An online detection method and system based on eddy current nondestructive metal flaw detection
Through an online detection method and system based on eddy current lossless metal flaw detection, the eddy current probe captures electromagnetic signals and builds an automatic detection channel, which solves the problems of low detection efficiency and low accuracy caused by manual participation in the prior art, and realizes the metal components efficient and reliable detection.
Patent Information
- Application Number
- CN202410675540.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-29
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-05-29
AI Technical Summary
Due to the cumbersome manual participation of existing metal flaw detection technology, the detection efficiency is low and the accuracy is low, and there is a risk of defect missed detection.
Using an online detection method and system based on eddy current lossless metal flaw detection, the target metal components are captured in real time, multiple eddy current probes are activated to induce eddy current, capture electromagnetic signals, and build an automatic online detection channel, including structural detection branch channels and defect detection branch channels. , realize automatic detection of metal components.
Improve detection efficiency and accuracy, avoid defect missed inspection, and make the detection process more reliable and comprehensive.
Smart Images

Figure CN118465054B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of quality inspection, and in particular to an online inspection method and system based on eddy current non-destructive metal flaw detection. Background Art
[0002] Nondestructive testing is a technology that detects the intrinsic or surface defects of a material or measures the properties of a material without damaging the material being tested. In industrial production, eddy current testing is the main means of testing various metal materials and a few non-metals, such as carbon fiber composite materials.
[0003] At present, existing metal flaw detection technologies mostly rely on offline detection or require manual participation, and cannot achieve efficient automated online detection. In addition, some traditional flaw detection methods may have certain limitations in accuracy, speed and reliability.
[0004] In summary, due to the tediousness and limitations of manual participation, the existing technology has low detection efficiency and accuracy, and there is a risk of missing defects. Summary of the invention
[0005] The purpose of this application is to provide an online detection method and system based on eddy current non-destructive metal flaw detection, so as to solve the problems of low detection efficiency and accuracy and the risk of missed defects in the prior art due to the tediousness and limitations of manual participation.
[0006] In view of the above problems, the present application provides an online detection method and system based on eddy current non-destructive metal flaw detection.
[0007] In the first aspect, the present application provides an online detection method based on eddy current non-destructive metal flaw detection, and the method is implemented by an online detection system based on eddy current non-destructive metal flaw detection, wherein the method includes: based on the production process of metal materials, real-time grabbing of target metal components; starting multiple eddy current probes to sense eddy currents of target metal components, and obtaining an eddy current detection parameter set, wherein the eddy current detection parameters include eddy current distribution parameters and eddy current change parameters; based on the eddy current change parameters, capturing multiple electromagnetic signals generated by the eddy current distribution parameters; constructing an automatic online detection channel based on the multiple electromagnetic signals, wherein the automatic online detection channel includes a structure detection branch channel and a defect detection branch channel; outputting structural information of the target metal component through the structure detection branch channel, synchronizing the structural information to the defect detection branch channel for automatic detection, and obtaining the detection result of the target metal component; activating the grabbing unit to remove the target metal component that has completed the detection, and associating and identifying the removed target metal component according to the detection result.
[0008] In the second aspect, the present application also provides an online detection system based on eddy current non-destructive metal flaw detection, which is used to execute an online detection method based on eddy current non-destructive metal flaw detection as described in the first aspect, wherein the system includes: a metal component acquisition module, the metal component acquisition module is used to capture the target metal component in real time based on the production process of the metal material; a detection parameter set acquisition module, the detection parameter set acquisition module is used to start multiple eddy current probes to sense the eddy current of the target metal component, and obtain an eddy current detection parameter set, the eddy current detection parameters include eddy current distribution parameters and eddy current change parameters; an electromagnetic signal capture module, the electromagnetic signal capture module is used to capture the eddy current distribution parameters based on the eddy current change parameters. a detection channel construction module, the detection channel construction module is used to construct an automatic online detection channel based on the multiple electromagnetic signals, and the automatic online detection channel includes a structure detection branch channel and a defect detection branch channel; a detection result acquisition module, the detection result acquisition module is used to output the structural information of the target metal component through the structure detection branch channel, synchronize the structural information to the defect detection branch channel for automatic detection, and obtain the detection result of the target metal component; an association identification module, the association identification module is used to activate the grasping unit to remove the target metal component that has completed the detection, and associate and identify the removed target metal component according to the detection result.
[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0010] Through the production process based on metal materials, the target metal component is grasped in real time; multiple eddy current probes are started to sense the eddy current of the target metal component to obtain an eddy current detection parameter set, wherein the eddy current detection parameter includes an eddy current distribution parameter and an eddy current change parameter; based on the eddy current change parameter, multiple electromagnetic signals generated by the eddy current distribution parameter are captured; based on the multiple electromagnetic signals, an automatic online detection channel is constructed, wherein the automatic online detection channel includes a structure detection branch channel and a defect detection branch channel; the structural information of the target metal component is output through the structure detection branch channel, and the structural information is synchronized to the defect detection branch channel for automatic detection to obtain the detection result of the target metal component; the grasping unit is activated to remove the target metal component that has been detected, and the removed target metal component is associated and identified according to the detection result, which effectively solves the problem that the existing technology is cumbersome and limited by manual participation, resulting in low detection efficiency and accuracy, and the risk of defect omission. The automatic detection channel is constructed through the electromagnetic signals generated by the eddy current detection technology, and the structure and potential defects of the target metal are accurately detected respectively, thereby avoiding defect omission and making the entire detection process more reliable and comprehensive.
