Appearance visual detection method and device for electronic control component of new energy automobile

Abstract

The invention discloses an appearance visual detection method and device for an electric control part of a new energy automobile. Structural stress distribution data of the electric control component are obtained and subjected to partition processing, and the surface of the electric control component is divided into a stress sensitive area and a conventional area; carrying out multi-wavelength spectrum scanning and polarization state detection on the stress sensitive area by adopting self-adaptive path planning, synchronously acquiring a polarization characteristic matrix, dual-wavelength interference data and depth spectrum data of the determined key attention area, and carrying out time domain synchronization and space registration processing to generate multi-dimensional detection characteristic data; extracting an optical anisotropy coefficient, a microcosmic deformation tensor and an interface reflection feature vector, and determining layering defect feature parameters in combination with multiple threshold judgment; and calculating local curvature distribution, an interface shear stress field and thermal stress distribution, and performing quantitative matching with the material characteristic parameters to generate quantitative evaluation indexes of the layering defects. According to the invention, accurate identification and rapid positioning of microscopic layering defects in the multi-layer composite insulation structure of the electric control component are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of visual appearance inspection of electric control components for new energy vehicles, and in particular to a method and device for visual appearance inspection of electric control components for new energy vehicles. Background Art

[0002] Electronic control components for new energy vehicles refer to key power electronic devices in the high-voltage drive system of new energy vehicles, mainly including power conversion units such as DC converters, inverters, and on-board chargers. These electronic control components adopt a unique multi-layer composite insulation structure design, usually consisting of an inner ceramic substrate, an intermediate organic insulation layer, and an outer nano-modified protective layer. Among them, the ceramic substrate provides basic mechanical strength and heat dissipation performance, the organic insulation layer ensures high-voltage electrical insulation performance, and the nano-modified protective layer provides environmental protection and stress buffering functions. This multi-layer composite structure design is designed to meet the reliability requirements of new energy vehicle electronic control components in high-voltage and high-current working environments.

[0003] However, in the manufacturing process of electronic control components, due to the differences in material properties between the ceramic substrate, organic insulation layer and nano-protective layer, especially the mismatch of thermal expansion coefficients during injection molding and curing, micro-scale delamination or bubbles are easily generated between the interfaces. Although this interface delamination defect is small in size at the initial stage, it is very easy to expand and eventually lead to insulation failure in the working environment where the electronic control components are subjected to high voltage electric fields and temperature cycles for a long time. Traditional appearance visual inspection methods mainly rely on direct imaging of diffuse reflection light sources. Since the insulation layer has a certain degree of transparency and the refractive index difference in the delamination area is small, it is difficult to accurately identify this potential insulation failure hazard. Especially in the contour transition area and curved surface of the component, the missed detection rate is significantly improved, which poses a potential threat to the safe operation of new energy vehicles. Therefore, it is of great significance to develop an appearance visual inspection method that can accurately identify and locate interface delamination defects in the multilayer composite insulation structure of electronic control components. Summary of the invention

[0004] The main purpose of the present invention is to solve the technical problem that the existing visual inspection method for the appearance of electric control components for new energy vehicles cannot accurately identify microscopic stratification defects in multi-layer composite insulation structures.

[0005] A first aspect of the present invention provides a method for visually inspecting the appearance of an electronic control component for a new energy vehicle, the method comprising: Acquire structural stress distribution data of the electric control component, and perform zoning processing on the electric control component according to the structural stress distribution data to obtain zoning data of stress sensitive areas and conventional areas, wherein the stress sensitive areas include the periphery of the power device mounting area, the heat sink contact area, the connector lead-out area, and the periphery of the fixing screw hole; Based on the partition data, adaptive path planning is used to perform multi-wavelength spectral scanning and polarization state detection on the stress sensitive area to obtain spectral characteristic parameters and polarization state changes, and a key focus area is determined according to the degree of deviation between the spectral characteristic parameters and the polarization state changes and a preset threshold value; For the key focus area, synchronously collect polarization feature matrix, dual-wavelength interference data and depth spectrum data, perform time domain synchronization and spatial registration processing on the polarization feature matrix, dual-wavelength interference data and depth spectrum data, and generate multi-dimensional detection feature data; Extracting an optical anisotropy coefficient, a microscopic deformation tensor and an interface reflection characteristic vector according to the multidimensional detection characteristic data, and determining a delamination defect characteristic parameter when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection characteristic vector exceeds a third preset threshold; According to the characteristic parameters of the delamination defect, the local curvature distribution, the interface shear stress field and the thermal stress distribution are calculated, and the local curvature distribution, the interface shear stress field and the thermal stress distribution are quantitatively matched with the material characteristic parameters of the interface type where the defect is located to generate a quantitative evaluation index of the delamination defect.

[0006] Optionally, the acquiring of structural stress distribution data of the electric control component, and performing zoning processing on the electric control component according to the structural stress distribution data to obtain zoning data of stress sensitive areas and conventional areas, includes: Collect the surface reflection images of the electronic control components at four polarization states of 0°, 45°, 90°, and 135°, and extract the polarization degree distribution data; A dual-wavelength laser with wavelengths of 850nm and 1050nm is used to perform periodic scanning heating on the electronic control components, and the surface temperature response sequence is collected during the heating-cooling cycle; According to the surface temperature response sequence, the temperature response time constant and phase delay of each measuring point are calculated to obtain the interface thermal diffusion characteristic function; The polarization degree distribution data is spatially mapped and fused with the interface thermal diffusion characteristic function to generate depth-resolved stress distribution data; According to the depth-resolved stress distribution data, adaptive threshold segmentation is performed in combination with the interface characteristics of the ceramic substrate, the organic insulating layer and the nano-protective layer to determine the partition data of the stress-sensitive area and the conventional area.

[0007] Optionally, collecting the surface reflection images of the electric control component at four polarization states of 0°, 45°, 90° and 135° to extract polarization degree distribution data includes: According to the layout of power devices of the electronic control component, differentiated light source incident angles are set around the power device installation area, the heat sink contact area, the connector lead-out area and the fixing screw hole to generate an incident light spot density distribution map; In each spot area, the incident light is polarized using a liquid crystal tunable polarizer to collect high dynamic range reflection images at four polarization states: 0°, 45°, 90°, and 135°. Performing local contrast enhancement on the high dynamic range reflection image to extract polarized reflection intensity data of the ceramic substrate surface, the organic insulating layer surface and the nano-protective layer surface; Calculating the Stokes parameters and polarization coefficient of each surface according to the polarized reflection intensity data, and establishing a surface-interface polarization characteristic mapping relationship; Based on the surface-interface polarization characteristic mapping relationship and in combination with the birefringence effect parameters of the interface material, the stress-induced birefringence component at the interface is calculated to generate polarization degree distribution data.

[0008] Optionally, based on the partition data, using adaptive path planning to perform multi-wavelength spectral scanning and polarization state detection on the stress sensitive area to obtain spectral characteristic parameters and polarization state changes includes: Performing stress gradient calculation on the stress sensitive area, and generating a scanning grid of dynamic density according to the stress gradient value; Using a wavelength tunable laser, continuously scan each node in the scanning grid in the 850nm-1050nm band, and record the spectral reflection intensity curve as a spectral characteristic parameter; Applying electrically controlled liquid crystal polarization modulation at each scanning node to obtain depth-resolved light intensity data in different polarization states; The light intensity-polarization tensor including the ceramic-organic layer interface and the organic layer-protective layer interface is constructed according to the depth-resolved light intensity data, and the lattice orientation parameters and the interface stress birefringence coefficient of each interface are calculated as the polarization state change.

[0009] Optionally, determining the focus area according to the degree of deviation between the spectral characteristic parameter and the polarization state change amount and a preset threshold value includes: Constructing a principal component analysis matrix of the spectral characteristic parameters, performing singular value decomposition on the principal component analysis matrix, and extracting a characteristic frequency band response of the 850nm-1050nm band; According to the light intensity attenuation law of the characteristic frequency band response at different depths, the position parameters of the ceramic-organic layer interface and the organic layer-protective layer interface are calculated to generate interface depth profile data; Establishing a photoelastic tensor matrix for the polarization state variation, and calculating the interface stress response based on a temperature-stress coupling model of the electric control component; Performing spatial position alignment and feature quantity normalization processing on the interface depth profile data and the interface stress response to generate a defect sensitivity evaluation index; According to the defect sensitivity evaluation index, threshold parameters related to the layering type are set, the stress sensitive area is segmented into layers, and the area where the defect sensitivity exceeds the corresponding threshold is marked as a key focus area.

