A detection method and system for the inside of a switch cabinet based on defect reconstruction technology

Abstract

The present invention provides a detection method and system for the inside of a switch cabinet based on defect reconstruction technology, which relates to the field of defect detection technology, including obtaining a cross-sectional boundary temperature curve at the position of the highest temperature point, and using the cross-sectional boundary temperature curve as the temperature input for defect reconstruction and positioning; using a multi-layer equivalent analysis module for processing, and using a multi-layer rectangular parallelepiped structure model as a model input for defect reconstruction and positioning; simulating defects occurring during material manufacturing to obtain prior information on material defects; obtaining preliminary positioning information on potential defects inside the switch cabinet; using known quantities to correct the result value of the piecewise polynomial singular value truncation decomposition method to obtain a corrected result; determining the reconstruction and positioning result of the defect inside the switch cabinet, and clarifying the location and type of the defect based on the reconstruction and positioning result. The beneficial effects of the present invention are to improve the intelligent level of defect detection and realize comprehensive monitoring and management of complex engineering structures.

Description

Technical Field

[0001] The present invention relates to the field of defect detection technology, and in particular to a method and system for detecting the inside of a switch cabinet based on defect reconstruction technology. Background Art

[0002] In the current field of power system maintenance, switchgear, as a key power distribution equipment, undertakes important functions of power distribution, control and protection, so the detection of its internal defects is crucial. Traditional detection methods mainly rely on regular offline inspections and certain online monitoring technologies, such as ultrasonic and infrared thermal imaging detection. Although these technologies can detect surface or near-surface defects to a certain extent, they are often limited by the detection depth and accuracy, and it is difficult to accurately evaluate internal defects. In addition, the existing technology usually adopts empirical formulas or qualitative analysis, which are relatively backward and cannot meet the needs of modern power systems for high reliability and accuracy. Summary of the invention

[0003] The purpose of the present invention is to provide a detection method and system for the inside of a switch cabinet based on defect reconstruction technology to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0004] In a first aspect, the present application provides a method for detecting the inside of a switch cabinet based on defect reconstruction technology, comprising:

[0005] Scan the switch cabinet surface with an infrared thermal imager to obtain the cross-sectional boundary temperature curve at the highest temperature point, and use the cross-sectional boundary temperature curve as the temperature input for defect reconstruction and positioning;

[0006] According to the internal structure and material characteristics of the switch cabinet, the multi-layer equivalent analysis module is used to process it and convert it into a multi-layer rectangular parallelepiped structure model, which is used as the model input for defect reconstruction and positioning.

[0007] Using multi-scale parametric modeling data of woven composite materials, through the progressive damage and failure analysis of thermal-mechanical-oxygen coupling, the defects that occur during the material manufacturing process are simulated to obtain the prior information of material defects;

[0008] Combining the temperature input and the model input, the preset defect reconstruction and positioning module is used to process the information to obtain preliminary positioning information of the potential defects inside the switch cabinet, wherein the preliminary positioning information includes the location, size or shape information of the defect;

[0009] The prior information and the preliminary positioning information are combined to obtain a comprehensive data set for defect positioning, the comprehensive data set is processed according to the differential method, and the processed comprehensive data set is input into the solution equation matrix as a known quantity, and the result value of the piecewise polynomial singular value truncation decomposition method is corrected by using the known quantity to obtain a corrected result;

[0010] According to the corrected results, the reconstructed positioning results of the internal defects of the switch cabinet are determined, and the location and type of the defects are clarified according to the reconstructed positioning results.

[0011] Preferably, the surface of the switch cabinet is scanned by an infrared thermal imager to obtain a cross-sectional boundary temperature curve at the position of the highest temperature point, and the cross-sectional boundary temperature curve is used as a temperature input for defect reconstruction and positioning, which includes:

[0012] Use an infrared thermal imager to scan the surface of the switch cabinet to obtain a temperature distribution map, and then use an image processing algorithm to process the temperature distribution map to obtain the coordinates of the highest temperature point;

[0013] Based on the coordinates, the edge detection algorithm is used to extract the cross-sectional boundary of the highest temperature point from the temperature distribution map, and the temperature curve on the boundary is calculated;

[0014] According to the temperature curve, the curve fitting method is applied to generate a continuous temperature curve as the temperature input for defect reconstruction and positioning.

[0015] Preferably, the switch cabinet internal structure and material properties are processed by using a multi-layer equivalent analysis module to be equivalent to a multi-layer rectangular parallelepiped structure model, and the multi-layer rectangular parallelepiped structure model is used as a model input for defect reconstruction and positioning, including:

[0016] According to the design drawings and material property data of the switch cabinet, a three-dimensional model of the internal structure of the switch cabinet is obtained through geometric modeling processing, wherein the material property data includes data describing the physical, chemical and mechanical properties of the material;

[0017] Based on the three-dimensional model, the material property equivalent algorithm is used to equate the thermal conductivity and mechanical properties of different materials to the parameters of the multi-layer cuboid structure to obtain an equivalent model;

[0018] According to the equivalent parameters, a multi-layer cuboid structure model is generated by the finite element analysis method, and the model is used as the model input for defect reconstruction and positioning.