[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.
[0013] Figure 1 A schematic diagram of a process of an online detection method based on eddy current non-destructive metal flaw detection in this application;
[0014] Figure 2 This is a structural schematic diagram of an online detection system based on eddy current non-destructive metal flaw detection in this application.
[0015] Description of reference numerals:
[0016] A metal component acquisition module 11, a detection parameter set acquisition module 12, an electromagnetic signal capture module 13, a detection channel construction module 14, a detection result acquisition module 15, and an association identification module 16. DETAILED DESCRIPTION
[0017] The present application provides an online detection method and system based on eddy current non-destructive metal flaw detection, which solves the problems of low detection efficiency and accuracy and the risk of missed defects in the prior art due to the cumbersomeness and limitations of manual participation. An automatic detection channel is constructed through the electromagnetic signals generated by eddy current detection technology to accurately detect the structure and potential defects of the target metal, thereby avoiding missed defects and making the entire detection process more reliable and comprehensive.
[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.
[0019] Embodiment 1
[0020] Please see attached Figure 1 The present application provides an online detection method based on eddy current non-destructive metal flaw detection, wherein the method is applied to an online detection system based on eddy current non-destructive metal flaw detection, and the method specifically comprises the following steps:
[0021] S1: Based on the production process of metal materials, real-time grasping of target metal components;
[0022] Specifically, high-precision cameras and sensors are installed on the production line to calibrate the robotic arm to ensure that it can accurately move to the predetermined position for grasping. The camera continuously captures images on the production line and analyzes them in real time through the image recognition algorithm. The recognition algorithm searches for objects that match the characteristics of the target metal component and determines its precise location. The target metal component is identified and the optimal path for the robotic arm to reach the component and grasp it is calculated. The robotic arm quickly moves above the target metal component according to the path planning. The gripper at the end of the robotic arm, such as a clamping claw, magnetic suction cup, etc., will accurately align the component and perform the grasping action.
[0023] S2: starting multiple eddy current probes to sense eddy currents of target metal components, and obtaining an eddy current detection parameter set, wherein the eddy current detection parameters include eddy current distribution parameters and eddy current variation parameters;
[0024] Specifically, the eddy current probe is placed on the surface of the target metal component and the probe is started. During the startup process, the probe generates an alternating electromagnetic field, which induces eddy currents on the surface of the metal component. When the eddy current flows in the metal component, if it encounters defects or structural changes, the distribution and intensity of the eddy current will change. The eddy current probe can capture these changes and convert them into electrical signals. These electrical signals contain information on eddy current distribution parameters and eddy current variation parameters. These include eddy current distribution parameters and eddy current variation parameters. The eddy current distribution parameters reflect the distribution of eddy currents on the surface of the metal component. For example, the density, direction and intensity of the eddy current. The eddy current variation parameters reflect the various changes encountered by the eddy current when it flows in the metal component. For example, when the eddy current encounters cracks, inclusions or other defects, its flow will be hindered, resulting in a weakening of the eddy current intensity or a change in direction.
[0025] S3: based on the eddy current variation parameter, capturing a plurality of electromagnetic signals generated by the eddy current distribution parameter;
[0026] Specifically, when eddy currents flow in metal components, they generate specific electromagnetic signals due to the conductive properties and structural characteristics of the metal. Signal receivers are used to capture these electromagnetic signals generated by eddy currents. The captured electromagnetic signals usually need to be converted. This includes converting analog signals into digital signals, or performing signal amplification, filtering, and other operations to improve the signal-to-noise ratio and resolution.
[0027] S4: constructing an automatic online detection channel based on the multiple electromagnetic signals, wherein the automatic online detection channel includes a structure detection branch channel and a defect detection branch channel;
[0028] Specifically, an automatic online detection channel is constructed based on the multiple electromagnetic signals. The channel can be designed to include two main branch channels. The structural detection branch channel is used to monitor and evaluate the structural integrity and stability of the object under test. By analyzing the changes in electromagnetic signals, key parameters such as slight deformation and stress distribution of the structure can be inferred. Using multi-channel electromagnetic signal detection technology, the electromagnetic signals emitted by the object under test are captured and analyzed in real time. These signals are processed by algorithms to extract characteristic information related to the structural state. The defect detection branch channel is used to detect possible defects or damage inside the object under test. By analyzing the abnormal changes in electromagnetic signals, the defect position can be discovered and located in time. Signals are captured using multi-channel electromagnetic signal detection technology. Advanced signal processing and pattern recognition algorithms are used to conduct in-depth analysis of the captured signals to identify the characteristics of the defects.
