Welding quality detection method for precise heat dissipation tooth-shaped plate

Abstract

The invention discloses a welding quality detection method for a precise heat dissipation tooth-shaped plate, and the method comprises the steps: determining a heat flow feature point on the tooth-shaped plate, and applying a thermal pulse sequence to obtain dynamic thermal response data; performing space-time decomposition on the thermal response data to generate thermal disturbance response fingerprints; applying a preset phase thermal pulse source according to the thermal disturbance response fingerprint to obtain thermal wave interference field data and generate a thermal wave interference defect map; periodic thermal excitation is applied according to the thermal wave interference defect map, a thermal attenuation time constant is obtained, and a resonance thermal attenuation spectrum and a spectrum consistency spatial distribution map are generated; and calculating a thermal defect severity index according to the spectral consistency spatial distribution map and the thermal wave interference defect map, and generating a welding quality evaluation result. According to the technical scheme, direct correlation between the welding quality and the heat dissipation performance can be established through thermal wave interference and thermal resonance analysis, the welding defect which affects the heat conduction efficiency but does not cause obvious structural failure is effectively recognized, and accurate evaluation of the welding quality is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of welding detection of heat dissipation devices, and particularly to a method for detecting the welding quality of a precision heat dissipation tooth-shaped plate. Background Art

[0002] A precision heat dissipation tooth-shaped plate is a key thermal management component applied in high-power electronic devices, power systems, and precision instruments. Its characteristic is that a large number of precisely designed tooth-shaped structures are distributed on its surface. These tooth shapes usually have an aspect ratio characteristic, that is, the tooth height is much larger than the tooth width, so that the heat dissipation area can be maximized in a limited space. During the manufacturing process of the precision heat dissipation tooth-shaped plate, the welding quality between the tooth shape and the substrate directly determines its heat dissipation performance and service life. The detection of welding quality mainly includes the integrity assessment of the welding structure, the interface continuity analysis, and the measurement of heat conduction efficiency. These detections are of great significance for ensuring the functionality and reliability of the heat dissipation tooth-shaped plate.

[0003] Currently, the commonly used detection methods in the industry mainly include X-ray fluoroscopy detection, ultrasonic detection, and mechanical property testing, etc. When X-ray detection faces tooth-shaped structures arranged in a high density, serious image overlap and signal interference will occur, resulting in a decrease in detection accuracy; ultrasonic detection is difficult to penetrate the complex geometric interfaces in the precision tooth-shaped structures, resulting in a large amount of signal attenuation and scattering; and mechanical property testing often cannot directly reflect the change in heat dissipation performance. In addition, none of the existing detection methods can establish a direct correlation between welding quality and heat dissipation function, and heat dissipation performance is precisely the core functional index of the precision heat dissipation tooth-shaped plate. When welding defects affect the heat conduction efficiency but do not cause obvious structural failure, the existing detection methods are often difficult to effectively identify, which leads to many potential performance problems gradually emerging during the product application process. Therefore, how to achieve a precise correlation assessment between welding quality and heat dissipation performance during the welding quality detection process of the precision heat dissipation tooth-shaped plate has become a technical problem to be solved urgently. Summary of the Invention

[0004] The main purpose of the present invention is: how to achieve an integrated detection of welding quality and heat dissipation function without damaging the structural integrity of the precision heat dissipation tooth-shaped plate, so as to ensure that the influence of welding defects on heat dissipation performance can be accurately identified and quantified.

[0005] The first aspect of the present invention provides a method for detecting the welding quality of a precision heat dissipation tooth-shaped plate, and the method for detecting the welding quality of the precision heat dissipation tooth-shaped plate includes: Taking the tooth tip points, tooth midpoints, tooth root points, tooth valley points between teeth, and tooth side points on the tooth-shaped plate as heat flow characteristic points, applying a spatially asymmetric distributed heat pulse sequence on the back of the tooth-shaped plate substrate, and obtaining the dynamic heat response data of the heat flow characteristic points; Based on the dynamic thermal response data, perform spatio-temporal decomposition of the thermal conduction time delay, thermal attenuation rate, and thermal echo characteristics of the thermal flow characteristic points to generate a thermal response feature vector. Through spatial clustering analysis, obtain a thermal response region distribution map and generate a thermal perturbation response fingerprint. According to the thermal perturbation response fingerprint, apply multiple thermal pulse sources with a preset phase relationship on the back of the substrate to obtain thermal wave interference field distribution data. Through phase difference mapping and thermal wave diffraction angle analysis, generate a thermal wave interference defect map. According to the thermal wave interference defect map, apply a periodic thermal excitation with a preset frequency to obtain the thermal attenuation time constant of each thermal flow characteristic point, generate a resonance thermal attenuation spectrum, and construct a spectrum consistency spatial distribution map. According to the spectrum consistency spatial distribution map and the thermal wave interference defect map, calculate the thermal defect severity index and generate a welding quality evaluation result.

[0006] Preferably, taking the tooth tip points, tooth midpoint, tooth root points, tooth valley points, and tooth side points on the toothed plate as thermal flow characteristic points, applying a spatially asymmetric distributed thermal pulse sequence on the back of the toothed plate substrate to obtain the dynamic thermal response data of the thermal flow characteristic points includes: Define hierarchical thermal flow nodes for the tooth tip points, tooth midpoint, tooth root points, tooth valley points, and tooth side points on the toothed plate. Define the tooth root points as primary thermal flow nodes, the tooth valley points as secondary thermal flow nodes, the tooth midpoint as a thermal flow bifurcation point, the tooth side points as thermal flow coupling points, and the tooth tip points as thermal flow terminal points, and construct a multi-level thermal conduction network. Based on the multi-level thermal conduction network, divide the thermal pulse sequence into a central excitation sequence and a peripheral excitation sequence. The central excitation sequence corresponds to the position of the primary thermal flow nodes, and the peripheral excitation sequence corresponds to the position of the secondary thermal flow nodes. Apply the central excitation sequence, measure the heat distribution ratio of the thermal flow bifurcation point, measure the transverse conduction coefficient of the thermal flow coupling point, measure the energy loss rate of the thermal flow terminal point, and obtain the first set of thermal response characteristic data. Apply the peripheral excitation sequence, measure the thermal flow interweaving intensity between the primary thermal flow nodes and the secondary thermal flow nodes, measure the energy exchange efficiency between the thermal flow bifurcation point and the thermal flow coupling point, and obtain the second set of thermal response characteristic data. According to the first set of thermal response characteristic data and the second set of thermal response characteristic data, calculate the overall thermal response characteristics of the thermal conduction network and obtain the dynamic thermal response data of the thermal flow characteristic points.

[0007] Preferably, the step of performing spatio-temporal decomposition of the thermal conduction time delay, thermal attenuation rate, and thermal echo characteristics of the thermal flow characteristic points based on the dynamic thermal response data to generate a thermal response feature vector, obtaining a thermal response region distribution map through spatial clustering analysis, and generating a thermal perturbation response fingerprint includes: Segment the dynamic thermal response data by time series, divide it into a heat conduction startup segment, a heat conduction stable segment, and a heat conduction decay segment, extract the temperature change curves of each segment, and generate segmented thermal response data; According to the segmented thermal response data, measure the conduction time from the first-level heat flow node to the second-level heat flow node in the heat conduction startup segment, calculate the conduction stability coefficient of the heat flow bifurcation point in the heat conduction stable segment, measure the decay time of the heat flow endpoint in the heat conduction decay segment, and generate a conduction time delay vector; Conduct a heat decay rate analysis on the conduction time delay vector and the segmented thermal response data, calculate the initial decay rate of the heat conduction startup segment, the steady-state decay rate of the heat conduction stable segment, and the final decay rate of the heat conduction decay segment, and generate a decay characteristic vector; Conduct a heat echo analysis based on the segmented thermal response data, extract the echo peak value and echo period of the heat conduction stable segment, calculate the echo attenuation coefficient and echo energy ratio of the heat conduction decay segment, and generate a heat echo characteristic vector; Perform weighted combination on the conduction time delay vector, the decay characteristic vector, and the heat echo characteristic vector according to the heat flow node level, generate a thermal response characteristic vector, obtain a thermal response region distribution map through Mahalanobis distance space clustering analysis, and generate a thermal disturbance response fingerprint.

[0008] Preferably, the conducting a heat echo analysis according to the segmented thermal response data, extracting the echo peak value and echo period of the heat conduction stable segment, calculating the echo attenuation coefficient and echo energy ratio of the heat conduction decay segment, and generating a heat echo characteristic vector includes: Perform interference noise filtering on the temperature response curve of each heat flow characteristic point on the toothed plate to obtain a pure thermal response signal; According to the pure thermal response signal, identify the heat echo signal in the temperature curve, extract the peak and valley values of the heat echo signal, and calculate the echo amplitude and duration to obtain the basic echo characteristic data; Perform cross-correlation analysis on the heat echo signals of the heat flow characteristic points, calculate the echo time delay difference and energy attenuation ratio between adjacent heat flow nodes, and obtain the echo propagation characteristics; According to the basic echo characteristic data and the echo propagation characteristics, reconstruct the heat echo reflection path in the toothed structure and generate a heat echo characteristic vector.