[0019] Preferably, the multi-scale parameterized modeling data of the braided composite material is used to simulate the defects occurring in the material manufacturing process through the progressive damage and failure analysis of thermal-mechanical-oxygen coupling to obtain the prior information of the material defects, which includes:

[0020] Based on the microstructure and physical property data of the woven composite material, a parametric model of the material is obtained through multi-scale modeling algorithm processing;

[0021] Based on the parameterized model, the thermal-mechanical-oxygen coupling analysis algorithm is used to simulate the stress, thermal distribution and oxygen permeation behavior of composite materials in actual working environments. Combined with material failure theory, the microcrack initiation and propagation paths inside the material are predicted to generate material damage evolution data.

[0022] Based on the damage evolution data, machine learning algorithms are applied to classify and identify potential defect patterns. Statistical analysis methods are used to extract the probability distribution of defect size, location and type from the damage evolution data, generate a priori information library for defect reconstruction and positioning, and then extract prior information on material defects.

[0023] Preferably, the step of extracting prior information of material defects then comprises:

[0024] Apply principal component analysis to filter the prior information and obtain the most representative and discriminative parameters for defect characteristics;

[0025] Based on the parameter selection, the genetic algorithm is used to optimize the defect characteristic parameters, determine the optimal defect characteristic parameters, and use the optimal defect characteristic parameters to train the defect detection model;

[0026] The determined optimal parameters and the trained defect detection model are integrated into the defect reconstruction and positioning module and used as input parameters.

[0027] In a second aspect, the present application also provides a detection system for the inside of a switch cabinet based on defect reconstruction technology, comprising:

[0028] Acquisition module: used to scan the switch cabinet surface with an infrared thermal imager, obtain the cross-sectional boundary temperature curve at the highest temperature point, and use the cross-sectional boundary temperature curve as the temperature input for defect reconstruction and positioning;

[0029] The first processing module is used to process the internal structure and material characteristics of the switch cabinet using a multi-layer equivalent analysis module to convert it into a multi-layer rectangular parallelepiped structure model, and use the multi-layer rectangular parallelepiped structure model as a model input for defect reconstruction and positioning;

[0030] Analysis module: It is used to utilize the multi-scale parametric modeling data of woven composite materials to simulate the defects that occur during the material manufacturing process through the progressive damage and failure analysis of thermal-mechanical-oxygen coupling, and obtain the prior information of material defects;

[0031] The second processing module is used to combine the temperature input and the model input, and process them through a preset defect reconstruction and positioning module to obtain preliminary positioning information of potential defects inside the switch cabinet, wherein the preliminary positioning information includes the location, size or shape information of the defect;

[0032] Solution correction module: used to combine the prior information and the preliminary positioning information to obtain a comprehensive data set for defect positioning, process the comprehensive data set according to the differential method, and input the processed comprehensive data set as a known quantity into the solution equation matrix, and use the known quantity to correct the result value of the piecewise polynomial singular value truncation decomposition method to obtain the corrected result;

[0033] Determination module: used to determine the reconstruction positioning results of the internal defects of the switch cabinet according to the corrected results, and to clarify the location and type of the defects according to the reconstruction positioning results.

[0034] The beneficial effects of the present invention are:

[0035] The present invention realizes accurate modeling and temperature field analysis of the switch cabinet surface and internal structure by combining the temperature distribution map obtained by infrared thermal imaging technology and the structural model of finite element analysis, which significantly improves the accuracy of defect detection; uses differential method to process comprehensive data sets, identifies abnormal changes in temperature and stress distribution, and corrects the results of piecewise polynomial singular value truncation decomposition method through optimization algorithm, further enhancing the accuracy of defect location; simulates the damage and failure behavior of woven composite materials under the action of multi-physical fields, and realizes a comprehensive analysis of materials from micro to macro; the defect reconstruction and positioning results provided by the present invention clarify the location and type of defects, help to formulate more accurate maintenance and overhaul plans, reduce unnecessary disassembly and inspection, and improve maintenance efficiency; the present invention integrates advanced data processing algorithms and machine learning technologies, promotes the application of intelligent and automated detection technologies in power system maintenance, and provides technical support for the development of smart grids; the present invention not only improves the performance of a single detection process, but also provides a solid technical foundation for the long-term stable operation and intelligent development of power systems.

[0036] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0038] Figure 1 It is a schematic flow chart of a detection method inside a switch cabinet based on defect reconstruction technology according to an embodiment of the present invention;

[0039] Figure 2 It is a schematic diagram of the structure of the detection system inside the switch cabinet based on the defect reconstruction technology described in an embodiment of the present invention.

[0040] In the figure: 701, acquisition module; 7011, acquisition unit; 7012, extraction unit; 7013, temperature generation unit; 702, first processing module; 7021, first processing unit; 7022, equivalent unit; 7023, model generation unit; 703, analysis module; 7031, second processing unit; 7032, simulation prediction unit; 7033, identification and analysis unit; 70331, screening unit; 70332, tuning unit; 70333, integration unit; 704, second processing module; 705, solution correction module; 706, determination module. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0042] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0043] Embodiment 1:

[0044] This embodiment provides a method for detecting the inside of a switch cabinet based on defect reconstruction technology.

[0045] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300, step S400, step S500 and step S600.

[0046] S100, scanning the switch cabinet surface with an infrared thermal imager to obtain a cross-sectional boundary temperature curve at the position of the highest temperature point, and using the cross-sectional boundary temperature curve as a temperature input for defect reconstruction and positioning.