[0010] Optionally, for the focus area, synchronously collecting polarization characteristic matrix, dual-wavelength interference data and deep spectrum data includes: Performing interface layering positioning on the key focus area, and setting polarization modulation parameters, interference wavelength parameters and spectrum sampling parameters according to the material properties of the ceramic substrate, the organic insulating layer and the nano protective layer; The incident light is polarized and the polarization state is modulated. Polarization reflection data is collected at intervals of 15° in the range of 0°-180° to construct a polarization characteristic matrix. Applying dual-wavelength phase modulated light to the focus area, measuring the phase difference between the ceramic-organic layer interface and the organic layer-protective layer interface, and obtaining dual-wavelength interference data; Confocal microscopy technology is used to scan different depth levels layer by layer to obtain the spectral response characteristics of the ceramic substrate surface, ceramic-organic layer interface, organic layer-protective layer interface and protective layer surface, and generate depth-resolved interface spectral data.

[0011] Optionally, the polarization feature matrix, dual-wavelength interference data and depth spectrum data are subjected to time domain synchronization and spatial registration processing to generate multi-dimensional detection feature data, including: The polarized reflection data is transformed by Stokes vector to calculate the polarization degree and polarization state angle of each interface, and generate an interface polarization characteristic diagram; According to the phase difference distribution of the dual-wavelength interference data, the optical path difference variation between the ceramic-organic layer interface and the organic layer-protective layer interface is calculated to obtain interface morphology data; Performing dispersion characteristic analysis on the deep spectral data, extracting the reflectivity spectrum and absorptivity spectrum of each interface in the 850nm-1050nm band, and building a characteristic spectrum database in combination with the optical dispersion characteristics of the interface material; The interface polarization characteristic map, interface morphology data and characteristic spectrum database are aligned layer by layer to establish a hierarchical characteristic mapping relationship; According to the hierarchical feature mapping relationship, an interface stress-optical response model is used to fuse and reconstruct multi-dimensional features to generate multi-dimensional detection feature data.

[0012] Optionally, extracting an optical anisotropy coefficient, a microscopic deformation tensor and an interface reflection feature vector according to the multidimensional detection feature data, and determining a delamination defect feature parameter when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection feature vector exceeds a third preset threshold, comprises: Constructing an interface Mueller matrix according to the multi-dimensional detection characteristic data, performing characteristic decomposition on the polarization state changes at different interfaces, and calculating the optical anisotropy coefficient of each layer interface; Based on the dual-wavelength interference phase distribution of the multi-dimensional detection characteristic data, a two-dimensional phase gradient algorithm is used to calculate the in-plane strain and out-of-plane displacement of the interface, and a microscopic deformation tensor is established through the strain-displacement relationship; Performing wavelength domain analysis on the multi-dimensional detection characteristic data to extract the interface reflection characteristic vectors of the ceramic-organic layer and the organic layer-protective layer respectively; According to the operating temperature range and voltage level of the electronic control component, a first preset threshold value of the optical anisotropy coefficient, a second preset threshold value of the microscopic deformation tensor, and a third preset threshold value of the interface reflection characteristic vector are set; When the detection parameters exceed the corresponding preset thresholds, the depth position, geometric dimensions and interface type information of the defect area are extracted to determine the characteristic parameters of the delamination defect.

[0013] Optionally, the local curvature distribution, interface shear stress field and thermal stress distribution are calculated according to the characteristic parameters of the delamination defect, and the local curvature distribution, interface shear stress field and thermal stress distribution are quantitatively matched with the material characteristic parameters of the interface type where the defect is located to generate a quantitative evaluation index of the delamination defect, including: According to the depth position and geometric size in the delamination defect characteristic parameters, a Gaussian curvature operator is used to calculate the local deformation of the interface to generate local curvature distribution data; Based on the interface type information and in combination with the difference in elastic modulus of the ceramic substrate, the organic insulating layer and the nano protective layer, an interface stress transfer function is constructed to calculate the interface shear stress field; Use infrared thermal imaging array to scan the temperature field of the surface of the electronic control components in real time, extract the temperature gradient data around the power device and the heat sink contact area, and calculate the thermal stress distribution; Performing tensor superposition on the local curvature distribution data, the interface shear stress field and the thermal stress field data to obtain a comprehensive interface stress state; According to the critical stress thresholds and fatigue life parameters of different interface types, the comprehensive stress state of the interface is normalized to generate a quantitative evaluation index of the delamination defect.

[0014] The second aspect of the present invention provides a visual inspection device for the appearance of an electric control component for a new energy vehicle, comprising: A stress partitioning module is used to obtain structural stress distribution data of the electric control component, and partition the electric control component according to the structural stress distribution data to obtain partition data of stress sensitive areas and conventional areas, wherein the stress sensitive area includes the periphery of the power device installation area, the heat sink contact area, the connector lead-out area and the periphery of the fixing screw hole; A feature detection module, for performing multi-wavelength spectral scanning and polarization state detection on the stress sensitive area based on the partition data by using adaptive path planning, obtaining spectral feature parameters and polarization state variation, and determining a key focus area according to the degree of deviation between the spectral feature parameters and polarization state variation and a preset threshold; A data acquisition module is used to synchronously collect polarization feature matrix, dual-wavelength interference data and depth spectrum data for the key focus area, perform time domain synchronization and spatial registration processing on the polarization feature matrix, dual-wavelength interference data and depth spectrum data, and generate multi-dimensional detection feature data; a defect recognition module, for extracting an optical anisotropy coefficient, a microscopic deformation tensor and an interface reflection feature vector according to the multidimensional detection feature data, and determining a delamination defect feature parameter when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection feature vector exceeds a third preset threshold; The evaluation processing module is used to calculate the local curvature distribution, interface shear stress field and thermal stress distribution according to the characteristic parameters of the delamination defect, quantitatively match the local curvature distribution, interface shear stress field and thermal stress distribution with the material characteristic parameters of the interface type where the defect is located, and generate a quantitative evaluation index for the delamination defect.

[0015] The technical solution provided by the embodiment of the present application has at least the following advantages: Firstly, by acquiring the structural stress distribution data of the electronic control components and performing zoning processing, the accurate division of stress-sensitive areas such as the periphery of the power device installation area and the heat sink contact area is achieved, effectively avoiding the problem of low detection efficiency caused by full-area scanning; secondly, adaptive path planning is used to perform multi-wavelength spectral scanning and polarization state detection on the stress-sensitive area, and the detection sensitivity of tiny delamination defects is improved through dual feature analysis of spectral feature parameters and polarization state changes; thirdly, the polarization feature matrix, dual-wavelength interference data and deep spectrum data are synchronously collected for the key focus area, and time domain synchronization and spatial registration processing are performed to achieve accurate extraction of delamination defect characteristics of different interface layers; then, by extracting the optical anisotropy coefficient, microscopic deformation tensor and interface reflection feature vector, and setting a multiple threshold judgment mechanism, the false alarm rate of detection is significantly reduced; finally, by calculating the local curvature distribution, interface shear stress field and thermal stress distribution, and quantitatively matching them with material characteristic parameters, a correlation analysis method between delamination defect characteristics and stress distribution is established, providing a reliable basis for evaluating the degree of damage of defects. This solution not only solves the problem of low sensitivity of traditional detection methods in identifying microscopic stratification defects, but also realizes high-reliability detection of contour transition areas and curved surface areas, providing technical guarantee for ensuring the insulation safety of electronic control components of new energy vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0017] Figure 1 A schematic diagram of an embodiment of a method for visually inspecting the appearance of an electric control component for a new energy vehicle according to an embodiment of the present invention; Figure 2 It is a schematic diagram of an embodiment of a device for visually inspecting the appearance of an electronic control component for a new energy vehicle in an embodiment of the present invention.

[0018] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0021] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, "and / or" in the full text includes three solutions. Taking A and / or B as an example, it includes technical solution A, technical solution B, and technical solution that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, which must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0022] An embodiment of the present application provides a method for visually inspecting the appearance of an electronic control component for a new energy vehicle. Figure 1 A flow chart of a method for visual inspection of the appearance of electric control components for new energy vehicles provided in one embodiment of the present application. In this embodiment, the method includes: See also Figure 1 , obtaining structural stress distribution data of the electric control component, partitioning the electric control component according to the structural stress distribution data, and obtaining partition data of the stress sensitive area and the conventional area, wherein the stress sensitive area includes the periphery of the power device installation area, the heat sink contact area, the connector lead-out area and the periphery of the fixing screw hole; In one embodiment of the present invention, the acquiring of structural stress distribution data of an electric control component, zoning the electric control component according to the structural stress distribution data, and obtaining zoning data of a stress sensitive area and a conventional area include: collecting surface reflection images of the electric control component at four polarization states of 0°, 45°, 90°, and 135°, and extracting polarization distribution data; using a dual-wavelength laser with wavelengths of 850nm and 1050nm to perform periodic scanning heating on the electric control component, and collecting a surface temperature response sequence during a heating-cooling cycle; calculating the temperature response time constant and phase delay of each measuring point according to the surface temperature response sequence, and obtaining an interface heat diffusion characteristic function; spatially mapping and fusing the polarization distribution data with the interface heat diffusion characteristic function to generate depth-resolved stress distribution data; and performing adaptive threshold segmentation according to the depth-resolved stress distribution data in combination with the interface characteristics of the ceramic substrate, the organic insulating layer, and the nano-protective layer to determine the zoning data of the stress sensitive area and the conventional area.

