vector, where the construction of the kinetic parameter vector is to combine the accumulation fitting results, discharge fitting results, time constant ratio and time constant reproducibility characteristics into a multidimensional feature vector, and the numerical standardization is to normalize each parameter to the range of 0-1 according to the maximum-minimum value method, and finally generate the accumulation-discharge kinetic parameters (a set of quantitative parameters describing the dynamic characteristics of the electrostatic accumulation and discharge process). The dynamic characteristics of the electrostatic accumulation-discharge process are accurately quantified through mathematical fitting methods. The time constant ratio can effectively distinguish the differences in electrostatic behavior under different matching states, providing a quantitative physical basis for matching judgment.

[0030] Accumulation-discharge kinetic parameters and periodic characteristic parameters were time-series aligned and normalized. This process unified the time bases of different parameter types onto the same time axis to ensure accurate temporal correspondence between parameters. Parameters with different dimensions and numerical ranges were converted to dimensionless standardized values. Maximum-minimum normalization was used to map parameter values ​​to the range 0-1, resulting in standardized characteristic data. A time-feature matrix was then constructed and parameter encoding was performed on the standardized characteristic data. The time-feature matrix is ​​a two-dimensional data matrix with time as rows and characteristic parameters as columns. Each element in the matrix represents the value of a specific characteristic parameter at a specific time point. For example, at 90 seconds, the accumulation time constant is 12.5 seconds (normalized to 0.35), the discharge time constant is 8.2 seconds (normalized to 0.28), and the time constant ratio is 1.52 (normalized to 0.65). Parameter encoding converts the numerical information in the time-feature matrix into a structured feature vector representation. Sequential encoding is used to encode the feature evolution process in the time dimension into a fixed-length feature vector. The electrostatic time series feature parameters are the final feature representation after encoding, containing the complete electrostatic evolution time series information and dynamic characteristics.

[0031] Based on a pre-set multi-point circumferential position of the cell (referring to 12 sensor positions equally spaced along the cell circumference, with an angular interval of 30 degrees, located at 0, 30, 60, and up to 330 degrees), the sensor array performs angular-to-static charge density mapping and spatial mapping on the multi-period electrostatic evolution data. This involves establishing a one-to-one correspondence between the angular coordinates of each sensor position and the corresponding measured static charge density value, forming discrete spatially distributed data points. A continuous static charge density distribution curve is then generated between adjacent sensors using cubic spline interpolation, expanding the 12 discrete measurement points into an electrostatic distribution at 360 continuous angular points, resulting in a spatially continuous electrostatic distribution. For example, if the sensor measurement value at 0 degrees is 1.2 μC / m² and the sensor measurement value at 30 degrees is 0.8 μC / m², the static charge density at 15 degrees is calculated to be 1.05 μC / m² via cubic spline interpolation. The spatially continuous electrostatic distribution is the 360-degree continuous electrostatic charge density distribution data obtained through interpolation. Then, spatial statistics and distribution correlation calculations are performed on the spatially continuous electrostatic distribution. Spatial statistics include calculating statistical quantities such as the mean, standard deviation, skewness, and kurtosis of the electrostatic charge density, while distribution correlation calculations analyze the correlation coefficient and spatial autocorrelation of the electrostatic charge density between adjacent angular regions to obtain spatial distribution statistical characteristics (including quantitative indicators of the central tendency, degree of dispersion, and spatial correlation of the electrostatic distribution). This converts discrete sensor measurement data into a continuous spatial distribution. Spatial interpolation effectively fills the data gaps between sensors, and statistical calculations comprehensively describe the spatial characteristics of the electrostatic distribution.

[0032] Based on the statistical characteristics of spatial distribution, spatial gradient calculation and anomaly detection are performed on the corresponding adjacent angular positions in the spatially continuous electrostatic distribution (referring to the angular points with adjacent 1-degree intervals in the spatially continuous electrostatic distribution). That is, the gradient value of each position is obtained by calculating the difference in electrostatic charge density between adjacent angular points and dividing it by the angular interval. The position points where the gradient value exceeds the normal fluctuation range are identified. The sliding window method is used to calculate the difference between the gradient value of each point and the average gradient within the surrounding 10-degree range. When the difference exceeds 3 times the standard deviation, it is determined to be a gradient mutation point, and the spatial gradient characteristic parameters (including the maximum gradient value, the number of gradient mutation points, the gradient change rate, and the gradient distribution uniformity index) are obtained. Based on the preset local charge density threshold (which is the abnormal judgment benchmark value determined by the average value plus 2 times the standard deviation in the spatial distribution statistical characteristics), local extreme value detection and outlier identification are performed on the spatial gradient characteristic parameters. That is, the local maximum and minimum points of the electrostatic charge density are found in the spatially continuous electrostatic distribution, and the extreme value points exceeding the local charge density threshold are marked as abnormal areas. For example, when the average electrostatic charge density in a battery cell test is 0.5μC / m² and the standard deviation is 0.15μC / m², the local charge density threshold is set to 0.8μC / m², and a measurement value of 1.1μC / m² at an angle of 120 degrees is identified as an outlier. This generates island marking data (a data set that records the angular position, degree of anomaly, and type of anomaly area). Gradient analysis effectively identifies discontinuities and sudden changes in the electrostatic distribution. Anomaly detection precisely locates the spatial location of poor contact between the heat shrink tubing and the battery cell. Island marking provides key anomaly indication information for matching assessment.