[0047] It can be understood that step S100 includes S101, S102 and S103, wherein:

[0048] S101, using an infrared thermal imager to scan the surface of the switch cabinet to obtain a temperature distribution map, and using an image processing algorithm to process the temperature distribution map to obtain the coordinates of the position of the highest temperature point;

[0049] S102, based on the coordinates, using an edge detection algorithm to extract the cross-sectional boundary of the highest temperature point from the temperature distribution map, and calculate the temperature curve on the boundary;

[0050] S103 . According to the temperature curve, a curve fitting method is applied to generate a continuous temperature curve as a temperature input for defect reconstruction and positioning.

[0051] It should be noted that in step S100, a high-resolution infrared thermal imager is used to comprehensively scan the surface of the switch cabinet to capture the temperature field distribution of the equipment during operation and generate an initial temperature distribution map. Image preprocessing algorithms, such as filtering and denoising, can be applied to eliminate random noise and image artifacts introduced during the scanning process and improve the quality of temperature data. Image enhancement techniques, such as histogram equalization or contrast enhancement, are then used to highlight subtle changes in the temperature distribution map, providing a clear image basis for subsequent hotspot identification. Finally, image segmentation algorithms, such as threshold segmentation or region growing, are used to identify temperature abnormality areas, provide preliminary information for determining the highest temperature point, and then accurately obtain the temperature distribution map of the switch cabinet surface, providing high-quality basic data for subsequent hotspot positioning and defect analysis.

[0052] In step S102, the precise coordinates of the highest temperature point are determined. These coordinates indicate the center position of the potential defect area. With the highest temperature point as the center, an edge detection algorithm, such as the Canny operator, is applied to identify the hot spot boundary in the temperature distribution map. This step can reveal the area where the temperature gradient changes most significantly. The cross-sectional boundary is extracted from the edge detection result. This boundary defines the geometric contour of the hot spot area. Finally, along the cross-sectional boundary, the temperature value is sampled and recorded to form a temperature curve. This curve reflects the temperature distribution characteristics of the hot spot area. The temperature distribution map in step S102 is calculated as follows:

[0053]

[0054] Where ρ is the material density, c p is the specific heat capacity, T is the temperature, t is the time, k is the thermal conductivity, Q is the thermal imaging temperature, ∈ is the emissivity of the material, σ is the Stefan-Boltzmann constant, A is the surface area of ​​the object, T is the surface temperature of the object, T env is the ambient temperature.

[0055] It should be noted that in step S103, the discrete temperature curve obtained in step S102 is analyzed, the main trend and characteristic points of the curve are identified, and the discrete data points are fitted using curve fitting techniques, such as polynomial fitting or nonlinear least squares method, to generate a smooth and continuous temperature curve. Appropriate fitting parameters and models are selected to ensure that the fitting results can reflect the characteristics of the original data and provide sufficient smoothness for subsequent analysis. Finally, the continuous curve obtained by fitting is used for further mathematical processing, such as derivation or integration, to extract the differential properties of temperature changes and provide deeper characteristic information for defect reconstruction. Through curve fitting, a continuous and smooth temperature curve can be obtained, which can not only accurately reflect the thermal state of the switch cabinet surface, but also provide an important basis for the quantitative analysis of defects through its differential properties.

[0056] S200, according to the internal structure and material characteristics of the switch cabinet, use the multi-layer equivalent analysis module to process it, and convert it into a multi-layer rectangular parallelepiped structure model, and use the multi-layer rectangular parallelepiped structure model as the model input for defect reconstruction and positioning.

[0057] It can be understood that step S200 includes S201, S202 and S203, wherein:

[0058] S201, obtaining a three-dimensional model of the internal structure of the switch cabinet through geometric modeling according to the design drawings and material property data of the switch cabinet, wherein the material property data includes data describing the physical, chemical and mechanical properties of the material;

[0059] S202, based on the three-dimensional model, using a material property equivalent algorithm, the thermal conductivity and mechanical properties of different materials are equivalent to parameters of a multi-layer rectangular parallelepiped structure to obtain an equivalent model;

[0060] S203. Generate a multi-layer cuboid structure model through a finite element analysis method according to the equivalent parameters, and use the model as a model input for defect reconstruction and positioning.

[0061] It should be noted that in step S201, the precise geometric dimensions and shape information of the switch cabinet are first extracted from the design drawings. A professional computer-aided design (CAD) tool is used to construct a three-dimensional geometric model of the switch cabinet based on this information. Next, combined with the material attribute data, which describe the physical properties (such as density, specific heat capacity), chemical properties (such as corrosion resistance, oxidation resistance) and mechanical properties (such as elastic modulus, yield strength) of the material, the corresponding material properties are assigned to different components in the CAD model. This process involves complex data processing and attribute mapping to ensure that the model of each component can truly reflect its performance in actual use. After completing this step, the obtained three-dimensional model is not only consistent with the actual switch cabinet in terms of geometry, but also accurately simulated in terms of material properties.

[0062] In step S202, based on the three-dimensional model obtained in step S201, the material property equivalent algorithm is applied to analyze the different material layers in the model, taking into account the anisotropy and inhomogeneity of the material, and simplifying the complex multi-layer material structure into an equivalent homogenized model. For example, using the homogenization theory or multi-scale homogenization method, the microstructural characteristics of the woven composite material are equivalent on a macro scale to obtain the equivalent thermal conductivity and mechanical characteristic parameters of each rectangular layer. This step involves complex mathematical calculations and material behavior simulations to ensure that the equivalent model can accurately reflect the physical behavior of the original model.