[0009] Preferably, the applying a plurality of heat pulse sources with a preset phase relationship on the back of the substrate according to the thermal disturbance response fingerprint, obtaining the thermal wave interference field distribution data, and generating a thermal wave interference defect map through phase difference mapping and thermal wave diffraction angle analysis includes: Perform phase configuration on the heat pulse sources according to the thermal disturbance response fingerprint, and set the phase difference between adjacent heat pulse sources to π / 4, π / 2, 3π / 4 to obtain a phase sequence of multiple heat pulse sources; Extract the amplitude of the thermal wave interference field for each group of phase sequences, calculate the interference intensity ratio between adjacent heat flux nodes, and obtain the intensity distribution of the thermal wave interference field; According to the intensity distribution of the thermal wave interference field, map the phase difference between adjacent heat flux nodes, measure the relative phase delay of each node, and generate a phase difference distribution map; Analyze the diffraction angle of the propagation path of the thermal wave in the tooth root region, calculate the diffraction deflection angle at each heat flux node, and obtain the diffraction angle distribution; According to the phase difference distribution map and the diffraction angle distribution, calculate the distortion degree of the thermal wave propagation path, and generate a thermal wave interference defect map.

[0010] Preferably, the step of mapping the phase difference between adjacent heat flux nodes according to the intensity distribution of the thermal wave interference field, measuring the relative phase delay of each node, and generating a phase difference distribution map includes: Subtract the initial phase value of the reference heat pulse source from the phase data of adjacent heat flux nodes according to the heat conduction path, limit the phase deviation within the range of -π to π, and obtain the reference phase distribution data; Calculate the phase difference between adjacent measurement points spaced 0.5 mm apart in the tooth root region of the reference phase distribution data to obtain the phase change curve at the tooth root, calculate the phase difference between adjacent measurement points spaced 0.2 mm apart in the tooth tip region to obtain the phase change curve at the tooth tip, and calculate the slopes of the phase change curves at the tooth root and the tooth tip respectively to obtain the phase change rate in the direction perpendicular to the propagation direction; According to the reference phase distribution data, pair the lateral measurement points between adjacent tooth profiles in pairs according to the central symmetry principle, calculate the phase difference of each pair of measurement points, and determine the phase jump value at the welding interface through the mutation position and mutation amplitude of the phase difference to obtain the phase distortion degree in the lateral propagation direction; Normalize the phase change rate and the phase distortion degree, set the weight coefficient in the direction perpendicular to the propagation direction to 0.6, and set the weight coefficient in the lateral propagation direction to 0.4, and generate the phase difference distribution map.

[0011] Preferably, the step of applying a periodic thermal excitation with a preset frequency according to the thermal wave interference defect map, obtaining the thermal decay time constant of each heat flux characteristic point, generating a resonance thermal decay spectrum, and constructing a spectrum consistency spatial distribution map includes: According to the thermal wave interference defect map, divide the heat flux characteristic points into a first-level response area, a second-level response area, and an edge response area according to the heat conduction level, and apply a periodic thermal excitation in the range of 0.01 - 1 Hz to the first-level response area to obtain the main frequency response data; Calculate the heat conduction delay time from the primary response region to the secondary response region according to the aspect ratio of the tooth profile structure, and apply a periodically thermal excitation with a phase lag to the secondary response region according to the heat conduction delay time to obtain secondary frequency response data; Apply a thermal excitation with the same frequency but opposite phase as the primary response region to the edge response region, obtain the thermal interference response data between adjacent tooth profiles, and measure the temperature decay curves of each response region after the thermal excitation stops; Based on the temperature decay curves of each response region, calculate the heat decay time constants in the vertical conduction direction and the transverse conduction direction, analyze the decay characteristics of the heat flow on different conduction paths, and generate a resonance heat decay spectrum; Perform a conduction path mapping on the resonance heat decay spectra of each heat flow characteristic point, calculate the spectral correlation degree between adjacent conduction paths, and construct a spectral consistency spatial distribution map.

[0012] Preferably, the calculating the heat decay time constants in the vertical conduction direction and the transverse conduction direction based on the temperature decay curves of each response region, analyzing the decay characteristics of the heat flow on different conduction paths, and generating a resonance heat decay spectrum includes: Take a sampling point every 0.2 mm from the tooth root to the tooth tip of the temperature decay curve in the vertical conduction direction, divide the temperature difference between adjacent sampling points by the time interval to obtain a linearization coefficient, divide the region where the linearization coefficient is greater than 0.5 K / s into a fast decay region according to the tooth profile height, and divide the region where the linearization coefficient is less than 0.5 K / s into a slow decay region to obtain the decay time constant in the vertical conduction direction; Perform a Fourier transform on the temperature decay curve in the transverse conduction direction, extract the main frequency component of the energy density spectrum, calculate the energy transfer coefficient between adjacent tooth profiles, divide the region where the energy transfer coefficient is greater than 0.7 into a strong coupling region, and divide the region where the energy transfer coefficient is less than 0.7 into a weak coupling region to obtain the decay time constant in the transverse conduction direction; According to the decay time constant in the vertical conduction direction and the decay time constant in the transverse conduction direction, construct a conduction matrix, calculate the coupling coefficient of each conduction region, optimize the minimum thermal resistance of the conduction path, and generate a resonance heat decay spectrum.

[0013] Preferably, the calculating the heat defect severity index according to the spectral consistency spatial distribution map and the thermal wave interference defect map, and generating a welding quality evaluation result includes: According to the spectral consistency spatial distribution map, calculate the standard deviation of the spectral consistency between adjacent tooth profiles, mark the region where the standard deviation is greater than 0.3 as the welding non-uniform region, and mark the region where the standard deviation is less than 0.1 as the welding stable region to generate welding uniformity evaluation data; Performing spatial gradient analysis on the phase distortion in the thermal wave interference defect map, marking the area with a gradient value greater than 0.5 rad / mm as a welding interface fracture area, and marking the area with a gradient value between 0.2-0.5 rad / mm as a welding interface weak connection area, to generate welding continuity evaluation data; According to the welding uniformity evaluation data and the welding continuity evaluation data, the degree of blocking of the heat dissipation channel by the welding defects is calculated, and the area where the thermal resistance increases by more than 50% is determined as an L4 defect, the area where the thermal resistance increases by more than 20% and less than 50% is determined as an L3 defect, the area where the thermal resistance increases by more than 5% and less than 20% is determined as an L2 defect, and the area where the thermal resistance increases by less than 5% is determined as an L1 defect, and the thermal defect severity index is obtained; The toothed plate is regionally graded and labeled according to the thermal defect severity index to generate a welding quality assessment result.

[0014] The technical solution provided in the embodiment of the present application obtains dynamic thermal response data by establishing a heat flow characteristic point mapping system and applying an asymmetric heat pulse sequence. This method can obtain thermal conductivity characteristic information without destroying the structure. When the heat pulse sequence is applied to the back of the substrate, the heat energy will propagate along the welding interface, forming a unique thermal response pattern at the heat flow characteristic points at different positions. A good welding interface will show regular thermal conductivity characteristics, while welding defects will cause abnormal thermal response.

[0015] The thermal response data is decomposed in time and space to generate thermal disturbance response fingerprints, which can capture the impact of welding defects on heat conduction from two dimensions: time and space. Based on the thermal disturbance response fingerprint, multiple heat pulse sources with preset phase relationships are applied, and the generated thermal wave interference field can reflect the continuity of the welding interface. The thermal wave will be modulated by the welding defect during propagation, and this modulation will change the phase and diffraction characteristics of the thermal wave. By analyzing these changes, the welding defects can be accurately located and characterized.

[0016] Applying periodic thermal excitation of a preset frequency and analyzing the thermal decay time constant can reveal the thermal response characteristics of the welding interface at different frequencies. A good welding interface will exhibit stable thermal decay characteristics, while welding defects will cause abnormal thermal decay phenomena at certain frequencies. By analyzing this frequency-dependent thermal response behavior, tiny defects that affect heat conduction but do not cause obvious structural damage can be identified. The generated resonant thermal decay spectrum and spectral consistency spatial distribution map comprehensively record these thermal response characteristics.

[0017] The thermal defect severity index is calculated based on the thermal wave interference defect map and the spectral consistency spatial distribution map, and a quantitative relationship between welding defects and heat dissipation performance is established. This method can not only detect the location and type of welding defects, but more importantly, it can evaluate the actual impact of these defects on heat dissipation performance. By analyzing the characteristics of thermal wave propagation and thermal attenuation behavior, it is possible to accurately determine whether welding defects will affect the core function of the heat dissipation tooth-shaped plate, even if these defects have not caused visible structural damage. Brief Description of the Drawings

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on the structures shown in these drawings.