[0033] For the isolated island marker data, adjacent abnormal marker points are grouped and clustered according to their spatial proximity. The density clustering algorithm is used to merge abnormal points with an angular interval less than 15 degrees into the same abnormal area, and the spatial anomaly distribution characteristics (including the number of abnormal areas, the angular span of a single abnormal area, the interval distribution between abnormal areas, and the spatial distribution pattern of anomaly intensity) are obtained. For example, during one inspection, three isolated islands were identified, located at 60-75 degrees, 150-165 degrees, and 240-270 degrees, respectively. The total angular span of the abnormal regions was 60 degrees, accounting for 16.7% of the total circumference. A multidimensional feature matrix was then constructed based on the spatial distribution statistical characteristics, spatial gradient characteristic parameters, and spatial anomaly distribution characteristics, and typical spatial distribution patterns were identified. This structure was organized into a multidimensional array structure, with each row representing an angular position and each column representing a characteristic parameter. A pattern matching algorithm was then used to identify characteristic pattern types in the electrostatic distribution, including uniform distribution patterns, concentrated distribution patterns, bimodal distribution patterns, and multi-point anomaly distribution patterns. This yielded electrostatic fingerprint data (a comprehensive feature description that integrates multiple features, similar to the feature point distribution map in fingerprint recognition, used to uniquely identify a specific heat shrink tubing-battery cell matching state). Spatial aggregation effectively integrated scattered anomaly information, resulting in a multidimensional feature matrix that comprehensively described the spatial characteristics of the electrostatic distribution. The electrostatic fingerprint provided unique spatial signatures for different matching states.

[0034] Time-frequency analysis is performed on the ESD fingerprint data to extract frequency domain features. The time-frequency analysis window for the heat shrink process is determined based on the temporal characteristics of the different heat shrink stages. A 5-second window length is used during the preheating phase to capture slowly changing features, a 2-second window length is used during the rapid shrinking phase to capture rapidly changing features, and an 8-second window length is used during the stabilization phase to analyze steady-state features. Frame segmentation and windowing are performed to segment the ESD fingerprint data according to the specified window lengths. A Hanning window function is applied to each data segment to reduce spectral leakage, resulting in a windowed time-frequency data matrix. For example, a 120-second heat shrink process is divided into 60 overlapping data frames, with an inter-frame overlap of 50% to ensure continuity in the time-frequency analysis. The windowed time-frequency data matrix is ​​a two-dimensional data array after framing and windowing, with rows representing time frames and columns representing the spatial location data of the ESD fingerprint. Then, for the windowed time-frequency data matrix, the short-time Fourier transform method is used to convert the electrostatic change signal in the time domain into the time-frequency domain to obtain the frequency component distribution at each time point, and the complex time-frequency transform result is decomposed into the amplitude part and the phase part. The amplitude part reflects the intensity of each frequency component, and the phase part reflects the time delay relationship of each frequency component. The time-frequency amplitude-phase distribution (which records the amplitude and phase information corresponding to each time-frequency point) is obtained, thereby providing complete frequency domain information for in-depth analysis of the dynamic characteristics of the friction process.