[0063] Specifically, in step S203, the multi-layer rectangular parallelepiped structure model is analyzed in detail using finite element analysis software. The continuous model is discretized into a series of finite elements, such as tetrahedron or hexahedron elements, by dividing the finite element grid. Then, appropriate boundary conditions and loading conditions are defined, including but not limited to heat flow, pressure or mechanical load, to simulate the stress and thermal state of the switch cabinet in the actual working environment. The FEA software will calculate the stress, strain and temperature distribution of each element according to the equivalent parameters and defined conditions, so as to obtain the response of the entire structure.

[0064] The calculation formula of the multi-layer cuboid structure model in step S203 is as follows:

[0065] Heat conduction equation for each layer:

[0066]

[0067] Boundary conditions between layers:

[0068]

[0069] Initial conditions:

[0070] T i (x,0)=T i,0 (x)

[0071] External surface boundary conditions:

[0072]

[0073] Where i=1,2,...,n is the number of layers, ρ i is the density of the i-th layer of material, c p,i is the specific heat capacity of the i-th layer material, T i is the temperature of the i-th layer of material, t is the time, k i is the thermal conductivity of the i-th layer of material, Q i is the internal defect heat source of the i-th layer of material, x is the spatial coordinate, L i is the thickness of the i-th layer of material, T i+1 is the temperature of the i+1th layer of material, T i,0 is the temperature distribution of the i-th layer of material at the initial moment, h ext is the external convective heat transfer coefficient, T ext is the external environment temperature, k n is the thermal conductivity of the nth layer of material, T1(x=0) is the temperature of the first layer at x=0, T n (x=L n ) is the nth layer at x = L n The temperature at n is the thickness of the nth layer.

[0074] In actual industrial applications, the application of the above steps can significantly improve the efficiency and accuracy of the design, analysis and maintenance processes of power equipment such as switchgear. For example, in the design stage, accurate 3D modeling and material property analysis can be used to predict and optimize the structural strength and heat dissipation performance of the switchgear. In the production process, equivalent models and finite element analysis can be used for quality control to ensure that each component meets the design specifications.

[0075] S300, using the multi-scale parametric modeling data of woven composite materials, through the progressive damage and failure analysis of thermal-mechanical-oxygen coupling, simulate the defects that occur in the material manufacturing process and obtain the prior information of material defects.

[0076] It can be understood that step S300 includes S301, S302 and S303, wherein:

[0077] S301, obtaining a parameterized model of the material based on the microstructure and physical property data of the woven composite material through multi-scale modeling algorithm processing;

[0078] S302. Based on the parameterized model, the thermal-mechanical-oxygen coupling analysis algorithm is used to simulate the stress, thermal distribution and oxygen permeation behavior of composite materials in actual working environments. Combined with material failure theory, the microcrack initiation and propagation paths inside the material are predicted to generate damage evolution data of the material.

[0079] S303. Based on the damage evolution data, a machine learning algorithm is applied to classify and identify potential defect patterns, and a statistical analysis method is used to extract the probability distribution of defect size, location and type from the damage evolution data, generate a priori information library for defect reconstruction and positioning, and then extract prior information of material defects.

[0080] It should be noted that in this step, based on the microscopic image and material composition information of the woven composite material, the microstructural features such as the single filament diameter, woven texture and fiber volume fraction are extracted through image processing and pattern recognition algorithms, and the image segmentation technology is used to distinguish different material phases, and the geometric parameters of the fiber bundle are extracted through morphological operations. According to the microstructural features extracted in step 1, the microscopic model is converted into a macroscopic model of a continuous medium through a homogenization method, a finite element model with multi-scale characteristics is constructed, and periodic boundary conditions are applied to simulate the repeating units of the woven composite material. The parametric representation of the geometry and material properties of the model is realized through finite element analysis software.

[0081] Specifically, according to the multi-scale model constructed in step 2, the heat-mechanical-oxygen coupled field simulation algorithm is used to simulate the stress, heat distribution and oxygen permeation behavior of the composite material in the actual working environment. Combined with material failure theory, such as the Paris crack growth law, the microcrack initiation and expansion path inside the material is predicted to generate time series data of damage evolution. According to the damage evolution data obtained in step 3, machine learning algorithms such as support vector machine (SVM) or random forest are applied to classify and identify potential defect patterns. Using statistical analysis methods, the probability distribution of defect size, location and type is extracted from the damage data to generate a priori information library for defect reconstruction and positioning. According to the priori information library generated in step 4, the optimal defect feature parameters are determined by optimization algorithms such as genetic algorithms or particle swarm optimization to maximize the accuracy and efficiency of defect detection, and the optimized priori information is integrated into the defect reconstruction and positioning module as an input parameter to guide the subsequent defect detection and analysis process.

[0082] Through these steps, we not only achieved multi-scale modeling from micro to macro, but also combined the physical properties of the material and environmental factors to perform coupled field simulation and damage prediction. Finally, through the application of machine learning and optimization algorithms, specific and implementable defect detection prior information was generated, providing strong data support for defect reconstruction and positioning of woven composite materials.

[0083] It can be understood that step S303 includes S3031, S3032 and S3033, wherein:

[0084] S3031. Apply principal component analysis to screen the prior information and obtain the most representative and discriminative parameters for defect characteristics;

[0085] S3032, optimizing the defect characteristic parameters by using a genetic algorithm based on the parameters, determining the optimal defect characteristic parameters, and training the defect detection model by using the optimal defect characteristic parameters;

[0086] S3033. Integrate the determined optimal parameters and the trained defect detection model into the defect reconstruction and positioning module as input parameters.

