Method and system for identifying cracks of pole piece of standby battery of communication base station by utilizing thermal imaging

Abstract

The invention belongs to the field of battery pole piece detection, and relates to a method and system for identifying cracks of a standby battery pole piece of a communication base station through thermal imaging, and the method comprises the steps: heating the battery pole piece to a target temperature range, and collecting heat distribution image data of the surface of the battery pole piece; continuously applying pulse type thermal excitation signals to the surface of the battery pole piece, and collecting dynamic thermal distribution image data of the surface of the battery pole piece; processing the benchmark reference thermal image and the dynamic thermal image sequence, and extracting characteristic intensity distribution of a suspected crack region; according to the characteristic intensity distribution of the suspected crack areas, constructing a three-dimensional thermal gradient model, and screening the suspected crack areas to obtain real crack areas; the three-dimensional thermal gradient model of the real crack area is compared with a standard crack library, and the crack type and the crack size are determined; the defects in small crack detection and large-scale production line rapid detection in the prior art are overcome, the overall performance of the battery is improved, and the safety of the base station standby battery is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of battery pole piece detection, and specifically discloses a method and system for identifying cracks in communication base station standby battery pole pieces using thermal imaging. BACKGROUND

[0002] In the production and manufacturing process of communication base station standby batteries, fast and efficient quality detection technology is needed, and high-quality control of battery pole pieces is a key link to ensure the overall performance and safety of the battery. With the continuous progress of communication base station standby battery technology, the quality control and crack detection of battery pole pieces have gradually become key factors affecting the performance and safety of the battery. In the current battery production process, the industry has increasingly prominent demand for detection solutions that combine efficiency and intelligence. However, existing crack detection methods mostly use electrical signal analysis or mechanical vibration monitoring, which often face the dilemma of insufficient precision and difficulty in achieving non-destructive testing in actual production environments. At the same time, these methods have weak discrimination ability for the specific location and morphology of cracks, making it difficult to fully meet the high standards of the production line. The existing technology discloses a technical method and supporting device for wing component crack monitoring and repair; the device mainly consists of an electrode, a single-chip microcomputer, a thermal imaging detector, a filter, and a battery, etc. Its working principle is to detect cracks through surface current and automatically repair in flight state. Specifically, this technology first uses micro-current detection to identify the crack location, then applies a strong current to the crack to make the metal material approach the melting point, and simultaneously uses a thermal imaging detector to monitor temperature changes in real time, thereby achieving precise control of the repair process. However, this existing technology relies on strong current to make the metal at the crack approach the melting point to achieve repair, which is not suitable for the material properties of battery pole pieces and may cause material damage or performance degradation. Moreover, the thermal imaging technology is only used to monitor temperature changes during the repair process, not directly for crack identification and positioning, and lacks precise analysis capability for crack morphology characteristics.

[0003] Therefore, the present application discloses a method and system for identifying cracks in communication base station standby battery pole pieces using thermal imaging, which can perform high-precision, non-destructive detection of cracks on the surface and inside of battery pole pieces and achieve accurate identification of crack morphology, effectively improving the efficiency and accuracy of communication base station standby battery crack detection, and being suitable for battery pole pieces made of complex materials. This solves the deficiencies of existing technology in micro-crack detection and rapid detection on large-scale production lines, improves the overall performance of the battery, and ensures the safety of the base station standby battery. SUMMARY

[0004] The purpose of the present application is to provide a method and system for identifying communication base station standby battery pole piece cracks using thermal imaging, which solves the problem of the deficiencies of the prior art in micro-crack detection and rapid detection in large-scale production lines, improves the overall performance of the battery, and ensures the safety of the base station standby battery. The specific scheme is as follows: The method for identifying communication base station standby battery pole piece cracks using thermal imaging comprises the following steps: step 1, heating the battery pole piece to a target temperature range and collecting thermal distribution image data of the surface of the battery pole piece to obtain a baseline reference thermal image; step 2, continuously applying a pulse-type thermal excitation signal to the surface of the battery pole piece and collecting dynamic thermal distribution image data of the surface of the battery pole piece to obtain a dynamic thermal image sequence; step 3, processing the baseline reference thermal image and the dynamic thermal image sequence through a multi-physical field coupling analysis algorithm to extract the characteristic intensity distribution of the suspected crack area; step 4, constructing a three-dimensional thermal gradient model based on the characteristic intensity distribution of the suspected crack area and screening the suspected crack area based on the three-dimensional thermal gradient model to obtain a real crack area; and step 5, comparing the three-dimensional thermal gradient model of the real crack area with a standard crack library to determine the crack type and crack size.