[0035] For the time-frequency amplitude-phase distribution, the instantaneous frequency value at each moment is obtained by time-differentiating the phase information, which reflects the speed of electrostatic change, and the mean and variance of the instantaneous frequency sequence are calculated. The mean reflects the central trend of the instantaneous frequency, and the variance reflects the degree of fluctuation of the instantaneous frequency, and the instantaneous frequency characteristic parameters (including the instantaneous frequency mean, instantaneous frequency variance, instantaneous frequency maximum and instantaneous frequency change rate) are obtained; and for the time-frequency amplitude-phase distribution, by calculating the power distribution of each frequency component, the Welch method is used to perform power spectrum estimation on the overlapping data segments to improve the estimation accuracy, and the spectrum is divided into three intervals: low frequency band (0-5Hz), medium frequency band (5-20Hz) and high frequency band (20-50Hz). The total energy proportion in each frequency band is calculated respectively, and the frequency band energy distribution characteristics (including the energy proportion of each frequency band, the energy ratio between frequency bands and the dominant frequency band identification result) are obtained. For example, when the heat shrink tubing is tightly fitted, the low-frequency energy accounts for 65%, the mid-frequency energy accounts for 25%, and the high-frequency energy accounts for 10%. However, in an overly loosely matched state, the high-frequency energy accounts for a significant 35%. Furthermore, by analyzing the modulation phenomena in the time-frequency amplitude-phase distribution, including amplitude modulation depth and frequency modulation depth, the amplitude modulation depth reflects the degree of periodic variation in the amplitude of the electrostatic signal, and the frequency modulation depth reflects the degree of periodic variation in the instantaneous frequency, multidimensional modulation characteristic parameters (including amplitude modulation depth, frequency modulation depth, modulation frequency, and modulation stability index) are obtained. Frequency domain analysis deeply reveals the frequency characteristics of electrostatic changes, power spectral density estimation accurately quantifies the relative strengths of different frequency components, and multidimensional modulation analysis effectively identifies the complex dynamic characteristics of interfacial friction during the heat shrink process.

[0036] The various extracted frequency-domain feature parameters are combined, optimized, and screened. Specifically, the instantaneous frequency feature parameters, frequency band energy distribution features, and multidimensional modulation feature parameters are combined based on relevance and importance to form a comprehensive feature vector containing 15 feature components. The mutual information method is then used to evaluate the correlation between each feature parameter and the matching status, and the feature parameters with the top 10 mutual information values ​​are selected as the core feature set. For example, feature parameters such as instantaneous frequency variance, high-frequency band energy proportion, and amplitude modulation depth have a high ability to distinguish matching states and are therefore prioritized for inclusion in the core feature set. The feature screening process also includes redundant feature removal. By calculating the correlation coefficient between features, highly correlated redundant features are identified, and feature parameters with greater information content are retained, ultimately resulting in a set of frequency-domain feature parameters (containing the most valuable frequency-domain feature information for matching judgment).

[0037] 103. Based on the preset battery cell and heat shrink tubing specifications, perform feature fusion on the electrostatic timing feature parameter and frequency domain feature parameter set and weight optimization calculation on the matching degree between the heat shrink tubing and the battery cell to obtain a comprehensive feature vector. Then, perform matching mode classification and membership function mapping on the comprehensive feature vector between the heat shrink tubing and the battery cell to obtain matching evaluation data. In this embodiment, the electrostatic timing feature parameters and the frequency domain feature parameter sets are aligned in feature dimensions and synchronized in time to obtain the original feature data set, and the feature correlation calculation and principal component dimensionality reduction of the heat shrink tube and battery cell matching are performed on the original feature data set to obtain the reduced-dimensional feature space; the weight coefficients of the electrostatic timing features and frequency domain features are optimized and searched for in the reduced-dimensional feature space to obtain a multi-objective optimization weight matrix, and based on the preset battery cell and heat shrink tube specification parameters, the multi-objective optimization weight matrix is ​​weighted compensated to obtain the compensated feature weight; based on the compensated feature weight, the weighted fusion calculation and matching pattern clustering recognition of the heat shrink tube and battery cell matching are performed on the reduced-dimensional feature space to obtain a comprehensive feature vector.

[0038] In practical applications, for the electrostatic time series feature parameters and frequency domain feature parameter sets, the feature parameters of different dimensions are unified into the same data structure. The electrostatic time series feature parameters contain 15 time domain feature components, and the frequency domain feature parameter set contains 10 frequency domain feature components. The two are aligned into a 20-dimensional feature vector by zero padding. In order to ensure the precise correspondence between the time series features and the frequency domain features on the time axis, the timestamp matching method is used to pair and combine the feature data with completely consistent sampling times to obtain the original feature data set (containing 25 feature dimensions and 120 time sampling points); then, for the original feature data set, the Pearson correlation coefficient is used to analyze the linear relationship strength between each feature parameter and the matching status. Features with an absolute value of the correlation coefficient greater than 0.6 are considered to have strong matching indication ability, and the principal component analysis method is used to compress the 25-dimensional original feature space into an 8-dimensional reduced dimensionality feature space, retaining the principal components with a cumulative contribution rate of 85% to reduce data redundancy while maintaining key information. For example, the eigenvalues ​​of the first eight principal components are 5.2, 3.8, 2.1, 1.9, 1.4, 1.1, 0.8, and 0.6, respectively, cumulatively explaining 85.3% of the variance in the original data. The reduced feature space is a low-dimensional representation of features after principal component analysis, effectively eliminating the effects of collinearity between features.