[0029] S5: outputting structural information of the target metal component through the structural detection branch channel, synchronizing the structural information to the defect detection branch channel for automatic detection, and obtaining a detection result of the target metal component;
[0030] Specifically, the structural detection branch channel will first analyze and process the electromagnetic signals of the target metal component to extract the structural information of the component. The structural information includes key parameters such as the component's geometry, size, material distribution, and stress state. When the structural detection branch channel completes the extraction of structural information, this information will be synchronously transmitted to the defect detection branch channel. After receiving the structural information, the defect detection branch channel will combine the electromagnetic signal data to perform automated defect detection. By comparing the structural information with the actual electromagnetic signal characteristics, the defect detection branch channel can identify abnormal signals that do not conform to the normal structure, thereby determining possible defects inside the component. After the defect detection branch channel completes the detection, it will output the detection results of the target metal component. Including detailed information such as the location, size, type, and severity of the defect.
[0031] S6: activating the grasping unit to take out the target metal component that has completed the inspection, and associating and marking the taken out target metal component according to the inspection result.
[0032] Specifically, when the target metal component is inspected, a command is sent to activate the corresponding grabbing unit. The grabbing unit can be an automated device such as a robotic arm, a suction cup, or a clamp, which is used to accurately remove the target metal component. The grabbing unit moves to the position of the target metal component according to the preset program and path. The grabbing unit accurately clamps or adsorbs the component, and then moves it from the inspection position to the designated storage area. While removing the component, the previously obtained inspection results are associated with the component. This is achieved by assigning a unique identification code to each component in the system, such as a serial number, barcode, or RFID tag. According to the associated inspection results, the corresponding information is printed or affixed to the label on or next to the component, such as qualified, unqualified, and requiring further processing.
[0033] Furthermore, step S2 of the present application also includes:
[0034] Applying high-frequency alternating current to the plurality of eddy current probes through an excitation power supply to generate eddy current data of the target metal component;
[0035] Based on the eddy current data, the plurality of eddy current probes are started for detection, and eddy current change sensing is performed through the started eddy current probes to generate eddy current change parameters;
[0036] Based on the eddy current variation parameters, the remaining eddy current probes are started for detection to capture eddy current distribution parameters;
[0037] The eddy current variation parameter and the eddy current distribution parameter are integrated to determine the eddy current detection parameter set.
[0038] Specifically, an excitation power supply is used to apply high-frequency alternating current to multiple eddy current probes. This current generates an alternating electromagnetic field around the eddy current probe. When the eddy current probe is close to the target metal component, the alternating electromagnetic field induces eddy currents in the component. These eddy currents generate their own electromagnetic fields, which in turn affect the current in the eddy current probe. By measuring the changes in the current in the eddy current probe, eddy current data of the target metal component can be generated. These data reflect the response of the metal component to the eddy current. Based on the generated eddy current data, multiple eddy current probes are started for detection. The started eddy current probes will perform eddy current change induction, that is, monitor the changes in eddy currents in the target metal component. These changes may be caused by changes in structural features, defects or material properties inside the metal component. Eddy current change parameters are generated based on the induced eddy current changes. These parameters describe the dynamic behavior of eddy currents in the target metal component, such as the intensity and frequency changes of eddy currents. Based on the eddy current change parameters, the remaining eddy current probes are started for detection to capture the distribution of eddy currents on the surface of the target metal component. The eddy current distribution parameters reflect the intensity and flow direction of eddy currents at different locations. Finally, the eddy current variation parameters and eddy current distribution parameters are integrated to form a comprehensive eddy current detection parameter set, which includes not only the dynamic variation information of eddy currents, but also the distribution of eddy currents on the surface of metal components.
[0039] Furthermore, step S4 of the present application also includes:
[0040] Extracting features of the plurality of electromagnetic signals using a signal processing algorithm to obtain metal structure features of the target metal component;
[0041] Constructing a feature vector based on the metal structure feature, performing supervised training using the feature vector as training data, and constructing the structure detection branch channel;
[0042] extracting a selected signal from the plurality of electromagnetic signals to determine a metal defect signal;
[0043] Analyze parameter changes of the metal defect signal to obtain multiple parameter change sets, wherein the multiple parameter change sets include frequency change parameters, amplitude change parameters, and phase change parameters;
[0044] Retrieving the historical defect types of the target metal component, performing data association in combination with the frequency change parameter, the amplitude change parameter, and the phase change parameter, and establishing a metal defect matching database;
[0045] Performing random selection based on the metal defect matching database to determine a first metal defect sample;
[0046] Perform unsupervised training using the first metal defect sample as training data to construct the defect detection branch channel;
[0047] The structure detection branch channel and the defect detection branch channel are integrated to obtain the automatic online detection channel.
