Power System Defect Location and Detection Methods and Systems

Abstract

This invention relates to the field of positioning and detection technology, specifically to a method and system for locating and detecting defects in power systems. The method includes: checking whether the interaction between the terminal and the power system is correct, and submitting normal form data when the interaction is correct; checking the processing logs of the corresponding interface on the server side, and connecting to the server to perform log operations; describing and determining the target point information at any time or point in time by receiving the amplitude signal; setting cutoff conditions based on the change in the mean fitness of individual individuals in the detection amplitude curve according to rules; and determining the defect location of the tested component in the power system by comparing the rate of change between the detected amplitude curve and the fitted amplitude curve. This invention solves the problem that existing methods for locating internal defects in power equipment are affected by environmental changes and internal facility anomalies, and the power equipment terminal cannot accurately reflect the relationship between the distance and amplitude of the defect location, thus failing to meet the needs of internal defect detection and positioning.

Description

Technical Field

[0001] This invention relates to the field of location detection technology, and specifically to a method and system for locating and detecting defects in power systems. Background Technology

[0002] A power system is an electrical energy production and consumption system composed of power plants, transmission and transformation lines, substations, and electricity consumers. Its function is to convert primary energy from nature into electrical energy through power generation devices, and then supply this electrical energy to users through transmission, transformation, and distribution. To achieve this function, the power system also has corresponding information and control systems at various stages and levels to measure, regulate, control, protect, communicate, and dispatch the electrical energy production process, ensuring that users receive safe and high-quality electrical energy. The main structure of a power system includes power sources (hydropower stations, thermal power plants, nuclear power plants, etc.), substations (step-up substations, load center substations, etc.), transmission and distribution lines, and load centers. These power sources are interconnected to achieve electrical energy exchange and regulation between different regions, thereby improving the security and economy of power supply. The network formed by transmission lines and substations is usually called a power network. The information and control systems of a power system consist of various detection equipment, communication equipment, safety protection devices, automatic control devices, and monitoring automation and dispatch automation systems. The structure of the power system should ensure a rational coordination between power production and consumption, based on advanced technology and equipment and high economic efficiency.

[0003] Partial discharge detection in power systems is a primary method for defect detection. However, its poor anti-interference capability and extremely high requirements for the testing environment, coupled with the complex operating environment of power systems and severe electromagnetic interference and environmental noise at the testing site, make partial discharge detection unsuitable for online inspection of vehicle-mounted cable terminals. Before repair, defect location is necessary. Existing technologies involve repairing the entire power system, which is inefficient. This invention describes a power system defect location and detection system and method that solves the above problems. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention discloses a method and system for locating and detecting defects in power systems. This method solves the problem that existing methods for locating internal defects in power equipment are affected by environmental changes and internal facility anomalies, and the power equipment terminal cannot accurately reflect the relationship between the distance and amplitude of the defect location, thus failing to meet the requirements for internal defect detection and location.

[0005] This invention is achieved through the following technical solution:

[0006] Firstly, a power system defect location and detection system includes:

[0007] The interaction module is used to check whether the terminal is interacting correctly with the power system, and to submit normal form data when the judgment is correct.

[0008] The working module is used to view the processing logs of the corresponding interface on the server side and connect to the server to perform log operations;

[0009] The sensing module is used to perform electromagnetic ultrasonic guided wave position access of the target point, and to describe and determine the target point information at any time or point in time by receiving the amplitude signal of the signal.

[0010] An integrated module is used to detect amplitude curves based on rules, retain the mean change in individual fitness of the population, and set cutoff conditions.

[0011] The signal fitting module fits a distance versus amplitude variation curve based on the amplitude signal of the received signal.

[0012] The positioning gateway module determines the location of defects in the tested components of the power system by comparing the rate of change of the detected amplitude curve with that of the fitted amplitude curve.

[0013] Furthermore, the working module comprises the following sub-modules, including:

[0014] The acquisition module is used to constrain the target range by preprocessing the prior probabilities associated with the candidate locations of the target points.

[0015] The monitoring module acquires the location data of the target point according to the parameter precision control, and processes the acquired location data using an algorithm to obtain parameters;

[0016] The initialization module is used to initialize the parameters in the defect detection algorithm designed based on the principle of swarm evolution.