[0039] The weight coefficients of the feature space after dimensionality reduction are optimized and the specification parameters are adapted. A genetic algorithm is used to search for the optimal weight combination in the range of 0-1. The fitness function is set as the weighted average of the matching classification accuracy and feature stability to obtain a multi-objective optimization weight matrix (recording the optimal ratio of the time series feature weights and the frequency domain feature weights. After 500 generations of genetic algorithm iterations, the optimal configuration of the time series feature weights being 0.65 and the frequency domain feature weights being 0.35 is obtained). Based on the preset battery cell and heat shrink tubing specification parameters (including physical parameters such as battery cell diameter, heat shrink tubing shrinkage ratio, wall thickness difference, and material elastic modulus, which directly affect the electrostatic accumulation characteristics during the heat shrinking process), the multi-objective optimization weight matrix is ​​corrected according to the differences in specification parameters. When the battery cell diameter is larger, the weight coefficient of the spatial feature is increased, and when the heat shrink tubing shrinkage ratio is higher, the weight coefficient of the time series feature is increased to obtain the compensated feature weight. For example, for large-size battery cells with a diameter of 32mm, the spatial feature weight was adjusted from 0.3 to 0.45; and for heat shrink tubing with a high shrinkage ratio of 3:1, the temporal feature weight was adjusted from 0.65 to 0.75. This enables adaptive adjustment of the detection method for products of different specifications.

[0040] Based on the compensated feature weights, the features of each dimension in the reduced dimensionality space are linearly weighted according to the compensated feature weights to obtain a single integrated feature value. The K-means clustering algorithm is then used to divide historical test data into three cluster centers: perfect match, too loose match, and too tight match. New test samples are then classified by calculating their Euclidean distance to each cluster center to obtain a comprehensive feature vector. A fuzzy logic evaluation system, constructed based on the comprehensive feature vector, maps continuous feature values ​​(such as static electricity accumulation rate, discharge efficiency, and spatial uniformity) to discrete matching levels. The static electricity accumulation rate is the inverse of the accumulation time constant extracted from the static electricity time series characteristic parameters, reflecting the speed of static electricity accumulation. A trapezoidal membership function is established to divide the accumulation rate into three fuzzy sets: slow, medium, and fast. Discharge efficiency and spatial uniformity are two key metrics that are calculated from the processed characteristic parameters. Discharge efficiency is calculated using the discharge time constant from the accumulation-discharge kinetic parameters. A smaller value indicates faster discharge and higher efficiency. A membership function categorizes discharge efficiency into low, medium, and high levels. Spatial uniformity is calculated using the spatial distribution statistics from the electrostatic fingerprint data. A smaller coefficient of variation indicates a more uniform distribution. A membership function categorizes spatial uniformity into poor, fair, and good levels. For example, an accumulation rate of 15 seconds corresponds to a medium-speed membership of 0.8, a discharge efficiency of 95% corresponds to a high-efficiency membership of 0.9, and a spatial uniformity coefficient of variation of 0.12 corresponds to a good-quality membership of 0.85. Matching evaluation data is the final match quantification derived from the membership values ​​of the three key metrics using fuzzy inference rules. Values ​​range from 0 to 1, with values ​​above 0.8 indicating an excellent match, 0.6-0.8 indicating a good match, and below 0.6 indicating an unsatisfactory match. This comprehensive evaluation of the three metrics comprehensively reflects the multidimensional nature of the matching status.

[0041] 104. Perform quantitative calculation and confidence calculation on the matching evaluation data to generate a matching test result.

[0042] In this embodiment, the matching evaluation data is quantitatively calculated and graded to obtain a quantitative matching score; based on the quantitative matching score, the matching evaluation data is confidence calculated and stability evaluated to obtain a confidence evaluation index; the quantitative matching score and the confidence evaluation index are used to comprehensively evaluate the matching performance of the heat shrink tubing and the battery cell to generate a matching test result.

[0043] In practical applications, the matching evaluation data is quantitatively calculated and graded. This is done by converting the continuous matching value output by the fuzzy mapping system into a discrete grade score. A linear mapping method is used to proportionally convert the matching value in the range of 0-1 into a 0-100 scoring system. Furthermore, the quantitative scores are divided into five grades based on the quality standards of the battery manufacturing industry: excellent (90-100 points), good (80-89 points), qualified (70-79 points), warning (60-69 points), and unqualified (0-59 points). For example, when the fuzzy mapping system outputs a matching degree of 0.85, the quantitative matching score is 85 points, corresponding to a good matching state; when the output matching degree is 0.65, the quantitative score is 65 points, corresponding to a warning matching state, ultimately resulting in a quantitative matching score.