[0005] Further, the target temperature range is 40℃-80℃, and the power density range of heating is 10W / cm²-50W / cm².

[0006] Further, the waveform of the pulse-type thermal excitation signal is a square wave or a sine wave, the frequency is 1Hz-50Hz, the amplitude is 10W-50W, and the duration is 1s-10s.

[0007] Further, the multi-physical field coupling analysis algorithm is as follows:

[0008] wherein F(x,y) represents a characteristic intensity distribution function; represents the start time of the application of the pulse-type thermal excitation signal; represents the end time of the application of the pulse-type thermal excitation signal; T(x,y,t) represents the temperature value of pixel point (x,y) at time t; x and y represent the horizontal and vertical coordinates of the thermal image, respectively; represents the baseline temperature value of pixel point (x,y) in the baseline reference thermal image; represents the highest temperature in the dynamic thermal image sequence; represents the lowest temperature in the dynamic thermal image sequence; represents the integral of time in the application time period of the pulse-type thermal excitation signal.

[0009] Further, step 4 comprises: establishing an associated mapping relationship between the two-dimensional image coordinates and the temperature values based on the feature intensity distribution; setting the crack detection area of the battery pole piece as an XY plane, and taking the Z axis vertical direction to represent the temperature gradient to obtain the spatial structure of the three-dimensional thermal gradient model; constructing the three-dimensional thermal gradient model based on the associated mapping relationship and the spatial structure of the three-dimensional thermal gradient model; and the expression of the three-dimensional thermal gradient model is: ; Wherein, G(x, y, z) represents the three-dimensional thermal gradient value of the pixel point with coordinates (x, y) in the two-dimensional image in the three-dimensional thermal gradient model (x, y, z); F(x, y) represents the spatial gradient of the feature intensity distribution F(x, y) at the two-dimensional plane coordinates (x, y); represents the derivation with respect to the time sequence t; T(x, y, t) represents the temperature value of the pixel point (x, y) at the time sequence t; t represents the time sequence variable; the three-dimensional thermal gradient model of the suspected crack area is screened based on the feature intensity threshold to obtain the real crack area.

[0010] Further, determining the crack type comprises: calculating the similarity of the three-dimensional gradient vector of the real crack area and the three-dimensional gradient vector of the crack sample to obtain the global thermal gradient similarity; the three-dimensional gradient vector refers to the vector obtained by gradient calculation on the three-dimensional thermal gradient model of the crack; evaluating the coincidence degree of the three-dimensional space structure of the real crack area and the three-dimensional space structure of the crack sample through the Hausdorff distance to obtain the three-dimensional morphological similarity; the three-dimensional space structure refers to the space structure of the three-dimensional thermal gradient model of the crack; calculating the divergence of the temperature in the three-dimensional thermal gradient model vector field to obtain the divergence feature of the thermal gradient vector field; quantifying the heat aggregation degree of the crack area through the energy calculation formula to obtain the heat aggregation effect at the crack; verifying the thermal response data and material characteristic data of the real crack and the crack sample based on the coupling equation group to obtain the data verification result; the crack sample refers to the crack sample stored in the crack standard database; the coupling equation group comprises the non-steady-state heat conduction equation and the elastic mechanics equation; and the crack type is output by integrating the global thermal gradient similarity, the three-dimensional morphological similarity, the divergence feature of the thermal gradient vector field, the heat aggregation effect at the crack and the data verification result.

[0011] Further, the similarity of the three-dimensional gradient vector of the real crack area and the three-dimensional gradient vector of the crack sample is:

[0012] Wherein, Sim represents the global thermal gradient similarity; represents the three-dimensional gradient vector of the real crack area; represents the three-dimensional gradient vector of the crack sample; represents the weight vector; The divergence of the temperature in the three-dimensional thermal gradient model vector field is calculated as:

[0013] Where divG represents the divergence characteristic of the thermal gradient vector field; 、 and represent the components of the three-dimensional thermal gradient model vector field in the x, y, and z directions, respectively; 、 and represent the partial derivatives with respect to the x, y, and z directions, respectively; The energy calculation formula is: ; Where E represents the thermal concentration effect at the crack; p represents the material density; c represents the specific heat capacity of the material; DT represents the temperature change; and dV represents the volume element; The unsteady heat conduction equation is: ; The elasticity equation is: ; Where represents the partial derivative of the temperature T on the battery pole piece surface with respect to the time sequence t; represents the heat conduction term; k represents the material thermal conductivity; and Q represents the internal heat source term; T represents the temperature gradient vector; represents the divergence operator; represents the fourth-order elastic tensor; represents the strain tensor; represents the displacement vector; represents the transpose operation; represents the strain caused by thermal expansion; a represents the material thermal expansion coefficient; and T represents the temperature value at the current time sequence t; represents the reference temperature value; represents the volume force density.