[0023] Specifically, in the multi-layer composite insulation structure of the electric control components of new energy vehicles, the detection of delamination defects needs to be combined with optical and thermal property analysis. First, the polarization state is detected, and the incident light is modulated using a liquid crystal polarizer, and four angles of 0°, 45°, 90°, and 135° are set respectively. The selection of these four angles can fully describe the polarization response characteristics of the material: 0° and 90° reflect the stress distribution in the main axis direction, and 45° and 135° are used to detect the influence of shear stress. During measurement, the power of the detection light source is set to 100 milliwatts, and the incident angle is controlled at 30 degrees to ensure sufficient reflection signal intensity. For example, when detecting the surface of the ceramic substrate in the IGBT module, a reflection image with a resolution of 200×200 pixels is collected at each polarization angle, and the physical spacing between adjacent pixels is 50 microns. Such a spatial resolution is sufficient to capture the surface stress distribution characteristics at the micron level.

[0024] When performing thermal property analysis, a dual-wavelength laser is used for scanning heating, where the 850-nanometer laser power is set to 200 milliwatts, mainly acting on the surface layer within a depth range of 300 microns; the 1050-nanometer laser power is set to 300 milliwatts, which can reach a detection depth of 800 microns. The scanning process adopts a grid path, and the spacing between adjacent scanning lines is set to 1 mm. At each scanning point, a 2-second heating-cooling cycle is performed: the laser irradiates for 1 second for heating, and then the laser is turned off for 1 second for natural cooling. For example, when detecting the insulating substrate of the DC converter, the sampling frequency of the infrared thermal imager is set to 50 Hz and the temperature resolution is 0.1 degrees Celsius, so that the temperature change characteristics at the interface can be accurately captured.

[0025] When analyzing the collected temperature response sequence, first perform exponential fitting on the temperature curve of each measuring point to extract the temperature response time constant. The typical time constant reflects the diffusion rate of heat in the material. For example, in the ceramic substrate area, the time constant is usually in the range of 0.5-1 second; in the organic insulating layer area, due to poor thermal conductivity, the time constant can reach 2-3 seconds. When the temperature response time constant difference between adjacent measuring points exceeds 0.5 seconds, it indicates the existence of a material interface, based on which the ceramic-organic layer interface position (about 300 microns) and the organic layer-protective layer interface position (about 500 microns) can be determined. At the same time, the phase delay of the temperature response is calculated. The phase delay of a normal interface is in the range of 15-20 degrees. If stratification occurs, the phase delay will increase sharply to 30-40 degrees.

[0026] The data fusion process first establishes a spatial correspondence. The 100×100 mm detection area is divided into 1000×1000 data points, and the polarization data and thermal characteristic data are mapped to this unified grid. Each grid point stores a data structure containing the polarization degree value (range 0-1) and the thermal diffusion characteristic value (time constant and phase delay). For example, the polarization degree of a certain measurement point is 0.8, the time constant is 1.5 seconds, and the phase delay is 25 degrees. These data together indicate that there may be an interface abnormality at that location.

[0027] When the area is finally segmented, the threshold parameters are set according to the material properties. The elastic modulus of the ceramic substrate is about 300GPa, and the thermal conductivity is 40W / (m·K); the elastic modulus of the organic insulating layer is about 3GPa, and the thermal conductivity is 0.2W / (m·K); the elastic modulus of the organic insulating layer is about 3GPa, and the thermal conductivity is 0.2W / (m·K); the elastic modulus of the nano protective layer is about 10GPa, and the thermal conductivity is 1W / (m·K). Based on these parameters, the partition threshold is set: when the comprehensive stress index of the measuring point (calculated by polarization degree and thermal characteristics) exceeds 70% of the ultimate strength of the material, it is divided into a stress-sensitive area; when the index is less than 30%, it is divided into a conventional area. Taking the IGBT module as an example, stress-sensitive areas usually appear within 2-3 mm around the power device, and these areas are often the locations where delamination defects are prone to occur.

[0028] In one embodiment of the present invention, the collecting of surface reflection images of the electric control component at four polarization states of 0°, 45°, 90°, and 135°, and extracting polarization distribution data, include: according to the layout of power devices of the electric control component, setting differentiated light source incident angles around the power device installation area, the heat sink contact area, the connector lead-out area, and the fixing screw hole, to generate an incident light spot density distribution diagram; in each light spot area, using a liquid crystal tunable polarizer to modulate the polarization state of the incident light, and collecting high dynamic range reflection images at four polarization states of 0°, 45°, 90°, and 135°; performing local contrast enhancement on the high dynamic range reflection image, and extracting polarization reflection intensity data of the ceramic substrate surface, the organic insulating layer surface, and the nano protective layer surface; according to the polarization reflection intensity data, calculating the Stokes parameters and polarization coefficients of each surface, and establishing a surface-interface polarization characteristic mapping relationship; based on the surface-interface polarization characteristic mapping relationship, combined with the birefringence effect parameters of the interface material, calculating the stress-induced birefringence component at the interface, and generating polarization distribution data.

[0029] Specifically, in the detection of electronic control components, there are significant differences in the characteristics of interface stress distribution in different functional areas. According to the typical structure of DC converters and inverters, the incident angle around the power device installation area is set to 45 degrees. This is because the stress gradient in this area is the largest, and 45-degree incidence can obtain the best reflection signal; the radiator contact area is set to a 30-degree incident angle to avoid shading by the heat sink; the connector lead-out area is set to a 60-degree incident angle to ensure that the stress distribution in the pin transition area can be detected; vertical incidence is used around the fixing screw hole to reduce the shadow effect at the edge of the screw hole. The spot size is controlled at 2 mm in diameter, and the overlap rate of adjacent spots is set to 30%. For example, in a 100×100 mm detection area, about 2,500 detection spots are formed to ensure that there are no detection blind spots.

[0030] In each spot area, a liquid crystal polarizer is used to modulate the polarization state of the incident light. The device modulates the incident light into four polarization states of 0°, 45°, 90°, and 135° by applying a control voltage of 0-5 volts to the liquid crystal molecules. The rise time of the control voltage is set to 10 milliseconds to ensure that the liquid crystal molecules respond fully. In each polarization state, an industrial camera with a 14-bit dynamic range is used to capture the reflected image, the exposure time is set to 1 millisecond, and the gain is set to 6 decibels. These parameter settings ensure that the image is neither oversaturated nor the signal is too weak. For example, for the detection of IGBT modules, a typical reflected image grayscale value range is between 3000-14000.

[0031] When performing local contrast enhancement on the collected reflection image, a 5×5 pixel processing window is used, and histogram equalization is performed within the window. In the enhanced image, the typical reflection intensity of the ceramic substrate surface is in the range of 12000-14000, the organic insulating layer surface is in the range of 8000-10000, and the nano-protective layer surface is in the range of 4000-6000. This difference is mainly due to the different reflectivity of the three layers of materials: about 85% for the ceramic substrate, about 60% for the organic layer, and about 30% for the nano-layer.

[0032] Based on the enhanced image, the polarization characteristics of each detection point are calculated. The Stokes parameters S0 (total intensity), S1 (intensity difference between 0° and 90°), and S2 (intensity difference between 45° and 135°) are calculated through the reflection intensity values ​​under four polarization states. For example, the reflection intensity of a certain detection point under four polarization states is 12000, 8000, 10000, and 9000, respectively. The calculated polarization degree is 0.75, indicating that there is a strong stress-induced birefringence effect at this location. The polarization degree in a normal area is usually less than 0.3. If it exceeds this value, it means that stress concentration may exist at this location.