[0017] Furthermore, the signal fitting module comprises the following sub-modules, including:

[0018] The classification module is used to classify the defect categories within the amplitude signal dataset to obtain a classification dataset;

[0019] The training module is used to train the classification dataset using an object detection classification algorithm to obtain a defect classification model;

[0020] The processing module is used to fit the distance and amplitude of the defect location information of the target point using the defect classification model, obtain the distance and amplitude change curve, and output the defect location detection result.

[0021] Furthermore, during the acquisition process, the monitoring module extracts the main peak position and the distance between the main and side peaks of the surface concavity and convexity of the target point from the location data distribution, and selects feature data related to the target point position to obtain the estimated location of the defect.

[0022] Furthermore, the target location-related feature data includes data on signal strength or propagation loss, and the target location-related feature data is optimally configured to coordinate the generation of all requests related to the target location-related data in the determination of the location served by the identified cellular location.

[0023] Secondly, a method for locating and detecting defects in a power system includes the following steps:

[0024] Step 1: Check whether the terminal is interacting correctly with the power system, and submit normal form data when it is determined to be correct;

[0025] Step 2: Check the processing logs of the corresponding interface on the server and connect to the server to perform log operations;

[0026] Step 3: Perform electromagnetic ultrasonic guided wave location access on the target point, and describe and determine the target point information at any time or point in time by receiving the amplitude signal.

[0027] Step 4: Based on the rules, detect the amplitude curve, retain the mean change in the fitness of individual populations, and set cutoff conditions;

[0028] Step 5: Filter the distance distribution and amplitude distribution of the target point respectively to obtain the filtered distance curve and the filtered amplitude curve;

[0029] Step 6: Fit a distance versus amplitude variation curve based on the amplitude signal of the received signal;

[0030] Step 7: Compare the rate of change of the detected amplitude curve and the fitted amplitude curve to determine the location of the defect in the tested component of the power system.

[0031] Furthermore, Step 2, when performing log operations, includes the following sub-steps:

[0032] Step 21: By preprocessing the prior probabilities associated with the candidate locations of the target points, extract the arrival time vectors of all defect detection signals along the detection paths to constrain the target range;

[0033] Step 22: Obtain the position data of the target point according to the parameter precision control, and process the obtained position data by algorithm to obtain parameters;

[0034] Step 23: Initialize individual screening and fitness mean calculation, and calculate the parameters of the defect detection algorithm based on the principle of group evolution algorithm.

[0035] Furthermore, based on the detection parameters determined in Step 22, near-surface defects in the target point are detected, the scanning range displays complete through wave, defect diffraction wave, and deformed wave signals, and the scanned image of the target point is stored.

[0036] Furthermore, Step 7, when performing data fitting, includes the following sub-steps:

[0037] Step 71: Classify the defect categories in the amplitude signal dataset to obtain a classification dataset;

[0038] Step 72: Train the classification dataset using an object detection and classification algorithm to obtain a defect classification model;

[0039] Step 73: Use the defect classification model to fit the distance and amplitude of the defect location information of the target point to obtain the distance and amplitude change curve, and output the defect location detection result.

[0040] Furthermore, the target point defect localization selects an appropriate electromagnetic ultrasonic probe frequency, electromagnetic ultrasonic probe angle, and wafer size based on the defect location, and adjusts the electromagnetic ultrasonic probe center spacing, time window range, detection sensitivity, pulse repetition frequency, and scan increment.

[0041] The beneficial effects of this invention are as follows:

[0042] 1. This invention can replace traditional manual chip defect detection methods and traditional machine learning defect detection methods, combining high precision and high flexibility. It boasts strong network expressive power, eliminates the need for manual feature design, and is easy to apply and transfer. This invention employs deep learning algorithms for chip defect target detection, improving the efficiency and quality of defect localization. The method for locating and detecting defects in near-surface blind zones is highly accurate, has strong applicability, and possesses significant engineering application value.