[0014] Further, the crack size is determined, including: using a linear regression model of the thermal gradient propagation characteristic to solve a length parameter; using a curve fitting model of the edge gradient change to solve a width parameter; and using an exponential model of the thermal attenuation characteristic to solve a depth parameter.

[0015] Further, the linear regression model of the thermal gradient propagation characteristic is: ; Where L represents the length parameter; represents a first constant coefficient; and b represents a second constant coefficient; represents the amplitude of the temperature gradient vector; denotes length infinitesimal; The curve fitting model of the edge gradient change is: wherein, W represents a width parameter; sigma represents a standard deviation; ln represents a logarithmic function; The exponential model of the thermal decay characteristic is: wherein, D represents a depth parameter; denotes maximum depth; exp represents an exponential function; lambda represents a material attenuation coefficient; delta T represents a surface temperature difference.

[0016] The application also provides a system for identifying cracks in communication base station standby battery pole pieces using thermal imaging according to the method described above, comprising: a constant temperature heating platform, a pulsed thermal excitation signal source, an infrared thermal imager and a data processing unit; the constant temperature heating platform is used to heat the battery pole piece to a target temperature range; the infrared thermal imager is used to collect thermal distribution image data of the surface of the battery pole piece to obtain a reference thermal image; dynamic thermal distribution image data of the surface of the battery pole piece is collected to obtain a dynamic thermal image sequence; the pulsed thermal excitation signal source is used to continuously apply a pulsed thermal excitation signal to the surface of the battery pole piece; the data processing unit is used to process the reference thermal image and the dynamic thermal image sequence through a multi-physical field coupling analysis algorithm, extract the characteristic intensity distribution of the suspected crack area; according to the characteristic intensity distribution of the suspected crack area, a three-dimensional thermal gradient model is constructed, and the suspected crack area is screened based on the three-dimensional thermal gradient model to obtain a real crack area; the three-dimensional thermal gradient model of the real crack area is compared with a standard crack library to determine the crack type and crack size.

[0017] The application has the following advantages and beneficial effects: The application uses thermal conduction characteristics to locate crack areas, can identify small cracks and complex morphology cracks in battery pole pieces, is also suitable for battery pole pieces with various material characteristics, solves the limitations of traditional detection techniques under complex working conditions, and provides an efficient and intelligent solution for communication base station standby battery manufacturing processes.

[0018] The application can realize high-precision and non-destructive detection of cracks on the surface and inside of the base station standby battery pole piece through the thermal imaging technology combined with the multi-physical field coupling analysis algorithm, and realize accurate identification of crack morphology.

[0019] The application realizes intelligent classification and size quantification of crack types by introducing a deep learning algorithm, and provides an efficient and intelligent detection solution for modern battery manufacturing processes. BRIEF DESCRIPTION OF DRAWINGS

[0020] ​​Figure 1 An exemplary flow chart of the method for identifying cracks in the electrode plate of a backup battery of a communication base station using thermal imaging according to the present application is shown in FIG. 1. Figure 2 An exemplary block diagram of the system for identifying cracks in the electrode plate of a backup battery of a communication base station using thermal imaging according to the present application is shown in FIG. 2. DETAILED DESCRIPTION

[0021] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings of the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0022] The present application discloses a method and system for identifying cracks in the electrode plate of a backup battery of a communication base station using thermal imaging, aiming to realize high-precision and non-destructive detection of cracks on the surface and inside of the battery electrode plate by fusing thermal imaging technology and multi-physical field coupling analysis algorithm, accurately identify the crack morphology and quantify the size, and be suitable for complex material electrode plate and large-scale production line rapid detection scene. The technical solutions of the present application will be described in detail below in connection with the drawings and specific embodiments.