[0033] Finally, combined with the birefringence effect parameters of the material (the stress-optical coefficient of the ceramic substrate is 3.5× , the organic insulating layer is 2.8× , the nano protective layer is 1.2× ), calculate the stress distribution. When a material is subjected to stress, its birefringence effect is linearly related to the stress magnitude. For example, at the edge of the IGBT module, if the measured birefringence component is 0.008, the stress at this location can be deduced to be about 200 MPa based on the stress optical coefficient, which far exceeds the safety threshold of 100 MPa for the organic insulating layer, indicating that delamination is very likely to occur there.

[0034] Please continue reading Figure 1 , based on the partition data, adopt adaptive path planning to perform multi-wavelength spectral scanning and polarization state detection on the stress sensitive area, obtain spectral characteristic parameters and polarization state change, and determine the key focus area according to the degree of deviation between the spectral characteristic parameters and polarization state change and a preset threshold; In one embodiment of the present invention, based on the partition data, adaptive path planning is used to perform multi-wavelength spectral scanning and polarization state detection on the stress sensitive area to obtain spectral characteristic parameters and polarization state changes, including: performing stress gradient calculation on the stress sensitive area, and generating a scanning grid with dynamic density according to the stress gradient value; using a wavelength tunable laser to continuously scan each node in the scanning grid in the 850nm-1050nm band, and recording the spectral reflection intensity curve as a spectral characteristic parameter; applying electrically controlled liquid crystal polarization modulation at each scanning node to obtain depth-resolved light intensity data under different polarization states; constructing a light intensity-polarization tensor including a ceramic-organic layer interface and an organic layer-protective layer interface according to the depth-resolved light intensity data, and calculating the lattice orientation parameters and the interface stress birefringence coefficient of each interface as the polarization state change.

[0035] Specifically, in the stress sensitive area detection of the electronic control component, it is first necessary to calculate the stress gradient according to the stress distribution data obtained in the previous step. The calculation process uses a spatial differential algorithm, with each pixel as the center, to calculate the stress difference between it and the adjacent 8 points. For example, in a detection area of ​​100×100 mm, if the stress value of a point is 150 MPa and the stress value of its adjacent point is 100 MPa, the stress gradient of the point is 50 MPa / mm. The scanning density is dynamically adjusted according to the magnitude of the stress gradient: in areas where the gradient is greater than 30 MPa / mm, the scanning point spacing is set to 0.5 mm; in areas where the gradient is between 10-30 MPa / mm, the spacing is set to 1 mm; in areas where the gradient is less than 10 MPa / mm, the spacing is set to 2 mm. In this way, higher density detection data can be obtained in areas where stress changes drastically, such as the edge of the power device of the IGBT module.

[0036] When using a wavelength tunable laser to perform spectral scanning, the wavelength range of the laser is set between 850-1050 nanometers, with a step interval of 10 nanometers, so that data at 21 wavelength points can be obtained at each scanning point. The output power of the laser is set to 200 milliwatts, and the spot diameter is controlled at 100 microns. Taking the detection of the DC converter as an example, the laser stays at each wavelength point for 20 milliseconds during the scanning process, and the reflected light intensity is recorded by the photodetector. The reflectivity of a normal ceramic substrate is about 85% at 850 nanometers, which gradually decreases to 75% with increasing wavelength; the reflectivity of the organic insulating layer decreases from 60% to 45%; and the nano protective layer decreases from 40% to 25%. If there is stratification at the interface, reflectivity anomalies will occur at specific wavelengths, such as a sudden change in reflectivity at 950 nanometers.

[0037] At each scanning node, an electrically controlled liquid crystal polarization modulator is used to modulate the incident light. The response time of the modulator is set to 5 milliseconds, and the control voltage range is 0-10 volts. Different polarization states are achieved by changing the applied voltage. At each wavelength point, the reflected light intensity at four polarization states of 0°, 45°, 90°, and 135° is collected. For example, at the ceramic-organic layer interface, if the reflected light intensity of the four polarization states is 100, 80, 60, and 85 (unit: microwatts), this uneven distribution indicates that there is stress on the interface; while the difference in the reflected light intensity of the four polarization states of a normal interface is usually less than 10%.

[0038] Based on the measured depth-resolved light intensity data, the light intensity-polarization tensor describing the interface characteristics is established. For the ceramic-organic layer interface, the lattice orientation parameter is first calculated. This parameter reflects the material structure changes caused by interface stress. The calculation method is: substitute the light intensity data of the four polarization states into the lattice orientation formula. The orientation parameter of a normal interface is in the range of 0.1-0.3. If it exceeds 0.5, it indicates that the interface stress is significant. At the same time, the interface stress birefringence coefficient is calculated. This coefficient directly reflects the stress magnitude. For example, in an IGBT module, if the birefringence coefficient of a certain point is measured to be 3× , the corresponding interface stress is about 150 MPa, which is close to the bearing limit of the organic insulating layer. The same method is used to analyze the interface between the organic layer and the protective layer, but due to different material properties, its critical value needs to be adjusted accordingly: the warning value of the orientation parameter is 0.4, and the warning value of the birefringence index is 2× .

[0039] In one embodiment of the present invention, determining the key focus area according to the degree of deviation of the spectral characteristic parameters and the polarization state change amount from a preset threshold includes: constructing a principal component analysis matrix of the spectral characteristic parameters, performing singular value decomposition on the principal component analysis matrix, and extracting the characteristic frequency band response of the 850nm-1050nm band; calculating the position parameters of the ceramic-organic layer interface and the organic layer-protective layer interface according to the light intensity attenuation law of the characteristic frequency band response at different depths, and generating interface depth profile data; establishing a photoelastic tensor matrix for the polarization state change amount, and calculating the interface stress response based on the temperature-stress coupling model of the electronically controlled component; spatially aligning the interface depth profile data with the interface stress response and normalizing the characteristic amount to generate a defect sensitivity evaluation index; setting a threshold parameter related to the layering type according to the defect sensitivity evaluation index, hierarchically segmenting the stress sensitive area, and marking the area where the defect sensitivity exceeds the corresponding threshold as the key focus area.

[0040] Specifically, when the principal component analysis is performed on the spectral feature data of the electronic control component, the reflection intensity values ​​of the 21 wavelength points of each measuring point in the 850-1050 nanometer band are first formed into a spectral matrix. For example, for a detection area of ​​100×100 mm, if the sampling point spacing is 1 mm, a 10000×21 data matrix is ​​formed. Singular value decomposition is performed on the matrix, and the first three singular values ​​with the largest contribution rate are selected, corresponding to the spectral response characteristics at wavelengths of 850nm, 950nm and 1050nm, which directly reflect the optical properties of the interface material. In actual detection, the first eigenvector (corresponding to the maximum singular value) reflects the average reflection characteristics of the material, and the second and third eigenvectors reflect the spectral anomalies at the interface. For example, when detecting an IGBT module, the first singular value in the normal area usually accounts for more than 85% of the total energy. If its proportion is less than 70%, it indicates that there may be an interface anomaly at that location.

[0041] When using the attenuation characteristics of light to determine the position of the interface, the Lambert-Beer law is used to analyze the variation of light intensity with depth. At the ceramic-organic layer interface, due to the sudden change in the refractive index of the material (the refractive index of the ceramic substrate is about 2.0, and that of the organic layer is about 1.6), an obvious reflection peak will be generated. By analyzing the depth distribution curve of the reflected light intensity, when the reflection intensity shows a sudden change of more than 20%, it is marked as the interface position. For example, in the light intensity attenuation curve from the surface to the inside of a certain measuring point, the reflection intensity jumps from 20% to 35% at a depth of 300 microns, which can be determined as the ceramic-organic layer interface; similarly, at the organic layer-protective layer interface at about 600 microns, there will also be a reflection intensity jump of about 15%.

[0042] When establishing the photoelastic tensor matrix for the change in polarization state, it is necessary to consider the anisotropic change of the material under stress. The main diagonal elements of the matrix represent the photoelastic response in the direction of the principal stress, and the non-diagonal elements represent the influence of shear stress. In the operating temperature range of 25-125 degrees Celsius of the electronic control component, the photoelastic coefficient of the material will change with temperature: the change rate of the ceramic substrate is about 0.1% / ℃, and the organic layer is about 0.3% / ℃. These temperature effects need to be compensated in the stress calculation. For example, when the device operates at 80 degrees Celsius, the measured photoelastic coefficient needs to be multiplied by the temperature correction factor (about 1.055 for the ceramic layer and about 1.165 for the organic layer) to obtain an accurate stress value.

