Radiator strength detection equipment and detection method

By acquiring and processing the multi-scale structure and multi-environment coupled material feature data of the radiator, a thermal-flow-structure multi-physics fusion data set is constructed, and the multi-physics coupling characteristics is analyzed, the problem of ignoring the multi-physics coupling effect in the prior art is solved, and more accurate and reliable radiator intensity detection and life prediction are achieved.

CN120105357AInactive Publication Date: 2025-06-06GUANGDONG ZHENXI PRECISION COMPONENTS CO LTD
View PDF 0 Cites 10 Cited by

Patent Information

Application Number
CN202510600590.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-05-12
Publication Date
2025-06-06
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing radiator intensity detection methods ignore the multi-physics coupling effect and the proprietary characteristics of the radiator, resulting in low accuracy of intensity evaluation results and reliability of predicted lifespan.

Method used

By acquiring the multi-scale structural feature data of the heat sink and the multi-environment coupled material feature data, adaptive wavelet noise reduction and structural correlation mapping are performed to generate a heat-flow-structure multi-physics fusion data set. Then, a multi-level feature vector set is extracted, and the spatiotemporal alignment and interactive mapping of multi-physics coupled features are performed to construct a multi-physics coupled data matrix. Finally, through the separation and iterative solution of multiple physical field control equation systems, the multi-physics space-time distribution state of the radiator is calculated, and the damage evolution state is identified and life-time calculation is performed to generate comprehensive intensity detection results.

Benefits of technology

It effectively improves the accuracy of radiator strength detection, especially in terms of thermal cycle fatigue, material thinning and stress distribution, and fully considers the multi-physical coupling characteristics and failure mechanism of the radiator, which improves the reliability of detection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120105357A_ABST
    Figure CN120105357A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of strength detection, and discloses radiator strength detection equipment and a detection method. The method comprises the following steps: carrying out adaptive wavelet denoising and structural association mapping on collected multi-scale structural feature data and multi-environment coupling material feature data of a radiator to be detected, and carrying out multi-dimensional feature extraction, space-time alignment and interactive mapping on an obtained fusion data set to obtain a multi-physical field coupling data matrix; carrying out physical field control equation set separation iteration solution and heat dissipation cyclic load and environmental action calculation on the coupling data matrix to obtain a damage evolution state; and performing multi-dimensional damage sensitive area identification based on the damage evolution state, generating a risk distribution map, performing failure comparison and life calculation based on historical failure data of the same type of radiators, and generating an intensity comprehensive detection result. Through multi-physical field coupling analysis and multi-dimensional damage evaluation, accurate detection of the strength state of the radiator is realized, and the accuracy of structural reliability evaluation is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of strength detection, and in particular to a radiator strength detection device and a detection method. Background Art

[0002] In the field of radiator technology, strength testing is an important part of product development and quality control, and damage evolution analysis in a multi-physics coupling environment is the most critical and challenging stage in strength testing. Accurately detecting the strength status of radiators is crucial for manufacturers, R&D institutions, and testing centers, as it directly affects product reliability assessment, performance optimization, and overall service life. Therefore, an effective strength testing method to monitor the damage evolution of radiators under complex working conditions is essential to ensure the structural stability and heat transfer performance of radiators.

[0003] Nowadays, the strength test of heat sinks is mainly evaluated by finite element analysis, experimental testing or a hybrid method combining the two. Some studies have tried to use digital image correlation technology to analyze the strain distribution of heat sinks under thermal cycling conditions, or to establish simplified strength prediction models to evaluate the remaining life, and even tried to apply machine learning technology to the damage identification process. However, these methods still face challenges in integrating multi-dimensional physical field data, dealing with thermal-fluid-structure coupling effects, and adapting to complex heat dissipation cycle conditions, and often ignore some key features of heat sinks, such as thermal cycle fatigue, temperature sensitivity, welding residual stress, and corrosive environment effects, which may have a significant impact on strength judgment and weak link identification. That is, the existing heat sink strength detection methods ignore the multi-physical field coupling effects and the proprietary characteristics of the heat sink, resulting in low accuracy of strength assessment results and low reliability of predicted life. Summary of the invention

[0004] The main purpose of the present invention is to solve the problem that the existing radiator strength detection method ignores the multi-physical field coupling effect and the unique characteristics of the radiator, resulting in low accuracy of the strength evaluation result and low reliability of the predicted life.

[0005] A first aspect of the present invention provides a radiator strength detection method, the radiator strength detection method comprising: obtaining multi-scale structural feature data and multi-environment coupling material feature data of a radiator to be detected, and performing adaptive wavelet denoising and structural correlation mapping on the multi-scale structural feature data and the multi-environment coupling material feature data to obtain a heat-fluid-structure multi-physics field fusion data set; extracting multi-dimensional radiator features from the heat-fluid-structure multi-physics field fusion data set to obtain a multi-level feature vector set, and performing spatiotemporal alignment and interactive mapping of multi-physics field coupling features on the multi-level feature vector set to obtain a multi-level feature vector set. Physical field coupling data matrix; performing separate iterative solutions to the multi-physical field coupling data matrix for multiple physical field control equations to obtain the multi-physical field spatiotemporal distribution state of the radiator, and performing heat dissipation cycle load and environmental action calculations on the multi-physical field spatiotemporal distribution state to obtain the damage evolution state of the radiator; based on the damage evolution state, identifying the multi-dimensional damage sensitive area of ​​the radiator to be tested, generating a multi-dimensional risk distribution map, and based on the historical failure data of the radiator to be tested corresponding to the same type of radiator, performing failure comparison and life calculation on the multi-dimensional risk distribution map to generate a comprehensive strength test result.

[0006] Optionally, in a first implementation method of the first aspect of the present invention, the multi-scale structural feature data includes a three-dimensional model of the outer surface of the radiator, internal three-dimensional structure data and internal micro-defect distribution data, and the multi-environment coupling material feature data includes temperature-related mechanical property data, thermal cycle fatigue feature data and environmental sensitivity data. The multi-scale structural feature data and multi-environment coupling material feature data of the radiator to be detected are obtained, including: performing a multi-spectral three-dimensional scan on the radiator to be detected to obtain surface geometric point cloud data of a preset spatial resolution, and performing multi-angle stereo stitching on the surface geometric point cloud data to obtain a three-dimensional model of the outer surface of the radiator, and performing a micro-tomography scan on the radiator to be detected to obtain internal three-dimensional structure data of a preset radiation scanning resolution, and performing high-frequency ultrasonic phased array detection on the radiator structure stress concentration area of ​​the radiator to be detected to obtain internal micro-defect distribution data; performing a multi-temperature domain quasi-static mechanical property test on the radiator to be detected to obtain temperature-related mechanical property data of the radiator material, and performing a thermal-mechanical coupling cyclic load test on the radiator to be detected to obtain thermal cycle fatigue feature data and environmental sensitivity data of the radiator material.

[0007] Optionally, in a second implementation manner of the first aspect of the present invention, the multi-scale structural feature data and the multi-environment coupling material feature data are subjected to adaptive wavelet denoising and structural association mapping to obtain a heat-fluid-structure multi-physics field fusion data set, including: adaptive wavelet filtering of the surface geometric point cloud data to obtain denoised surface data corresponding to structural edge features, and extracting surface roughness distribution and local curvature parameters in the three-dimensional model of the outer surface of the radiator to obtain surface quality evaluation index data, and iteratively reconstructing and multi-scale contrast enhancing the internal three-dimensional structure data to obtain a resolution-enhanced internal structure three-dimensional. A three-dimensional voxel model is prepared, and phase correction and time inversion are performed on the internal micro-defect distribution data to obtain an enhanced internal defect distribution map, and spatial registration processing is performed on the denoised surface data, the surface quality evaluation index data, the enhanced internal structure three-dimensional voxel model and the enhanced internal defect distribution map to obtain a multi-scale structural model from macro to micro; anomaly detection and smooth spline fitting are performed on the multi-environment coupled material characteristic data to obtain a material property distribution function, and correlation mapping is performed between the multi-scale structural model and the material property distribution function to obtain a thermal-fluid-structure multi-physics field fusion data set associated with the structure and the material.

[0008] Optionally, in a third implementation method of the first aspect of the present invention, the multi-dimensional heat sink features are extracted from the thermal-fluid-structure multi-physics field fusion data set to obtain a multi-level feature vector set, including: geometrically decomposing the thermal-fluid-structure multi-physics field fusion data set to obtain a heat sink structure characteristic function, and topological skeleton extracting the internal structure data in the thermal-fluid-structure multi-physics field fusion data set to obtain a flow channel characteristic distribution function, and spatially clustering and probability distribution fitting the defect data in the thermal-fluid-structure multi-physics field fusion data set to obtain a micro-defect characteristic distribution function, and extracting key parameters of the material property data in the thermal-fluid-structure multi-physics field fusion data set to obtain a material temperature-related characteristic parameter set; organizing and integrating the heat sink structure characteristic function, the flow channel characteristic distribution function, the micro-defect characteristic distribution function and the material temperature-related characteristic parameter set to obtain a multi-level feature vector set.