[0019] Figure 1 It is a schematic diagram of an embodiment of the welding quality detection method for the precision heat dissipation tooth-shaped plate in the embodiment of the present invention.

[0020] The realization of the object, functional characteristics, and advantages of the present invention will be further described in conjunction with the embodiments and with reference to the drawings. Detailed Embodiments

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0022] It should be noted that if there are directional indications (such as up, down, left, right, front, back...) involved in the embodiments of the present invention, the directional indications are only used to explain the relative position relationship and movement conditions between components in a specific posture (as shown in the drawings). If the specific posture changes, the directional indications will also change accordingly.

[0023] In addition, the descriptions involving "first", "second", etc. in the present invention are for descriptive purposes only, and should not be construed as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In addition, "and / or" throughout the text includes three scenarios. Taking A and / or B as an example, it includes the technical solution of A, the technical solution of B, and the technical solution where both A and B are satisfied simultaneously. In addition, the technical solutions between various embodiments can be combined with each other, which must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions results in contradictions or inability to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0024] An embodiment of the present application provides a method for detecting the welding quality of a precision heat dissipation tooth-shaped plate. Figure 1 It is a flowchart of a method for detecting the welding quality of a precision heat dissipation tooth-shaped plate provided by an embodiment of the present application. In this embodiment, the method includes: Please refer to Figure 1 , taking the tooth tip points, tooth midpoint points, tooth root points, tooth valley points between teeth, and tooth side points on the tooth-shaped plate as heat flow characteristic points, applying a spatially asymmetrically distributed thermal pulse sequence on the back of the tooth-shaped plate substrate, and obtaining the dynamic thermal response data of the heat flow characteristic points; In an embodiment of the present invention, the step of taking the tooth tip points, tooth midpoint points, tooth root points, tooth valley points between teeth, and tooth side points on the tooth-shaped plate as heat flow characteristic points, applying a spatially asymmetrically distributed thermal pulse sequence on the back of the tooth-shaped plate substrate, and obtaining the dynamic thermal response data of the heat flow characteristic points includes: Defining hierarchical heat flow nodes for the tooth tip points, tooth midpoint points, tooth root points, tooth valley points between teeth, and tooth side points on the tooth-shaped plate, defining the tooth root points as first-level heat flow nodes, the tooth valley points between teeth as second-level heat flow nodes, the tooth midpoint points as heat flow bifurcation points, the tooth side points as heat flow coupling points, and the tooth tip points as heat flow terminal points, and constructing a multi-level heat flow conduction network; Based on the multi-level heat flow conduction network, dividing the thermal pulse sequence into a central excitation sequence and a peripheral excitation sequence, where the central excitation sequence corresponds to the position of the first-level heat flow nodes, and the peripheral excitation sequence corresponds to the position of the second-level heat flow nodes; Applying the central excitation sequence, measuring the heat distribution ratio of the heat flow bifurcation point, measuring the transverse conduction coefficient of the heat flow coupling point, measuring the energy loss rate of the heat flow terminal point, and obtaining the first set of thermal response characteristic data; Applying the peripheral excitation sequence, measuring the heat flow interweaving intensity between the first-level heat flow nodes and the second-level heat flow nodes, measuring the energy exchange efficiency between the heat flow bifurcation point and the heat flow coupling point, and obtaining the second set of thermal response characteristic data; Calculate the overall thermal response characteristics of the heat conduction network based on the first set and the second set of thermal response characteristic data to obtain the dynamic thermal response data of the heat flow characteristic points.

[0025] The following specifically describes the steps involved in the above embodiments: The distribution of heat flow characteristic points on the precision heat dissipation tooth-shaped plate corresponds to the key positions of heat conduction. The tooth root point is located at the connection between the tooth shape and the substrate and is the main inlet for heat to enter the tooth shape from the substrate, which is defined as a primary heat flow node; the tooth valley point is located at the intersection of adjacent tooth shapes and undertakes the secondary distribution function of heat, which is defined as a secondary heat flow node; the tooth midpoint is located in the middle of the tooth shape height and is responsible for the heat shunt in the vertical direction, which is defined as a heat flow bifurcation point; the tooth side point is located on the side of the tooth shape and is responsible for the heat exchange between adjacent tooth shapes, which is defined as a heat flow coupling point; the tooth tip point is located at the top of the tooth shape and is the final outlet for heat to dissipate to the environment, which is defined as a heat flow terminal point. Calibrate these characteristic points through an infrared thermal imager to establish a multi-level network structure reflecting the heat conduction path. This hierarchical definition fully considers the structural characteristics of the heat dissipation tooth-shaped plate, enabling each characteristic point to have a clear heat conduction functional attribute.

[0026] Based on the multi-level heat conduction network, the heat pulse sequences are divided into two categories: the central excitation sequence is applied at the position of the primary heat flow node, that is, the area where the tooth root point is located; the peripheral excitation sequence is applied at the position of the secondary heat flow node, that is, the area where the tooth valley point is located. The excitation sequence is generated by a heat pulse generator, and the distribution method of the sequence matches the heat conduction path of the heat dissipation tooth-shaped plate. For example, on a 10mm×10mm heat dissipation tooth-shaped plate, the central excitation sequence covers the central 6mm×6mm area, and the peripheral excitation sequence covers the peripheral area. This distribution method can accurately reflect the flow characteristics of heat between different conduction levels.

[0027] After applying the central excitation sequence, collect the temperature field data through a high-speed thermal imager. The heat distribution ratio is calculated by measuring the temperature difference between the upper and lower sides of the heat flow bifurcation point, reflecting the heat distribution in the vertical direction; the lateral conduction coefficient is obtained by measuring the temperature change rate of the heat flow coupling points of adjacent tooth shapes, characterizing the lateral diffusion ability of heat; the energy loss rate is determined by the slope of the temperature decay curve of the heat flow terminal point, reflecting the heat dissipation efficiency. For example, in a good welding state, the temperature difference between the upper and lower sides of the heat flow bifurcation point is usually 1 - 2K, the lateral conduction coefficient is in the range of 0.8 - 0.9K, and the energy loss rate is about 0.1K / s.

[0028] When applying the peripheral excitation sequence, the temperature field data is also collected. The heat flux intertwining intensity is calculated through the temperature cross-correlation function of the primary heat flux nodes and the secondary heat flux nodes, reflecting the degree of heat interaction between different levels; the energy exchange efficiency is obtained by calculating the temperature response time difference between the heat flux bifurcation point and the heat flux coupling point, characterizing the heat transfer efficiency in space. For example, when the welding interface is intact, the peak value of the temperature cross-correlation function is usually greater than 0.85, and the temperature response time difference is less than 100 ms.

[0029] Finally, the two sets of thermal response characteristic data are transformed into the frequency domain through Fourier transform, the spectral energy distribution is calculated, and the dynamic thermal response data reflecting the overall heat transfer characteristics is obtained. The data processing process includes: sampling the temperature curve at 50 Hz, selecting a 60 s data segment for FFT transformation, and extracting the energy distribution characteristics in the 0 - 5 Hz frequency band. This processing method not only retains the dynamic characteristics of heat conduction but also filters out the influence of environmental noise. For a heat dissipation tooth-shaped plate with a size of 10 mm × 10 mm, the main energy is concentrated in the 0.1 - 1 Hz frequency band, which corresponds to the thermal time constant of the tooth-shaped structure.

[0030] Please continue to refer to Figure 1 , according to the dynamic thermal response data, perform spatio-temporal decomposition of the heat conduction time delay, heat attenuation rate, and heat echo characteristics of the heat flux characteristic points to generate a thermal response characteristic vector, obtain a thermal response region distribution map through spatial clustering analysis, and generate a thermal perturbation response fingerprint; In an embodiment of the present invention, the performing spatio-temporal decomposition of the heat conduction time delay, heat attenuation rate, and heat echo characteristics of the heat flux characteristic points according to the dynamic thermal response data to generate a thermal response characteristic vector, obtaining a thermal response region distribution map through spatial clustering analysis, and generating a thermal perturbation response fingerprint includes: Perform time series segmentation on the dynamic thermal response data, divide it into a heat conduction start segment, a heat conduction stable segment, and a heat conduction attenuation segment, extract the temperature change curves of each segment, and generate segmented thermal response data; According to the segmented thermal response data, measure the conduction time from the primary heat flux node to the secondary heat flux node in the heat conduction start segment, calculate the conduction stability coefficient of the heat flux bifurcation point in the heat conduction stable segment, measure the attenuation time of the heat flux terminal point in the heat conduction attenuation segment, and generate a conduction time delay vector; Perform heat attenuation rate analysis on the conduction time delay vector and the segmented thermal response data, calculate the initial attenuation rate in the heat conduction start segment, the steady-state attenuation rate in the heat conduction stable segment, and the final attenuation rate in the heat conduction attenuation segment, and generate an attenuation characteristic vector; Perform heat echo analysis according to the segmented thermal response data, extract the echo peak value and echo period in the heat conduction stable segment, calculate the echo attenuation coefficient and echo energy ratio in the heat conduction attenuation segment, and generate a heat echo characteristic vector; The conduction time-delay vector, attenuation feature vector, and thermal echo feature vector are weighted and combined according to the thermal flow node hierarchy to generate a thermal response feature vector. Through Mahalanobis distance space clustering analysis, a thermal response region distribution map is obtained, and a thermal disturbance response fingerprint is generated.