[0043] When performing spatial registration of the interface depth profile and stress response, the two sets of data are first unified to the same spatial grid. For a detection area of ​​100×100 mm, a spatial resolution of 0.1 mm is used to form a data array of 1000×1000×2 (2 represents two interface layers). The measured data is standardized according to the actual physical quantity: the depth data is normalized based on the maximum detection depth of 800 microns, and the stress data is normalized based on the yield strength of the material of 200 MPa. For example, if the original interface depth of a point is 350 microns and the maximum depth is 800 microns, the normalized value is 0.4375; if the stress value of the point is 120 MPa and the critical stress is 200 MPa, the normalized value is 0.6.

[0044] Finally, the defect sensitivity evaluation index is calculated based on the normalized data. For the ceramic-organic layer interface, when the normalized depth deviation exceeds 0.15 and the normalized stress value exceeds 0.7, the area is marked as a high-risk area; for the organic layer-protective layer interface, considering its weak bearing capacity, the corresponding thresholds are adjusted to 0.12 and 0.6. For example, in the detection of a DC converter, if the depth deviation of a certain area is 0.18 and the stress value is 0.85, the area will be marked as a key focus area and requires more detailed inspection or rework. This layered and graded method can effectively identify potential failure hazards and prevent insulation breakdown during device use.

[0045] Please continue reading Figure 1 , for the key focus area, synchronously collect polarization feature matrix, dual-wavelength interference data and depth spectrum data, perform time domain synchronization and spatial registration processing on the polarization feature matrix, dual-wavelength interference data and depth spectrum data, and generate multi-dimensional detection feature data; In one embodiment of the present invention, the polarization characteristic matrix, dual-wavelength interference data and depth spectrum data are synchronously collected for the key focus area, including: performing interface layering positioning on the key focus area, setting polarization modulation parameters, interference wavelength parameters and spectrum sampling parameters according to the material properties of the ceramic substrate, the organic insulating layer and the nano protective layer; modulating the polarization state of the incident light, collecting polarization reflection data at intervals of 15° in the range of 0°-180°, and constructing a polarization characteristic matrix; applying dual-wavelength phase modulated light to the key focus area, measuring the phase difference between the ceramic-organic layer interface and the organic layer-protective layer interface, and obtaining dual-wavelength interference data; and using confocal microscopy technology to scan different depth levels layer by layer, obtaining the spectral response characteristics of the ceramic substrate surface, the ceramic-organic layer interface, the organic layer-protective layer interface and the protective layer surface, and generating depth-resolved interface spectrum data.

[0046] Specifically, in the detection of the key areas of interest of the electronic control components, the detection parameters are first set according to the optical properties of the three layers of materials. The refractive index of the ceramic substrate is 2.0, the thickness is usually 300 microns, and its polarization modulation frequency is set to 1 kHz; the refractive index of the organic insulating layer is 1.6, the thickness is 200 microns, and the modulation frequency is set to 800 Hz; the refractive index of the nano-protective layer is 1.8, the thickness is 100 microns, and the modulation frequency is set to 600 Hz. These different modulation frequencies allow the signals of each layer to be separated in the frequency domain. Interferometry uses a two-color light source with wavelengths of 850 nanometers and 1050 nanometers, respectively, with a beam diameter of 50 microns and a power of 100 milliwatts each. The spectral sampling is in the range of 400-1100 nanometers, and the step size is set to 2 nanometers to obtain sufficient spectral resolution.

[0047] In polarization state measurement, a voltage-adjustable liquid crystal polarizer is used with a voltage range of 0-10 volts. Starting from 0°, polarization reflection data is collected every 15° until 180°, and data is collected at a total of 13 angles. The dwell time at each angle is 50 milliseconds to ensure signal stability. For example, for a certain detection point in the IGBT module, the reflection intensity may be measured as 100 microwatts at 0°, 95 microwatts at 15°, and so on, forming a 13×1 polarization reflection data vector. Combining the data of all points in the detection area, assuming that the detection area is 10×10 mm and the sampling interval is 0.1 mm, a 100×100×13 polarization feature matrix is ​​finally formed.

[0048] In dual-wavelength interferometry, two beams of light are simultaneously irradiated onto the sample surface after passing through a beam splitter, and interfere on the detector surface after reflection. The optical path of one optical arm is adjusted by piezoelectric ceramics so that the phase difference is scanned in the range of 0-2π, with a step value of π / 8. At the ceramic-organic layer interface, due to the refractive index difference (△n=0.4), the optical path difference is about 120 nanometers, which will result in a phase difference of 0.89π for 850 nanometer light and 0.72π for 1050 nanometer light. Similarly, at the organic layer-protective layer interface, the optical path difference caused by the refractive index difference (△n=0.2) is about 40 nanometers, and the corresponding phase differences are 0.30π and 0.24π, respectively. By recording these phase difference values, the interface position can be accurately located.

[0049] When using confocal microscopy for depth scanning, an objective lens with a numerical aperture of 0.7 and a working distance of 3 mm are used. The depth of field is set to 2 microns, and data is collected every 1 micron from the surface until the scanning depth reaches 1 mm. At each depth position, the reflectance spectrum in the 400-1100 nanometer band is recorded. For example, on the surface of the ceramic substrate, the typical reflectivity is around 90%; when scanning to the ceramic-organic layer interface at a depth of 300 microns, the reflectivity drops to 60%; at the organic layer-protective layer interface at 500 microns, the reflectivity further drops to 40%; the reflectivity of the outermost protective layer surface is about 30%. Through this layer-by-layer scanning, a complete interface spectral response feature can be obtained, and any anomalies that deviate from these typical values ​​may indicate interface delamination or other defects.

[0050] In one embodiment of the present invention, the polarization characteristic matrix, dual-wavelength interference data and deep spectrum data are subjected to time-domain synchronization and spatial registration processing to generate multi-dimensional detection characteristic data, including: performing Stokes vector transformation on the polarized reflection data, calculating the polarization degree and polarization state angle of each interface, and generating an interface polarization characteristic diagram; calculating the optical path difference variation of the ceramic-organic layer interface and the organic layer-protective layer interface according to the phase difference distribution of the dual-wavelength interference data, and obtaining interface morphology data; performing dispersion characteristic analysis on the deep spectrum data, extracting the reflectivity spectrum and absorptivity spectrum of each interface in the 850nm-1050nm band, and constructing a characteristic spectrum database in combination with the optical dispersion characteristics of the interface material; performing layer-by-layer corresponding registration on the interface polarization characteristic diagram, the interface morphology data and the characteristic spectrum database, and establishing a hierarchical feature mapping relationship; and according to the hierarchical feature mapping relationship, using the interface stress-optical response model to fuse and reconstruct the multi-dimensional features to generate multi-dimensional detection characteristic data.

[0051] Specifically, in the multi-layer interface analysis of the electronically controlled components, the polarized reflection data is first subjected to Stokes vector transformation. For each measurement point, the reflection intensity values ​​at 13 polarization angles (0°-180°, interval 15°) are substituted into the Stokes parameter calculation formula. For example, the reflection intensity at 0° and 90° of a certain measurement point is 100 microwatts and 60 microwatts, respectively, and the calculated linear polarization degree is 0.25; the reflection intensity at 45° and 135° is 85 microwatts and 75 microwatts, respectively, and the circular polarization degree is 0.06. These parameters can be used to determine the total polarization degree (value range 0-1) and polarization state angle (0°-360°) of the point. In the 10×10 mm detection area, if the sampling interval is 0.1 mm, an interface polarization feature map of 100×100 measurement points is formed, and each point contains two parameters: polarization degree and polarization state angle.

[0052] When analyzing dual-wavelength interference data, the optical path difference is calculated using the phase difference between the two wavelengths of 850 nanometers and 1050 nanometers at the interface. Taking the ceramic-organic layer interface as an example, if the phase difference measured for 850 nanometer light is 0.89π and for 1050 nanometer light is 0.72π, the optical path difference is calculated to be 120 nanometers according to the phase difference formula. By performing such calculations point by point in the detection area, the microscopic morphology distribution of the interface can be obtained. The change in the optical path difference of a normal interface is usually less than 50 nanometers. If a mutation of more than 100 nanometers occurs in a certain area, there is a high probability of stratification or bubbles. The same method is used to analyze the organic layer-protective layer interface, and the normal optical path difference fluctuation range is even smaller, usually not exceeding 30 nanometers.

[0053] In the deep spectral data analysis, the reflectance spectrum in the 850-1050 nanometer band is processed. In this band, data at 41 wavelength points are obtained at intervals of 5 nanometers. For each interface, its reflectance spectrum and absorptivity spectrum are recorded. For example, the normal ceramic-organic layer interface has a reflectance of 60% and an absorptivity of 15% at 850 nanometers. As the wavelength increases, the reflectance slowly decreases to 50% and the absorptivity increases to 25%; while the reflectance of the organic layer-protective layer interface decreases from 40% to 35%, and the absorptivity increases from 20% to 30%. These spectral features are directly related to the intrinsic optical dispersion properties of the material (the relationship between the refractive index and the wavelength), and any deviation from this law may indicate an interface defect.