[0009] Optionally, in a fourth implementation method of the first aspect of the present invention, the multi-level feature vector set is subjected to spatiotemporal alignment and interactive mapping of multi-physical field coupling features to obtain a multi-physical field coupling data matrix, including: synchronizing the time base of various features in the multi-level feature vector set to obtain a time-aligned feature data sequence, and spatially registering the time-aligned feature data sequence to obtain a spatially registered feature mapping; quantifying the interaction of multiple physical fields on the spatially registered feature mapping to obtain a multi-physical field coupling coefficient matrix, and performing tensor decomposition and reconstruction of the multi-physical field coupling coefficient matrix to obtain a multi-physical field coupling data matrix.

[0010] Optionally, in a fifth implementation of the first aspect of the present invention, the multi-physical field coupling data matrix is ​​subjected to separate iterative solutions of multiple physical field control equations to obtain the multi-physical field spatiotemporal distribution state of the radiator, including: extracting radiator structural parameters and thermophysical property parameters from the multi-physical field coupling data matrix to obtain fin configuration data and temperature-related thermal conductivity distribution of the radiator to be detected, and performing multi-scale adaptive grid division on the fin configuration data to obtain a multi-physical field calculation grid; based on the set of thermal cycle boundary conditions corresponding to the radiator to be detected, solving and calculating the heat conduction equations for the multi-physical field calculation grid and the temperature-related thermal conductivity distribution to obtain the temperature gradient distribution and fluid pressure distribution of each part of the radiator to be detected, performing thermal strain calculation on the temperature gradient distribution to obtain the initial strain field of thermal expansion, and performing surface force conversion on the fluid pressure distribution to obtain the mechanical load distribution corresponding to the inner surface of the radiator to be detected; The elastic-plastic mechanics equations are solved and calculated for the initial strain field of thermal expansion and the mechanical load distribution to obtain the multi-physical field spatiotemporal distribution state of the radiator.

[0011] Optionally, in a sixth implementation manner of the first aspect of the present invention, the heat dissipation cycle load and environmental action calculation of the multi-physical field spatiotemporal distribution state to obtain the damage evolution state of the radiator includes: performing cyclic parameter extraction on the stress distribution in the multi-physical field spatiotemporal distribution state to obtain a cyclic stress characteristic parameter distribution, and performing thermal-mechanical coupling calculation on the cyclic stress characteristic parameter distribution and the temperature field distribution in the multi-physical field spatiotemporal distribution state to obtain the cyclic damage accumulation rate of each part of the radiator to be tested; performing time integral calculation on the cyclic damage accumulation rate to obtain a fatigue damage distribution map corresponding to the thermal cycle, and performing electrochemical corrosion calculation on the temperature field distribution and the stress distribution to obtain the corrosion current density distribution on the surface of the radiator; performing material thinning rate calculation on the corrosion current density distribution to obtain a material loss distribution map on the surface of the radiator, and performing damage contribution evaluation and weighted combination on the fatigue damage distribution map and the material loss distribution map to obtain the damage evolution state of the radiator.

[0012] Optionally, in a seventh implementation manner of the first aspect of the present invention, based on the damage evolution state, the multi-dimensional damage sensitive area of ​​the radiator to be tested is identified to generate a multi-dimensional risk distribution map, including: performing strength risk decomposition on the damage evolution state to obtain the static strength damage distribution of the radiator, and performing thermal fatigue risk decomposition on the damage evolution state to obtain the thermal cycle fatigue damage distribution of the radiator, and performing corrosion risk decomposition on the damage evolution state to obtain the material thinning damage distribution of the radiator; performing threshold judgment on the static strength damage distribution to obtain the high stress risk area of ​​the radiator, and performing remaining life calculation on the thermal cycle fatigue damage distribution to obtain the fatigue sensitivity area of ​​the radiator, and performing penetration time prediction on the material thinning damage distribution to obtain the corrosion sensitive area of ​​the radiator, and spatially superimposing the high stress risk area, the fatigue sensitivity area and the corrosion sensitive area to obtain a multi-dimensional risk distribution map of the radiator.

[0013] Optionally, in an eighth implementation manner of the first aspect of the present invention, based on the historical failure data of the same type of radiators corresponding to the radiator to be detected, the multi-dimensional risk distribution map is subjected to failure comparison and life calculation to generate a comprehensive strength detection result, including: extracting strength characteristic parameters from the multi-dimensional risk distribution map to obtain a strength risk characteristic vector of the radiator, and performing regional risk value statistics on the multi-dimensional risk distribution map to obtain a risk level distribution of each region of the radiator, and based on the historical failure data of the same type of radiators corresponding to the radiator to be detected, similarity matching is performed on the strength risk characteristic vector to obtain a potential strength failure mode, and position marking is performed on high-risk areas in the risk level distribution to obtain the coordinates of key weak parts of the radiator; the damage accumulation equation is solved for the potential strength failure mode to obtain the remaining life prediction values ​​corresponding to multiple strength failure types, and the minimum value is screened for each of the remaining life prediction values ​​to obtain a strength life assessment result of the radiator; strength level determination is performed on the risk level distribution and the strength life assessment result to obtain a comprehensive strength level of the radiator, and the comprehensive strength level and the coordinates of the key weak parts are reported and integrated to obtain a strength detection report of the radiator.

[0014] A second aspect of the present invention provides a radiator strength detection device, the radiator strength detection device comprising: a data acquisition module, used to acquire multi-scale structural feature data and multi-environment coupling material feature data of the radiator to be detected, and to perform adaptive wavelet denoising and structural correlation mapping on the multi-scale structural feature data and the multi-environment coupling material feature data to obtain a heat-fluid-structure multi-physics field fusion data set; a feature extraction module, used to extract multi-dimensional radiator features from the heat-fluid-structure multi-physics field fusion data set to obtain a multi-level feature vector set, and to perform spatiotemporal alignment and interactive mapping of multi-physics field coupling features on the multi-level feature vector set to obtain a multi-level feature vector set. A physical field coupling data matrix; a physical field analysis module, used to perform separate iterative solutions to multiple physical field control equations on the multi-physical field coupling data matrix to obtain the multi-physical field spatiotemporal distribution state of the radiator, and perform heat dissipation cycle load and environmental action calculations on the multi-physical field spatiotemporal distribution state to obtain the damage evolution state of the radiator; a strength assessment module, used to identify the multi-dimensional damage sensitive area of ​​the radiator to be tested based on the damage evolution state, generate a multi-dimensional risk distribution map, and perform failure comparison and life calculation on the multi-dimensional risk distribution map based on the historical failure data of the radiator to be tested corresponding to the same type of radiator, to generate a comprehensive strength test result.

[0015] The above-mentioned radiator strength detection equipment and detection method. In the embodiment of the present invention, by acquiring and processing the multi-scale structural features and multi-environment coupling material features of the radiator to be detected, data denoising and structural correlation mapping are realized, and a heat-flow-structure multi-physics field fusion data set is generated; then the data set is subjected to feature extraction and spatiotemporal alignment processing to construct a multi-physics field coupling data matrix; then the separation iteration method is used to solve the multi-physics field control equation group, and the heat dissipation cycle load calculation is performed to obtain the damage evolution state of the radiator; finally, the sensitive area is identified based on the damage state, and the life calculation is performed in combination with the historical failure data, and the comprehensive strength detection result is output. Through multi-level data processing and feature analysis, the problem of accurate evaluation of radiator strength detection is solved, especially in terms of thermal cycle fatigue, material thinning and stress distribution, the multi-physics field coupling characteristics and failure mechanism of the radiator are fully considered, and the detection accuracy is effectively improved; and the multi-dimensional feature extraction and risk distribution mapping strategy are adopted, which not only realizes the coupling analysis between physical fields, but also enhances the reliability of life prediction; in addition, through the multi-dimensional risk distribution map and failure mode matching, the weak links and life limiting factors are accurately identified, thereby realizing the high-precision detection of radiator strength as a whole.

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

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

[0018] Figure 1 It is a schematic diagram of a first embodiment of a heat sink strength detection method according to an embodiment of the present invention; Figure 2 Schematic diagram of an embodiment of a radiator strength detection device in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device end including a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or device ends.