[0043] 2. This invention calculates the location and depth of internal defects by analyzing the position of the main peak of surface concavity and the spacing between the main and secondary peaks in the multi-order differential distribution of the curve. Due to the gradient information of the target point extracted during the differentiation process, this invention is unaffected by environmental changes, making it an accurate and efficient online method for locating and detecting internal thermal defects in vehicle-mounted cable terminals. Utilizing the geometric relationship between the sound paths of different types of waves, a solution model is obtained by constructing the defect endpoint position. Combining the characteristics of the defect detection signal with classical intelligent analysis algorithms, the above problems are analyzed to achieve precise quantitative determination of the defect endpoint position. When locating and detecting defects at distances, the detection results are output.

[0044] 3. This invention addresses the problem of defect location, primarily aiming to improve the efficiency of defect modification, helping developers find defects faster and saving time. Its sensitivity to structural defects and changes in material properties allows for long-distance, rapid detection and monitoring of defects in power systems. With the rise of artificial intelligence theory and application technologies, and the increasing demands for testing in industry, the multifunctionality, automation, and intelligence of non-destructive testing (NDT) technology have become a trend. This is of great significance for realizing intelligent defect detection in power systems. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 A schematic diagram of the system structure of a power system defect location and detection system;

[0047] Figure 2 A flowchart illustrating the method for locating and detecting defects in power systems;

[0048] Figure 3 A flowchart illustrating the log operation process for a power system defect location and detection method.

[0049] Figure 4 A flowchart illustrating the data fitting process for a power system defect location and detection method.

[0050] The labels in the diagram represent: 1. Interaction module; 2. Working module; 3. Sensing module; 4. Integration module; 5. Signal fitting module; 6. Positioning gateway module; 21. Acquisition module; 22. Monitoring module; 23. Initialization module; 51. Classification module; 52. Training module; 53. Processing module. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Example 1

[0053] This embodiment provides a power system defect location and detection system. Please refer to [link / reference]. Figure 1 ,include:

[0054] Interaction module 1 is used to check whether the terminal interacts with the power system correctly, and submits normal form data when it is determined to be correct;

[0055] Module 2 is used to view the processing logs of the corresponding interface on the server side and connect to the server to perform log operations;

[0056] Sensing module 3 is used to perform electromagnetic ultrasonic guided wave position access of the target point, and to describe and determine the target point information at any time or point in time by receiving the amplitude signal of the signal.

[0057] Integrated module 4 is used to detect the amplitude curve according to rules, retain the mean change of individual fitness in the population, and set cutoff conditions.

[0058] The signal fitting module 5 fits a distance versus amplitude change curve based on the amplitude signal of the received signal.

[0059] The positioning gateway module 6 determines the location of defects in the tested components of the power system by comparing the rate of change of the detected amplitude curve with that of the fitted amplitude curve.

[0060] The working module 2 consists of the following sub-modules, including:

[0061] The acquisition module 21 is used to constrain the target range by preprocessing the prior probabilities associated with the candidate locations of the target points.

[0062] The monitoring module 22 acquires the location data of the target point according to the parameter precision control, and processes the acquired location data into parameters using an algorithm.

[0063] Initialization module 23 is used to initialize the parameters in the defect detection algorithm designed based on the principle of swarm evolution algorithm.

[0064] The signal fitting module 5 consists of the following sub-modules, including:

[0065] Classification module 51 is used to classify the defect categories in the amplitude signal dataset to obtain a classification dataset;

[0066] Training module 52 is used to train the classification dataset using an object detection classification algorithm to obtain a defect classification model;

[0067] The processing module 53 is used to fit the distance and amplitude of the defect location information of the target point using the defect classification model, obtain the distance and amplitude change curve, and output the defect location detection result.

[0068] During the acquisition process, the monitoring module 22 extracts the position of the main peak and the distance between the main and side peaks of the surface of the target point from the location data distribution, and selects the feature data related to the position of the target point to obtain the estimated position for defect location.

[0069] The target location-related feature data includes data on signal strength or propagation loss, and the target location-related feature data is optimally configured to coordinate the generation of all requests related to the target location-related data in the determination of the location served by the identified cellular location.