[0048] Specifically, signal processing algorithms such as Fourier transform and wavelet transform are used to extract features from multiple captured electromagnetic signals. Key information that can represent the metal structure characteristics of the target metal component is extracted, such as the frequency component, amplitude, waveform, etc. of the signal. The extracted features are combined into feature vectors. Each feature vector represents a sample and contains the metal structure information of the sample. Metal component samples with known structural characteristics are used as training data for supervised learning. Through machine learning algorithms such as support vector machines, decision trees, neural networks, etc., the model is trained to enable it to identify the structural characteristics of metal components based on the characteristics of electromagnetic signals. After training, this model constitutes a structural detection branch channel. Specific signals are selected from multiple electromagnetic signals, which are related to potential defects in metal components. The selected metal defect signals are deeply analyzed to extract frequency change parameters, amplitude change parameters, phase change parameters, etc. The historical defect type data of the target metal component is retrieved. The historical defect type is associated with the extracted frequency, amplitude, phase and other parameter changes. A metal defect matching database is constructed for subsequent defect identification and classification. Defect samples are randomly selected from the metal defect matching database as training data. Unsupervised learning algorithms, such as cluster analysis and anomaly detection, are used to train these samples to identify defect patterns that are different from normal patterns. After training, this model forms a defect detection branch channel. The constructed structure detection branch channel and defect detection branch channel are integrated to ensure that the two channels can work together to complete the comprehensive detection of the target metal components.
[0049] Furthermore, the present application also includes:
[0050] Dividing the first metal defect sample into data based on metal defect characteristics to generate a training data set and a verification data set;
[0051] Performing unsupervised training on the defect detection branch channel based on deep learning using the training data set;
[0052] Performing verification and evaluation on the output result of the defect detection branch channel according to the verification data set to generate an evaluation result;
[0053] The defect detection branch channel is iteratively optimized according to the evaluation result. When the preset requirements are met for consecutive preset times, it is considered to be converged, and the defect detection branch channel is output.
[0054] Specifically, based on the metal defect characteristics, the first metal defect sample is divided into a training data set and a validation data set. The training data set is used to train the deep learning model of the defect detection branch channel, while the validation data set is used to verify the performance and accuracy of the model. The training data set contains most of the metal defect samples and is used to train the deep learning model so that it can learn the characteristics and patterns of metal defects. The validation data set contains a part of the metal defect samples, which does not participate in the training process of the model, but is used to evaluate the generalization ability and performance of the model. Using the training data set, unsupervised training is performed through a deep learning algorithm to build a model of the defect detection branch channel. In unsupervised training, the model attempts to automatically discover and learn the structure and association in the data from the training data without relying on label information. After the training is completed, the output results of the defect detection branch channel are verified and evaluated using the validation data set. By comparing the prediction results of the model with the true labels in the validation data set, the performance indicators of the model on the validation data set, such as accuracy, recall rate, F1 score, etc., can be calculated. According to the results of the validation evaluation, the defect detection branch channel is iteratively optimized. This includes adjusting the parameters of the model, the network structure or the learning strategy, etc., to improve the performance and accuracy of the model. In each iteration, the model is adjusted and optimized according to the evaluation results. When the preset requirements are met for consecutive preset times, such as when the performance improvement is less than a certain threshold, the model can be considered to have converged. When the model converges, the final defect detection branch channel is output.
[0055] Furthermore, step S5 of the present application also includes:
[0056] Setting a detection boundary value based on the structural information, and setting a defect detection threshold based on the detection boundary value;
[0057] Decomposing the structural information to obtain the outer surface structural information and the inner surface structural information of the target metal component;
[0058] Determining whether the outer surface structure information is greater than or equal to the defect detection threshold, and whether the inner surface structure information is greater than or equal to the defect detection threshold;
[0059] If the outer surface structure information is less than the defect detection threshold, and the inner surface structure information is less than or equal to the defect detection threshold, it is considered that the target metal component has no defects, and a first detection result is generated;
[0060] If the outer surface structure information is greater than or equal to the defect detection threshold, and the inner surface structure information is less than the defect detection threshold, it is considered that there is a defect on the outer surface of the target metal component, the outer surface of the target metal component is traversed to locate the defect, generate a first defect report, and determine a second detection result;
[0061] If the outer surface structure information is less than the defect detection threshold, and the inner surface structure information is greater than or equal to the defect detection threshold, it is considered that there is a defect on the inner surface of the target metal component, and the inner surface of the target metal component is traversed to locate the defect, generate a second defect report, and determine a third detection result;
[0062] If the outer surface structural information is greater than or equal to the defect detection threshold, and the inner surface structural information is greater than or equal to the defect detection threshold, it is deemed that both the outer surface and the inner surface of the target metal component have defects, a defect traceability label is marked on the target metal component, a third defect report is generated, and a fourth detection result is determined;
[0063] The detection result of the target metal component is output based on the first detection result, the second detection result, the third detection result, and the fourth detection result.