[0023] Figure 1 An exemplary flow chart of the method for identifying cracks in the electrode plate of a backup battery of a communication base station using thermal imaging according to the present application is shown in FIG. 1. Figure 1 As shown in FIG. 1, the method for identifying cracks in the electrode plate of a backup battery of a communication base station using thermal imaging includes the following contents: Step 1, heat the battery electrode plate to a target temperature range, and collect thermal distribution image data of the surface of the battery electrode plate to obtain a baseline reference thermal image. For example, place the battery electrode plate to be detected on a constant temperature heating platform, and heat it at a power density of 10 W / cm²~50 W / cm² to make the temperature of the electrode plate rise to a target temperature interval of 40℃~80℃. The selection of the target temperature needs to be optimized according to the thermal conductivity and thermal expansion coefficient characteristics of the battery electrode plate material to ensure that the thermal state of the electrode plate is stable and the material performance is not damaged. After the temperature of the battery electrode plate is maintained stable, start the infrared thermal imager to collect the thermal distribution image data of the surface of the battery electrode plate to obtain a baseline reference thermal image. The single-frame collection time is set to <10ms, and each pixel point (x, y) in the image corresponds to a baseline temperature value T0(x, y), where x and y represent the horizontal and vertical coordinates of the thermal image, respectively, providing a baseline reference for subsequent dynamic thermal data comparison.

[0024] Step 2, continuously apply the pulsed thermal excitation signal to the surface of the battery electrode sheet, and collect the dynamic thermal distribution image data of the surface of the battery electrode sheet to obtain a dynamic thermal image sequence. For example, the pulsed thermal excitation signal is continuously applied to the surface of the battery electrode sheet through a pulsed thermal excitation signal source. The waveform of the signal is a square wave or a sine wave, the frequency range is 1 Hz to 50 Hz, the power amplitude range is 10 W to 50 W, and the action duration is controlled between 1 s to 10 s. The determination of the signal parameters needs to consider the thermal diffusion coefficient and crack depth characteristics of the electrode sheet material to ensure that the thermal excitation can effectively highlight the thermal response difference of the crack area. During the whole process of applying the thermal excitation signal, the infrared thermal imager records the thermal distribution changes of the surface of the electrode sheet in real time to generate a dynamic thermal image sequence. The sequence contains thermal images at different time sequences t, and each pixel point (x, y) corresponds to a temperature value T(x, y, t) at time sequence t, which fully reflects the dynamic thermal response process of the electrode sheet under the action of thermal excitation.

[0025] Step 3, processing the reference thermal image and the dynamic thermal image sequence by a multi-physical field coupling analysis algorithm to extract the characteristic intensity distribution of the suspected crack area. For example, the data processing unit calls the multi-physical field coupling analysis algorithm to jointly process the reference thermal image and the dynamic thermal image sequence to extract the characteristic intensity distribution of the suspected crack area. The multi-physical field coupling analysis algorithm is: ; Wherein, F(x, y) represents a characteristic intensity distribution function, which is used to quantify the abnormal degree of thermal response at pixel point (x, y); represents the start time of applying the pulsed thermal excitation signal; represents the end time of applying the pulsed thermal excitation signal; T(x, y, t) represents the temperature value of pixel point (x, y) at time sequence t; x and y represent the horizontal coordinate and vertical coordinate of the thermal image, respectively; represents the reference temperature value of pixel point (x, y) in the reference thermal image; represents the highest temperature in the dynamic thermal image sequence; represents the lowest temperature in the dynamic thermal image sequence; represents the integral of time in the application time period of the pulsed thermal excitation signal. represents the cumulative effect of the temperature difference of pixel point (x, y) from the reference temperature in the thermal excitation application time period.

[0026] Step 4, based on the characteristic intensity distribution of the suspected crack area, a three-dimensional thermal gradient model is constructed, and the suspected crack area is screened based on the three-dimensional thermal gradient model to obtain the real crack area. For example, based on the characteristic intensity distribution F(x, y) obtained in step 3, a correlation mapping relationship between the two-dimensional image coordinates (x, y) and the temperature value T(x, y, t) is established, and the correspondence between the pixel coordinates and the thermal response data is determined. The crack detection area of the battery pole piece is set as the XY plane (maintaining the same resolution as the image captured by the infrared thermal imager), and the Z axis represents the temperature gradient (each level of gradient is 0.1°C) in the vertical direction, forming the spatial structure of the three-dimensional thermal gradient model. Based on the correlation mapping relationship and the spatial structure of the three-dimensional thermal gradient model, a three-dimensional thermal gradient model is constructed; the expression of the three-dimensional thermal gradient model is: ; where G(x, y, z) represents the three-dimensional thermal gradient value of the pixel point with coordinates (x, y) in the two-dimensional image in the three-dimensional thermal gradient model (x, y, z); F(x, y) represents the spatial gradient of the characteristic intensity distribution F(x, y) at the two-dimensional plane coordinates (x, y); denotes the derivative with respect to time t; T(x, y, t) represents the temperature value of the pixel point (x, y) at time t; t represents the time variable.