[0070] This invention can replace traditional artificial chip defect detection methods and traditional machine learning defect detection methods, combining high precision and high flexibility. It boasts strong network expressive power, eliminates the need for manual feature design, and is easy to apply and transfer. This invention employs deep learning algorithms for chip defect target detection, improving the efficiency and quality of defect localization. The method for locating and detecting defects in near-surface blind zones is highly accurate, widely applicable, and has significant engineering application value.

[0071] This invention calculates the location and depth of internal defects by analyzing the position of the main peak of surface concavity and the spacing between the main and secondary peaks in the multi-order differential distribution of the curve. Due to the gradient information of the target point extracted during the differentiation process, this invention is unaffected by environmental changes, making it an accurate and efficient online method for locating and detecting internal thermal defects in vehicle-mounted cable terminals. Utilizing the geometric relationship between the sound paths of different types of waves, a solution model is obtained by constructing the defect endpoint position. Combining the characteristics of the defect detection signal with classical intelligent analysis algorithms, the above problems are analyzed to achieve precise quantification of the defect endpoint position. When locating and detecting defects at distances, the detection results are output.

[0072] This invention addresses the problem of defect location, primarily aiming to improve the efficiency of defect modification, helping developers find defects faster and saving time. Its sensitivity to structural defects and changes in material properties allows for long-distance, rapid detection and monitoring of defects in power systems. With the rise of artificial intelligence theory and application technologies, and the increasing demands for testing from industry, the multifunctionality, automation, and intelligence of non-destructive testing (NDT) technology have become a trend. This is of great significance for realizing intelligent defect detection in power systems.

[0073] Example 2

[0074] At the implementation level, based on Example 1, this example refers to... Figure 2 , Figure 3 and Figure 4 The following is a further detailed description of the power system defect location and detection system in Example 1: A power system defect location and detection method includes the following steps:

[0075] Step 1: Check whether the terminal is interacting correctly with the power system, and submit normal form data when it is determined to be correct;

[0076] Step 2: Check the processing logs of the corresponding interface on the server and connect to the server to perform log operations;

[0077] Step 2, when performing log operations, consists of the following steps:

[0078] Step 21: By preprocessing the prior probabilities associated with the candidate locations of the target points, extract the arrival time vectors of all defect detection signals along the detection paths to constrain the target range;

[0079] Step 22: Obtain the position data of the target point according to the parameter precision control, and process the obtained position data by algorithm to obtain parameters;

[0080] Step 23: Initialize individual screening and fitness mean calculation, and calculate the parameters of the defect detection algorithm based on the principle of group evolution algorithm.

[0081] Step 3: Perform electromagnetic ultrasonic guided wave location access on the target point, and describe and determine the target point information at any time or point in time by receiving the amplitude signal.

[0082] Step 4: Based on the rules, detect the amplitude curve, retain the mean change in the fitness of individual populations, and set cutoff conditions;

[0083] Step 5: Filter the distance distribution and amplitude distribution of the target point respectively to obtain the filtered distance curve and the filtered amplitude curve;

[0084] Step 6: Fit a distance versus amplitude variation curve based on the amplitude signal of the received signal;

[0085] Step 7: Compare the rate of change of the detected amplitude curve and the fitted amplitude curve to determine the location of the defect in the tested component of the power system.

[0086] Step 7, when performing data fitting, includes the following sub-steps:

[0087] Step 71: Classify the defect categories in the amplitude signal dataset to obtain a classification dataset;

[0088] Step 72: Train the classification dataset using an object detection and classification algorithm to obtain a defect classification model;

[0089] Step 73: Use the defect classification model to fit the distance and amplitude of the defect location information of the target point to obtain the distance and amplitude change curve, and output the defect location detection result.

[0090] Based on the detection parameters determined in Step 22, near-surface defects in the target point are detected, and the scanning range displays complete through wave, defect diffraction wave, and deformed wave signals, and the scanned image of the target point is stored.

[0091] The target point defect localization selects an appropriate electromagnetic ultrasonic probe frequency, electromagnetic ultrasonic probe angle, and wafer size based on the defect location, and adjusts the electromagnetic ultrasonic probe center spacing, time window range, detection sensitivity, pulse repetition frequency, and scan increment.