[0054] The data registration process uses two steps: spatial registration and feature registration. First, in the spatial coordinate system, the interface polarization feature map, morphology data, and spectral features are matched one by one. In the 10×10 mm area, a three-dimensional data grid with a resolution of 0.1 mm is established. Each grid point contains: polarization degree (0-1), polarization state angle (0°-360°), optical path difference (0-200 nanometers), reflectivity and absorptivity data at 41 wavelengths. For example, the data of a grid point may be: polarization degree 0.35, polarization state angle 125°, optical path difference 85 nanometers, reflectivity 55% at 850 nanometers, etc.

[0055] When the features are finally fused, the interface stress-optical response model is used to associate the features. The model considers the effect of stress on the optical properties of the material: for every 10 MPa increase in stress, the degree of polarization changes by about 0.02, the optical path difference increases by about 5 nanometers, and the reflectivity changes by about 2%. For example, at a certain detection point of the IGBT module, if the measured polarization degree increases by 0.1, the optical path difference increases by 25 nanometers, and the reflectivity decreases by 10%, it can be inferred from these changes that the stress at that point is about 50 MPa. Through this feature association, the multi-dimensional detection feature data finally generated not only reflects the geometric characteristics of the interface, but also contains stress state information, providing a comprehensive basis for defect assessment.

[0056] Please continue reading Figure 1 , extracting the optical anisotropy coefficient, the microscopic deformation tensor and the interface reflection characteristic vector according to the multidimensional detection feature data, and determining the delamination defect characteristic parameters when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection characteristic vector exceeds a third preset threshold; In one embodiment of the present invention, the optical anisotropy coefficient, the microscopic deformation tensor and the interface reflection characteristic vector are extracted according to the multidimensional detection feature data, and when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection characteristic vector exceeds a third preset threshold, the delamination defect characteristic parameter is determined, including: constructing an interface Mueller matrix according to the multidimensional detection feature data, characteristically decomposing the polarization state changes at different interfaces, and calculating the optical anisotropy coefficient of each layer interface; based on the dual-wavelength interference phase distribution of the multidimensional detection feature data, using a two-dimensional The phase gradient algorithm calculates the in-plane strain and out-of-plane displacement of the interface, and establishes a microscopic deformation tensor through the strain-displacement relationship; the wavelength domain analysis is performed on the multi-dimensional detection feature data to extract the interface reflection feature vectors of the ceramic-organic layer and the organic layer-protective layer respectively; according to the operating temperature range and voltage level of the electronically controlled component, the first preset threshold value of the optical anisotropy coefficient, the second preset threshold value of the microscopic deformation tensor and the third preset threshold value of the interface reflection feature vector are set; when the detection parameter exceeds the corresponding preset threshold value, the depth position, geometric size and interface type information of the defect area are extracted to determine the delamination defect feature parameters.

[0057] Specifically, in the interface characteristic analysis of the electronic control components, a 4×4 interface Mueller matrix is ​​first constructed through multi-dimensional detection feature data. The input data includes the reflection intensity values ​​obtained at four polarization states of 0°, 45°, 90°, and 135°. For example, at a certain measuring point on the ceramic-organic layer interface, the reflection intensities of the four polarization states are 100, 85, 70, and 80 microwatts, respectively. These data are used to fill the corresponding elements of the Mueller matrix. The matrix is ​​eigenvalue decomposed, and the ratio of the maximum eigenvalue to the minimum eigenvalue is defined as the optical anisotropy coefficient. This ratio of a normal interface is usually between 1.1 and 1.3. If it exceeds 1.5, it indicates that there is obvious stress-induced anisotropy at the interface.

[0058] When dual-wavelength interferometry is used to measure interface deformation, a dual-color light source of 850 nanometers and 1050 nanometers is used. Phase distribution data is obtained at a spacing of 0.1 mm in a detection area of ​​10×10 mm. For each measuring point, the phase difference between it and the adjacent point is calculated to obtain the phase gradient distribution. For example, if the phase value of a point is 0.5π and its adjacent point is 0.55π, then the phase gradient is 0.05π / 0.1 mm. The in-plane strain is calculated by the phase gradient: each change of 0.01π / mm in the phase gradient corresponds to a strain of approximately 0.01%. At the same time, the out-of-plane displacement is calculated based on the absolute value of the phase: a phase difference of 2π corresponds to a displacement of half a wavelength. In an IGBT module, the in-plane strain of a normal interface is usually less than 0.05%, and the out-of-plane displacement is less than 100 nanometers.

[0059] When the wavelength domain analysis of the test data is performed, the interface reflection characteristics are extracted in the range of 850-1050 nanometers. For the ceramic-organic layer interface, the reflection spectrum normally shows a smooth decreasing trend, from 65% at 850 nanometers to 55% at 1050 nanometers; while the reflectivity of the organic layer-protective layer interface drops from 45% to 35%. These reflectivity values ​​are combined into feature vectors, and each interface forms a 201-dimensional vector (wavelength interval 1 nanometer). If the interface is stratified, the reflection spectrum will fluctuate abnormally, such as a sudden change in reflectivity at a certain wavelength or an abnormal absorption peak.

[0060] The threshold setting needs to take into account the working conditions of the electronic control components. Taking the 1200 volt IGBT module as an example, the operating temperature range is -40 to 125 degrees Celsius. Within this range, the first preset threshold of the optical anisotropy coefficient is set to 1.5, which has taken into account the temperature-induced change in the intrinsic anisotropy of the material (about 0.1% / °C). The second preset threshold of the microscopic deformation tensor is set to: 0.1% in-plane strain and 150 nanometers out-of-plane displacement. These values ​​are 80% of the deformation that the material can withstand under normal operating temperature cycles. The third preset threshold of the interface reflection eigenvector is defined as: the reflectivity deviation does not exceed ±15% of the standard spectrum.

[0061] When the detected parameters exceed the preset threshold, the system starts to extract defect features. For example, it is detected somewhere in the DC converter: the optical anisotropy coefficient is 1.8 (exceeding the threshold of 1.5), the in-plane strain is 0.15% (exceeding the threshold of 0.1%), and the reflectance spectrum deviates from the standard spectrum by 20% at 950 nanometers (exceeding the threshold of 15%). This position is precisely located, and it is found that the defect is located at a depth of 350 microns from the surface (ceramic-organic layer interface). The defect area is elliptical, with a long axis of 2 mm and a short axis of 1.2 mm. These parameters together constitute the characteristic description of the delamination defect, providing a quantitative basis for subsequent defect evaluation and processing.

[0062] Please continue reading Figure 1According to the characteristic parameters of the delamination defect, the local curvature distribution, the interface shear stress field and the thermal stress distribution are calculated, and the local curvature distribution, the interface shear stress field and the thermal stress distribution are quantitatively matched with the material characteristic parameters of the interface type where the defect is located to generate a quantitative evaluation index of the delamination defect.

[0063] In one embodiment of the present invention, the local curvature distribution, interface shear stress field and thermal stress distribution are calculated according to the characteristic parameters of the delamination defect, and the local curvature distribution, interface shear stress field and thermal stress distribution are quantitatively matched with the material characteristic parameters of the interface type where the defect is located to generate a quantitative evaluation index of the delamination defect, including: according to the depth position and geometric dimensions in the characteristic parameters of the delamination defect, the local deformation of the interface is calculated using a Gaussian curvature operator to generate local curvature distribution data; based on the interface type information, combined with the difference in elastic modulus of the ceramic substrate, the organic insulating layer and the nano protective layer, an interface stress transfer function is constructed to calculate the interface shear stress field; an infrared thermal imaging array is used to perform real-time temperature field scanning on the surface of the electric control component, and the temperature gradient data of the power device periphery and the heat sink contact area are extracted to calculate the thermal stress distribution; the local curvature distribution data, the interface shear stress field and the thermal stress field data are tensor-superimposed to obtain the interface comprehensive stress state; according to the critical stress threshold and fatigue life parameters of different interface types, the interface comprehensive stress state is normalized to generate a quantitative evaluation index of the delamination defect.

[0064] Specifically, in the interface stress analysis of the electronic control components, the Gaussian curvature operator is first used to calculate the local deformation of the interface. For a 10×10 mm detection area, a height data matrix is ​​established with a grid spacing of 0.1 mm. For example, when a defect depth of 350 microns and a lateral size of 2×1.2 mm is detected, a 5×5 mm calculation window is selected to calculate the principal curvature value for each grid point. The curvature is calculated by the central difference method: if the height of a point is h and the distance between adjacent points is △x, then the curvature K≈(h(x+△x)+h(x-△x)-2h(x)) / The curvature value of a normal interface is usually less than 0.01 / mm. When obvious deformation occurs, the local curvature can reach 0.1 / mm.