[0031] The following specifically describes the steps involved in the above embodiments: The time-series segmentation of the dynamic thermal response data is based on the temperature change characteristics. The heat conduction startup segment refers to the stage when the temperature starts to rise after the heat pulse is applied. Since the heat just starts to enter the tooth-shaped structure, the temperature changes violently in this stage; the heat conduction stable segment refers to the stage when the temperature reaches stability. At this time, a stable heat conduction channel is formed in the tooth-shaped structure; the heat conduction attenuation segment refers to the stage when the temperature gradually decreases after the heat pulse stops, reflecting the heat dissipation process. A high-speed data acquisition system is used to sample the temperature data at 100 Hz, and the temperature-time curve is recorded. The selection of the 100 Hz sampling frequency is based on the thermal response characteristics of the heat dissipation tooth-shaped plate and can accurately capture temperature changes at the millisecond level.

[0032] The generation process of the conduction time-delay vector includes the measurement of three key parameters. The conduction time from the primary heat flow node to the secondary heat flow node is measured in the heat conduction startup segment, which reflects the heat conduction speed at the welding interface; the ratio of the standard deviation to the average value of the temperature fluctuation at the heat flow bifurcation point is calculated in the heat conduction stable segment, and this ratio is the conduction stability coefficient, which reflects the uniformity of heat distribution; the time required for the temperature at the heat flow end point to drop to 37% of the peak value is measured in the heat conduction attenuation segment, and this time is the attenuation time, which reflects the heat dissipation ability. The selection of the 37% threshold is based on the characteristic time definition of the thermal system.

[0033] The thermal decay rate analysis focuses on the temperature change rate. The initial decay rate reflects the conduction characteristics in the initial stage of heat input and is obtained by calculating the change rate of the temperature curve in the startup segment; the steady-state decay rate characterizes the heat loss rate during stable conduction and is obtained by calculating the average temperature drop rate in the stable segment; the final decay rate reflects the characteristics of the final heat dissipation stage and is obtained by calculating the exponential fitting slope of the temperature curve in the attenuation segment. These three decay rates constitute the attenuation feature vector, which completely describes the dynamic characteristics of the entire heat conduction process.

[0034] The thermal echo feature analysis includes the extraction of four parameters. The echo peak value refers to the maximum amplitude of the periodic fluctuation in the temperature signal, and the echo period refers to the repeated time interval of the temperature fluctuation. These two parameters are extracted from the heat conduction stable segment; the echo attenuation coefficient refers to the attenuation rate of the echo amplitude, and the echo energy ratio refers to the proportion of the echo signal energy in the total energy. These two parameters are extracted from the heat conduction attenuation segment. These echo features are separated from the temperature signal through the wavelet transform method to generate the thermal echo feature vector.

[0035] The weighted combination of eigenvectors takes into account the hierarchical relationship of heat flow nodes. The first-level heat flow nodes have the highest weight because they directly reflect the heat transfer performance of the welding interface; the second-level heat flow nodes have the next highest weight, reflecting the lateral heat diffusion ability; the other nodes have the lowest weight and are used to supplement the detailed characteristics of heat conduction. Mahalanobis distance is used for spatial clustering to group regions with similar heat response characteristics into one category, and finally a heat response region distribution map reflecting the welding quality distribution is formed.

[0036] In one embodiment of the present invention, the heat echo analysis is performed according to the segmented heat response data, the echo peak value and echo period of the stable section of heat conduction are extracted, the echo attenuation coefficient and echo energy ratio of the attenuation section of heat conduction are calculated, and a heat echo feature vector is generated, including: The temperature response curves of each heat flow feature point on the toothed plate are processed to filter out interference noise, and a pure heat response signal is obtained; According to the pure heat response signal, the heat echo signal in the temperature curve is identified, the peak and valley values of the heat echo signal are extracted, the echo amplitude and duration are calculated, and the basic echo feature data are obtained; The cross-correlation analysis is performed on the heat echo signals of the heat flow feature points, the echo time delay difference and energy attenuation ratio between adjacent heat flow nodes are calculated, and the echo propagation characteristics are obtained; According to the basic echo feature data and echo propagation characteristics, the heat echo reflection path in the toothed structure is reconstructed to generate a heat echo feature vector.

[0037] The following specifically describes the steps involved in the above embodiment: The interference noise filtering of the temperature response curve adopts a multi-stage filtering method. First, median filtering is used to remove the pulse noise in the temperature data, and the filtering window is set to 5 data points. The window size is selected based on the minimum time scale of the temperature change of the heat dissipation toothed plate; then the wavelet transform method is used to decompose the signal, a fourth-order wavelet basis function is selected for 4-layer decomposition, and the effective signal components in the 3-10 Hz frequency band are extracted. This frequency band range covers the main frequency characteristics of the heat conduction process; finally, Wiener filtering is used to suppress random noise to obtain a pure heat response signal. This combination of multi-stage filtering fully considers the characteristics of the temperature signal of the heat dissipation toothed plate, retains the heat echo information and effectively removes various interferences.

[0038] The recognition of the thermal echo signal adopts the peak detection method. In the pure thermal response signal, the peak detection threshold is set to 1.5 times the root mean square of the signal, and this threshold selection is based on the contrast between the thermal echo signal intensity and the background temperature fluctuation; at the same time, it is required that the interval between adjacent peaks is greater than 0.1 second to filter out false peaks. Pair the detected peak and valley values, calculate the amplitude difference between each pair of peak and valley values to obtain the echo amplitude, and calculate the duration to obtain the time characteristics of the echo. These parameters constitute the basic characteristic data of the echo, reflecting the energy change law of the thermal echo when propagating at the welding interface.

[0039] The cross-correlation analysis of the thermal echo signal is based on the signal comparison of adjacent heat flux nodes. Perform cross-correlation calculation on the echo signals of adjacent nodes. The peak position of the cross-correlation function represents the echo time delay difference, and the peak amplitude ratio represents the energy attenuation ratio. The time delay difference reflects the time required for the thermal wave to propagate between adjacent nodes, and the energy attenuation ratio reflects the energy loss during the propagation process. This analysis method reveals the spatio-temporal characteristics of the thermal echo when propagating at the welding interface. When the thermal echo propagates on a sound welding interface, it shows regular time delay and attenuation characteristics.

[0040] The reconstruction of the thermal echo reflection path maps the echo characteristic data onto the spatial distribution of the tooth-shaped structure. Determine the propagation direction of the thermal echo based on the basic characteristic data of the echo, and calculate the geometric parameters of the propagation path according to the propagation characteristics of the echo. The echo will be reflected at the welding interface during the propagation process, and the integrity of the reflection path directly reflects the welding quality. By calculating the continuity coefficient and energy transfer efficiency of the reflection path, a thermal echo feature vector containing spatial distribution information is generated. This reconstruction method establishes a direct correlation between the thermal echo characteristics and the physical state of the welding interface.

[0041] Please continue to refer to Figure 1 , according to the thermal disturbance response fingerprint, apply multiple thermal pulse sources with a preset phase relationship on the back of the substrate, obtain the thermal wave interference field distribution data, and generate a thermal wave interference defect map through phase difference mapping and thermal wave diffraction angle analysis; In an embodiment of the present invention, the step of applying multiple thermal pulse sources with a preset phase relationship on the back of the substrate according to the thermal disturbance response fingerprint, obtaining the thermal wave interference field distribution data, and generating a thermal wave interference defect map through phase difference mapping and thermal wave diffraction angle analysis includes: Perform phase configuration on the thermal pulse sources according to the thermal disturbance response fingerprint, and set the phase difference between adjacent thermal pulse sources to π / 4, π / 2, 3π / 4 to obtain multiple groups of phase sequences of the thermal pulse sources; Extract the amplitude of the thermal wave interference field under each group of phase sequences, calculate the interference intensity ratio between adjacent heat flux nodes, and obtain the thermal wave interference field intensity distribution; According to the intensity distribution of the thermal wave interference field, map the phase difference between adjacent heat flux nodes, measure the relative phase delay of each node, and generate a phase difference distribution map; Analyze the diffraction angle of the propagation path of the thermal wave in the tooth root region, calculate the diffraction deflection angle at each heat flux node, and obtain the diffraction angle distribution; According to the phase difference distribution map and the diffraction angle distribution, calculate the distortion degree of the thermal wave propagation path, and generate a thermal wave interference defect map.