[0064] Specifically, the defect detection threshold is a critical value set during quality inspection or fault detection, which is used to determine whether the inspected object, such as products, materials, equipment, etc., has defects or abnormalities. When a certain parameter or index detected exceeds this threshold, it is considered that the inspected object has defects or does not meet the quality standards. The defect detection threshold is set based on historical data and industry standards. For example, in eddy current detection of metal components, the defect detection threshold may be set based on the electromagnetic signal characteristics of normal metal components. If the detected signal strength, frequency change or other parameters exceed this normal range, that is, exceed the set threshold, the system will determine that the component has defects. According to the structural information of the metal component, a detection boundary value is set. This boundary value is a benchmark for judging whether the component structure is normal. Based on this detection boundary value, a defect detection threshold is set. When the structural information exceeds this threshold, it is considered that there is a defect. The structural information of the target metal component is decomposed into outer surface structural information and inner surface structural information. In this way, defect detection can be performed on the outer surface and inner surface respectively. Use the set defect detection threshold to determine whether the structural information of the outer surface and inner surface exceeds this threshold. If the structural information of both the outer surface and the inner surface is less than the defect detection threshold, it means that there are no defects in the target metal component, and the first test result is generated, indicating that the component is normal. If only the structural information of the outer surface is greater than or equal to the defect detection threshold, it means that there are defects on the outer surface. At this time, traverse the outer surface to locate the defects, generate the first defect report, and determine the second test result. If only the structural information of the inner surface is greater than or equal to the defect detection threshold, it means that there are defects on the inner surface. Traverse the inner surface to locate the defects, generate the second defect report, and determine the third test result. If the structural information of both the outer surface and the inner surface is greater than or equal to the defect detection threshold, it means that there are defects on both the inner and outer surfaces of the target metal component. At this time, mark the defect traceability label for the component, generate the third defect report, and determine the fourth test result. Finally, based on the first, second, third and fourth test results, the test result of the target metal component is comprehensively output. This result will include information such as whether the component has defects, the location of the defects, and related defect reports.
[0065] Furthermore, step S6 of the present application also includes:
[0066] Based on the detection results, the target metal components that have been detected are marked and the grabbing decision logic is preset;
[0067] generating a grasping signal according to the grasping decision logic, and sending the grasping signal to the grasping unit for activation;
[0068] The placement position is determined by activating the grabbing unit, and the target metal component is removed according to the placement position.
[0069] Specifically, the components are identified according to the test results. For example, color coding, labels or digital codes can be used to identify whether the components have defects and the type and severity of the defects. According to the identification of the test results, a grasping decision logic is preset. This logic determines how to control the subsequent grasping and placement operations according to the identification results of the components. For example, if the component is identified as defect-free, it may be placed in the qualified product area; if it is defective, it may be placed in the repair or scrap area. According to the grasping decision logic, the corresponding grasping signal is generated. This signal contains specific instructions on how to grasp the component and how to deal with it after grasping. The generated grasping signal is sent to the unit responsible for grasping, such as a robotic arm, an automated grasping system, etc., for activation. After the grasping unit is activated, it will determine the placement position of the component according to the preset logic and the on-site situation. This position may be a specific spatial coordinate or a relative position, such as a specific area on a conveyor belt. Finally, the grasping unit will take out the target metal component according to the determined placement position.
[0070] Furthermore, the present application also includes:
[0071] Processing the detection result by image drawing to generate a visual detection result;
[0072] generating a retrieval record identifier based on a target metal component retrieval by the retrieval operation;
[0073] The removal record identifier is associated and matched with the visual detection result according to the placement position to obtain an associated identifier of the target metal component removed by the removal operation.
[0074] Specifically, the inspection results are presented in the form of images using image processing software or specific visualization tools. This includes marking the location, size and type of defects in the image, and using different colors or marks to indicate defects of different degrees. The generated visualization inspection results can be saved as image files, such as JPEG, PNG and other formats, for easy viewing and sharing. When the target metal component is taken out through the removal operation, a unique removal record identifier is generated for it. This identifier can be a number, a letter, or a combination of the two, which is used to uniquely identify the removed component. According to the placement position information, the removal record identifier is associated and matched with the corresponding visualization inspection result. Each removed metal component has its corresponding visualization inspection result. The association matching can be achieved through database query, such as using SQL statements to find and match the corresponding records in the database. Once the removal record identifier is successfully associated with the visualization inspection result, the association identifier of the target metal component removed through the removal operation can be obtained. This association identifier contains the unique identifier of the component and its corresponding inspection result information, providing important data support for quality control and process management.