[0027] Based on the characteristic intensity threshold ΔF, the three-dimensional thermal gradient model of the suspected crack area is screened to obtain the real crack area. The value range of ΔF is 0.1-0.5, and the specific value needs to be determined through experiments, and its setting is based on the adaptive adjustment of the thermal gradient difference between the crack area and the non-crack area to ensure the accuracy and system robustness of the detection results.

[0028] Step 5, compare the three-dimensional thermal gradient model of the real crack area with the standard crack library to determine the crack type and crack size. For example, the data processing unit compares and analyzes the three-dimensional thermal gradient model of the real crack area with the samples in the standard crack database, and uses a deep learning algorithm (using a convolutional neural network (CNN) structure, inputting the three-dimensional thermal gradient model, and outputting the crack type label and size parameters) to complete the crack type classification and size quantification, and the specific process is as follows: The five-dimensional feature integration realizes accurate identification of the crack type, and the global thermal gradient similarity is calculated as follows: the similarity between the three-dimensional gradient vector G1 of the real crack area and the three-dimensional gradient vector G2 of the crack sample is calculated to obtain the global thermal gradient similarity; the three-dimensional gradient vector is a vector obtained by gradient calculation on the three-dimensional thermal gradient model of the crack. The similarity between the three-dimensional gradient vector of the real crack area and the three-dimensional gradient vector of the crack sample is: ; where Sim represents the global thermal gradient similarity; The three-dimensional gradient vector representing the actual crack region is obtained by performing gradient calculations on the three-dimensional thermal gradient model of the actual crack. The three-dimensional gradient vector of the crack sample is obtained by performing gradient calculation on the three-dimensional thermal gradient model of the sample in the standard database. This represents the weight vector, used to assign different importance to different components of the gradient vector.

[0029] 3D morphological similarity: The degree of agreement between the 3D spatial structure of the actual crack region and the 3D spatial structure of the crack sample is evaluated using the Hausdorff distance to obtain 3D morphological similarity. The Hausdorff distance is a metric for quantifying the spatial similarity between two point sets; the smaller the distance, the better the spatial geometric structures (such as the length, width, and depth distribution of the crack) match. The 3D spatial structure refers to the spatial structure of the 3D thermal gradient model of the crack.

[0030] Divergence characteristics of the thermal gradient vector field: The divergence of temperature in the three-dimensional thermal gradient model vector field is calculated to obtain the divergence characteristics of the thermal gradient vector field. The divergence of temperature in the three-dimensional thermal gradient model vector field is calculated as follows: ; Wherein, divG represents the divergence characteristic of the thermal gradient vector field; , and These represent the components of the vector field of the three-dimensional thermal gradient model in the x, y, and z directions, respectively. , and These represent partial derivatives with respect to the x, y, and z directions, respectively, reflecting the divergence or convergence characteristics of the thermal gradient in space. Crack regions exhibit divergence anomalies due to impeded heat conduction.

[0031] Thermal accumulation effect at cracks: The degree of thermal accumulation in the crack region is quantified using an energy calculation formula to obtain the thermal accumulation effect at cracks; the energy calculation formula is: ; Where E represents the heat accumulation effect at the crack; ρ represents the material density; c represents the material specific heat capacity; ΔT represents the temperature change; dV represents the volume element; and the triple integral represents the cumulative calculation of the heat accumulation effect over the entire crack region.

[0032] Data verification result: based on the coupled equation set, the thermal response data and material property data of the real crack and crack sample are verified to obtain the data verification result; the crack sample refers to the crack sample stored in the crack standard database; the coupled equation set includes the unsteady heat conduction equation and the elastic mechanics equation. In the verification process, the thermal response data (time-varying temperature field T(t, x, y, z)) and material property data (p, c, k, a, C, etc.) of the real crack are collected, and the time-varying thermal response and material properties of the corresponding sample in the standard database are compared to confirm the consistency of the characteristics.

[0033] The unsteady heat conduction equation is: ; The elastic mechanics equation is: ; Wherein, represents the partial derivative of the temperature T on the time sequence t on the surface of the battery pole piece; represents the heat conduction term; k represents the material thermal conductivity; Q represents the internal heat source term; T represents the temperature gradient vector; represents the divergence operator; represents the fourth-order elastic tensor; represents the strain tensor; represents the displacement vector; represents the transpose operation; represents the strain caused by thermal expansion; a represents the material thermal expansion coefficient; T represents the temperature value at the current time sequence t; represents the reference temperature value; represents the volume force density.