[0042] The following specifically describes the steps involved in the above embodiments: The phase configuration of the thermal pulse source is controlled by a multi-channel pulse signal generator, and each thermal pulse source is driven through an independent control channel. The trigger times of the pulse signals are staggered by 4 ms, 8 ms, and 12 ms in sequence, corresponding to phase differences of π / 4, π / 2, and 3π / 4 respectively, forming three different phase sequences. In each phase sequence, the pulse duration is set to 20 ms, the pulse interval is 100 ms, and the thermal pulse is excited 5 times under each phase difference configuration to ensure the repeatability of the measurement. The setting of the trigger time interval is based on the thermal diffusion characteristics of the heat dissipation tooth-shaped plate. The minimum interval of 4 ms ensures that obvious phase differences can be formed between adjacent measurement points by the thermal wave, while the maximum interval of 12 ms avoids signal distortion caused by excessive attenuation of the thermal wave.

[0043] The analysis of the thermal wave interference field uses a high-precision infrared thermal imager to collect the temperature field, and the sampling rate is set to 200 Hz to capture rapid temperature changes. Fourier transform is performed on the collected temperature data, and the amplitude of the main frequency component is extracted as the interference intensity index. The interference intensity ratio between adjacent heat flux nodes is obtained by calculating the ratio of the amplitudes of the two points, and this ratio reflects the energy change law of the thermal wave during propagation. In the area with good welding, the interference intensity ratio between adjacent nodes fluctuates within the range of 0.8 - 1.2, and this range is obtained based on a large number of experimental data statistics; when the ratio is lower than 0.8, it indicates that the heat conduction is blocked, suggesting the existence of defects at the welding interface; when the ratio exceeds 1.2, it means that abnormal heat accumulation occurs, also indicating welding abnormalities.

[0044] The phase difference mapping uses the Hilbert transform method to calculate the instantaneous phase of the temperature signal. First, perform the Hilbert transform on the temperature signal to obtain the analytic signal, and then calculate the phase angle of the analytic signal to obtain the instantaneous phase. The relative phase delay between adjacent heat flux nodes is calculated by the difference between the instantaneous phases of the two points, and the calculation result is converted into a standard phase difference within the range of 0 - 2π through 2π normalization processing. This processing method eliminates the influence of phase jumps and makes the phase difference distribution more continuous and smooth. On the intact welding interface, the phase delay between adjacent nodes shows a linear increasing pattern; when welding defects occur, the phase delay will show a mutation at the defect position, deviating from the normal increasing trend.

[0045] The diffraction angle analysis is based on the calculation of the spatial gradient of the temperature field. The two-dimensional gradient vector of the temperature field is calculated using the central difference scheme. The propagation direction of the thermal wave is determined by the gradient vector and is perpendicular to the isotherm. By calculating the angle between the gradient vector and the vertical direction, the diffraction deflection angle at each heat flux node is obtained. In the normal welding area, the propagation path of the thermal wave is basically along the axial direction of the tooth-shaped structure, and the diffraction deflection angle is less than 5°. When encountering welding defects, the thermal wave will deflect significantly, and the diffraction angle will increase sharply, exceeding 15°. These angle thresholds are determined according to the geometric characteristics of the tooth-shaped structure and the physical laws of heat conduction.

[0046] The generation of the thermal wave interference defect map adopts a multi-parameter weighted superposition method. First, the spatial gradient values and diffraction angle values of the phase difference are normalized to make the two types of data comparable. Then, weighted superposition is carried out according to a weight ratio of 3:2 to obtain a comprehensive index characterizing the distortion degree of the thermal wave propagation path. The selection of the weight ratio is based on the sensitivity analysis results of the phase difference and diffraction angle to defect detection. The areas with a distortion degree greater than 0.5 are marked as severe defects, and the heat conduction in these areas is significantly affected. The areas with a distortion degree between 0.3 and 0.5 are marked as minor defects, indicating the existence of welding problems that affect the heat dissipation performance but have not led to complete failure. The areas with a distortion degree less than 0.3 are marked as intact areas, indicating that the welding quality meets the requirements.

[0047] In an embodiment of the present invention, mapping the phase difference between adjacent heat flux nodes according to the intensity distribution of the thermal wave interference field, measuring the relative phase delay of each node, and generating a phase difference distribution map includes: Subtracting the initial phase value of the reference heat pulse source from the phase data of adjacent heat flux nodes along the heat conduction path, and limiting the phase deviation within the range of -π to π to obtain the reference phase distribution data; Calculating the phase difference between adjacent measurement points spaced 0.5 mm apart in the tooth root region of the reference phase distribution data to obtain the phase change curve at the tooth root, and calculating the phase difference between adjacent measurement points spaced 0.2 mm apart in the tooth tip region to obtain the phase change curve at the tooth tip. Respectively calculating the slopes of the phase change curves at the tooth root and the tooth tip to obtain the phase change rate in the vertical propagation direction; According to the reference phase distribution data, pairing the transverse measurement points between adjacent tooth shapes in pairs according to the central symmetry principle, calculating the phase difference value of each pair of measurement points, and determining the phase jump value at the welding interface through the mutation position and mutation amplitude of the phase difference to obtain the phase distortion degree in the transverse propagation direction; Normalize the phase change rate and phase distortion degree, set the weight coefficient in the vertical propagation direction to 0.6, and the weight coefficient in the horizontal propagation direction to 0.4 to generate the phase difference distribution map.

[0048] The following specifically describes the steps involved in the above embodiments: For the benchmarking process of phase data, first determine the reference heat pulse source, and select the heat pulse source closest to the area to be measured as the reference source. Process the phase data of each heat flow node along the heat conduction path in sequence, subtract the initial phase value of the reference source from the measured phase value to obtain the relative phase value. The relative phase value is limited to the range between -π and π through the phase unwrapping algorithm to eliminate the influence of periodic jumps. The benchmarking process is similar to aligning all waveforms to the same starting point when multiple sound waves are synchronized, enabling subsequent phase comparisons to have a unified reference benchmark. The reference phase distribution data reflects the phase change law of the thermal wave during propagation and excludes the interference caused by inconsistent initial phases.

[0049] For the phase difference analysis in the vertical propagation direction, different spatial sampling intervals are used. In the root region of the tooth, a measurement point is set every 0.5 mm, and this interval value is selected based on the relatively large heat flux density at the tooth root; in the tip region of the tooth, a measurement point is set every 0.2 mm, and the denser sampling interval is used to capture the faster heat diffusion changes at the tooth tip. Use an infrared thermal imager to collect the temperature data of these measurement points, and obtain the phase information through Fourier transform. Continuously plot the phase differences between adjacent measurement points to obtain the phase change curves at the tooth root and tooth tip respectively. Perform linear fitting on these two curves and calculate their slope values. The slope reflects the rate of phase change in space. Different sampling intervals are used in the root region and tip region of the tooth to meet the measurement accuracy requirements of different regions.

[0050] The phase analysis in the horizontal propagation direction is based on the principle of central symmetry. Select horizontal measurement point pairs between adjacent tooth profiles, and the measurement point pairs are symmetrically distributed about the center line of the tooth profile. Calculate the phase difference for each pair of measurement points and record the spatial distribution of the phase difference. When the thermal wave passes through the welding interface, a continuous phase change occurs at a sound welding interface; while at a defective position, a sudden change in the phase difference occurs. By setting a threshold for the phase difference change rate (a change greater than π / 4 per millimeter is considered a sudden change), identify the position and amplitude of the phase jump. The magnitude of the phase jump value reflects the severity of the welding defect, and the jump position indicates the specific location of the defect.

[0051] The generation of the phase difference distribution map adopts a weighted combination method. First, the phase change rate in the vertical propagation direction and the phase distortion degree in the lateral propagation direction are normalized so that the numerical range is unified between 0 and 1. The weight coefficient in the vertical direction is set to 0.6, and the weight coefficient in the lateral direction is set to 0.4. This weight ratio is determined based on the physical characteristic that the main direction of heat conduction is the vertical direction. The data after weighted combination forms a phase difference distribution map, which intuitively shows the spatial distribution characteristics of the welding quality. In this way, both the heat transfer characteristics mainly based on vertical conduction and the contribution of lateral conduction to the welding quality assessment are considered.