[0087] It should be noted that in this step, statistical and machine learning techniques are used to analyze the prior information library to identify the most critical feature parameters for defect detection. This process may involve dimensionality reduction technology to reduce computational complexity and exclude irrelevant or redundant information; the selected parameters are used to guide the search process of the optimization algorithm to find the optimal parameter settings. At the same time, these parameters are also used to train the defect detection model so that it can accurately identify and classify different types of defects based on these features; finally, the optimized parameters and trained model are integrated into the defect reconstruction and positioning module. This may require the development or adjustment of the module's internal algorithm to ensure that it can effectively use this optimized prior information for more accurate defect positioning.

[0088] S400, combining the temperature input and the model input, and processing them through a preset defect reconstruction and positioning module to obtain preliminary positioning information of potential defects inside the switch cabinet, wherein the preliminary positioning information includes the location, size or shape information of the defect.

[0089] It can be understood that in this step, based on the temperature distribution map captured by the infrared thermal imager, the abnormal temperature area on the surface of the switch cabinet is accurately identified through image segmentation and feature extraction technology. At the same time, the structural stress distribution data obtained by finite element analysis is aligned with the temperature distribution map to ensure that the two correspond in the same coordinate system. The temperature data is spatially smoothed using the nearest neighbor interpolation or three-dimensional bilinear interpolation algorithm to reduce noise and fill possible data gaps. Based on the fused data set, a physical property analysis algorithm is used, such as calculating the thermal gradient and stress gradient of each grid node, as well as their rate of change, and a clustering algorithm, such as K-means or DBSCAN, is used to group nodes with similar physical properties to identify abnormal clusters that may indicate defects.

[0090] Further, machine learning classifiers, including but not limited to random forests or gradient boosting machines, are applied to classify these clusters to distinguish between normal and abnormal areas. Then, defect reconstruction algorithms, such as topology optimization based on the level set method, are used to model the shape and size of potential defects, and combined with material mechanical behavior models, such as damage mechanics or fracture mechanics models, to predict the development and impact of defects. Through optimization techniques, such as particle swarm optimization or genetic algorithms, the parameters of the defect model are adjusted to maximize the consistency with the actual measurement data, thereby achieving accurate reconstruction and positioning of potential defects, providing a detailed geometric description of the defects and possible physical impacts, and providing a scientific basis for subsequent repairs and improvements.

[0091] S500. Combine the prior information and the preliminary positioning information to obtain a comprehensive data set for defect positioning, process the comprehensive data set according to the differential method, and input the processed comprehensive data set as a known quantity into the solution equation matrix, and use the known quantity to correct the result value of the piecewise polynomial singular value truncation decomposition method to obtain a corrected result.

[0092] It should be noted that the combination of prior information and preliminary positioning information is based on the prior information obtained from the multi-scale parameterized model and finite element analysis, such as the mechanical properties of woven composite materials and the expected damage mode, and the preliminary positioning information obtained by infrared thermal imagers and structural model analysis, such as surface temperature anomalies and stress concentration areas, to perform data-level fusion, and use data fusion techniques, such as Kalman filtering or Bayesian methods, to integrate these two types of information to form a comprehensive data set, and then apply the differential method to the comprehensive data set, using numerical differentiation or analytical differentiation to calculate the rate of change and directional information of the key features in the data set. This step can reveal the temperature gradient and stress changes in the defect area, providing a dynamic perspective for further defect analysis.

[0093] It can be understood that the comprehensive data set processed by the differential method is used as a known quantity and input into the constructed solution equation matrix, where these known quantities contain the detailed physical characteristics of the defect area and can be used to drive the solution process of the mathematical model. In this process, the piecewise polynomial singular value truncation decomposition method is used to numerically solve the model, and then the solution result is corrected according to the known quantity to improve the accuracy and reliability of defect location.

[0094] In the calculation process, the calculation formula of the piecewise polynomial singular value truncation decomposition method in step S500 is as follows:

[0095] T A =T p +β(TT p )

[0096] Where, T A is the output value of the piecewise polynomial singular value truncation decomposition method, T p is the temperature distribution calculated by the piecewise polynomial singular value truncation decomposition method, T is the temperature distribution calculated by differentiation, and β is the modified proportional correction coefficient.

[0097] In actual industrial inspection, the application of step S500 can greatly improve the defect detection accuracy of power equipment such as switch cabinets. For example, in the regular maintenance of the power system, through the automated data collection and processing process, individuals with potential defects can be quickly identified from a large number of devices, and detailed analysis and evaluation can be performed on them.

[0098] S600: Determine a reconstructed positioning result of the internal defect of the switch cabinet according to the corrected result, and clarify the position and type of the defect according to the reconstructed positioning result.

[0099] It is understandable that in this step, data analysis techniques, such as threshold determination or cluster analysis, are used to identify and distinguish abnormal patterns in the data, and spatial positioning algorithms, such as nearest neighbor search or pattern matching technology, are used to associate the identified abnormal patterns with the three-dimensional model of the switch cabinet to determine the exact location of the defect. This step involves locating the spatial coordinates of the abnormal data points in the model to ensure the accuracy of the defect location, and then judging the defect type based on the physical properties of the defect area, such as temperature distribution, stress state and material properties, combined with an expert system or machine learning classifier, including but not limited to cracks, corrosion, material degradation or other forms of damage. Finally, the specific location of the defect inside the switch cabinet is accurately pointed out, and the defect type is classified, providing a detailed basis for maintenance, replacement or further analysis.