[0034] Integrating the global thermal gradient similarity, the three-dimensional morphological similarity, the divergence characteristics of the thermal gradient vector field, the thermal aggregation effect at the crack, and the data verification result, the crack type (such as linear crack, network crack, point crack, etc.) is output through a deep learning algorithm. The deep learning algorithm adopts a convolutional neural network (CNN) structure design, the input data of which is a three-dimensional thermal gradient model of the crack area, and the output result is a crack type label and its size parameter, and the training data is derived from a mixed data set of experimental sample and simulation generated data.

[0035] Three types of mathematical models are used to solve the length, width and depth parameters of the crack, including: a linear regression model of thermal gradient extension characteristics is used to solve the length parameter; a curve fitting model of edge gradient change is used to solve the width parameter; an exponential model based on thermal attenuation characteristics is used to solve the depth parameter. The linear regression model of thermal gradient extension characteristics is: ; Wherein, L represents the length parameter, i.e. the actual length of the crack; represents a first constant coefficient; β represents a second constant coefficient, both of which are determined by experimental calibration; T| represents the amplitude of the temperature gradient vector, reflecting the intensity of the spatial variation of temperature; represents a length infinitesimal, a differential length unit along the crack propagation path, used to integrate and accumulate T| ; The curve fitting model of the edge gradient variation is: ; Wherein, W represents the width parameter, i.e. the actual width of the crack; σ represents the standard deviation, representing the dispersion degree of the distribution of the crack edge gradient variation data, which is calculated by the temperature gradient data of the crack edge pixels in the dynamic thermal image sequence; ln represents the logarithmic function; The exponential model of the thermal decay characteristic is: ; Wherein, D represents the depth parameter, i.e. the actual depth of the crack; represents the maximum possible depth, which is determined by the thickness of the pole material and the resolution of the detection system; exp represents the exponential function; λ represents the material attenuation coefficient; ΔT represents the surface temperature difference, i.e. the surface temperature difference between the crack area and the normal area.

[0036] Figure 2 is an exemplary block diagram of the system for identifying the pole crack of the standby battery of the communication base station by using thermal imaging according to the present application. The present application also provides a system for identifying the pole crack of the standby battery of the communication base station by using thermal imaging, as shown in Figure 2 The system comprises a constant temperature heating platform, a pulsed thermal excitation signal source, an infrared thermal imager, a data processing unit and a standard crack database, and the functions of each component are as follows: Constant temperature heating platform: used for heating the battery pole to be detected to a target temperature range and maintaining the temperature stable, and the heating power density can be adjusted in the range of 10 W / cm²~50 W / cm².

[0037] Pulsed thermal excitation signal source: used for continuously applying a pulsed thermal excitation signal to the surface of the battery pole.

[0038] Infrared thermal imager: used for collecting thermal distribution image data of the pole surface, including a reference thermal image and a dynamic thermal image sequence, and the single-frame collection time is set to <10 ms to ensure the timeliness of data collection.

[0039] Data processing unit: used for executing multi-physical field coupling analysis algorithm and deep learning algorithm to complete feature extraction, model construction, crack screening and comparison analysis.

[0040] Standard crack database: store three-dimensional thermal gradient models of various crack types and their corresponding classification labels and size parameters, providing a benchmark for crack recognition.

[0041] The detection method of the present application is combined with the core technologies of non-contact thermal imaging detection and pulse thermal excitation signal, and through the GPU accelerated multi-physical field coupling analysis algorithm, the millisecond level data processing speed can be realized, and the surface thermal distribution image acquisition and crack area feature extraction of the pole piece can be efficiently completed, which fully meets the demand of large-scale production line for detection efficiency; At the same time, the training data of the deep learning algorithm uses a mixed data set of experimental sample and simulation generated data, which can cover different types and sizes of crack features, effectively guaranteeing the accuracy of crack recognition, laying a foundation for subsequent accurate classification and quantification. In addition, the method relies on the comparison and analysis of the established three-dimensional thermal gradient model and the preset standard crack database, which can further improve the efficiency and accuracy of crack detection, and can obtain the thermal distribution image data of the entire pole piece at one time, avoiding the limitations of local detection; Its non-contact detection characteristics do not damage the performance of the battery pole piece material throughout the process, and is suitable for communication base station standby battery pole pieces with various material characteristics, effectively solving the shortcomings of the prior art in micro crack detection and complex working conditions, and can be combined with the production line MES system to realize the rapid identification, comparison and analysis and continuous detection of the pole piece crack, and completely solve the current technical problems in crack recognition accuracy and large-scale production line rapid detection adaptability.