[0044] Based on the quantitative matching score, the matching assessment data is subjected to confidence calculation and stability assessment. The confidence calculation is performed by quantitatively evaluating the consistency and credibility of each component in the matching assessment data, and analyzing the membership distribution characteristics of the three key indicators of electrostatic accumulation rate, discharge efficiency and spatial uniformity in the fuzzy mapping process to calculate the confidence value. The specific method is to calculate the degree of dispersion of the membership value of each indicator. When the membership value of an indicator is relatively evenly distributed in multiple fuzzy sets, it indicates that the judgment uncertainty is high and the confidence is low; when the membership value dominates in a fuzzy set, it indicates that the judgment certainty is high and the confidence is high. The stability assessment is to analyze the fluctuation of each characteristic parameter in the matching assessment data throughout the detection cycle, and evaluate the data stability by calculating the time series variance of the accumulation-discharge dynamics parameters, spatial distribution statistical characteristics and frequency domain characteristic parameters. For example, when the membership value corresponding to the electrostatic accumulation rate is 0.85 in the medium-speed set and 0.15 in the fast set, it indicates that the judgment certainty is high and the confidence coefficient is 0.85; when the coefficient of variation of the spatial distribution statistical characteristics during the detection process is less than 5%, it indicates that the data stability is good, and finally the confidence evaluation index is obtained (including the judgment certainty coefficient of each key indicator, the characteristic parameter stability coefficient and the comprehensive confidence coefficient).

[0045] The final test result is generated by comprehensively analyzing the quantitative match score and confidence assessment indicators. Specifically, the quantitative match score is used as the primary evaluation basis, and the confidence assessment indicators are used as correction factors for the reliability of the results. A weighted calculation is performed to obtain the revised final score. When the combined confidence coefficient is greater than 0.8, the final score is equal to the quantitative match score. When the confidence coefficient is between 0.6 and 0.8, the final score is reduced by 5-15 points to reflect the uncertainty. When the confidence coefficient is below 0.6, the test result is marked as unreliable and a retest is recommended. Finally, a complete report (including the final score, match level, confidence coefficient, test timestamp, and anomaly indications) is generated. For example, a test with a quantitative match score of 88 and a combined confidence coefficient of 0.92 would produce the following output: "Match: 88, Grade: Good, Confidence: 92%, Test Time: 2 minutes 35 seconds, Status: Normal." The abnormal indication information records the spatial distribution abnormalities, time series characteristic abnormalities or frequency domain characteristic abnormalities found during the detection process, providing reference information for subsequent process optimization.

[0046] In an embodiment of the present invention, through hierarchical electrostatic data processing and feature analysis, the problem of accurate identification of the matching degree detection of heat shrink tubing and battery cells is solved, especially in terms of static electricity accumulation-discharge cycle, interface friction characteristics and spatial distribution uniformity, the dynamic characteristics and matching characteristics of the heat shrink wrapping process are fully considered, and the detection accuracy is effectively improved; and a multi-dimensional electrostatic feature fusion and pattern classification strategy is adopted to realize the correlation analysis between time domain, space domain and frequency domain features, and enhance the reliability of matching degree evaluation; in addition, through electrostatic fingerprint map construction and membership function mapping, the matching pattern and confidence level are accurately identified, thereby realizing efficient and accurate detection of the matching degree between heat shrink tubing and battery cells as a whole.

[0047] The above describes the method for detecting the matching degree between the battery heat shrink tube and the battery cell according to the embodiment of the present invention. The following describes the device for detecting the matching degree between the battery heat shrink tube and the battery cell according to the embodiment of the present invention. Figure 2 In one embodiment of the present invention, a device for detecting the degree of matching between a battery heat shrink tube and a battery cell includes: The data acquisition module 201 is used to collect the electrostatic evolution data of multiple periods generated when the target heat shrink tube is wrapped around the target battery cell based on the preset standardized electrostatic detection parameters; Feature extraction module 202, for performing time domain segmentation of the heat shrink process and identification of the electrostatic accumulation-discharge cycle on the multi-period electrostatic evolution data to obtain electrostatic time series characteristic parameters, and extracting electrostatic spatiotemporal distribution characteristics and calculating spatial distribution of the multi-period electrostatic evolution data based on preset multiple points around the battery cell to obtain electrostatic fingerprint data, and performing frequency domain transformation on the electrostatic fingerprint data and calculating the spectrum characteristics of the heat shrink tube-battery cell interface friction frequency to obtain a set of frequency domain characteristic parameters; A feature fusion module 203 is configured to perform feature fusion on the electrostatic time series feature parameters and the frequency domain feature parameter set based on preset battery cell and heat shrink tubing specification parameters, and perform weighted optimization calculation of the matching degree between the heat shrink tubing and the battery cell to obtain a comprehensive feature vector, and to perform matching mode classification and membership function mapping on the comprehensive feature vector between the heat shrink tubing and the battery cell to obtain matching evaluation data; The result generation module 204 is configured to perform quantitative calculation and confidence calculation on the matching evaluation data to generate a matching detection result.