[0021] To facilitate understanding of this embodiment, the specific process of the embodiment of the present invention is described below. Figure 1 , a first embodiment of the heat sink strength detection method in the embodiment of the present invention includes: 101. Acquire multi-scale structural feature data and multi-environment coupling material feature data of the radiator to be tested, and perform adaptive wavelet denoising and structural correlation mapping on the multi-scale structural feature data and the multi-environment coupling material feature data to obtain a heat-fluid-structure multi-physics field fusion data set; The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0022] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0023] In this embodiment, a multi-spectral three-dimensional scan is performed on the radiator to be inspected to obtain surface geometric point cloud data of a preset spatial resolution, and the surface geometric point cloud data is stereoscopically spliced ​​from multiple angles to obtain a three-dimensional model of the outer surface of the radiator, and a micro-tomography scan is performed on the radiator to be inspected to obtain internal three-dimensional structure data of a preset radiation scanning resolution, and a high-frequency ultrasonic phased array detection is performed on the radiator to be inspected in the stress concentration area of ​​the radiator structure to obtain internal micro-defect distribution data; a quasi-static mechanical property test is performed on the radiator to be inspected in multiple temperature domains to obtain temperature-related mechanical property data of the radiator material, and a thermal-mechanical coupled cyclic load test is performed on the radiator to be inspected to obtain thermal cycle fatigue characteristic data and environmental sensitivity data of the radiator material, wherein the multi-scale structural characteristic data includes the three-dimensional model of the outer surface of the radiator, the internal three-dimensional structure data and the internal micro-defect distribution data, and the multi-environment coupling material characteristic data includes temperature-related mechanical property data, thermal cycle fatigue characteristic data and environmental sensitivity data. Environmental sensitivity data; adaptive wavelet filtering is performed on the surface geometric point cloud data to obtain denoised surface data corresponding to the structural edge features, and the surface roughness distribution and local curvature parameters in the three-dimensional model of the radiator outer surface are extracted to obtain surface quality evaluation index data, and the internal three-dimensional structure data are iteratively reconstructed and multi-scale contrast enhanced to obtain a resolution-enhanced internal structure three-dimensional voxel model, and the internal micro-defect distribution data are phase corrected and time inverted to obtain an enhanced internal defect distribution map, and the denoised surface data, surface quality evaluation index data, enhanced internal structure three-dimensional voxel model and enhanced internal defect distribution map are spatially registered to obtain a multi-scale structural model from macro to micro; anomaly detection and smooth spline fitting are performed on the multi-environment coupled material feature data to obtain the material property distribution function, and the multi-scale structure model and the material property distribution function are correlated and mapped to obtain a thermal-fluid-structure multi-physics field fusion data set associated with the structure and the material. The spatial resolution here refers to the accuracy index when the 3D laser scanning system acquires surface geometric point cloud data, which means the minimum distinguishable distance between two adjacent sampling points (such as 0.05mm); the radiation scanning resolution here refers to the accuracy index when the industrial CT system performs micro-tomography scanning inside the radiator (such as 0.1mm).

[0024] In practical applications, a multispectral 3D scanning system equipped with visible light, near-infrared and far-infrared sensors is first used to perform a full-scale scan of the radiator to be inspected. During the scanning process, the laser spot diameter is controlled to be less than 0.05mm to ensure that the surface geometry point cloud data of the preset spatial resolution is obtained. At the same time, the reflectivity information at different wavelengths is recorded for subsequent analysis of the material properties of the radiator surface. The point cloud data obtained contains the complete geometry information and material reflection characteristics of the outer surface of the radiator. Subsequently, the iterative closest point (ICP) registration algorithm and multi-view fusion technology are used to perform stereo stitching of point cloud data from different angles to eliminate data loss caused by occlusion, and a high-precision mesh reconstruction algorithm is used to generate a high-precision The 3D model of the outer surface of the radiator is obtained, thereby fully characterizing the external geometric features of the radiator; and the radiator to be tested is placed in the industrial CT system for micro-tomography. The energy and beam intensity of the X-ray source are adjusted to adapt to the transmission characteristics of the radiator material, and the pixel size of the detector is set to 0.1mm to obtain the internal 3D structure data of the preset radiation scanning resolution. The collected CT projection data is reconstructed into a 3D voxel model through the filtered back-projection algorithm, which clearly shows the internal flow channel structure, wall thickness distribution and structural connection status of the radiator; and according to the typical structural characteristics and engineering experience of the radiator, the structural stress concentration areas of the radiator, such as the root of the fin, the welding joint and the flow channel corner, are identified. The key areas are finely inspected using 5-15MHz high-frequency ultrasonic phased array technology. By adjusting the incident angle and focusing depth of the ultrasonic wave, high-sensitivity detection of internal micro-defects is achieved, and internal micro-defect distribution data containing defect location, size and type information is obtained; then the material properties of the radiator to be inspected are characterized, that is, samples of standard size are cut from the same type of radiator in advance, and multi-temperature domain quasi-static mechanical properties tests are performed using an electronic universal testing machine equipped with a temperature-controlled environmental chamber. The test temperature range covers -40°C to 250°C, and test points are set at intervals of 20°C. Tensile, compression and shear tests are performed at each temperature point to obtain the elastic modulus and yield strength of the radiator material at different temperatures. The temperature-related mechanical property data such as strength, tensile strength and Poisson's ratio are collected, and the samples are subjected to thermal-mechanical fatigue tests. By applying temperature cycles and mechanical loads at the same time, the temperature cycle range is set from room temperature to 200°C, the stress ratio range is -1 to 0.1, and three loading modes of temperature and stress are designed: in-phase, anti-phase and 90° phase difference. The strain evolution and damage accumulation process of the samples under cyclic loads are recorded in real time, so as to obtain the thermal cycle fatigue characteristic data of the radiator material. At the same time, the samples are placed in a corrosion test chamber simulating the actual working environment. The environmental parameters include humidity, pH value and electrolyte concentration. Electrochemical impedance spectroscopy test is carried out under temperature cycling conditions to evaluate the environmental sensitivity data of the material.

[0025] Secondly, an adaptive wavelet filtering algorithm is used to process the surface geometric point cloud data, that is, by selecting appropriate wavelet basis functions (such as Daubechies wavelets) and adaptive threshold determination methods, accurate denoising of noisy point cloud data is achieved, and high-quality denoised surface data is obtained that retains key geometric features such as the fin edges and weld contours of the radiator and removes random measurement noise; and the local neighborhood analysis method is used to calculate the roughness parameters of each point on the surface of the radiator outer surface three-dimensional model, including the arithmetic mean roughness Ra and the ten-point average roughness Rz. At the same time, the differential geometry method is used to calculate the principal curvature and Gaussian curvature distribution, and these parameters are mapped to the radiator surface to form surface quality evaluation index data (wherein these data can effectively reflect the surface processing quality of the radiator and potential stress concentration areas); and the super-resolution iterative reconstruction technology is used for the internal three-dimensional structure data to improve the spatial resolution, that is, the three-dimensional volume data is decomposed into different frequency components, and then adaptive enhancement is applied to each frequency component, the high-frequency component enhances the internal microstructure, the medium-frequency component enhances the tissue boundary, and the low-frequency component maintains the overall contrast distribution, and finally the enhanced After the multi-frequency components are reorganized, a three-dimensional voxel model of the internal structure with significantly improved resolution and contrast is obtained, which clearly shows the detailed features of the micro-flow channels, wall thickness changes and connection structures inside the radiator; and the phase correction algorithm is applied to the internal micro-defect distribution data to compensate for the phase distortion caused by the propagation of sound waves in heterogeneous materials, so as to improve the accuracy of defect imaging. Subsequently, the time reversal technology is used to use the principle of reverse propagation of wave field characteristics and focusing at the defect position to refocus the scattered waves to the scattering source position, significantly enhancing the signal intensity and spatial resolution of micro-defects, and obtaining an enhanced image that can accurately characterize the distribution of internal defects such as microcracks, pores and inclusions; thereby, these data sets from different detection methods and with different spatial resolutions are spatially registered, that is, the spatial correspondence between the denoised surface data and the internal structure voxel model is established based on geometric feature points and shape descriptors, and then the internal defect distribution map is mapped to a unified coordinate system using rigid registration and non-rigid deformation technology, and the spatial association between surface quality evaluation indicators and internal structures is established, finally forming a multi-scale structural model from centimeter-level macro-geometric features to micron-level internal defects;At the same time, the anomaly detection algorithm based on Mahalanobis distance and local density is applied to the characteristic data of multi-environment coupling materials to identify and eliminate outliers in material testing. Then, cubic B-spline fitting is used to establish a continuous mechanical property-temperature relationship function for temperature-related mechanical property data. The Chaboche nonlinear cumulative damage model is used for parameter fitting of thermal cycle fatigue characteristic data. At the same time, multivariate regression analysis is used to establish the relationship function between corrosion rate and temperature and pH value for environmental sensitivity data. The material property distribution function that describes the mechanical behavior of materials under multi-environment conditions is formed comprehensively. Then, by establishing a mapping relationship between structural position and material properties, the multi-scale structural model is associated and integrated with the material property distribution function to form a heat-fluid-structure multi-physics field fusion data set containing geometric features, internal structure, defect distribution and local material properties. ;