[0075] In summary, the online detection method based on eddy current non-destructive metal flaw detection provided by the present application has the following technical effects:
[0076] Through the production process based on metal materials, the target metal component is grasped in real time; multiple eddy current probes are started to sense the eddy current of the target metal component to obtain an eddy current detection parameter set, wherein the eddy current detection parameter includes an eddy current distribution parameter and an eddy current change parameter; based on the eddy current change parameter, multiple electromagnetic signals generated by the eddy current distribution parameter are captured; based on the multiple electromagnetic signals, an automatic online detection channel is constructed, wherein the automatic online detection channel includes a structure detection branch channel and a defect detection branch channel; the structural information of the target metal component is output through the structure detection branch channel, and the structural information is synchronized to the defect detection branch channel for automatic detection to obtain the detection result of the target metal component; the grasping unit is activated to remove the target metal component that has been detected, and the removed target metal component is associated and identified according to the detection result, which effectively solves the problem that the existing technology is cumbersome and limited by manual participation, resulting in low detection efficiency and accuracy, and the risk of defect omission. The automatic detection channel is constructed through the electromagnetic signals generated by the eddy current detection technology, and the structure and potential defects of the target metal are accurately detected respectively, thereby avoiding defect omission and making the entire detection process more reliable and comprehensive.
[0077] Embodiment 2
[0078] Based on the same inventive concept as the online detection method based on eddy current non-destructive metal flaw detection in the aforementioned embodiment, the present application also provides an online detection system based on eddy current non-destructive metal flaw detection, please refer to the attached Figure 2 , the system comprising:
[0079] A metal component acquisition module 11, wherein the metal component acquisition module 11 is used to capture a target metal component in real time based on a production process of metal materials;
[0080] A detection parameter set acquisition module 12, wherein the detection parameter set acquisition module 12 is used to start multiple eddy current probes to sense eddy currents of a target metal component and obtain an eddy current detection parameter set, wherein the eddy current detection parameters include eddy current distribution parameters and eddy current variation parameters;
[0081] An electromagnetic signal capturing module 13, the electromagnetic signal capturing module 13 is used to capture a plurality of electromagnetic signals generated by the eddy current distribution parameters based on the eddy current variation parameters;
[0082] A detection channel construction module 14, wherein the detection channel construction module 14 is used to construct an automatic online detection channel based on the multiple electromagnetic signals, wherein the automatic online detection channel includes a structure detection branch channel and a defect detection branch channel;
[0083] A detection result acquisition module 15, the detection result acquisition module 15 is used to output the structural information of the target metal component through the structure detection branch channel, synchronize the structural information to the defect detection branch channel for automatic detection, and obtain the detection result of the target metal component;
[0084] The associated identification module 16 is used to activate the grasping unit to take out the target metal component that has completed the detection, and associate and identify the taken out target metal component according to the detection result.
[0085] Furthermore, the detection parameter set acquisition module 12 in the system is also used for:
[0086] Applying high-frequency alternating current to the plurality of eddy current probes through an excitation power supply to generate eddy current data of the target metal component;
[0087] Based on the eddy current data, the plurality of eddy current probes are started for detection, and eddy current change sensing is performed through the started eddy current probes to generate eddy current change parameters;
[0088] Based on the eddy current variation parameters, the remaining eddy current probes are started for detection to capture eddy current distribution parameters;
[0089] The eddy current variation parameter and the eddy current distribution parameter are integrated to determine the eddy current detection parameter set.
[0090] Furthermore, the detection channel construction module 14 in the system is also used for:
[0091] Extracting features of the plurality of electromagnetic signals using a signal processing algorithm to obtain metal structure features of the target metal component;
[0092] Constructing a feature vector based on the metal structure feature, performing supervised training using the feature vector as training data, and constructing the structure detection branch channel;
[0093] extracting a selected signal from the plurality of electromagnetic signals to determine a metal defect signal;
[0094] Analyze parameter changes of the metal defect signal to obtain multiple parameter change sets, wherein the multiple parameter change sets include frequency change parameters, amplitude change parameters, and phase change parameters;
[0095] Retrieving the historical defect types of the target metal component, performing data association in combination with the frequency change parameter, the amplitude change parameter, and the phase change parameter, and establishing a metal defect matching database;
[0096] Performing random selection based on the metal defect matching database to determine a first metal defect sample;
[0097] Perform unsupervised training using the first metal defect sample as training data to construct the defect detection branch channel;
[0098] The structure detection branch channel and the defect detection branch channel are integrated to obtain the automatic online detection channel.
[0099] Furthermore, the system further comprises an iterative optimization module, wherein the iterative optimization module is used to:
[0100] Dividing the first metal defect sample into data based on metal defect characteristics to generate a training data set and a verification data set;
[0101] Performing unsupervised training on the defect detection branch channel based on deep learning using the training data set;
[0102] Performing verification and evaluation on the output result of the defect detection branch channel according to the verification data set to generate an evaluation result;
[0103] The defect detection branch channel is iteratively optimized according to the evaluation result. When the preset requirements are met for consecutive preset times, it is considered to be converged, and the defect detection branch channel is output.