[0065] The interface stress transfer analysis needs to consider the differences in the elastic properties of the materials. The elastic modulus of the ceramic substrate is 300GPa, the organic insulating layer is 3GPa, and the nano protective layer is 10GPa. Due to the huge difference in elastic modulus (up to 100 times), stress concentration will occur at the interface. When calculating the interface shear stress, the stress transfer equation of the layered structure is used: when a stress of 1MPa is applied to the upper layer, the ceramic-organic layer interface will produce a shear stress of about 0.8MPa, and the organic layer-protective layer interface will produce a shear stress of about 0.5MPa. For example, in the edge area of ​​the IGBT module, if the surface stress is measured to be 10MPa, it can be calculated that the shear stress of the ceramic-organic layer interface is about 8MPa.

[0066] The temperature field analysis uses a 320×256 pixel infrared thermal imager with a temperature resolution of 0.05°C. The surface of the electronic control component is scanned in real time, and the sampling frequency is set to 30Hz. When the power device is working, the temperature around the device can reach 85°C, while the temperature at a distance of 20 mm drops to 45°C, resulting in a significant temperature gradient. The thermal stress is calculated using the Fourier heat conduction equation: for every 10°C increase in temperature, the ceramic substrate generates a thermal stress of about 2MPa, and the organic layer generates a thermal stress of about 0.5MPa. For example, when the local temperature gradient is 2°C / mm, it can be calculated that the thermal stress distribution gradually decreases from 4MPa in the center to 0.5MPa at the edge.

[0067] When the three stress field data are tensor superimposed, their directionality needs to be considered. Taking a 2×2 mm area as an example, if the local curvature there produces a stress of 5 MPa, the interface shear stress is 8 MPa, the thermal stress is 4 MPa, and the angles of these stress directions are 30° and 60° respectively, then the combined stress amplitude can be calculated to be about 15 MPa and the principal stress direction is 45° through stress tensor superposition. According to the stress tensor theory, the stress components and shear stress components in the principal stress direction are calculated respectively, and the von Mises criterion is used to comprehensively evaluate the interface stress state.

[0068] When performing the final stress assessment, the bearing capacity of different interfaces is considered. The critical shear stress of the ceramic-organic layer interface is 20MPa, and the fatigue life parameter is set to 10^6 cycles (under 15MPa stress); the critical value of the organic layer-protective layer interface is lower, 12MPa, and the fatigue life parameter is 10^5 cycles (under 10MPa stress). The measured comprehensive stress value is divided by the critical stress of the corresponding interface to obtain the normalized stress value. For example, if the comprehensive stress measured at a certain point is 15MPa, located at the ceramic-organic layer interface, the normalized value is 0.75; if the stress acts on the organic layer-protective layer interface, the normalized value will reach 1.25, indicating that the safety threshold has been exceeded. This quantitative assessment method can intuitively reflect the degree of danger of the defect and provide a basis for maintenance decisions.

[0069] The above describes the visual inspection method for the appearance of the electric control components for new energy vehicles in the embodiment of the present invention. The following describes the visual inspection device for the appearance of the electric control components for new energy vehicles in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, an appearance visual inspection device for electric control components for new energy vehicles includes: A stress partitioning module 201 is used to obtain structural stress distribution data of the electric control component, and partition the electric control component according to the structural stress distribution data to obtain partition data of stress sensitive areas and conventional areas, wherein the stress sensitive area includes the periphery of the power device installation area, the heat sink contact area, the connector lead-out area and the periphery of the fixing screw hole; A feature detection module 202 is used to perform multi-wavelength spectral scanning and polarization state detection on the stress sensitive area based on the partition data by using adaptive path planning to obtain spectral feature parameters and polarization state changes, and determine a key focus area according to the degree of deviation between the spectral feature parameters and the polarization state changes and a preset threshold; The data acquisition module 203 is used to synchronously acquire polarization characteristic matrix, dual-wavelength interference data and depth spectrum data for the key focus area, perform time domain synchronization and spatial registration processing on the polarization characteristic matrix, dual-wavelength interference data and depth spectrum data, and generate multi-dimensional detection characteristic data; The defect recognition module 204 is used to extract the optical anisotropy coefficient, the microscopic deformation tensor and the interface reflection characteristic vector according to the multi-dimensional detection feature data, and determine the delamination defect characteristic parameters when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection characteristic vector exceeds a third preset threshold; The evaluation processing module 205 is used to calculate the local curvature distribution, interface shear stress field and thermal stress distribution according to the characteristic parameters of the delamination defect, quantitatively match the local curvature distribution, interface shear stress field and thermal stress distribution with the material characteristic parameters of the interface type where the defect is located, and generate a quantitative evaluation index for the delamination defect.

[0070] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A method for visual inspection of the appearance of electronic control components for new energy vehicles, characterized in that: include: Acquire structural stress distribution data of the electric control component, and perform zoning processing on the electric control component according to the structural stress distribution data to obtain zoning data of stress sensitive areas and conventional areas, wherein the stress sensitive areas include the periphery of the power device mounting area, the heat sink contact area, the connector lead-out area, and the periphery of the fixing screw hole; Based on the partition data, adaptive path planning is used to perform multi-wavelength spectral scanning and polarization state detection on the stress sensitive area to obtain spectral characteristic parameters and polarization state changes, and a key focus area is determined according to the degree of deviation between the spectral characteristic parameters and the polarization state changes and a preset threshold value; For the key focus area, synchronously collect polarization feature matrix, dual-wavelength interference data and depth spectrum data, perform time domain synchronization and spatial registration processing on the polarization feature matrix, dual-wavelength interference data and depth spectrum data, and generate multi-dimensional detection feature data; Extracting an optical anisotropy coefficient, a microscopic deformation tensor and an interface reflection characteristic vector according to the multidimensional detection characteristic data, and determining a delamination defect characteristic parameter when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection characteristic vector exceeds a third preset threshold; According to the characteristic parameters of the delamination defect, the local curvature distribution, the interface shear stress field and the thermal stress distribution are calculated, and the local curvature distribution, the interface shear stress field and the thermal stress distribution are quantitatively matched with the material characteristic parameters of the interface type where the defect is located to generate a quantitative evaluation index of the delamination defect.

2. The visual inspection method for the appearance of electric control components for new energy vehicles according to claim 1, characterized in that: The step of acquiring structural stress distribution data of the electric control component and performing zoning processing on the electric control component according to the structural stress distribution data to obtain zoning data of stress sensitive areas and conventional areas includes: Collect the surface reflection images of the electronic control components at four polarization states of 0°, 45°, 90°, and 135°, and extract the polarization degree distribution data; A dual-wavelength laser with wavelengths of 850nm and 1050nm is used to perform periodic scanning heating on the electronic control components, and the surface temperature response sequence is collected during the heating-cooling cycle; According to the surface temperature response sequence, the temperature response time constant and phase delay of each measuring point are calculated to obtain the interface thermal diffusion characteristic function; The polarization degree distribution data is spatially mapped and fused with the interface thermal diffusion characteristic function to generate depth-resolved stress distribution data; According to the depth-resolved stress distribution data, adaptive threshold segmentation is performed in combination with the interface characteristics of the ceramic substrate, the organic insulating layer and the nano-protective layer to determine the partition data of the stress-sensitive area and the conventional area.

3. The visual inspection method for the appearance of electric control components for new energy vehicles according to claim 2 is characterized in that: The collecting of the surface reflection images of the electric control component at four polarization states of 0°, 45°, 90° and 135° to extract polarization degree distribution data includes: According to the layout of power devices of the electronic control component, differentiated light source incident angles are set around the power device installation area, the heat sink contact area, the connector lead-out area and the fixing screw hole to generate an incident light spot density distribution map; In each spot area, the incident light is polarized using a liquid crystal tunable polarizer to collect high dynamic range reflection images at four polarization states: 0°, 45°, 90°, and 135°. Performing local contrast enhancement on the high dynamic range reflection image to extract polarized reflection intensity data of the ceramic substrate surface, the organic insulating layer surface and the nano-protective layer surface; Calculating the Stokes parameters and polarization coefficient of each surface according to the polarized reflection intensity data, and establishing a surface-interface polarization characteristic mapping relationship; Based on the surface-interface polarization characteristic mapping relationship and in combination with the birefringence effect parameters of the interface material, the stress-induced birefringence component at the interface is calculated to generate polarization degree distribution data.