[0042] The above is only the preferred embodiment of the present application and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for identifying cracks in the electrode plates of backup batteries in communication base stations using thermal imaging, characterized in that, include: Step 1: Heat the battery electrode to the target temperature range and collect thermal distribution image data on the surface of the battery electrode to obtain a reference thermal image; Step 2: Continuously apply a pulsed thermal excitation signal to the surface of the battery electrode and collect dynamic thermal distribution image data of the surface of the battery electrode to obtain a dynamic thermal image sequence. Step 3: Process the reference thermal image and dynamic thermal image sequence using a multiphysics coupling analysis algorithm to extract the feature intensity distribution of the suspected crack area; Step 4: Based on the characteristic intensity distribution of the suspected crack area, construct a three-dimensional thermal gradient model, and screen the suspected crack area based on the three-dimensional thermal gradient model to obtain the real crack area. Step 5: Compare the three-dimensional thermal gradient model of the actual crack region with the standard crack library to determine the crack type and crack size.

2. The method for identifying cracks in the electrode plates of a backup battery in a communication base station using thermal imaging according to claim 1, characterized in that, The target temperature range is 40℃~80℃; the heating power density range is 10W / cm²~50W / cm².

3. The method for identifying cracks in the electrode plates of a backup battery in a communication base station using thermal imaging according to claim 1, characterized in that, The waveform of the pulsed thermal excitation signal is a square wave or a sine wave, with a frequency of 1Hz to 50Hz, an amplitude of 10W to 50W, and a duration of 1s to 10s.

4. The method for identifying cracks in the electrode plates of a backup battery in a communication base station using thermal imaging according to claim 1, characterized in that, The multiphysics coupling analysis algorithm is as follows: ； Where F(x,y) represents the characteristic intensity distribution function; Indicates the start time of the pulsed thermal excitation signal; The pulsed thermal excitation signal ends at the application time; T(x,y,t) represents the temperature value of pixel (x,y) at time t; x and y represent the horizontal and vertical coordinates of the thermal image, respectively. This represents the reference temperature value of pixel (x,y) in the reference thermal image. This indicates the highest temperature in the dynamic thermal image sequence; This indicates the lowest temperature in a dynamic thermal image sequence. This represents the integral over time during the application period of the pulsed thermal excitation signal.

5. The method for identifying cracks in the electrode plates of a backup battery in a communication base station using thermal imaging according to claim 1, characterized in that, Step 4 includes: Based on the feature intensity distribution, establish the correlation mapping relationship between two-dimensional image coordinates and temperature values; The crack detection area of ​​the battery electrode is set as the XY plane, and the temperature gradient is represented by the vertical direction of the Z axis, so as to obtain the spatial structure of the three-dimensional thermal gradient model. Based on the correlation mapping relationship and the spatial structure of the three-dimensional thermal gradient model, a three-dimensional thermal gradient model is constructed; the expression of the three-dimensional thermal gradient model is: ； Where G(x,y,z) represents the three-dimensional thermal gradient value of the pixel with coordinates (x,y) in the two-dimensional image in the three-dimensional thermal gradient model (x,y,z); F(x,y) represents the spatial gradient of the feature intensity distribution F(x,y) at the two-dimensional plane coordinates (x,y); This indicates the derivative with respect to time t; T(x,y,t) represents the temperature value of pixel (x,y) at time t; t represents the time series variable; The three-dimensional thermal gradient model of the suspected crack region is screened based on the characteristic intensity threshold to obtain the real crack region.

6. The method for identifying cracks in the electrode plates of a communication base station backup battery using thermal imaging according to claim 1, characterized in that, Determine the type of crack, including: The similarity between the three-dimensional gradient vector of the real crack region and the three-dimensional gradient vector of the crack sample is calculated to obtain the global thermal gradient similarity; the three-dimensional gradient vector refers to the vector obtained by performing gradient calculation on the three-dimensional thermal gradient model of the crack. The Hausdorff distance is used to assess the degree of agreement between the three-dimensional spatial structure of the real crack region and the three-dimensional spatial structure of the crack sample, thus obtaining the three-dimensional morphological similarity; the three-dimensional spatial structure refers to the spatial structure of the three-dimensional thermal gradient model of the crack. The divergence of temperature in the three-dimensional thermal gradient model vector field is calculated to obtain the divergence characteristics of the thermal gradient vector field; The degree of heat accumulation in the crack region is quantified by energy calculation formula, and the heat accumulation effect at the crack is obtained. Based on the coupled equation set, the thermal response data and material property data of real cracks and crack samples are verified to obtain data verification results; crack samples refer to crack samples stored in the crack standard database; the coupled equation set includes unsteady heat conduction equations and elasticity equations. By integrating global thermal gradient similarity, three-dimensional morphological similarity, thermal gradient vector field divergence characteristics, thermal aggregation effect at the crack, and data verification results, the crack type is output.