[0100] Embodiment 2:

[0101] like Figure 2 As shown, this embodiment provides a detection system inside the switch cabinet based on defect reconstruction technology, see Figure 2 The system comprises:

[0102] Acquisition module 701: used to scan the switch cabinet surface with an infrared thermal imager, obtain the cross-sectional boundary temperature curve at the position of the highest temperature point, and use the cross-sectional boundary temperature curve as the temperature input for defect reconstruction and positioning;

[0103] The first processing module 702 is used to process the internal structure and material characteristics of the switch cabinet using the multi-layer equivalent analysis module 703 to convert it into a multi-layer rectangular parallelepiped structure model, and use the multi-layer rectangular parallelepiped structure model as a model input for defect reconstruction and positioning;

[0104] Analysis module 703: used to utilize multi-scale parametric modeling data of braided composite materials, simulate defects occurring in the material manufacturing process through progressive damage and failure analysis of thermal-mechanical-oxygen coupling, and obtain prior information of material defects;

[0105] The second processing module 704 is used to combine the temperature input and the model input, and process them through a preset defect reconstruction and positioning module to obtain preliminary positioning information of potential defects inside the switch cabinet, wherein the preliminary positioning information includes the location, size or shape information of the defect;

[0106] Solution correction module 705: used to combine the prior information and the preliminary positioning information to obtain a comprehensive data set for defect positioning, process the comprehensive data set according to the differential method, and input the processed comprehensive data set as a known quantity into the solution equation matrix, and use the known quantity to correct the result value of the piecewise polynomial singular value truncation decomposition method to obtain a corrected result;

[0107] Determination module 706: is used to determine the reconstructed positioning result of the internal defect of the switch cabinet according to the corrected result, and to clarify the location and type of the defect according to the reconstructed positioning result.

[0108] Specifically, the acquisition module 701 includes:

[0109] The acquisition unit 7011 is used to scan the surface of the switch cabinet using an infrared thermal imager to obtain a temperature distribution map, and process the temperature distribution map using an image processing algorithm to obtain the coordinates of the position of the highest temperature point;

[0110] Extraction unit 7012: used to extract the cross-sectional boundary of the highest temperature point from the temperature distribution map based on the coordinates and using an edge detection algorithm, and calculate the temperature curve on the boundary;

[0111] The temperature generation unit 7013 is used to generate a continuous temperature curve as the temperature input for defect reconstruction and positioning by applying a curve fitting method according to the temperature curve.

[0112] Specifically, the first processing module 702 includes:

[0113] The first processing unit 7021 is used to obtain a three-dimensional model of the internal structure of the switch cabinet through geometric modeling according to the design drawings and material property data of the switch cabinet, wherein the material property data includes data describing the physical, chemical and mechanical properties of the material;

[0114] Equivalent unit 7022: used to convert the thermal conductivity and mechanical properties of different materials into parameters of a multi-layer cuboid structure based on a three-dimensional model and using a material property equivalent algorithm to obtain an equivalent model;

[0115] Model generation unit 7023: used to generate a multi-layer rectangular parallelepiped structure model according to equivalent parameters through a finite element analysis method, and the model is used as a model input for defect reconstruction and positioning.

[0116] Specifically, the analysis module 703 includes:

[0117] The second processing unit 7031 is used to obtain a parameterized model of the material according to the microstructure and physical property data of the braided composite material through a multi-scale modeling algorithm;

[0118] Simulation prediction unit 7032: used to simulate the stress, heat distribution and oxygen permeation behavior of composite materials in actual working environment based on parameterized model and using thermal-mechanical-oxygen coupling analysis algorithm, and predict the microcrack initiation and propagation path inside the material in combination with material failure theory, and generate damage evolution data of the material;

[0119] Identification and analysis unit 7033: used to classify and identify potential defect patterns based on damage evolution data by applying machine learning algorithms, and to extract the probability distribution of defect size, location and type from the damage evolution data using statistical analysis methods, to generate a priori information library for defect reconstruction and positioning, and further to extract prior information of material defects.

[0120] Specifically, the identification and analysis unit 7033 then includes:

[0121] Screening unit 70331: used to screen the prior information by applying principal component analysis to obtain the most representative and discriminative parameters for defect characteristics;

[0122] The tuning unit 70332 is used to tune the defect characteristic parameters based on the parameter selection genetic algorithm, determine the optimal defect characteristic parameters, and train the defect detection model using the optimal defect characteristic parameters;

[0123] Integration unit 70333: used to integrate the determined optimal parameters and the trained defect detection model into the defect reconstruction and positioning module as input parameters.

[0124] In summary, the present invention first uses an infrared thermal imager to obtain the detailed temperature distribution of the switch cabinet surface, and then extracts the cross-sectional boundary temperature curve of the highest temperature point through image processing technology, and combines the microstructure and physical property data of the woven composite material, and uses a multi-scale modeling algorithm to construct a macro model of the continuous medium, and uses a thermal-mechanical-oxygen coupling field simulation algorithm to simulate the stress, thermal distribution and oxygen permeation behavior of the material in an actual working environment. Furthermore, the present invention uses a machine learning algorithm to classify and identify potential defect modes, and determines the optimal defect characteristic parameters through an optimization algorithm; finally, the optimized prior information is integrated into the defect reconstruction and positioning module to achieve high-precision detection of internal defects of the switch cabinet, and not only improves the accuracy and efficiency of defect detection, but also is suitable for the evaluation of complex structural materials, providing strong technical support for the stable operation of the power system.