[0026] 102. Extract multi-dimensional heat sink features from the heat-fluid-structure multi-physics field fusion data set to obtain a multi-level feature vector set, and perform spatiotemporal alignment and interactive mapping of multi-physics field coupling features on the multi-level feature vector set to obtain a multi-physics field coupling data matrix; In this embodiment, the heat-fluid-structure multi-physics field fusion data set is geometrically decomposed to obtain the heat sink structure characteristic function, and the internal structure data in the heat-fluid-structure multi-physics field fusion data set is topologically extracted to obtain the flow channel characteristic distribution function, and the defect data in the heat-fluid-structure multi-physics field fusion data set is spatially clustered and probability distribution fitting is performed to obtain the microscopic defect characteristic distribution function, and the material property data in the heat-fluid-structure multi-physics field fusion data set is key parameter extracted to obtain the material temperature-related characteristic parameter set; the heat sink structure characteristic function, the flow channel characteristic distribution function, the microscopic defect characteristic distribution function and the material temperature-related characteristic parameter set are organized and integrated to obtain a multi-level feature vector set; various features in the multi-level feature vector set are synchronized with the time base to obtain a time-aligned feature data sequence, and the time-aligned feature data sequence is spatially aligned to obtain a feature map after spatial alignment; the interaction of multiple physical fields is quantified for the feature map after spatial alignment to obtain a multi-physics field coupling coefficient matrix, and the multi-physics field coupling coefficient matrix is ​​tensor decomposed and reconstructed to obtain a multi-physics field coupling data matrix.

[0027] In practical applications, the heat-fluid-structure multi-physics field fusion data set is first geometrically decomposed. The complex geometric shape of the radiator is decomposed into a reference shape and a series of characteristic deformation modes by combining principal component analysis (PCA) and eigenmode decomposition. These modes can efficiently express the key geometric features of the radiator. Then, the cross-sectional profile is extracted at preset intervals (usually 1 mm) along the main axis of the radiator, and the fin height distribution function H(z, θ), fin thickness distribution function T(z, θ) and fin spacing distribution function S(z, θ) on the cross section are calculated. At the same time, Fourier transform is used to obtain the fin height distribution function H(z, θ), fin thickness distribution function T(z, θ) and fin spacing distribution function S(z, θ). The leaf shape descriptor and Gaussian curvature distribution function are used to analyze the surface morphological characteristics of the radiator, and finally these parameterized expressions are integrated into the radiator structure characteristic function (wherein, this function can compactly and accurately describe the overall and local geometric characteristics of the radiator); and the topological skeleton extraction algorithm is applied to the internal structure data in the fused data set. The voxel refinement algorithm is used to gradually corrode the internal flow channel voxel model to the center line to keep the topology unchanged, and a one-dimensional center line skeleton is obtained. Then, the Voronoi diagram analysis method is used to calculate the distance distribution between the center line of the flow channel and the wall, forming the hydraulic radius distribution function Rh(l). Where l is the cumulative length along the skeleton line, and then the shape factor distribution function Sf(l) of each cross section is calculated. This function is defined as the ratio of the square of the cross-sectional perimeter to the area, which quantifies the influence of the flow channel cross-sectional shape on the flow resistance, and the wall thickness change rate function dW / dl(l) is obtained by differential calculation. These functions together constitute the flow channel characteristic distribution function that describes the internal flow channel geometric characteristics and fluid heat transfer characteristics of the radiator; and for the defect data in the multi-physics field fusion data set, the density-based spatial clustering algorithm is applied to identify the spatial distribution pattern of defects, so as to effectively distinguish between clustered and random distribution in the radiator type defects, while avoiding presetting the number of defect categories. After clustering, the spatial density distribution function of defects ρd(x,y,z) is calculated to describe the spatial variation of the number of defects per unit volume, and the defect size distribution function fs(d) is fitted by the kernel density estimation method to quantify the probability density of defects of different sizes. At the same time, the defect orientation distribution function fo(θ,φ) is fitted by Fisher distribution to characterize the spatial orientation characteristics of directional defects such as microcracks. These functions are combined to form a microscopic defect characteristic distribution function to comprehensively characterize the type, distribution, size and orientation characteristics of defects inside the radiator.And extract key parameters of the material property data in the fused data set, that is, by using the sensitivity analysis method to identify the material parameters that have the most significant impact on the strength of the radiator, and then extract the elastic modulus temperature gradient, yield strength temperature gradient and thermal expansion coefficient from the temperature-related mechanical property data, extract the fatigue strength coefficient, fatigue strength index, fatigue ductility coefficient and fatigue ductility index and other Coffin-Manson equation parameters from the thermal cycle fatigue characteristic data, and extract the functional relationship between the corrosion current density and temperature and stress and the stress corrosion cracking sensitivity coefficient from the environmental sensitivity data. These parameters together constitute the material temperature-related characteristic parameter set, which comprehensively represents the Characterize the evolution of mechanical properties of materials under multi-field coupling conditions; then organize and integrate the heat sink structure characteristic function, flow channel characteristic distribution function, micro defect characteristic distribution function and material temperature-related characteristic parameter set, that is, adopt a hierarchical feature organization structure, the macro layer contains the overall geometric parameters and main structural characteristics of the heat sink, the meso layer contains flow channel characteristics and regional material characteristics, and the micro layer contains defect characteristics and local material parameters. By establishing the correlation between the features, a complete feature topology network is formed, and the tensor representation method is used to organize the features of each layer into a unified mathematical form, and a feature vector set that comprehensively characterizes the multi-level characteristics of the heat sink structure-material-defect is obtained. ;

[0028] Secondly, the time reference of various features in the multi-level feature vector set is synchronized, that is, the dynamic time warping algorithm (DTW) is used to perform nonlinear time expansion and contraction transformation on various feature data sequences, and the heat sink structure characteristic function, flow channel characteristic distribution function, micro defect characteristic distribution function and material temperature-related characteristic parameter set are mapped to the same time axis, while retaining their respective time evolution characteristics, and compensating for the system delay between different measuring devices. The optimal time offset is determined by cross-correlation analysis, thereby solving the asynchronous problem of the data acquisition system, and finally obtaining a feature data sequence aligned in the time dimension, and then using multimodal image. The registration technology performs spatial registration on the time-aligned feature data sequence, establishes a spatial reference coordinate system based on the geometric marking points of the radiator, and then maps the feature data of different spatial resolutions and scales to a unified spatial domain through affine transformation and non-rigid registration algorithm. In particular, for the spatial correspondence between micro-defect distribution and macro-structural features, a multi-resolution B-spline deformation field is used to describe the non-rigid registration mapping, and the registration is optimized by minimizing similarity metrics such as mutual information or normalized cross-correlation, thereby obtaining a feature map that is accurately aligned in the spatial dimension; then, based on the feature map after spatial registration, the interaction of multiple physical fields is quantified. That is, by establishing the coupling relationship between thermal field, flow field, structural field and material field, firstly, the thermal-fluid coupling coefficient is constructed using physical theory and numerical simulation to characterize the influence of temperature field on fluid properties and the contribution of fluid to heat transfer; secondly, the thermal-structural coupling coefficient is constructed to describe the relationship between thermal strain caused by temperature field and structural deformation; then the flow-structure coupling coefficient is constructed to characterize the effect of fluid pressure on structural stress and the feedback of structural deformation on flow channel shape; finally, the material-structure coupling coefficient is constructed to describe the influence of material properties on structural response with temperature change. These coupling coefficients together constitute a multi-dimensional coupling coefficient matrix, which fully characterizes The interaction strength and direction between the physical fields are obtained. In order to reduce the computational complexity and extract the main coupling modes, the multi-physics field coupling coefficient matrix is ​​subjected to tensor decomposition. The high-order tensor decomposition methods such as Tucker decomposition or CP decomposition are used to decompose the original coupling matrix into the product of a core tensor and a series of factor matrices. The data is compressed and reduced in dimension by retaining the main modes with an energy share of more than 95%. Subsequently, the tensor is reconstructed based on the retained main modes to obtain the multi-physics field coupling data matrix after dimensionality reduction but retaining the key coupling information. This matrix can accurately capture the interaction mechanism of various physical fields of the radiator under complex working conditions.