[0048] In an embodiment of the present invention, through hierarchical electrostatic data processing and feature analysis, the problem of accurate identification of the matching degree detection of heat shrink tubing and battery cells is solved, especially in terms of static electricity accumulation-discharge cycle, interface friction characteristics and spatial distribution uniformity, the dynamic characteristics and matching characteristics of the heat shrink wrapping process are fully considered, and the detection accuracy is effectively improved; and a multi-dimensional electrostatic feature fusion and pattern classification strategy is adopted to realize the correlation analysis between time domain, space domain and frequency domain features, and enhance the reliability of matching degree evaluation; in addition, through electrostatic fingerprint map construction and membership function mapping, the matching pattern and confidence level are accurately identified, thereby realizing efficient and accurate detection of the matching degree between heat shrink tubing and battery cells as a whole.

[0049] above Figure 2 The battery heat shrink tube and battery cell matching detection device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The battery heat shrink tube and battery cell matching detection device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0050] Figure 3This is a schematic diagram of the structure of a device for testing the fit of a battery heat shrink tubing to a battery cell, provided in an embodiment of the present invention. The device 300 can vary significantly depending on configuration or performance. It may include one or more central processing units (CPUs) 310 (e.g., one or more processors), memory 320, and one or more storage media 330 (e.g., one or more mass storage devices) storing application programs 333 or data 332. The memory 320 and storage media 330 may be either transient or persistent storage. The program stored in the storage medium 330 may include one or more modules (not shown), each of which may include a series of instructions for operating on the device 300. Furthermore, the processor 310 may be configured to communicate with the storage medium 330 to execute the instructions stored in the storage medium 330 on the device 300.

[0051] The battery heat shrink tube and battery cell matching detection device 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 3 The structure of the battery heat shrink tube and battery cell matching detection device shown does not constitute a limitation of the battery heat shrink tube and battery cell matching detection device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0052] The present invention also provides a device for detecting the degree of matching between a battery heat shrink tube and a battery cell. The computer device includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor executes each step of the method for detecting the degree of matching between a battery heat shrink tube and a battery cell in the above-mentioned embodiments.

[0053] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions. When the instructions are run on a computer, the computer executes each step of the method for detecting the matching degree between the battery heat shrink tube and the battery cell.

[0054] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0055] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0056] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0057] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting the matching degree between a battery heat shrink tube and a battery cell, characterized in that: The method for detecting the matching degree between the battery heat shrink tube and the battery cell includes: Based on preset standardized electrostatic detection parameters, the multi-period electrostatic evolution data generated when the target heat shrink tubing is wrapped around the target battery cell is collected; The multi-period electrostatic evolution data is segmented in the time domain of the heat shrink process and the electrostatic accumulation-discharge cycle is identified to obtain electrostatic time series characteristic parameters, and based on the preset multi-point positions of the battery cell circumference, the electrostatic spatiotemporal distribution characteristics of the multi-period electrostatic evolution data are extracted and the spatial distribution is calculated to obtain electrostatic fingerprint spectrum data, and the electrostatic fingerprint spectrum data is transformed in the frequency domain and the spectral characteristics of the heat shrink tube-battery cell interface friction frequency are calculated to obtain a frequency domain characteristic parameter set; Based on preset battery cell and heat shrink tubing specification parameters, feature fusion and weight optimization calculation of the heat shrink tubing and battery cell matching are performed on the electrostatic timing feature parameters and the frequency domain feature parameter set to obtain a comprehensive feature vector, and matching mode classification and membership function mapping of the heat shrink tubing and battery cell are performed on the comprehensive feature vector to obtain matching evaluation data; Performing quantitative calculation and confidence calculation on the matching evaluation data to generate a matching detection result.