[0092] A loss function is established based on the network's predicted output and the actual output. The parameters are continuously updated through backpropagation to make the network's predicted output approximate the actual output. Three types of loss functions are established for each predicted bounding box: a loss function representing whether the target object is contained, a loss function representing the target's location, and a loss function representing the target's category.

[0093] This invention can replace traditional artificial chip defect detection methods and traditional machine learning defect detection methods, combining high precision and high flexibility. It boasts strong network expressive power, eliminates the need for manual feature design, and is easy to apply and transfer. This invention employs deep learning algorithms for chip defect target detection, improving the efficiency and quality of defect localization. The method for locating and detecting defects in near-surface blind zones is highly accurate, widely applicable, and has significant engineering application value.

[0094] This invention calculates the location and depth of internal defects by analyzing the position of the main peak of surface concavity and the spacing between the main and secondary peaks in the multi-order differential distribution of the curve. Due to the gradient information of the target point extracted during the differentiation process, this invention is unaffected by environmental changes, making it an accurate and efficient online method for locating and detecting internal thermal defects in vehicle-mounted cable terminals. Utilizing the geometric relationship between the sound paths of different types of waves, a solution model is obtained by constructing the defect endpoint position. Combining the characteristics of the defect detection signal with classical intelligent analysis algorithms, the above problems are analyzed to achieve precise quantification of the defect endpoint position. When locating and detecting defects at distances, the detection results are output.

[0095] This invention addresses the problem of defect location, primarily aiming to improve the efficiency of defect modification, helping developers find defects faster and saving time. Its sensitivity to structural defects and changes in material properties allows for long-distance, rapid detection and monitoring of defects in power systems. With the rise of artificial intelligence theory and application technologies, and the increasing demands for testing from industry, the multifunctionality, automation, and intelligence of non-destructive testing (NDT) technology have become a trend. This is of great significance for realizing intelligent defect detection in power systems.

[0096] Example 3

[0097] At the implementation level, based on Embodiment 2, this embodiment further describes the power system defect location and detection method in Embodiment 2 in more detail. This embodiment also provides a computer terminal device, including one or more processors, a memory, and a readable storage medium, on which a computer program is stored. The memory is coupled to the processor and is used to store one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the power system defect location and detection system and method as described in any of the above embodiments.

[0098] This invention can replace traditional artificial chip defect detection methods and traditional machine learning defect detection methods, combining high precision and high flexibility. It boasts strong network expressive power, eliminates the need for manual feature design, and is easy to apply and transfer. This invention employs deep learning algorithms for chip defect target detection, improving the efficiency and quality of defect localization. The method for locating and detecting defects in near-surface blind zones is highly accurate, widely applicable, and has significant engineering application value.

[0099] This invention calculates the location and depth of internal defects by analyzing the position of the main peak of surface concavity and the spacing between the main and secondary peaks in the multi-order differential distribution of the curve. Due to the gradient information of the target point extracted during the differentiation process, this invention is unaffected by environmental changes, making it an accurate and efficient online method for locating and detecting internal thermal defects in vehicle-mounted cable terminals. Utilizing the geometric relationship between the sound paths of different types of waves, a solution model is obtained by constructing the defect endpoint position. Combining the characteristics of the defect detection signal with classical intelligent analysis algorithms, the above problems are analyzed to achieve precise quantification of the defect endpoint position. When locating and detecting defects at distances, the detection results are output.

[0100] This invention addresses the problem of defect location, primarily aiming to improve the efficiency of defect modification, helping developers find defects faster and saving time. Its sensitivity to structural defects and changes in material properties allows for long-distance, rapid detection and monitoring of defects in power systems. With the rise of artificial intelligence theory and application technologies, and the increasing demands for testing from industry, the multifunctionality, automation, and intelligence of non-destructive testing (NDT) technology have become a trend. This is of great significance for realizing intelligent defect detection in power systems.