[0029] 103. Perform separate iterative solutions to multiple physical field control equations on the multi-physics field coupling data matrix to obtain the multi-physics field spatiotemporal distribution state of the radiator, and calculate the heat dissipation cycle load and environmental effects on the multi-physics field spatiotemporal distribution state to obtain the damage evolution state of the radiator; In this embodiment, the radiator structural parameters and thermophysical property parameters are extracted from the multi-physics field coupling data matrix to obtain the fin configuration data and temperature-related thermal conductivity distribution of the radiator to be tested, and the fin configuration data is subjected to multi-scale adaptive grid division to obtain a multi-physics field calculation grid; based on the thermal cycle boundary condition set corresponding to the radiator to be tested, the heat conduction equation is solved and calculated for the multi-physics field calculation grid and the temperature-related thermal conductivity distribution to obtain the temperature gradient distribution and fluid pressure distribution of each part of the radiator to be tested, and the thermal strain calculation is performed on the temperature gradient distribution to obtain the initial strain field of thermal expansion, and the surface force conversion is performed on the fluid pressure distribution to obtain the mechanical load distribution corresponding to the inner surface of the radiator to be tested; the elastic-plastic mechanics equation is solved and calculated for the initial strain field of thermal expansion and the mechanical load distribution to obtain. to the multi-physics field spatio-temporal distribution state of the radiator; extract the cyclic parameters of the stress distribution in the multi-physics field spatio-temporal distribution state to obtain the cyclic stress characteristic parameter distribution, and perform thermal-mechanical coupling calculation on the cyclic stress characteristic parameter distribution and the temperature field distribution in the multi-physics field spatio-temporal distribution state to obtain the cyclic damage accumulation rate of each part of the radiator to be tested; perform time integral calculation on the cyclic damage accumulation rate to obtain the fatigue damage distribution map corresponding to the thermal cycle, and perform electrochemical corrosion calculation on the temperature field distribution and stress distribution to obtain the corrosion current density distribution on the surface of the radiator; perform material thinning rate calculation on the corrosion current density distribution to obtain the material loss distribution map on the surface of the radiator, and perform damage contribution evaluation and weighted combination on the fatigue damage distribution map and the material loss distribution map to obtain the damage evolution state of the radiator. The above-mentioned thermal cycle boundary condition set refers to a data set that describes the time-varying changes in parameters such as temperature, flow rate and pressure experienced by the radiator in the actual working environment (such as inlet temperature conditions, flow rate change conditions, ambient temperature conditions, thermal cycle frequency, pressure boundary conditions, etc.).

[0030] In practical applications, the key geometric features and material parameters are first extracted from the coupled data matrix through a feature decoding algorithm. The structural parameter extraction is carried out by combining principal component analysis and non-negative matrix decomposition to recover the key configuration data of the radiator, such as the fin spacing function, fin thickness distribution function and fin shape coefficient, from the spatial mode of the coupled data matrix. At the same time, the distribution function of the material's thermal conductivity varying with temperature is extracted from the temperature-related mode. These functions are usually expressed in the form of piecewise polynomials or exponential functions, which accurately describe the thermal conductivity characteristics of the radiator material at different temperatures. These structural parameters and thermophysical parameters together constitute the basic data for subsequent multi-physics field calculations. Based on the extracted fin configuration data, a basic grid model is constructed according to the geometric characteristics of the radiator. Then, an adaptive grid refinement strategy based on an error estimator is used to refine the grid in the case of large structural stress gradients, drastic temperature changes and fluid Local mesh encryption is performed in key areas such as boundary layers, especially in stress concentration areas such as fin roots, welded joints and flow channel bends. The mesh size can be refined to 0.05mm, while coarser meshes are used in areas with gentle stress gradients to improve computational efficiency. At the same time, boundary layer mesh technology is used near the boundary of the fluid domain to ensure that the interaction between the fluid and the solid surface can be accurately captured, and finally a multi-physics field computational grid system including structural field, thermal field and flow field is generated; based on the generated multi-physics field computational grid and the extracted temperature-dependent thermal conductivity distribution, combined with the thermal cycle boundary condition set of the radiator in the actual working environment, including cosine fluctuations of inlet temperature of 80-110℃, step changes of flow rate of 0.5-2L / min and random fluctuations of ambient temperature of -40-60℃, the finite element method is used to solve the non-steady-state heat conduction equation. The process uses a group of partial differential equations containing time-varying boundary conditions: ; Where ρ is the material density, is the specific heat capacity, T is the temperature field, t is the time, k(T) is the temperature-related thermal conductivity, Q is the internal heat source term, and the equation is iteratively solved by numerical methods such as the conjugate gradient method or the multigrid method. At the same time, the boundary conditions of the solid-fluid interface are considered. The fluid pressure distribution and velocity field distribution are obtained by solving the fluid dynamics equation, and the temperature gradient distribution and fluid pressure distribution of each part of the radiator that change with time are obtained. Then, based on the calculated temperature gradient distribution and the thermal expansion coefficient of the radiator material, the thermal strain field is calculated. , where α(T) is the temperature-dependent thermal expansion coefficient, Tref is the reference temperature (usually 20°C), and T is the current temperature. The thermal strain field that characterizes the initial deformation state of the radiator under the action of temperature change is obtained. At the same time, the fluid pressure distribution is converted into the force distribution acting on the inner surface of the radiator through the fluid-solid coupling algorithm to obtain the mechanical load distribution that characterizes the fluid dynamics load; the initial strain field of thermal expansion and the mechanical load distribution are then used as the input conditions of the elastic-plastic mechanics equation, and the incremental elastic-plastic constitutive equation considering material nonlinearity and cyclic hardening effects is used to solve it: ; Where dε is the total strain increment, is the elastic strain increment, is the plastic strain increment, The von Mises yield criterion and kinematic hardening rule are used to determine the plastic deformation behavior of the material. The Newton-Raphson iteration method is used to solve the nonlinear equations to calculate the stress field σ(x, y, z, t), plastic strain field εp(x, y, z, t) and total deformation field u(x, y, z, t) of the radiator under the combined action of thermal cycle and fluid load. The distribution of these field quantities over time and space together constitutes the multi-physical field spatiotemporal distribution state of the radiator. Furthermore, by analyzing the key moments (such as temperature peak point, maximum temperature change rate point, pressure fluctuation point, etc.), the most unfavorable working condition and stress concentration area of ​​the radiator under thermal cycle conditions are identified, providing basic data for subsequent damage evolution analysis and strength assessment, especially for identifying the local stress peak area caused by the superposition of thermal stress and mechanical stress. These areas are usually the potential failure starting points of the radiator and need to be focused on and processed in the strength assessment.

[0031] Secondly, the rain flow counting method is used to extract the cyclic parameters of the stress distribution in the multi-physical field spatiotemporal distribution state from the time-varying stress history. That is, for the stress time history σ(x, y, z, t) of each finite element node in the radiator, a "rain flow matrix" is constructed to record the number of cycles of different amplitude and average value combinations, and then the stress amplitude distribution σa(x, y, z), average stress distribution σm(x, y, z) and stress ratio distribution R(x, y, z) of each node are calculated. These distributions together constitute the cyclic stress characteristic parameter distribution, and then the obtained cyclic stress characteristic parameter distribution is combined with the temperature field distribution in the multi-physical field spatiotemporal distribution state to perform thermal-mechanical coupled fatigue damage calculation. This process is based on the modified Coffin-Manson equation and Smith-Watson-Topper parameters, considering the influence of temperature on the fatigue properties of the material. For each node in the radiator, the fatigue damage increment per unit cycle is calculated. : ; Among them, Nf(T) is the fatigue life at a given temperature T, b and c are the temperature-related fatigue strength index and fatigue ductility index, respectively, and εf'(T) is the temperature-related fatigue ductility coefficient. The cyclic damage accumulation rate distribution of each part of the radiator under thermal cycle conditions is obtained, which characterizes the fatigue damage rate at different positions of the radiator; then the cyclic damage accumulation rate is calculated by time integration, and based on Miner's linear cumulative damage theory, the cumulative fatigue damage value of each node during the expected service life is calculated: ; Among them, f(τ) is the thermal cycle frequency function, which characterizes the frequency of cyclic loads at different time points. The integral is solved by the numerical integration method to obtain the thermal cycle fatigue damage distribution diagram that characterizes the cumulative fatigue damage degree of each part of the radiator. Based on the temperature field distribution and stress distribution in the multi-physical field space-time distribution state, combined with the electrochemical corrosion kinetics theory, the corrosion current density distribution on the radiator surface is calculated. This process takes into account the comprehensive effects of temperature, stress and electrochemical environment (such as pH value, oxygen concentration, etc.) on the corrosion process, and uses the modified Butler-Volmer equation to describe the electrochemical reaction kinetics: ; in, is the exchange current density, is the anode Tafel slope, V-Veq is the overpotential, γ is the stress influence coefficient, σ(x, y, z) is the stress distribution, R is the gas constant, T is the absolute temperature, and f(T) is the temperature correction function. The corrosion current density distribution at each point on the radiator surface is obtained by solving the equation. Based on the calculated corrosion current density distribution, the material thinning rate is calculated using Faraday's law to obtain the material thinning rate distribution at each point on the radiator surface. The material loss amount within the expected service life is calculated by time extrapolation to obtain the material loss distribution diagram on the radiator surface. The contribution of fatigue damage and corrosion damage to the overall damage is then determined based on typical failure statistics of the radiator. Relative contribution weights. For example, for a specific type of radiator, the fatigue damage weight may be determined to be 0.7 and the corrosion damage weight to be 0.3 based on historical failure case analysis. The two damage distributions are then weighted and combined according to these weights to obtain the overall damage distribution of the radiator. The time evolution characteristics of damage accumulation are considered to construct a complete description of the damage evolution state, which includes not only the spatial distribution information of the damage, but also the development trend of the damage over time. This is especially important for identifying damage hotspots under the combined effects of thermal cycles and corrosive environments. These areas are usually the locations where failure is most likely to occur in the actual service of the radiator and require special attention in strength assessment.