2. The method for detecting the matching degree between a battery heat shrink tube and a battery cell according to claim 1, wherein: The electrostatic evolution data of the multiple time periods is subjected to time-domain segmentation of the heat shrink process and identification of the electrostatic accumulation-discharge cycle to obtain electrostatic time series characteristic parameters, including: Based on the shrinkage temperature characteristics of the target heat shrinkable tube, the multi-period electrostatic evolution data is segmented in the time domain according to multiple heat shrinkage process stages to obtain segmented electrostatic data, and the segmented electrostatic data is filtered, denoised, and subjected to multiple data preprocessing to obtain a preprocessed electrostatic signal sequence; Calculating the electrostatic accumulation rate of the preprocessed electrostatic signal sequence when the target heat shrink tubing is wrapped around the target battery cell and identifying the electrostatic accumulation peak on the battery cell surface to obtain key characteristic points of accumulation, and performing periodic pattern recognition and periodic consistency calculation on the preprocessed electrostatic signal sequence to obtain periodic characteristic parameters, and performing piecewise exponential function fitting on the preprocessed electrostatic signal sequence based on the key characteristic points of accumulation to obtain accumulation-discharge kinetic parameters; The accumulation-discharge kinetic parameters and the periodic characteristic parameters are time-series aligned and parameter normalized to obtain standardized characteristic data, and the standardized characteristic data are time-feature matrix constructed and parameter encoded to obtain electrostatic time series characteristic parameters.

3. The method for detecting the matching degree between a battery heat shrink tube and a battery cell according to claim 2, wherein: The step of performing piecewise exponential function fitting on the pre-processed electrostatic signal sequence based on the accumulated key characteristic points to obtain accumulation-discharge kinetic parameters includes: Determining the accumulation peak value in the accumulation kinetic parameter and the accumulation segment boundary and the discharge segment boundary corresponding to the peak moment, and extracting data points and eliminating outliers for the accumulation segment and the discharge segment in the preprocessed electrostatic signal sequence based on the accumulation segment boundary and the discharge segment boundary to obtain a segmented data set; performing exponential function fitting and determining an accumulation time constant on the accumulated segment data in the segmented data set to generate an accumulation fitting result, and determining a discharge segment starting position and a discharge time constant on the discharge segment data in the segmented data set based on the accumulation fitting result to generate a discharge fitting result; performing a time constant ratio calculation on the accumulation fitting result and the discharge fitting result to obtain a time constant ratio, and identifying a plurality of complete accumulation-discharge cycles in the preprocessed electrostatic signal sequence based on the time constant ratio; The time constant variation coefficient and reproducibility index are calculated for the accumulation time constant and the discharge time constant corresponding to each accumulation-discharge cycle to obtain the time constant reproducibility characteristic, and the accumulation fitting result, the discharge fitting result, the time constant ratio and the time constant reproducibility characteristic are used to construct a kinetic parameter vector and perform numerical standardization to obtain the accumulation-discharge kinetic parameters.

4. The method for detecting the matching degree between a battery heat shrink tube and a battery cell according to claim 1, wherein: The electrostatic spatiotemporal distribution feature extraction and spatial distribution calculation are performed on the electrostatic evolution data of multiple time periods based on the preset multiple circumferential positions of the battery cell to obtain electrostatic fingerprint map data, including: Based on the preset multi-point positions of the battery cell circumference, the sensor array angle-static charge density mapping and spatial interpolation of the electrostatic evolution data of the multi-period are performed to obtain a spatially continuous electrostatic distribution, and the spatially continuous electrostatic distribution is subjected to spatial statistical calculation and distribution correlation calculation to obtain spatial distribution statistical characteristics; Based on the spatial distribution statistical characteristics, spatial gradient calculation and gradient mutation detection are performed on the corresponding adjacent angular positions in the spatially continuous electrostatic distribution to obtain spatial gradient characteristic parameters, and local extreme value detection and outlier identification are performed on the spatial gradient characteristic parameters based on a preset local charge density threshold to obtain island marking data; A spatial aggregation calculation is performed on the island mark data to obtain spatial anomaly distribution characteristics, and a multi-dimensional feature matrix is ​​constructed and a typical spatial distribution pattern is identified for the spatial distribution statistical characteristics, the spatial gradient feature parameters and the spatial anomaly distribution characteristics to obtain electrostatic fingerprint map data.