[0101] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power system defect location and detection system, characterized in that, include: The interaction module (1) is used to check whether the terminal interacts with the power system correctly, and submits normal form data when it is determined to be correct; The working module (2) is used to view the processing logs of the corresponding interface on the server side and connect to the server to perform log operations; The sensing module (3) is used to perform electromagnetic ultrasonic guided wave position access of the target point and describe and determine the target point information at any time or point in time by receiving the amplitude signal of the signal. The integrated module (4) is used to detect the amplitude curve according to the rules, retain the mean change of individual fitness in the population, and set the cutoff condition. The signal fitting module (5) fits the distance and amplitude change curve based on the amplitude signal of the received signal; The positioning gateway module (6) determines the defect location of the power system test component by comparing the rate of change of the detection amplitude curve and the fitted amplitude curve; The working module (2) consists of the following sub-modules, including: The acquisition module (21) is used to constrain the target range by preprocessing the prior probabilities associated with the candidate locations of the target points; The monitoring module (22) acquires the location data of the target point according to the parameter precision control, and processes the acquired location data by algorithm to obtain parameters; The initialization module (23) is used to initialize the parameters in the defect detection algorithm designed based on the principle of group evolution algorithm; The signal fitting module (5) consists of the following sub-modules, including: The classification module (51) is used to classify the defect categories in the amplitude signal dataset to obtain a classification dataset; Training module (52) is used to train the classification dataset using an object detection classification algorithm to obtain a defect classification model; The processing module (53) is used to fit the distance and amplitude of the defect location information of the target point using the defect classification model, obtain the distance and amplitude change curve, and output the defect location detection result; The monitoring module (22) extracts the position of the main peak and the distance between the main and side peaks of the target point surface concavity and convexity in the acquisition process, and selects the feature data related to the target point position to obtain the estimated position for defect location. The feature data related to the target point location includes data on signal strength or propagation loss.

2. A method for locating and detecting defects in a power system, wherein the method is an implementation method of the power system defect location and detection system as described in claim 1, characterized in that, Includes the following steps: Step 1: Check whether the terminal is interacting correctly with the power system, and submit normal form data when it is determined to be correct; Step 2: Check the processing logs of the corresponding interface on the server and connect to the server to perform log operations; Step 3: Perform electromagnetic ultrasonic guided wave location access on the target point, and describe and determine the target point information at any time or point in time by receiving the amplitude signal. Step 4: Based on the rules, detect the amplitude curve, retain the mean change in the fitness of individual populations, and set cutoff conditions; Step 5: Filter the distance distribution and amplitude distribution of the target point respectively to obtain the filtered distance curve and the filtered amplitude curve; Step 6: Fit a distance versus amplitude variation curve based on the amplitude signal of the received signal; Step 7: Compare the rate of change of the detection amplitude curve and the fitted amplitude curve to determine the location of the defect in the tested component of the power system.

3. The power system defect location and detection method according to claim 2, characterized in that, Step 2, when performing log operations, consists of the following steps: Step 21: By preprocessing the prior probabilities associated with the candidate locations of the target points, extract the arrival time vectors of all defect detection signals along the detection paths to constrain the target range; Step 22: Obtain the position data of the target point according to the parameter precision control, and process the obtained position data by algorithm to obtain parameters; Step 23: Initialize individual screening and fitness mean calculation, and calculate the parameters of the defect detection algorithm based on the principle of group evolution algorithm.

4. The power system defect location and detection method according to claim 3, characterized in that, Based on the detection parameters determined in Step 22, near-surface defects in the target point are detected. The scanning range displays complete direct wave, defect diffraction wave, and deformed wave signals, and the power system defect location detection system stores the scanned image of the target point.

5. The power system defect location and detection method according to claim 2, characterized in that, Step 7, when performing data fitting, includes the following sub-steps: Step 71: Classify the defect categories in the amplitude signal dataset to obtain a classification dataset; Step 72: Train the classification dataset using an object detection and classification algorithm to obtain a defect classification model; Step 73: Use the defect classification model to fit the distance and amplitude of the defect location information of the target point to obtain the distance and amplitude change curve, and output the defect location detection result.

6. The power system defect location and detection method according to claim 2, characterized in that: The target point defect localization selects an appropriate electromagnetic ultrasonic probe frequency, electromagnetic ultrasonic probe angle, and wafer size based on the defect location, and adjusts the electromagnetic ultrasonic probe center spacing, time window range, detection sensitivity, pulse repetition frequency, and scan increment.

Images