[0052] Please continue to refer to Figure 1 , according to the thermal wave interference defect map, apply a periodic thermal excitation with a preset frequency, obtain the thermal decay time constants of each heat flow characteristic point, generate a resonance thermal decay spectrum, and construct a spectrum consistency spatial distribution map; In an embodiment of the present invention, the step of applying a periodic thermal excitation with a preset frequency according to the thermal wave interference defect map, obtaining the thermal decay time constants of each heat flow characteristic point, generating a resonance thermal decay spectrum, and constructing a spectrum consistency spatial distribution map includes: According to the thermal wave interference defect map, divide the heat flow characteristic points into a primary response area, a secondary response area, and an edge response area according to the heat conduction level. Apply a periodic thermal excitation in the range of 0.01 - 1 Hz to the primary response area to obtain the main frequency response data; According to the aspect ratio of the tooth shape structure, calculate the heat conduction delay time from the primary response area to the secondary response area, and apply a periodically thermal excitation with a phase lag to the secondary response area according to the heat conduction delay time to obtain the secondary frequency response data; Apply a thermal excitation with the same frequency but opposite phase as the primary response area to the edge response area, obtain the thermal interference response data between adjacent tooth shapes, and measure the temperature decay curves of each response area after the thermal excitation stops; Based on the temperature decay curves of each response area, calculate the thermal decay time constants in the vertical conduction direction and the lateral conduction direction, analyze the decay characteristics of the heat flow on different conduction paths, and generate a resonance thermal decay spectrum; Perform a conduction path mapping on the resonance thermal decay spectra of each heat flow characteristic point, calculate the spectral correlation degree between adjacent conduction paths, and construct a spectrum consistency spatial distribution map.

[0053] The following specifically describes the steps involved in the above - mentioned embodiments: When partitioning the heat flux characteristic points according to the thermal wave interference defect map, the area closest to the heat source with the highest thermal wave interference intensity is defined as the primary response area, mainly including the heat flux nodes around the tooth root; the area adjacent to the primary response area with the second highest thermal wave interference intensity is defined as the secondary response area, including the heat flux nodes in the middle of the tooth; the area with the weakest thermal wave interference intensity is defined as the edge response area, including the heat flux nodes at the tooth tip and tooth side. The periodic thermal excitation applied to the primary response area is generated by a function generator, and the frequency range is set to 0.01 - 1 Hz. A set of response data is recorded every 0.01 Hz within this frequency range. 0.01 Hz corresponds to the slow response process of heat conduction, and 1 Hz corresponds to the fast response process. This frequency range covers the main thermal response characteristics of the heat dissipation tooth-shaped plate.

[0054] The heat conduction delay time is directly related to the aspect ratio of the tooth shape structure. The heat conduction delay time from the primary response area to the secondary response area is calculated by the formula t = h² / α, where h is the tooth height and α is the thermal diffusivity of the material. Taking the heat dissipation tooth-shaped plate with an aspect ratio of 5:1 as an example, when h = 5 mm, the delay time is about 25 ms. Based on this delay time, a thermally excited signal with a phase lag is applied to the secondary response area, and the phase lag angle φ = 2πft, where f is the excitation frequency and t is the delay time. This delay setting considering the structural parameters synchronizes the thermal excitation with the actual heat conduction process.

[0055] The thermal excitation of the edge response area uses a waveform with the same frequency as the primary response area but a phase difference of π. The thermal interference response data between adjacent tooth shapes are collected by an infrared thermal imager, and the sampling frequency is set to 100 Hz. After the thermal excitation stops, the temperature decay process of each response area over time is recorded to obtain the temperature-time curve. The recording time is set to 60 seconds after the thermal excitation stops to ensure capturing the complete temperature decay process. This anti-phase excitation method generates the mutual interference of heat fluxes and enhances the detection sensitivity to welding defects.

[0056] The analysis of the temperature decay curve is divided into the vertical conduction direction and the transverse conduction direction. The thermal decay time constant in the vertical direction is obtained by fitting the exponential function T(t) = T 0 exp(-t / τ), where T 0 represents the initial temperature value at the start of the temperature decay, that is, the temperature peak when the thermal excitation stops; t represents the time variable starting from when the thermal excitation stops, in seconds; τ is the decay time constant, representing the time elapsed when the temperature drops to 1 / e (about 37%) of the initial value, in seconds. The decay time constant between adjacent tooth shapes in the transverse direction is calculated in the same way. The decay time constants at different frequencies are plotted as a frequency response curve to form a resonance thermal decay spectrum. This two-way analysis method comprehensively reflects the spatial characteristics of heat conduction.

[0057] The conduction path mapping of the resonance thermal attenuation spectrum uses a spatial interpolation method. The spectral data of each thermal flow characteristic point is projected onto the actual conduction path, and the spectral correlation between adjacent conduction paths is obtained by calculating the cross-correlation coefficient of the spectra. The spectral correlation reflects the spatial continuity of the heat conduction characteristics. A high correlation indicates good continuity of the welding interface, while a low correlation indicates the presence of welding defects. By reconstructing the spectral correlation data according to the spatial position, a spectral consistency spatial distribution map characterizing the welding quality distribution is generated.

[0058] In one embodiment of the present invention, based on the temperature decay curves of each response region, calculating the thermal decay time constants in the vertical conduction direction and the lateral conduction direction, analyzing the decay characteristics of the heat flow on different conduction paths, and generating a resonance thermal attenuation spectrum, including: Taking a sampling point every 0.2 mm from the tooth root to the tooth tip of the temperature decay curve in the vertical conduction direction, dividing the temperature difference between adjacent sampling points by the time interval to obtain a linearization coefficient, and dividing the region where the linearization coefficient is greater than 0.5 K / s into a rapid decay region and the region where the linearization coefficient is less than 0.5 K / s into a slow decay region according to the tooth profile height, to obtain the decay time constant in the vertical conduction direction; Performing a Fourier transform on the temperature decay curve in the lateral conduction direction, extracting the main frequency component of the energy density spectrum, calculating the energy transfer coefficient between adjacent tooth profiles, dividing the region where the energy transfer coefficient is greater than 0.7 into a strong coupling region and the region where the energy transfer coefficient is less than 0.7 into a weak coupling region, to obtain the decay time constant in the lateral conduction direction; According to the decay time constant in the vertical conduction direction and the decay time constant in the lateral conduction direction, constructing a conduction matrix, calculating the coupling coefficient of each conduction region, and performing minimum thermal resistance optimization on the conduction path to generate a resonance thermal attenuation spectrum.

[0059] The following specifically describes the steps involved in the above embodiment: The temperature decay analysis in the vertical conduction direction uses a high-density sampling method. Using an infrared thermal imager to record a temperature sampling point every 0.2 mm along the direction from the tooth root to the tooth tip. The selection of the sampling point interval is based on the thermal diffusion length of the heat dissipation tooth profile plate, which is defined as the distance that heat propagates in one cycle in the material. Performing a difference calculation on the temperature data of adjacent sampling points, and dividing the temperature difference by the measurement time interval to obtain a linearization coefficient, which characterizes the speed of heat conduction. The linearization coefficient of 0.5 K / s is the critical value for distinguishing the rapid decay region and the slow decay region, and this threshold is determined by analyzing the thermal response characteristics of the heat dissipation tooth profile plate under various welding processes. When the heat conduction is hindered by welding defects, the linearization coefficient will drop sharply.

[0060] The temperature decay analysis in the lateral conduction direction adopts the frequency domain decomposition method. Fourier transform is performed on the temperature decay curve to obtain the energy density spectrum, which reflects the energy distribution of temperature fluctuations in different frequency components. The main frequency component refers to the frequency component with the maximum energy density, and this component carries the most important heat conduction information. The energy transfer coefficient between adjacent tooth profiles is obtained by calculating the ratio of the energies of the main frequency components. The coefficient value of 0.7 is the threshold for distinguishing the strong and weak coupling regions, and this threshold is determined based on the lateral heat conduction efficiency of the heat dissipation tooth profile plate. The strong coupling region indicates good continuity of the welding interface, while the weak coupling region indicates problems with the welding quality.

[0061] The construction of the conduction matrix adopts the grid mapping method. The decay time constants in the vertical and lateral conduction directions are arranged according to the spatial position to form a two-dimensional matrix. The coupling coefficient of the conduction region is obtained by calculating the correlation degree of adjacent matrix elements, and the coupling coefficient reflects the conduction relationship of heat in different directions. The minimum thermal resistance optimization is based on the principle of the minimum thermal resistance of the conduction path, and the main heat conduction channel is determined by finding the path with the minimum thermal resistance. This optimization method conforms to the physical law that heat conduction tends to select the path with the minimum resistance, and the generated resonance heat decay spectrum intuitively shows the heat transfer characteristics of different regions.

[0062] Please continue to refer to Figure 1 , and according to the spectral consistency spatial distribution map and the thermal wave interference defect map, calculate the thermal defect severity index to generate the welding quality evaluation result.