[0125] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0126] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0127] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A detection method for the inside of a switch cabinet based on defect reconstruction technology, characterized in that: include: Scan the switch cabinet surface with an infrared thermal imager to obtain the cross-sectional boundary temperature curve at the highest temperature point, and use the cross-sectional boundary temperature curve as the temperature input for defect reconstruction and positioning; According to the internal structure and material characteristics of the switch cabinet, the multi-layer equivalent analysis module is used to process it and convert it into a multi-layer rectangular parallelepiped structure model, which is used as the model input for defect reconstruction and positioning. Using multi-scale parametric modeling data of woven composite materials, through the progressive damage and failure analysis of thermal-mechanical-oxygen coupling, the defects that occur during the material manufacturing process are simulated to obtain the prior information of material defects; Combining the temperature input and the model input, the preset defect reconstruction and positioning module is used to process the information to obtain preliminary positioning information of the potential defects inside the switch cabinet, wherein the preliminary positioning information includes the location, size or shape information of the defect; The prior information and the preliminary positioning information are combined to obtain a comprehensive data set for defect positioning, the comprehensive data set is processed according to the differential method, and the processed comprehensive data set is input into the solution equation matrix as a known quantity, and the result value of the piecewise polynomial singular value truncation decomposition method is corrected by using the known quantity to obtain a corrected result; According to the corrected results, the reconstruction positioning results of the internal defects of the switch cabinet are determined, and the location and type of the defects are clarified according to the reconstruction positioning results; The method of using a multi-layer equivalent analysis module to process the internal structure and material properties of the switch cabinet to convert it into a multi-layer rectangular parallelepiped structure model, and using the multi-layer rectangular parallelepiped structure model as a model input for defect reconstruction and positioning includes: According to the design drawings and material property data of the switch cabinet, a three-dimensional model of the internal structure of the switch cabinet is obtained through geometric modeling processing, wherein the material property data includes data describing the physical, chemical and mechanical properties of the material; Based on the three-dimensional model, the material property equivalent algorithm is used to equate the thermal conductivity and mechanical properties of different materials to the parameters of the multi-layer cuboid structure to obtain an equivalent model; According to the equivalent parameters, a multi-layer cuboid structure model is generated by finite element analysis method, and the model is used as the model input for defect reconstruction and positioning; The above method is equivalent to a multi-layer rectangular parallelepiped structure model, wherein the calculation formula of the multi-layer rectangular parallelepiped structure model is as follows: Heat conduction equation for each layer: Boundary conditions between layers: Initial conditions: T i (x,0)=T i,0 (x) External surface boundary conditions: Where i=1,2,...,n is the number of layers, ρ i is the density of the i-th layer of material, c p,i is the specific heat capacity of the i-th layer material, T i is the temperature of the i-th layer of material, t is the time, k i is the thermal conductivity of the i-th layer of material, Q i is the internal defect heat source of the i-th layer of material, x is the spatial coordinate, L i is the thickness of the i-th layer of material, T i+1 is the temperature of the i+1th layer of material, T i,0 is the temperature distribution of the i-th layer of material at the initial moment, h ext is the external convective heat transfer coefficient, T ext is the external environment temperature, k n is the thermal conductivity of the nth layer of material, T1(x=0) is the temperature of the first layer at x=0, T n (x=L n ) is the nth layer at x = L n The temperature at n is the thickness of the nth layer.

2. The detection method for the inside of a switch cabinet based on defect reconstruction technology according to claim 1 is characterized in that: The switch cabinet surface is scanned by an infrared thermal imager to obtain a cross-sectional boundary temperature curve at the highest temperature point, and the cross-sectional boundary temperature curve is used as a temperature input for defect reconstruction and positioning, including: Use an infrared thermal imager to scan the surface of the switch cabinet to obtain a temperature distribution map, and then use an image processing algorithm to process the temperature distribution map to obtain the coordinates of the highest temperature point; Based on the coordinates, the edge detection algorithm is used to extract the cross-sectional boundary of the highest temperature point from the temperature distribution map, and the temperature curve on the boundary is calculated; According to the temperature curve, the curve fitting method is applied to generate a continuous temperature curve as the temperature input for defect reconstruction and positioning.

3. The detection method for the inside of a switch cabinet based on defect reconstruction technology according to claim 2 is characterized in that: The surface of the switch cabinet is scanned using an infrared thermal imager to obtain a temperature distribution map, wherein the temperature distribution map is calculated as follows: Where ρ is the material density, c p is the specific heat capacity, T is the temperature, t is the time, k is the thermal conductivity, Q is the thermal imaging temperature, is the emissivity of the material, σ is the Stefan-Boltzmann constant, A is the surface area of ​​the object, T is the surface temperature of the object, T env is the ambient temperature.

4. The detection method for the inside of a switch cabinet based on defect reconstruction technology according to claim 1 is characterized in that: The multi-scale parametric modeling data of the braided composite material is used to simulate the defects that occur during the material manufacturing process through the progressive damage and failure analysis of thermal-mechanical-oxygen coupling, and obtain the prior information of material defects, including: Based on the microstructure and physical property data of the woven composite material, a parametric model of the material is obtained through multi-scale modeling algorithm processing; Based on the parameterized model, the thermal-mechanical-oxygen coupling analysis algorithm is used to simulate the stress, thermal distribution and oxygen permeation behavior of composite materials in actual working environments. Combined with material failure theory, the microcrack initiation and propagation paths inside the material are predicted to generate material damage evolution data. Based on the damage evolution data, machine learning algorithms are applied to classify and identify potential defect patterns. Statistical analysis methods are used to extract the probability distribution of defect size, location and type from the damage evolution data, generate a priori information library for defect reconstruction and positioning, and then extract prior information on material defects.