4. The visual inspection method for the appearance of electric control components for new energy vehicles according to claim 1, characterized in that: The method of performing multi-wavelength spectral scanning and polarization state detection on the stress sensitive area based on the partition data by using adaptive path planning to obtain spectral characteristic parameters and polarization state variation includes: Performing stress gradient calculation on the stress sensitive area, and generating a scanning grid of dynamic density according to the stress gradient value; Using a wavelength tunable laser, continuously scan each node in the scanning grid in the 850nm-1050nm band, and record the spectral reflection intensity curve as a spectral characteristic parameter; Applying electrically controlled liquid crystal polarization modulation at each scanning node to obtain depth-resolved light intensity data in different polarization states; The light intensity-polarization tensor including the ceramic-organic layer interface and the organic layer-protective layer interface is constructed according to the depth-resolved light intensity data, and the lattice orientation parameters and the interface stress birefringence coefficient of each interface are calculated as the polarization state change.

5. The visual inspection method for the appearance of electric control components for new energy vehicles according to claim 4, characterized in that: Determining the key focus area according to the degree of deviation between the spectral characteristic parameter and the polarization state change amount and a preset threshold value includes: Constructing a principal component analysis matrix of the spectral characteristic parameters, performing singular value decomposition on the principal component analysis matrix, and extracting a characteristic frequency band response of the 850nm-1050nm band; According to the light intensity attenuation law of the characteristic frequency band response at different depths, the position parameters of the ceramic-organic layer interface and the organic layer-protective layer interface are calculated to generate interface depth profile data; Establishing a photoelastic tensor matrix for the polarization state variation, and calculating the interface stress response based on a temperature-stress coupling model of the electric control component; Performing spatial position alignment and feature quantity normalization processing on the interface depth profile data and the interface stress response to generate a defect sensitivity evaluation index; According to the defect sensitivity evaluation index, threshold parameters related to the layering type are set, the stress sensitive area is segmented into layers, and the area where the defect sensitivity exceeds the corresponding threshold is marked as a key focus area.

6. The visual inspection method for the appearance of electric control components for new energy vehicles according to claim 1, characterized in that: For the focus area, synchronously collecting polarization characteristic matrix, dual-wavelength interference data and deep spectrum data includes: Performing interface layering positioning on the key focus area, and setting polarization modulation parameters, interference wavelength parameters and spectrum sampling parameters according to the material properties of the ceramic substrate, the organic insulating layer and the nano protective layer; The incident light is polarized and the polarization state is modulated. Polarization reflection data is collected at intervals of 15° in the range of 0°-180° to construct a polarization characteristic matrix. Applying dual-wavelength phase modulated light to the focus area, measuring the phase difference between the ceramic-organic layer interface and the organic layer-protective layer interface, and obtaining dual-wavelength interference data; Confocal microscopy technology is used to scan different depth levels layer by layer to obtain the spectral response characteristics of the ceramic substrate surface, ceramic-organic layer interface, organic layer-protective layer interface and protective layer surface, and generate depth-resolved interface spectral data.

7. The visual inspection method for the appearance of electric control components for new energy vehicles according to claim 6, characterized in that: The step of performing time domain synchronization and spatial registration processing on the polarization feature matrix, dual-wavelength interference data and depth spectrum data to generate multi-dimensional detection feature data includes: The polarized reflection data is transformed by Stokes vector to calculate the polarization degree and polarization state angle of each interface, and generate an interface polarization characteristic diagram; According to the phase difference distribution of the dual-wavelength interference data, the optical path difference variation between the ceramic-organic layer interface and the organic layer-protective layer interface is calculated to obtain interface morphology data; Performing dispersion characteristic analysis on the deep spectral data, extracting the reflectivity spectrum and absorptivity spectrum of each interface in the 850nm-1050nm band, and building a characteristic spectrum database in combination with the optical dispersion characteristics of the interface material; The interface polarization characteristic map, interface morphology data and characteristic spectrum database are aligned layer by layer to establish a hierarchical characteristic mapping relationship; According to the hierarchical feature mapping relationship, an interface stress-optical response model is used to fuse and reconstruct multi-dimensional features to generate multi-dimensional detection feature data.

8. The visual inspection method for the appearance of electric control components for new energy vehicles according to claim 1, characterized in that: The method extracts an optical anisotropy coefficient, a microscopic deformation tensor and an interface reflection characteristic vector according to the multidimensional detection characteristic data, and determines a delamination defect characteristic parameter when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection characteristic vector exceeds a third preset threshold, including: Constructing an interface Mueller matrix according to the multi-dimensional detection characteristic data, performing characteristic decomposition on the polarization state changes at different interfaces, and calculating the optical anisotropy coefficient of each layer interface; Based on the dual-wavelength interference phase distribution of the multi-dimensional detection characteristic data, a two-dimensional phase gradient algorithm is used to calculate the in-plane strain and out-of-plane displacement of the interface, and a microscopic deformation tensor is established through the strain-displacement relationship; Performing wavelength domain analysis on the multi-dimensional detection characteristic data to extract the interface reflection characteristic vectors of the ceramic-organic layer and the organic layer-protective layer respectively; According to the operating temperature range and voltage level of the electronic control component, a first preset threshold value of the optical anisotropy coefficient, a second preset threshold value of the microscopic deformation tensor, and a third preset threshold value of the interface reflection characteristic vector are set; When the detection parameters exceed the corresponding preset thresholds, the depth position, geometric dimensions and interface type information of the defect area are extracted to determine the characteristic parameters of the delamination defect.

9. The visual inspection method for the appearance of electric control components for new energy vehicles according to claim 1, characterized in that: The method of calculating the local curvature distribution, the interface shear stress field and the thermal stress distribution according to the characteristic parameters of the delamination defect, quantitatively matching the local curvature distribution, the interface shear stress field and the thermal stress distribution with the material characteristic parameters of the interface type where the defect is located, and generating a quantitative evaluation index of the delamination defect includes: According to the depth position and geometric size in the delamination defect characteristic parameters, a Gaussian curvature operator is used to calculate the local deformation of the interface to generate local curvature distribution data; Based on the interface type information and the difference in elastic modulus between the ceramic substrate, organic insulating layer and nano-protective layer, the interface stress transfer function is constructed and the interface shear stress field is calculated. Use infrared thermal imaging array to scan the temperature field of the surface of the electronic control components in real time, extract the temperature gradient data around the power device and the heat sink contact area, and calculate the thermal stress distribution; Performing tensor superposition on the local curvature distribution data, the interface shear stress field and the thermal stress field data to obtain a comprehensive interface stress state; According to the critical stress thresholds and fatigue life parameters of different interface types, the comprehensive stress state of the interface is normalized to generate a quantitative evaluation index of the delamination defect.

10. A visual inspection device for the appearance of electric control components for new energy vehicles, characterized in that: The visual appearance inspection device for electric control components for new energy vehicles is applied to the visual appearance inspection method for electric control components for new energy vehicles as claimed in any one of claims 1 to 9, comprising: A stress partitioning module is used to obtain structural stress distribution data of the electric control component, and partition the electric control component according to the structural stress distribution data to obtain partition data of stress sensitive areas and conventional areas, wherein the stress sensitive area includes the periphery of the power device installation area, the heat sink contact area, the connector lead-out area and the periphery of the fixing screw hole; A feature detection module, for performing multi-wavelength spectral scanning and polarization state detection on the stress sensitive area based on the partition data by using adaptive path planning, obtaining spectral feature parameters and polarization state variation, and determining a key focus area according to the degree of deviation between the spectral feature parameters and polarization state variation and a preset threshold; A data acquisition module is used to synchronously collect polarization feature matrix, dual-wavelength interference data and depth spectrum data for the key focus area, perform time domain synchronization and spatial registration processing on the polarization feature matrix, dual-wavelength interference data and depth spectrum data, and generate multi-dimensional detection feature data; a defect recognition module, for extracting an optical anisotropy coefficient, a microscopic deformation tensor and an interface reflection feature vector according to the multidimensional detection feature data, and determining a delamination defect feature parameter when the optical anisotropy coefficient exceeds a first preset threshold, the microscopic deformation tensor exceeds a second preset threshold and the interface reflection feature vector exceeds a third preset threshold; The evaluation processing module is used to calculate the local curvature distribution, interface shear stress field and thermal stress distribution according to the characteristic parameters of the delamination defect, quantitatively match the local curvature distribution, interface shear stress field and thermal stress distribution with the material characteristic parameters of the interface type where the defect is located, and generate a quantitative evaluation index for the delamination defect.

Images