[0063] In an embodiment of the present invention, the calculating the thermal defect severity index according to the spectral consistency spatial distribution map and the thermal wave interference defect map to generate the welding quality evaluation result includes: According to the spectral consistency spatial distribution map, calculate the standard deviation of the spectral consistency between adjacent tooth profiles. Mark the area with a standard deviation greater than 0.3 as the welding non-uniform area, and mark the area with a standard deviation less than 0.1 as the welding stable area to generate the welding uniformity evaluation data; Perform spatial gradient analysis on the phase distortion degree in the thermal wave interference defect map. Mark the area with a gradient value greater than 0.5 rad / mm as the welding interface fracture area, and mark the area with a gradient value between 0.2 - 0.5 rad / mm as the welding interface weak connection area to generate the welding continuity evaluation data; According to the welding uniformity evaluation data and the welding continuity evaluation data, calculate the blocking degree of the welding defect to the heat dissipation channel. Determine the area with a thermal resistance increase greater than 50% as the L4-level defect, the area with a thermal resistance increase greater than 20% and less than 50% as the L3-level defect, the area with a thermal resistance increase greater than 5% and less than 20% as the L2-level defect, and the area with a thermal resistance increase less than 5% as the L1-level defect to obtain the thermal defect severity index; Perform regional grading and labeling on the toothed plate according to the thermal defect severity index to generate the welding quality evaluation result.

[0064] The following specifically describes the steps involved in the above embodiments: The analysis of the spectral consistency spatial distribution map adopts a statistical evaluation method. Symmetrical analysis windows are selected between adjacent tooth profiles, and the window size is set to 5mm×5mm, which is selected based on the tooth profile spacing of the heat dissipation toothed plate. Calculate the standard deviation of the spectral consistency data within each analysis window. The standard deviation characterizes the spatial fluctuation degree of the welding quality. The setting of the standard deviation threshold is based on the statistical analysis of a large number of welding samples: the area where the standard deviation is greater than 0.3 corresponds to the area with severely uneven heat conduction, and this non-uniformity leads to local heat accumulation; the area where the standard deviation is less than 0.1 exhibits stable heat conduction characteristics and corresponds to the area with good welding quality. Through the evaluation of all analysis windows, spatial distribution data reflecting the welding uniformity of the entire heat dissipation toothed plate is generated.

[0065] The gradient analysis of the phase distortion degree adopts a spatial differentiation method. Use the central difference format to calculate the change rate of the phase distortion degree in the x and y directions, and synthesize the total spatial gradient value. The gradient value reflects the spatial change rate of the phase distortion degree, with the unit of radians per millimeter (rad / mm). The selection of the spatial gradient threshold is based on the physical characteristics of the welding interface: a gradient value exceeding 0.5 rad / mm indicates a sharp change in the phase distortion, corresponding to a complete fracture of the welding interface; a gradient value between 0.2 - 0.5 rad / mm indicates an obvious jump in the phase distortion, corresponding to a partial connection of the welding interface. This grading and determination method establishes a quantitative relationship between the phase distortion and the welding continuity.

[0066] The severity evaluation of welding defects adopts a thermal resistance increment analysis method. Convert the welding uniformity evaluation data and continuity evaluation data into equivalent thermal resistance values, and calculate the thermal resistance increment based on the thermal resistance of the intact welding interface. The grading threshold of the thermal resistance increment is set based on the influence degree on the heat dissipation performance: a thermal resistance increase exceeding 50% leads to a serious failure of the heat dissipation function, and it is determined as a L4-level defect; a thermal resistance increase between 20% and 50% significantly affects the heat dissipation efficiency, and it is determined as a L3-level defect; a thermal resistance increase between 5% and 20% slightly affects the heat dissipation performance, and it is determined as a L2-level defect; a thermal resistance increase below 5% has a negligible impact on the heat dissipation, and it is determined as a L1-level defect. This grading standard directly correlates welding defects with the actual heat dissipation performance.

[0067] The area grading annotation adopts a color coding method. On the three-dimensional model of the heat dissipation tooth-shaped plate, different colors are used to identify defect areas of different levels: the L4-level defect area is marked in red, the L3-level defect area is marked in orange, the L2-level defect area is marked in yellow, and the L1-level defect area is marked in green. The color coding visually shows the spatial distribution of the welding quality, facilitating the quick identification of problem areas and the assessment of the overall welding quality. This visual expression method converts complex detection data into intuitive quality assessment results.

[0068] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structural transformation made under the inventive concept of the present invention by using the content of the specification and drawings of the present invention, or any direct / indirect application in other related technical fields, is included in the patent protection scope of the present invention.

Claims

1. A method for detecting the welding quality of a precision heat dissipation toothed plate, characterized in that: include: The tooth tip point, tooth midpoint, tooth root point, tooth valley point and tooth side point on the toothed plate are used as heat flow characteristic points, and a spatially asymmetrically distributed heat pulse sequence is applied to the back of the toothed plate substrate to obtain dynamic thermal response data of the heat flow characteristic points; According to the dynamic thermal response data, the heat flow characteristic points are subjected to time-space decomposition of heat conduction delay, thermal attenuation rate and thermal echo characteristics to generate a thermal response characteristic vector, a thermal response regional distribution map is obtained through spatial clustering analysis, and a thermal disturbance response fingerprint is generated; According to the thermal disturbance response fingerprint, a plurality of heat pulse sources with a preset phase relationship are applied to the back of the substrate to obtain thermal wave interference field distribution data, and a thermal wave interference defect map is generated through phase difference mapping and thermal wave diffraction angle analysis; According to the thermal wave interference defect map, a periodic thermal excitation of a preset frequency is applied to obtain the thermal decay time constant of each heat flow characteristic point, generate a resonant thermal decay spectrum, and construct a spectral consistency spatial distribution map; According to the spectral consistency spatial distribution diagram and the thermal wave interference defect spectrum, the thermal defect severity index is calculated to generate a welding quality evaluation result.

2. The welding quality detection method of the precision heat dissipation toothed plate according to claim 1 is characterized in that: The tooth tip point, tooth midpoint, tooth root point, tooth valley point and tooth side point on the toothed plate are used as heat flow characteristic points, and a spatially asymmetrically distributed heat pulse sequence is applied to the back side of the toothed plate substrate to obtain dynamic thermal response data of the heat flow characteristic points, including: The tooth tip point, tooth midpoint, tooth root point, tooth valley point and tooth side point on the tooth plate are defined as hierarchical heat flow nodes, the tooth root point is defined as a primary heat flow node, the tooth valley point is defined as a secondary heat flow node, the tooth midpoint is defined as a heat flow bifurcation point, the tooth side point is defined as a heat flow coupling point, and the tooth tip point is defined as a heat flow terminal point, so as to construct a multi-level heat flow conduction network; Based on the multi-level heat flow conduction network, the heat pulse sequence is divided into a central excitation sequence and a peripheral excitation sequence, wherein the central excitation sequence corresponds to the first-level heat flow node position, and the peripheral excitation sequence corresponds to the second-level heat flow node position; Applying the central excitation sequence, measuring the heat distribution ratio at the heat flow bifurcation point, measuring the lateral conduction coefficient at the heat flow coupling point, measuring the energy loss rate at the heat flow terminal point, and obtaining a first set of thermal response characteristic data; Applying the peripheral excitation sequence, measuring the heat flow interweaving intensity between the primary heat flow node and the secondary heat flow node, measuring the energy exchange efficiency between the heat flow bifurcation point and the heat flow coupling point, and obtaining a second set of thermal response characteristic data; The overall thermal response characteristics of the heat flow conduction network are calculated according to the first group of thermal response characteristic data and the second group of thermal response characteristic data to obtain the dynamic thermal response data of the heat flow characteristic points.

3. The welding quality detection method of the precision heat dissipation toothed plate according to claim 1 is characterized in that: According to the dynamic thermal response data, the heat flow characteristic points are subjected to time-space decomposition of heat conduction delay, thermal attenuation rate and thermal echo characteristics to generate a thermal response characteristic vector, a thermal response regional distribution map is obtained through spatial clustering analysis, and a thermal disturbance response fingerprint is generated, including: The dynamic thermal response data is segmented in time series into a thermal conduction start-up segment, a thermal conduction stable segment, and a thermal conduction decay segment, and the temperature change curve of each segment is extracted to generate segmented thermal response data; According to the segmented thermal response data, the conduction time from the primary heat flow node to the secondary heat flow node in the heat conduction startup segment is measured, the conduction stability coefficient of the heat flow bifurcation point in the heat conduction stability segment is calculated, and the decay time of the heat flow terminal point in the heat conduction decay segment is measured to generate a conduction delay vector; Performing thermal attenuation rate analysis on the conduction time delay vector and the segmented thermal response data, calculating the initial attenuation rate of the thermal conduction startup segment, the steady-state attenuation rate of the thermal conduction stable segment, and the final-state attenuation rate of the thermal conduction attenuation segment, and generating an attenuation characteristic vector; Performing thermal echo analysis according to the segmented thermal response data, extracting the echo peak value and echo period of the thermal conduction stable segment, calculating the echo attenuation coefficient and echo energy ratio of the thermal conduction attenuation segment, and generating a thermal echo feature vector; The conduction delay vector, attenuation feature vector and thermal echo feature vector are weightedly combined according to the heat flow node level to generate a thermal response feature vector. The thermal response area distribution map is obtained through Mahalanobis distance spatial clustering analysis, and a thermal disturbance response fingerprint is generated.