5. The method for detecting the inside of a switch cabinet based on defect reconstruction technology according to claim 4 is characterized in that: The method further extracts the prior information of material defects, and then includes: Apply principal component analysis to filter the prior information and obtain the most representative and discriminative parameters for defect characteristics; Based on the parameter selection, the genetic algorithm is used to optimize the defect characteristic parameters, determine the optimal defect characteristic parameters, and use the optimal defect characteristic parameters to train the defect detection model; The determined optimal parameters and the trained defect detection model are integrated into the defect reconstruction and positioning module and used as input parameters.

6. The method for detecting the inside of a switch cabinet based on defect reconstruction technology according to claim 1 is characterized in that: The known quantity is used to correct the result value of the piecewise polynomial singular value truncated decomposition method, wherein the calculation formula of the piecewise polynomial singular value truncated decomposition method is as follows: T A =T p +β(T-T p ) Where, T A is the output value of the piecewise polynomial singular value truncation decomposition method, T p is the temperature distribution calculated by the piecewise polynomial singular value truncation decomposition method, T is the temperature distribution calculated by differentiation, and β is the modified proportional correction coefficient.

7. A detection system for the inside of a switch cabinet based on defect reconstruction technology, based on the detection method for the inside of a switch cabinet based on defect reconstruction technology according to claim 1, characterized in that: include: Acquisition module: used to scan the switch cabinet surface with an infrared thermal imager, obtain the cross-sectional boundary temperature curve at the highest temperature point, and use the cross-sectional boundary temperature curve as the temperature input for defect reconstruction and positioning; The first processing module is used to process the internal structure and material characteristics of the switch cabinet using a multi-layer equivalent analysis module to convert it into a multi-layer rectangular parallelepiped structure model, and use the multi-layer rectangular parallelepiped structure model as a model input for defect reconstruction and positioning; Analysis module: It is used to utilize the multi-scale parametric modeling data of braided composite materials to simulate the defects that occur during the material manufacturing process through the progressive damage and failure analysis of thermal-mechanical-oxygen coupling, and obtain the prior information of material defects; The second processing module is used to combine the temperature input and the model input, and process them through a preset defect reconstruction and positioning module to obtain preliminary positioning information of potential defects inside the switch cabinet, wherein the preliminary positioning information includes the location, size or shape information of the defect; Solution correction module: used to combine the prior information and the preliminary positioning information to obtain a comprehensive data set for defect positioning, process the comprehensive data set according to the differential method, and input the processed comprehensive data set as a known quantity into the solution equation matrix, and use the known quantity to correct the result value of the piecewise polynomial singular value truncation decomposition method to obtain the corrected result; Determination module: used to determine the reconstruction positioning result of the internal defect of the switch cabinet according to the corrected result, and to clarify the location and type of the defect according to the reconstruction positioning result; Wherein, the first processing module includes: The first processing unit is used to obtain a three-dimensional model of the internal structure of the switch cabinet through geometric modeling according to the design drawings and material property data of the switch cabinet, wherein the material property data includes data describing the physical, chemical and mechanical properties of the material; Equivalent unit: Based on the three-dimensional model, using the material property equivalent algorithm, the thermal conductivity and mechanical properties of different materials are equivalent to the parameters of the multi-layer cuboid structure to obtain an equivalent model; Generate model unit: used to generate a multi-layer cuboid structure model through finite element analysis method according to equivalent parameters, and the model is used as the model input for defect reconstruction and positioning; The calculation formula of the multi-layer cuboid structure model in the first processing module is as follows: Heat conduction equation for each layer: Boundary conditions between layers: Initial conditions: T i (x,0)=T i,0 (x) External surface boundary conditions: Where i=1,2,...,n is the number of layers, ρ i is the density of the i-th layer of material, c p,i is the specific heat capacity of the i-th layer material, T i is the temperature of the i-th layer of material, t is the time, k i is the thermal conductivity of the i-th layer of material, Q i is the internal defect heat source of the i-th layer of material, x is the spatial coordinate, L i is the thickness of the i-th layer of material, T i+1 is the temperature of the i+1th layer of material, T i,0 is the temperature distribution of the i-th layer of material at the initial moment, h ext is the external convective heat transfer coefficient, T ext is the external environment temperature, k n is the thermal conductivity of the nth layer of material, T1(x=0) is the temperature of the first layer at x=0, T n (x=L n ) is the nth layer at x = L n The temperature at n is the thickness of the nth layer.

8. The switch cabinet internal detection system based on defect reconstruction technology according to claim 7 is characterized in that: The acquisition module includes: Acquisition unit: used to scan the surface of the switch cabinet with an infrared thermal imager to obtain a temperature distribution map, and process the temperature distribution map using an image processing algorithm to obtain the coordinates of the highest temperature point; Extraction unit: used to extract the cross-sectional boundary of the highest temperature point from the temperature distribution map based on the coordinates and using the edge detection algorithm, and calculate the temperature curve on the boundary; Generate temperature unit: It is used to generate a continuous temperature curve as the temperature input for defect reconstruction and positioning by applying the curve fitting method according to the temperature curve.

Images