[0104] Furthermore, the detection result acquisition module 15 in the system is also used for:
[0105] Setting a detection boundary value based on the structural information, and setting a defect detection threshold based on the detection boundary value;
[0106] Decomposing the structural information to obtain the outer surface structural information and the inner surface structural information of the target metal component;
[0107] Determining whether the outer surface structure information is greater than or equal to the defect detection threshold, and whether the inner surface structure information is greater than or equal to the defect detection threshold;
[0108] If the outer surface structure information is less than the defect detection threshold, and the inner surface structure information is less than or equal to the defect detection threshold, it is considered that the target metal component has no defects, and a first detection result is generated;
[0109] If the outer surface structure information is greater than or equal to the defect detection threshold, and the inner surface structure information is less than the defect detection threshold, it is considered that there is a defect on the outer surface of the target metal component, the outer surface of the target metal component is traversed to locate the defect, generate a first defect report, and determine a second detection result;
[0110] If the outer surface structure information is less than the defect detection threshold, and the inner surface structure information is greater than or equal to the defect detection threshold, it is considered that there is a defect on the inner surface of the target metal component, and the inner surface of the target metal component is traversed to locate the defect, generate a second defect report, and determine a third detection result;
[0111] If the outer surface structural information is greater than or equal to the defect detection threshold, and the inner surface structural information is greater than or equal to the defect detection threshold, it is deemed that both the outer surface and the inner surface of the target metal component have defects, a defect traceability label is marked on the target metal component, a third defect report is generated, and a fourth detection result is determined;
[0112] The detection result of the target metal component is output based on the first detection result, the second detection result, the third detection result, and the fourth detection result.
[0113] Furthermore, the association identification module 16 in the system is also used for:
[0114] Based on the detection results, the target metal components that have been detected are marked and the grabbing decision logic is preset;
[0115] generating a grasping signal according to the grasping decision logic, and sending the grasping signal to the grasping unit for activation;
[0116] The placement position is determined by activating the grabbing unit, and the target metal component is removed according to the placement position.
[0117] Furthermore, the system further comprises an association matching module, wherein the association matching module is used for:
[0118] Processing the detection result by image drawing to generate a visual detection result;
[0119] generating a retrieval record identifier based on a target metal component retrieval by the retrieval operation;
[0120] The removal record identifier is associated and matched with the visual detection result according to the placement position to obtain an associated identifier of the target metal component removed by the removal operation.
[0121] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The online detection method based on eddy current nondestructive metal flaw detection and the specific examples in the first embodiment are also applicable to the online detection system based on eddy current nondestructive metal flaw detection in the present embodiment. Through the above detailed description of the online detection method based on eddy current nondestructive metal flaw detection, those skilled in the art can clearly know the online detection system based on eddy current nondestructive metal flaw detection in the present embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.
[0122] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
[0123] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.
Claims
1. An online detection method based on eddy current non-destructive metal flaw detection, It is characterized in that The method comprises: Based on the production process of metal materials, the target metal components can be captured in real time; Starting a plurality of eddy current probes to sense eddy currents of a target metal component to obtain an eddy current detection parameter set, wherein the eddy current detection parameter set includes an eddy current distribution parameter and an eddy current variation parameter; Based on the eddy current variation parameters, capturing a plurality of electromagnetic signals generated by the eddy current distribution parameters; Building an automatic online detection channel based on the multiple electromagnetic signals, wherein the automatic online detection channel includes a structure detection branch channel and a defect detection branch channel; Outputting structural information of the target metal component through the structural detection branch channel, synchronizing the structural information to the defect detection branch channel for automatic detection, and obtaining the detection result of the target metal component; Activate the grabbing unit to take out the target metal component that has completed the inspection, and associate and mark the taken out target metal component according to the inspection result; A plurality of eddy current probes are started to sense eddy currents of a target metal component to obtain an eddy current detection parameter set, the method comprising: Applying high-frequency alternating current to the plurality of eddy current probes through an excitation power supply to generate eddy current data of the target metal component; Based on the eddy current data, the plurality of eddy current probes are started for detection, and eddy current change sensing is performed through the started eddy current probes to generate eddy current change parameters; Based on the eddy current variation parameters, the remaining eddy current probes are started for detection to capture eddy current distribution parameters; Integrating the eddy current variation parameter and the eddy current distribution parameter to determine the eddy current detection parameter set; Based on the multiple electromagnetic signals, an automatic online detection channel is constructed, and the method includes: Extracting features of the plurality of electromagnetic signals using a signal processing algorithm to obtain metal structure features of the target metal component; Constructing a feature vector based on the metal structure feature, performing supervised training using the feature vector as training data, and constructing the structure detection branch channel; extracting a selected signal from the plurality of electromagnetic signals to determine a metal defect signal; Analyze parameter changes of the metal defect signal to obtain multiple parameter change sets, wherein the multiple parameter change sets include frequency change parameters, amplitude change parameters, and phase change parameters; Retrieving historical defect types of target metal components, performing data association based on the frequency change parameter, the amplitude change parameter, and the phase change parameter, and establishing a metal defect matching database; Performing random selection based on the metal defect matching database to determine a first metal defect sample; Perform unsupervised training using the first metal defect sample as training data to construct the defect detection branch channel; Integrate the structure detection branch channel with the defect detection branch channel to obtain the automatic online detection channel; The structural detection branch channel will first analyze and process the electromagnetic signals of the target metal component to extract the structural information of the component; when the structural detection branch channel completes the extraction of the structural information, this information will be synchronously transmitted to the defect detection branch channel; after receiving the structural information, the defect detection branch channel will combine the electromagnetic signal data to perform automated defect detection; by comparing the structural information with the actual electromagnetic signal characteristics, the defect detection branch channel can identify abnormal signals that do not conform to the normal structure, thereby determining the defects inside the component; after completing the detection, the defect detection branch channel will output the detection results of the target metal component.