5. The method for detecting the matching degree between a battery heat shrink tube and a battery cell according to claim 1, wherein: The frequency domain transformation of the electrostatic fingerprint spectrum data and the spectrum characteristic calculation of the heat shrink tube-battery core interface friction frequency are performed to obtain a frequency domain characteristic parameter set, including: The electrostatic fingerprint data is subjected to time-frequency analysis window setting and frame windowing in the heat shrink process stage to obtain a windowed time-frequency data matrix, and the windowed time-frequency data matrix is ​​subjected to time-frequency transformation and amplitude-phase separation of the heat shrink tube-battery core interface friction signal to obtain a time-frequency amplitude-phase distribution; Performing instantaneous frequency calculation and mean-variance calculation of electrostatic changes on the cell surface on the time-frequency amplitude-phase distribution to obtain instantaneous frequency characteristic parameters, performing power spectrum density estimation and frequency band energy distribution calculation in the heat shrink process stage on the time-frequency amplitude-phase distribution to obtain frequency band energy distribution characteristics, and performing multidimensional modulation depth calculation of the heat shrink tube-cell interface on the time-frequency amplitude-phase distribution to obtain multidimensional modulation characteristic parameters; The instantaneous frequency characteristic parameters, the frequency band energy distribution characteristics and the multi-dimensional modulation characteristic parameters are subjected to characteristic vector combination and characteristic screening of heat shrink tube-battery core matching to obtain a frequency domain characteristic parameter set.

6. The method for detecting the matching degree between a battery heat shrink tube and a battery cell according to claim 1, wherein: Based on the preset battery cell and heat shrink tube specification parameters, the electrostatic time series feature parameters and the frequency domain feature parameter set are subjected to feature fusion and weight optimization calculation of the heat shrink tube and battery cell matching degree to obtain a comprehensive feature vector, including: Performing feature dimension alignment and time synchronization on the electrostatic time series feature parameters and the frequency domain feature parameter set to obtain an original feature data set, and performing feature correlation calculation and principal component dimensionality reduction for matching the heat shrink tubing with the battery cell on the original feature data set to obtain a feature space after dimensionality reduction; Performing an optimization search on the weight coefficients of electrostatic time series features and frequency domain features in the feature space after dimensionality reduction to obtain a multi-objective optimization weight matrix, and performing a weight compensation calculation on the multi-objective optimization weight matrix based on preset battery cell and heat shrink tubing specification parameters to obtain compensated feature weights; Based on the compensated feature weights, weighted fusion calculation of the heat shrink tube and battery cell matching and matching pattern clustering recognition are performed on the feature space after dimensionality reduction to obtain a comprehensive feature vector.

7. The method for detecting the matching degree between a battery heat shrink tube and a battery cell according to claim 1, wherein: The performing of quantitative calculation and confidence calculation on the matching evaluation data to generate a matching detection result includes: Quantitatively calculating and grading the matching level between the heat shrink tubing and the battery cell on the matching evaluation data to obtain a quantitative matching score; Based on the quantitative matching score, performing confidence calculation and stability evaluation on the matching evaluation data to obtain a confidence evaluation index; A comprehensive evaluation of the matching performance between the heat shrink tubing and the battery cell is performed on the quantitative matching score and the confidence evaluation index to generate a matching test result.

8. A device for detecting the matching degree between a battery heat shrink tube and a battery cell, characterized in that: The battery heat shrink tube and battery cell matching detection device includes: A data acquisition module is used to collect multi-period electrostatic evolution data generated when the target heat shrink tubing is wrapped around the target battery cell based on preset standardized electrostatic detection parameters; A feature extraction module is used to perform time domain segmentation of the heat shrink process and electrostatic accumulation-discharge cycle identification on the multi-period electrostatic evolution data to obtain electrostatic timing characteristic parameters, and based on preset multi-point positions around the battery cell, extract electrostatic spatiotemporal distribution features and calculate spatial distribution of the multi-period electrostatic evolution data to obtain electrostatic fingerprint data, and perform frequency domain transformation on the electrostatic fingerprint data and calculate the spectrum characteristics of the heat shrink tube-battery cell interface friction frequency to obtain a set of frequency domain characteristic parameters; A feature fusion module is used to perform feature fusion on the electrostatic time series feature parameters and the frequency domain feature parameter set based on preset battery cell and heat shrink tubing specification parameters, and perform weighted optimization calculation on the matching degree between the heat shrink tubing and the battery cell to obtain a comprehensive feature vector, and to perform matching mode classification and membership function mapping on the comprehensive feature vector between the heat shrink tubing and the battery cell to obtain matching evaluation data; The result generation module is used to perform quantitative calculation and confidence calculation on the matching evaluation data to generate a matching detection result.

9. A battery heat shrink tube and battery cell matching detection device, characterized in that: The battery heat shrink tube and battery cell matching detection device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instruction in the memory to enable the battery heat shrink tube and battery cell matching detection device to perform each step of the battery heat shrink tube and battery cell matching detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the method for detecting the matching degree between the battery heat shrink tube and the battery cell as described in any one of claims 1 to 7 are implemented.

Images