4. The welding quality detection method of the precision heat dissipation toothed plate according to claim 3 is characterized in that: The step of performing thermal echo analysis according to the segmented thermal response data, extracting the echo peak value and echo period of the thermal conduction stable segment, calculating the echo attenuation coefficient and echo energy ratio of the thermal conduction attenuation segment, and generating a thermal echo feature vector includes: The temperature response curve of each heat flow characteristic point on the toothed plate is subjected to interference noise filtering to obtain a pure thermal response signal; According to the pure thermal response signal, a thermal echo signal in the temperature curve is identified, peak and valley values ​​of the thermal echo signal are extracted, echo amplitude and duration are calculated, and basic echo characteristic data is obtained; Performing cross-correlation analysis on the thermal echo signals of the thermal flow characteristic points, calculating the echo delay difference and energy attenuation ratio between adjacent thermal flow nodes, and obtaining the echo propagation characteristics; According to the echo basic characteristic data and the echo propagation characteristics, the thermal echo reflection path in the toothed structure is reconstructed to generate a thermal echo characteristic vector.

5. The welding quality detection method of the precision heat dissipation toothed plate according to claim 1 is characterized in that: According to the thermal disturbance response fingerprint, a plurality of heat pulse sources with a preset phase relationship are applied to the back of the substrate to obtain thermal wave interference field distribution data, and a thermal wave interference defect map is generated through phase difference mapping and thermal wave diffraction angle analysis, including: According to the thermal disturbance response fingerprint, the phase of the thermal pulse source is configured, and the phase difference of adjacent thermal pulse sources is set to π / 4, π / 2, and 3π / 4 to obtain phase sequences of multiple groups of thermal pulse sources; The amplitude of the thermal wave interference field under each phase sequence is extracted, and the interference intensity ratio between adjacent heat flow nodes is calculated to obtain the intensity distribution of the thermal wave interference field. According to the intensity distribution of the thermal wave interference field, the phase difference of adjacent heat flow nodes is mapped, the relative phase delay of each node is measured, and a phase difference distribution diagram is generated; Perform a diffraction angle analysis on the propagation path of the heat wave in the tooth root area, calculate the diffraction deflection angle at each heat flow node, and obtain the diffraction angle distribution; According to the phase difference distribution diagram and the diffraction angle distribution, the distortion degree of the thermal wave propagation path is calculated to generate a thermal wave interference defect spectrum.

6. The welding quality detection method of the precision heat dissipation toothed plate according to claim 5 is characterized in that: The method maps the phase difference of adjacent heat flow nodes according to the intensity distribution of the thermal wave interference field, measures the relative phase delay of each node, and generates a phase difference distribution diagram, including: According to the heat conduction path, the phase data of the adjacent heat flow nodes are subtracted from the initial phase value of the reference heat pulse source, and the phase deviation is limited to the range of -π to π to obtain the reference phase distribution data; Calculate the phase difference of adjacent measurement points at intervals of 0.5 mm in the tooth root region in the reference phase distribution data to obtain a phase change curve at the tooth root, calculate the phase difference of adjacent measurement points at intervals of 0.2 mm in the tooth tip region to obtain a phase change curve at the tooth tip, calculate the slopes of the phase change curve at the tooth root and the phase change curve at the tooth tip respectively, and obtain the phase change rate in the vertical propagation direction; According to the reference phase distribution data, the lateral measurement points between adjacent tooth shapes are paired in pairs according to the principle of central symmetry, the phase difference value of each pair of measurement points is calculated, and the phase jump value at the welding interface is determined by the mutation position and mutation amplitude of the phase difference, so as to obtain the phase distortion degree in the lateral propagation direction; The phase change rate and the phase distortion are normalized, the weight coefficient in the vertical propagation direction is set to 0.6, and the weight coefficient in the lateral propagation direction is set to 0.4, and the phase difference distribution diagram is generated.

7. The welding quality detection method of the precision heat dissipation toothed plate according to claim 1 is characterized in that: According to the thermal wave interference defect map, a periodic thermal excitation of a preset frequency is applied to obtain the thermal decay time constant of each heat flow characteristic point, generate a resonant thermal decay spectrum, and construct a spectrum consistency spatial distribution map, including: According to the thermal wave interference defect map, the heat flow characteristic points are divided into the primary response area, the secondary response area and the edge response area according to the heat conduction level. Periodic thermal excitation in the range of 0.01-1Hz is applied to the primary response area to obtain the main frequency response data. According to the aspect ratio of the tooth-shaped structure, the heat conduction delay time from the primary response area to the secondary response area is calculated, and a phase-delayed periodic thermal excitation is applied to the secondary response area according to the heat conduction delay time to obtain secondary frequency response data; Apply thermal excitation of the same frequency but opposite phase to that of the primary response area to the edge response area, obtain thermal interference response data between adjacent tooth shapes, and measure the temperature attenuation curve of each response area after the thermal excitation stops; Based on the temperature decay curve of each response area, the thermal decay time constants in the vertical and lateral conduction directions are calculated, the attenuation characteristics of heat flow in different conduction paths are analyzed, and the resonant thermal decay spectrum is generated; The resonant thermal attenuation spectrum of each heat flow characteristic point is mapped for conduction paths, the spectrum correlation between adjacent conduction paths is calculated, and a spectral consistency spatial distribution map is constructed.

8. The welding quality detection method of the precision heat dissipation toothed plate according to claim 7 is characterized in that: The method calculates the thermal decay time constants in the vertical and lateral conduction directions based on the temperature decay curves of each response area, analyzes the attenuation characteristics of the heat flow in different conduction paths, and generates a resonant thermal decay spectrum, including: For the temperature decay curve in the vertical conduction direction, a sampling point is taken at every 0.2 mm interval from the tooth root to the tooth tip, and the temperature difference between adjacent sampling points is divided by the time interval to obtain the linearization coefficient. According to the tooth height, the area with a linearization coefficient greater than 0.5 K / s is divided into a fast decay area, and the area with a linearization coefficient less than 0.5 K / s is divided into a slow decay area, and the decay time constant in the vertical conduction direction is obtained; The temperature decay curve in the transverse conduction direction is Fourier transformed to extract the main frequency component of the energy density spectrum, and the energy transfer coefficient between adjacent tooth shapes is calculated. The area with an energy transfer coefficient greater than 0.7 is divided into a strong coupling area, and the area with an energy transfer coefficient less than 0.7 is divided into a weak coupling area, and the decay time constant in the transverse conduction direction is obtained; According to the attenuation time constant in the vertical conduction direction and the attenuation time constant in the lateral conduction direction, a conduction matrix is ​​constructed, the coupling coefficient of each conduction area is calculated, the conduction path is optimized for minimum thermal resistance, and a resonant thermal attenuation spectrum is generated.

9. The welding quality detection method of the precision heat dissipation toothed plate according to claim 1 is characterized in that: The step of calculating the thermal defect severity index based on the spectral consistency spatial distribution map and the thermal wave interference defect map to generate a welding quality evaluation result includes: According to the spectral consistency spatial distribution diagram, the standard deviation of the spectral consistency between adjacent tooth shapes is calculated, and the area with a standard deviation greater than 0.3 is marked as a welding uneven area, and the area with a standard deviation less than 0.1 is marked as a welding stable area, so as to generate welding uniformity evaluation data; Performing spatial gradient analysis on the phase distortion in the thermal wave interference defect map, marking the area with a gradient value greater than 0.5 rad / mm as a welding interface fracture area, and marking the area with a gradient value between 0.2-0.5 rad / mm as a welding interface weak connection area, to generate welding continuity evaluation data; According to the welding uniformity evaluation data and the welding continuity evaluation data, the degree of blocking of the heat dissipation channel by the welding defects is calculated, and the area where the thermal resistance increases by more than 50% is determined as an L4 defect, the area where the thermal resistance increases by more than 20% and less than 50% is determined as an L3 defect, the area where the thermal resistance increases by more than 5% and less than 20% is determined as an L2 defect, and the area where the thermal resistance increases by less than 5% is determined as an L1 defect, and the thermal defect severity index is obtained; The toothed plate is regionally graded and labeled according to the thermal defect severity index to generate a welding quality assessment result.

Images