2. The method according to claim 1, It is characterized in that The method of using the first metal defect sample as training data for unsupervised training to construct the defect detection branch channel includes: Dividing the first metal defect sample into data based on metal defect characteristics to generate a training data set and a verification data set; Performing unsupervised training on the defect detection branch channel based on deep learning using the training data set; Performing verification and evaluation on the output result of the defect detection branch channel according to the verification data set to generate an evaluation result; The defect detection branch channel is iteratively optimized according to the evaluation result. When a preset requirement is met for a preset number of consecutive times, it is considered to be converged, and the defect detection branch channel is output.
3. The method according to claim 1, It is characterized in that The structural information is synchronized to the defect detection branch channel for automatic detection to obtain the detection result of the target metal component, the method comprising: Setting a detection boundary value based on the structural information, and setting a defect detection threshold based on the detection boundary value; Decomposing the structural information to obtain the outer surface structural information and the inner surface structural information of the target metal component; Determining whether the outer surface structure information is greater than or equal to the defect detection threshold, and whether the inner surface structure information is greater than or equal to the defect detection threshold; If the outer surface structure information is smaller than the defect detection threshold, and the inner surface structure information is smaller than the defect detection threshold, it is considered that the target metal component has no defects, and a first detection result is generated; If the outer surface structure information is greater than or equal to the defect detection threshold, and the inner surface structure information is less than the defect detection threshold, it is considered that there is a defect on the outer surface of the target metal component, the outer surface of the target metal component is traversed to locate the defect, generate a first defect report, and determine a second detection result; If the outer surface structure information is less than the defect detection threshold, and the inner surface structure information is greater than or equal to the defect detection threshold, it is considered that there is a defect on the inner surface of the target metal component, and the inner surface of the target metal component is traversed to locate the defect, generate a second defect report, and determine a third detection result; If the outer surface structural information is greater than or equal to the defect detection threshold, and the inner surface structural information is greater than or equal to the defect detection threshold, it is deemed that there are defects on both the outer surface and the inner surface of the target metal component, a defect traceability label is marked on the target metal component, a third defect report is generated, and a fourth detection result is determined; The detection result of the target metal member is output based on the first detection result and the second detection result and the third detection result and the fourth detection result.
4. The method according to claim 1, It is characterized in that Activate the grabbing unit to take out the target metal component that has been inspected, and the method includes: Based on the detection results, the target metal components that have been detected are marked and the grabbing decision logic is preset; generating a grasping signal according to the grasping decision logic, and sending the grasping signal to the grasping unit for activation; The placement position is determined by activating the grabbing unit, and the target metal component is removed according to the placement position.
5. The method according to claim 4, It is characterized in that According to the detection result, an associated identification is performed with the removed target metal component, and the method includes: Processing the detection result by image drawing to generate a visual detection result; generating a retrieval record identifier based on a target metal component retrieval by the retrieval operation; The removal record identifier is associated and matched with the visual detection result according to the placement position to obtain an associated identifier of the target metal component removed by the removal operation.
6. An online detection system based on eddy current non-destructive metal flaw detection, It is characterized in that For implementing the steps of the method according to any one of claims 1 to 5, the system comprises: A metal component acquisition module, which is used to capture target metal components in real time based on the production process of metal materials; A detection parameter set acquisition module, the detection parameter set acquisition module is used to start multiple eddy current probes to sense the eddy current of the target metal component and obtain an eddy current detection parameter set, the eddy current detection parameters including eddy current distribution parameters and eddy current change parameters; An electromagnetic signal capturing module, the electromagnetic signal capturing module being used to capture a plurality of electromagnetic signals generated by the eddy current distribution parameters based on the eddy current variation parameters; A detection channel construction module, wherein the detection channel construction module is used to construct an automatic online detection channel based on the multiple electromagnetic signals, wherein the automatic online detection channel includes a structure detection branch channel and a defect detection branch channel; A detection result acquisition module, the detection result acquisition module is used to output the structural information of the target metal component through the structure detection branch channel, synchronize the structural information to the defect detection branch channel for automatic detection, and obtain the detection result of the target metal component; An association identification module is used to activate the grasping unit to perform a removal operation on the target metal component that has completed the detection, and to associate and identify the removed target metal component according to the detection result.