[0032] 104. Based on the damage evolution state, the multi-dimensional damage sensitive areas of the radiator to be tested are identified, and a multi-dimensional risk distribution map is generated. Based on the historical failure data of the radiator to be tested corresponding to the same type of radiator, the multi-dimensional risk distribution map is compared with the failure and life calculation is performed to generate a comprehensive strength test result.

[0033] In this embodiment, the damage evolution state is subjected to strength risk decomposition to obtain the static strength damage distribution of the radiator, the damage evolution state is subjected to thermal fatigue risk decomposition to obtain the thermal cycle fatigue damage distribution of the radiator, and the damage evolution state is subjected to corrosion risk decomposition to obtain the material thinning damage distribution of the radiator; a threshold judgment is performed on the static strength damage distribution to obtain the high stress risk area of ​​the radiator, the remaining life calculation is performed on the thermal cycle fatigue damage distribution to obtain the fatigue sensitivity area of ​​the radiator, and the penetration time prediction is performed on the material thinning damage distribution to obtain the corrosion sensitive area of ​​the radiator, and the high stress risk area, fatigue sensitivity area and corrosion sensitive area are spatially superimposed to obtain a multi-dimensional risk distribution map of the radiator; the multi-dimensional risk distribution map is subjected to strength feature parameter extraction processing to obtain the strength risk of the radiator. Characteristic vector, and regional risk value statistics of the multi-dimensional risk distribution map, to obtain the risk level distribution of each area of ​​the radiator, and based on the historical failure data of the radiator to be tested corresponding to the same type of radiator, similarity matching is performed on the strength risk characteristic vector to obtain the potential strength failure mode, and the high-risk area in the risk level distribution is marked to obtain the coordinates of the key weak parts of the radiator; the damage accumulation equation is solved for the potential strength failure mode to obtain the remaining life prediction values ​​corresponding to various strength failure types, and the minimum value of each remaining life prediction value is screened to obtain the strength life assessment result of the radiator; the strength level is determined for the risk level distribution and the strength life assessment results to obtain the comprehensive strength level of the radiator, and the comprehensive strength level and the coordinates of the key weak parts are reported and integrated to obtain the strength test report of the radiator.

[0034] In practical applications, the stress intensity factor and safety factor calculation are used to compare the stress components in the damage evolution state with the yield strength and ultimate strength of the material, and the static safety factor of each node is calculated. : , in is the temperature-dependent yield strength, The static strength damage distribution is constructed based on the inverse of the safety factor. This distribution characterizes the strength margin of each part of the radiator under the static load. Based on the cycle counting statistics and cumulative damage theory, the fatigue damage caused by temperature fluctuations and cyclic stress is separated to generate a thermal cycle fatigue damage distribution. This distribution characterizes the fatigue accumulation degree of the radiator under the temperature cycle load. The corrosion risk decomposition of the damage evolution state is carried out to separate the damage components related to the electrochemical corrosion of the material. The material thinning damage distribution is constructed by the time integral of the corrosion current density and the material thinning rate. This distribution reflects the degree of material loss caused by corrosion at each point on the surface of the radiator. Then, the static strength damage distribution is judged by the threshold value. Based on the critical threshold value of the preset safety factor (usually 1.5), the area with a safety factor lower than the threshold is identified as a high stress risk area. Yield failure may occur in this area under extreme working conditions. The thermal cycle fatigue damage distribution is calculated for the remaining life. Based on the modified Miner linear cumulative damage theory, the remaining number of cycles required for each node to reach the critical damage value (usually 1.0) is calculated. : ; Where D(x,y,z) is the current position The cumulative damage value at the point, ΔD(x,y,z) is the damage increment per unit cycle, and the area with the remaining number of cycles less than the preset threshold (usually 50% of the design life) is identified as a fatigue sensitive area, which is prone to fatigue cracks under long-term thermal cycling. The penetration time of the material thinning damage distribution is predicted. Based on the linear corrosion model, the time required for the wall thickness of each node to be thinned to the critical value is calculated, and the area with a penetration time shorter than the preset threshold (usually one-third of the design life) is identified as a corrosion sensitive area, which may leak or rupture under the action of a corrosive environment. Then, the multi-layer risk layer overlay technology is used to map the three risk areas to In a unified three-dimensional spatial coordinate system, a comprehensive risk index is then constructed according to the degree of risk overlap. For areas where risk areas overlap, the comprehensive risk index is a combined function of each individual risk index (such as maximum value, weighted sum or product). Special attention is paid to the intersection of the three risk areas, which have the highest failure risk under the combined action of multiple failure mechanisms. Through this spatial superposition analysis, a multi-dimensional risk distribution map of the radiator is generated. The map intuitively displays the distribution of various risk areas of the radiator in the form of three-dimensional or layered slices, and indicates the risk level by color depth, providing an intuitive and effective visual expression for the comprehensive strength assessment of the radiator and the identification of key weak points.

[0035] Secondly, image processing and feature engineering techniques are used to extract a series of quantitative feature parameters from the risk distribution map, including the volume percentage of high-risk areas, statistical moments of risk values ​​(mean, standard deviation, skewness and kurtosis), spatial distribution entropy of risk areas, risk gradient characteristics and topological characteristics. These parameters constitute a multi-dimensional intensity risk feature vector, which comprehensively characterizes the risk distribution characteristics of the radiator, and performs regional risk value statistics on the multi-dimensional risk distribution map, dividing the radiator into meaningful functional areas (such as fin area, joint area, flow channel area, etc.), and then calculating the average risk value, maximum risk value and risk area percentage in each area, and according to the preset risk level thresholds (such as low risk <0.3, medium risk 0.3-0.7, high risk >0.7) Classify each area and generate the risk level distribution of each area of ​​the radiator. The distribution intuitively displays the risk level of different parts of the radiator in the form of a heat map or layered coloring. Then, based on the historical failure database of the same type of radiators corresponding to the radiator to be tested, the nearest neighbor algorithm or support vector machine classification method in machine learning is used to calculate the similarity between the risk feature vector of the current radiator and various failure modes in the historical cases. The similarity measurement can be Euclidean distance, cosine similarity or Mahalanobis distance, etc. Through similarity sorting and threshold screening, the most likely potential strength failure mode of the current radiator is identified, such as thermal cycle fatigue failure, corrosion penetration failure, thermal stress cracking failure, etc., and the high-risk area in the risk level distribution is marked. The local maximum detection algorithm is used to identify the peak points of the risk value, and the three-dimensional coordinates of these peak points and the corresponding risk values ​​are recorded. These locations are marked as the key weak parts of the radiator and are the focus of subsequent monitoring. In particular, for the locations that are at the intersection of the high stress risk area and the fatigue sensitive area, the failure risk is the highest and needs to be marked first; then the damage accumulation equation is solved for the identified potential strength failure mode. Different failure modes use corresponding physical models for life prediction: for thermal cycle fatigue mode, the modified Miner cumulative damage equation is used to consider the relationship between temperature, stress and number of cycles; for corrosion penetration mode, the electrochemical corrosion rate equation is used to consider the influence of temperature and environmental factors on the corrosion rate; for thermal stress cracking mode, the Paris crack growth equation is used to consider the influence of temperature cycle on crack growth rate. By solving the crack growth equation, the remaining life prediction values ​​under each failure mode are obtained, such as the remaining number of cycles under fatigue failure mode, the penetration time under corrosion failure mode, and the critical crack length reaching time under crack failure mode, etc., and then the remaining life prediction values ​​of each failure mode are screened for the minimum value, and the shortest predicted life is selected as the controlled life of the radiator. At the same time, considering the uncertainty of the prediction, the confidence interval is calculated to obtain the strength life assessment result of the radiator; then based on the risk level distribution and the strength life assessment results, the overall strength of the radiator is graded, and a comprehensive scoring mechanism is usually adopted to calculate the strength comprehensive score S by combining the highest risk value, the percentage of high-risk areas, the ratio of predicted life to design life and other factors:. ; Among them, Rmax is the maximum risk value, Ahigh / Atotal is the volume percentage of the high-risk area, Tpred / Tdesign is the ratio of the predicted life to the design life, and w 1 、w 2 and w 3As the corresponding weight coefficient, the strength level of the radiator is divided into A (excellent, S≥0.85), B (good, 0.70≤S<0.85), C (qualified, 0.60≤S<0.70) and D (unqualified, S<0.60) according to the comprehensive score range, and a structured strength test report is generated in combination with the comprehensive strength level and the coordinates of the key weak parts. The report contains the basic information of the radiator, test conditions, risk distribution map, list of key weak parts, failure mode prediction, life prediction results, strength grade assessment and usage recommendations, etc., to provide relevant basis for the safe use and maintenance decision of the radiator.