7. The method for identifying cracks in the electrode plates of a backup battery in a communication base station using thermal imaging according to claim 6, characterized in that, The similarity between the three-dimensional gradient vector of the real crack region and the three-dimensional gradient vector of the crack sample is calculated as follows: ； Where Sim represents the global thermal gradient similarity; A three-dimensional gradient vector representing the actual crack region; Represents the three-dimensional gradient vector of the crack sample; Represents the weight vector; The divergence of temperature in the vector field of the three-dimensional thermal gradient model is calculated as follows: ； Wherein, divG represents the divergence characteristic of the thermal gradient vector field; , and These represent the components of the vector field of the three-dimensional thermal gradient model in the x, y, and z directions, respectively. , and Let x, y, and z represent the partial derivatives with respect to the x, y, and z directions, respectively. The energy calculation formula is: ； Where E represents the heat accumulation effect at the crack; ρ represents the material density; c represents the specific heat capacity of the material; ΔT represents the temperature change; and dV represents the volume element. The unsteady-state heat conduction equation is: ； The equation of elasticity is: ； in, This represents the partial derivative of the temperature T on the surface of the battery electrode with respect to time t; The term represents heat conduction; k represents the thermal conductivity of the material; Q represents the internal heat source term. T represents the temperature gradient vector; Represents the divergence operator; Represents the fourth-order elastic tensor; Represents the strain tensor; Represents the displacement vector; Indicates the transpose operation; The strain represents the thermal expansion; α represents the coefficient of thermal expansion of the material; T represents the temperature value at the current time t. Indicates the reference temperature value; This represents the volumetric force density.

8. The method for identifying cracks in the electrode plates of a backup battery in a communication base station using thermal imaging according to claim 1, characterized in that, Determining the crack size includes: The length parameter is solved using a linear regression model based on the thermal gradient extension characteristics. The width parameter is solved by fitting a curve model of edge gradient changes. The depth parameter is solved using an exponential model based on thermal decay characteristics.

9. The method for identifying cracks in the electrode plates of a backup battery in a communication base station using thermal imaging according to claim 8, characterized in that, The linear regression model for the thermal gradient extension characteristics is as follows: ； Where L represents the length parameter; β represents the first constant coefficient; β represents the second constant coefficient; | T| represents the magnitude of the temperature gradient vector; Represents a length infinitesimal element; The curve fitting model for edge gradient changes is as follows: ； Where W represents the width parameter; σ represents the standard deviation; and ln represents the logarithmic function. The exponential model for thermal decay characteristics is as follows: ； Where D represents the depth parameter; λ represents the maximum depth; exp represents the exponential function; λ represents the material attenuation coefficient; ΔT represents the surface temperature difference.

10. The system for identifying cracks in the electrode plates of a backup battery for a communication base station using thermal imaging, as described in any one of claims 1-9, is characterized in that... include: Thermostatic heating platform, pulsed thermal excitation signal source, infrared thermal imager and data processing unit; The constant temperature heating platform is used to heat the battery electrodes to the target temperature range; Infrared thermal imagers are used to acquire thermal distribution image data of the battery electrode surface to obtain a reference thermal image; and to acquire dynamic thermal distribution image data of the battery electrode surface to obtain a dynamic thermal image sequence. A pulsed thermal excitation signal source is used to continuously apply a pulsed thermal excitation signal to the surface of the battery electrode. The data processing unit is used to process the reference thermal image and dynamic thermal image sequence through a multiphysics coupling analysis algorithm to extract the characteristic intensity distribution of suspected crack areas; based on the characteristic intensity distribution of suspected crack areas, a three-dimensional thermal gradient model is constructed, and suspected crack areas are screened based on the three-dimensional thermal gradient model to obtain the real crack areas; the three-dimensional thermal gradient model of the real crack areas is compared with the standard crack library to determine the crack type and crack size.

Images