[0036] In the embodiment of the present invention, by acquiring and processing the multi-scale structural features and multi-environment coupling material features of the radiator to be tested, data denoising and structural correlation mapping are realized, and a heat-flow-structure multi-physics field fusion data set is generated; then the data set is subjected to feature extraction and spatiotemporal alignment processing to construct a multi-physics field coupling data matrix; then the separation iteration method is used to solve the multi-physics field control equations, and the heat dissipation cycle load is calculated to obtain the damage evolution state of the radiator; finally, the sensitive area is identified based on the damage state, and the life calculation is performed in combination with the historical failure data, and the strength comprehensive test results are output. Through multi-level data processing and feature analysis, the problem of accurate evaluation of radiator strength detection is solved, especially in terms of thermal cycle fatigue, material thinning and stress distribution, the multi-physics field coupling characteristics and failure mechanism of the radiator are fully considered, and the detection accuracy is effectively improved; and the multi-dimensional feature extraction and risk distribution mapping strategy are adopted, which not only realizes the coupling analysis between physical fields, but also enhances the reliability of life prediction; in addition, through the multi-dimensional risk distribution map and failure mode matching, the weak links and life limiting factors are accurately identified, thereby realizing the high-precision detection of the radiator strength as a whole.

[0037] The above describes the heat sink strength detection method in the embodiment of the present invention. The following describes the heat sink strength detection device in the embodiment of the present invention. Figure 2 , an embodiment of a radiator strength detection device in an embodiment of the present invention includes: The data acquisition module 201 is used to acquire multi-scale structural feature data and multi-environment coupling material feature data of the radiator to be detected, and to perform adaptive wavelet denoising and structural correlation mapping on the multi-scale structural feature data and the multi-environment coupling material feature data to obtain a heat-fluid-structure multi-physics field fusion data set; A feature extraction module 202 is used to extract multi-dimensional heat sink features from the heat-fluid-structure multi-physics field fusion data set to obtain a multi-level feature vector set, and perform spatiotemporal alignment and interactive mapping of multi-physics field coupling features on the multi-level feature vector set to obtain a multi-physics field coupling data matrix; The physical field analysis module 203 is used to perform separate iterative solutions to the multi-physical field coupling data matrix for multiple physical field control equations to obtain the multi-physical field spatiotemporal distribution state of the radiator, and to perform heat dissipation cycle load and environmental action calculations on the multi-physical field spatiotemporal distribution state to obtain the damage evolution state of the radiator; The strength assessment module 204 is used to identify the multi-dimensional damage sensitive area of ​​the radiator to be tested based on the damage evolution state, generate a multi-dimensional risk distribution map, and perform failure comparison and life calculation on the multi-dimensional risk distribution map based on the historical failure data of the radiator of the same type corresponding to the radiator to be tested, so as to generate a comprehensive strength test result.

[0038] In the embodiment of the present invention, by acquiring and processing the multi-scale structural features and multi-environment coupling material features of the radiator to be tested, data denoising and structural correlation mapping are realized, and a heat-flow-structure multi-physics field fusion data set is generated; then the data set is subjected to feature extraction and spatiotemporal alignment processing to construct a multi-physics field coupling data matrix; then the separation iteration method is used to solve the multi-physics field control equations, and the heat dissipation cycle load is calculated to obtain the damage evolution state of the radiator; finally, the sensitive area is identified based on the damage state, and the life calculation is performed in combination with the historical failure data, and the strength comprehensive test results are output. Through multi-level data processing and feature analysis, the problem of accurate evaluation of radiator strength detection is solved, especially in terms of thermal cycle fatigue, material thinning and stress distribution, the multi-physics field coupling characteristics and failure mechanism of the radiator are fully considered, and the detection accuracy is effectively improved; and the multi-dimensional feature extraction and risk distribution mapping strategy are adopted, which not only realizes the coupling analysis between physical fields, but also enhances the reliability of life prediction; in addition, through the multi-dimensional risk distribution map and failure mode matching, the weak links and life limiting factors are accurately identified, thereby realizing the high-precision detection of the radiator strength as a whole.

[0039] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0040] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A radiator strength detection method, characterized in that: The radiator strength detection method comprises: Acquire multi-scale structural feature data and multi-environment coupling material feature data of the radiator to be detected, and perform adaptive wavelet denoising and structural correlation mapping on the multi-scale structural feature data and the multi-environment coupling material feature data to obtain a heat-fluid-structure multi-physics field fusion data set; Extracting multi-dimensional heat sink features from the heat-fluid-structure multi-physics field fusion data set to obtain a multi-level feature vector set, and performing spatiotemporal alignment and interactive mapping of multi-physics field coupling features on the multi-level feature vector set to obtain a multi-physics field coupling data matrix; Separately iterate and solve the multi-physics field coupling data matrix by multiple physics field control equations to obtain the multi-physics field spatiotemporal distribution state of the radiator, and calculate the heat dissipation cycle load and environmental effect on the multi-physics field spatiotemporal distribution state to obtain the damage evolution state of the radiator; Based on the damage evolution state, the multi-dimensional damage sensitive areas of the radiator to be tested are identified, and a multi-dimensional risk distribution map is generated. Based on the historical failure data of the radiator to be tested corresponding to the same type of radiator, failure comparison and life calculation are performed on the multi-dimensional risk distribution map to generate a comprehensive strength test result.

2. The heat sink strength detection method according to claim 1, characterized in that: The multi-scale structural characteristic data includes a three-dimensional model of the outer surface of the radiator, internal three-dimensional structural data and internal micro-defect distribution data; the multi-environment coupling material characteristic data includes temperature-related mechanical property data, thermal cycle fatigue characteristic data and environmental sensitivity data; the multi-scale structural characteristic data and multi-environment coupling material characteristic data of the radiator to be tested are obtained, including: Performing a multi-spectral three-dimensional scan on the radiator to be inspected to obtain surface geometric point cloud data of a preset spatial resolution, and performing multi-angle stereoscopic stitching on the surface geometric point cloud data to obtain a three-dimensional model of the radiator's outer surface, and performing a micro-tomographic scan on the radiator to be inspected to obtain internal three-dimensional structure data of a preset radiation scanning resolution, and performing a high-frequency ultrasonic phased array inspection on the radiator structure stress concentration area to obtain internal micro-defect distribution data; The radiator to be tested is subjected to multi-temperature domain quasi-static mechanical property tests to obtain the temperature-related mechanical property data of the radiator material. The radiator to be tested is also subjected to a thermal-mechanical coupling cyclic load test to obtain the thermal cycle fatigue characteristic data and environmental sensitivity data of the radiator material.

3. The heat sink strength detection method according to claim 2, characterized in that: The step of performing adaptive wavelet denoising and structural correlation mapping on the multi-scale structural feature data and the multi-environment coupling material feature data to obtain a thermal-fluid-structure multi-physics field fusion data set includes: Adaptively performing wavelet filtering on the surface geometric point cloud data to obtain denoised surface data corresponding to the structural edge features, extracting the surface roughness distribution and local curvature parameters in the three-dimensional model of the outer surface of the radiator to obtain surface quality assessment index data, iteratively reconstructing and multi-scale contrast enhancement on the internal three-dimensional structure data to obtain a resolution-enhanced three-dimensional voxel model of the internal structure, and performing phase correction and time inversion on the internal micro-defect distribution data to obtain an enhanced internal defect distribution map, and performing spatial registration processing on the denoised surface data, the surface quality assessment index data, the enhanced three-dimensional voxel model of the internal structure and the enhanced internal defect distribution map to obtain a multi-scale structural model from macro to micro; Anomaly detection and smooth spline fitting are performed on the multi-environment coupled material characteristic data to obtain a material property distribution function, and correlation mapping is performed between the multi-scale structural model and the material property distribution function to obtain a thermal-fluid-structure multi-physics field fusion data set that associates the structure with the material.

4. The heat sink strength detection method according to claim 1, characterized in that: The multi-dimensional heat sink feature is extracted from the heat-fluid-structure multi-physics field fusion data set to obtain a multi-level feature vector set, including: Performing geometric morphological decomposition on the thermal-fluid-structure multi-physics field fusion data set to obtain a heat sink structural characteristic function, performing topological skeleton extraction on the internal structure data in the thermal-fluid-structure multi-physics field fusion data set to obtain a flow channel characteristic distribution function, performing spatial clustering and probability distribution fitting on the defect data in the thermal-fluid-structure multi-physics field fusion data set to obtain a microscopic defect characteristic distribution function, and performing key parameter extraction on the material property data in the thermal-fluid-structure multi-physics field fusion data set to obtain a material temperature-related characteristic parameter set; The heat sink structure characteristic function, the flow channel characteristic distribution function, the microscopic defect characteristic distribution function and the material temperature-related characteristic parameter set are organized and integrated to obtain a multi-level characteristic vector set.

5. The heat sink strength detection method according to claim 1, characterized in that: The step of performing spatiotemporal alignment and interactive mapping of multi-physics field coupling features on the multi-level feature vector set to obtain a multi-physics field coupling data matrix includes: Performing time base synchronization on various features in the multi-level feature vector set to obtain a time-aligned feature data sequence, and performing spatial registration on the time-aligned feature data sequence to obtain a feature map after spatial registration; The interaction of multiple physical fields is quantified for the spatially registered feature map to obtain a multi-physical field coupling coefficient matrix, and the multi-physical field coupling coefficient matrix is ​​tensor-decomposed and reconstructed to obtain a multi-physical field coupling data matrix.

6. The heat sink strength detection method according to claim 1, characterized in that: The multi-physics field coupling data matrix is ​​subjected to separate iterative solutions of multiple physics field control equations to obtain the multi-physics field spatiotemporal distribution state of the radiator, including: Extracting the heat sink structural parameters and thermophysical property parameters from the multi-physics field coupling data matrix to obtain the fin configuration data and temperature-related thermal conductivity distribution of the heat sink to be tested, and performing multi-scale adaptive grid division on the fin configuration data to obtain a multi-physics field calculation grid; Based on the thermal cycle boundary condition set corresponding to the radiator to be tested, the multi-physics field calculation grid and the temperature-related thermal conductivity distribution are solved and calculated by heat conduction equation to obtain the temperature gradient distribution and fluid pressure distribution of each part of the radiator to be tested, and the thermal strain calculation is performed on the temperature gradient distribution to obtain the initial strain field of thermal expansion, and the surface force conversion is performed on the fluid pressure distribution to obtain the mechanical load distribution corresponding to the inner surface of the radiator to be tested; The elastic-plastic mechanics equations are solved and calculated for the initial strain field of thermal expansion and the mechanical load distribution to obtain the multi-physical field spatiotemporal distribution state of the radiator.

7. The heat sink strength detection method according to claim 1, characterized in that: The heat dissipation cycle load and environmental action calculation is performed on the spatiotemporal distribution state of the multi-physical field to obtain the damage evolution state of the radiator, including: Performing cyclic parameter extraction on the stress distribution in the multi-physical field spatiotemporal distribution state to obtain a cyclic stress characteristic parameter distribution, and performing thermal-mechanical coupling calculation on the cyclic stress characteristic parameter distribution and the temperature field distribution in the multi-physical field spatiotemporal distribution state to obtain a cyclic damage accumulation rate of each part of the radiator to be tested; Performing time-integrated calculation on the cycle damage accumulation rate to obtain a fatigue damage distribution diagram corresponding to the thermal cycle, and performing electrochemical corrosion calculation on the temperature field distribution and the stress distribution to obtain a corrosion current density distribution on the surface of the radiator; The material thinning rate of the corrosion current density distribution is calculated to obtain a material loss distribution map on the surface of the radiator, and the damage contribution evaluation and weighted combination of the fatigue damage distribution map and the material loss distribution map are performed to obtain the damage evolution state of the radiator.

8. The heat sink strength detection method according to claim 1, characterized in that: The method of identifying the multi-dimensional damage sensitive area of ​​the radiator to be inspected based on the damage evolution state and generating a multi-dimensional risk distribution map includes: Performing strength risk decomposition on the damage evolution state to obtain static strength damage distribution of the radiator, performing thermal fatigue risk decomposition on the damage evolution state to obtain thermal cycle fatigue damage distribution of the radiator, and performing corrosion risk decomposition on the damage evolution state to obtain material thinning damage distribution of the radiator; A threshold judgment is performed on the static strength damage distribution to obtain the high stress risk area of ​​the radiator, and the remaining life is calculated on the thermal cycle fatigue damage distribution to obtain the fatigue sensitivity area of ​​the radiator. The penetration time of the material thinning damage distribution is predicted to obtain the corrosion sensitive area of ​​the radiator, and the high stress risk area, the fatigue sensitivity area and the corrosion sensitive area are spatially superimposed to obtain a multi-dimensional risk distribution map of the radiator.

9. The heat sink strength detection method according to claim 1, characterized in that: Based on the historical failure data of the radiator to be tested corresponding to the same type of radiator, the multi-dimensional risk distribution map is subjected to failure comparison and life calculation to generate a comprehensive strength test result, including: Performing strength characteristic parameter extraction processing on the multi-dimensional risk distribution map to obtain a strength risk characteristic vector of the radiator, and performing regional risk value statistics on the multi-dimensional risk distribution map to obtain a risk level distribution of each area of ​​the radiator, and performing similarity matching on the strength risk characteristic vector based on historical failure data of the radiator to be detected corresponding to the same type of radiator to obtain a potential strength failure mode, and marking the position of the high-risk area in the risk level distribution to obtain the coordinates of the key weak parts of the radiator; Solving the damage accumulation equation for the potential strength failure mode to obtain the remaining life prediction values ​​corresponding to various strength failure types, and performing minimum value screening on each of the remaining life prediction values ​​to obtain the strength life assessment result of the radiator; The strength level is determined based on the risk level distribution and the strength life assessment results to obtain the comprehensive strength level of the radiator, and the comprehensive strength level and the coordinates of the key weak parts are reported and integrated to obtain a strength test report of the radiator.

10. A radiator strength testing device, characterized in that: The radiator strength detection device comprises: A data acquisition module is used to acquire multi-scale structural feature data and multi-environment coupling material feature data of the radiator to be detected, and to perform adaptive wavelet denoising and structural correlation mapping on the multi-scale structural feature data and the multi-environment coupling material feature data to obtain a heat-flow-structure multi-physics field fusion data set; A feature extraction module is used to extract multi-dimensional heat sink features from the heat-fluid-structure multi-physics field fusion data set to obtain a multi-level feature vector set, and to perform spatiotemporal alignment and interactive mapping of multi-physics field coupling features on the multi-level feature vector set to obtain a multi-physics field coupling data matrix; A physical field analysis module is used to perform separate iterative solutions to the multi-physical field coupling data matrix for multiple physical field control equations to obtain the multi-physical field spatiotemporal distribution state of the radiator, and to perform heat dissipation cycle load and environmental action calculations on the multi-physical field spatiotemporal distribution state to obtain the damage evolution state of the radiator; The strength assessment module is used to identify the multi-dimensional damage sensitive area of ​​the radiator to be tested based on the damage evolution state, generate a multi-dimensional risk distribution map, and perform failure comparison and life calculation on the multi-dimensional risk distribution map based on the historical failure data of the radiator of the same type corresponding to the radiator to be tested, so as to generate a comprehensive strength test result.

Citation Information

Cited By

  • Automobile radiator surface corrosion degree evaluation method based on image processing

    CN120318224A

  • Crack detection device and method for precise automobile parts

    CN120404946A

  • Unmanned aerial vehicle electric power inspection image intelligent analysis method and system based on deep learning and multi-modal fusion and medium of unmanned aerial vehicle electric power inspection image intelligent analysis method and system

    CN120726041A

  • Polar ultralow temperature detection method and system based on temperature sensor

    CN120782426A

  • Intelligent path planning system and optimization method for large structure robot welding

    CN120985664A