CN115015394A - Composite material defect ultrasonic detection method based on convolution network and trajectory tracking - Google Patents
Composite material defect ultrasonic detection method based on convolution network and trajectory tracking Download PDFInfo
- Publication number
- CN115015394A CN115015394A CN202210810586.2A CN202210810586A CN115015394A CN 115015394 A CN115015394 A CN 115015394A CN 202210810586 A CN202210810586 A CN 202210810586A CN 115015394 A CN115015394 A CN 115015394A
- Authority
- CN
- China
- Prior art keywords
- ultrasonic
- probe
- data
- defect
- composite
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
- G01N29/04—Analysing solids
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
- G01N29/04—Analysing solids
- G01N29/048—Marking the faulty objects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
- G01N29/22—Details, e.g. general constructional or apparatus details
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Acoustics & Sound (AREA)
- Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
Abstract
The invention discloses a composite material defect ultrasonic detection method based on a convolution network and trajectory tracking, which comprises the following steps of 1, establishing an ultrasonic database; step 2, acquiring a data set from an ultrasonic database, constructing a deep learning neural network algorithm model, and determining parameters of the neural network model; step 3, training the constructed neural network deep learning algorithm model by using the acquired and constructed data set; and 4, embedding the trained neural network model parameters into ultrasonic equipment, and performing real-time diagnosis classification and defect visualization marking on the ultrasonic signals acquired at the current probe position. The method can realize real-time composite material defect diagnosis and marking. The defect visualization method based on the ultrasonic C-scan is provided to solve the problem that the existing composite defect or damage detection depends on the knowledge of the material characteristics of a test part or the expert priori knowledge of the extraction of the predetermined signal characteristics based on physics.
Description
Technical Field
The invention belongs to the technical field of composite material defect detection, and particularly relates to a composite material defect ultrasonic detection method based on a convolution network and trajectory tracking.
Background
The composite material has higher specific strength and specific rigidity, can effectively reduce the weight of the airplane structure, and also has the advantages of corrosion resistance and high damage safety. The composite material has excellent performance, is represented by higher and higher application ratio in the airplane structure, and has been developed from a secondary structure to a primary structure and a complex stressed structure. Composite components may be defective during production and various types of damage may occur during use, with the types of damage and defects including cracking, scratching, ablation, lightning strikes, pits, perforations, delamination, debonding, and the like. The method is crucial to the detection and analysis of defects and damages. There are many methods of evaluating composite materials or components, non-destructive testing being one of the important categories, which refers to the evaluation and testing of materials or parts to characterize or discover defects and damage without altering the original properties or harming the object being tested. The non-destructive testing technique provides a cost-effective testing means that can be used to investigate samples individually, or to inspect the entire material during the production process, and to inspect the entire material in a production quality control system. At present, the method for detecting the structural damage of the composite material of the airplane mainly comprises a visual detection method, a knocking detection method, an ultrasonic detection method, an infrared thermal wave method and an X-ray detection method.
The nondestructive detection of the instrument and equipment is mainly used for practical detection of infield maintenance and quality evaluation in the production process, wherein the ultrasonic detection technology has the characteristics of strong penetrating power and high sensitivity, and meanwhile, the detection equipment is light and low in cost, so that the nondestructive detection method is widely applied to detection of composite materials, and is suitable for detection of damages or defects such as layering, degumming, adhesive bonding air holes and the like of composite material structures.
The existing ultrasonic detection method is taken as a wide composite material nondestructive detection technology, the traditional ultrasonic nondestructive detection technology relies on the understanding of the material characteristics of a test part or the extraction of preset signal characteristics based on physics, the priori knowledge of ultrasonic detection experts damaged by different composite material parts is difficult to obtain, and meanwhile, the problems of high labor cost and low detection efficiency exist in manual signal interpretation.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a composite material defect ultrasonic detection method based on a convolution network and trajectory tracking so as to realize real-time composite material defect diagnosis and marking. The defect visualization method based on the ultrasonic C-scan is provided to solve the problem that the existing composite material defect or damage detection depends on the knowledge of the material characteristics of a test part or the expert priori knowledge of the extraction of the preset signal characteristics based on physics.
In order to achieve the purpose, the invention provides the following technical scheme:
a composite material defect ultrasonic detection method based on convolution network and trajectory tracking comprises the following steps,
step 3, training the constructed neural network deep learning algorithm model by using the acquired and constructed data set;
and 4, embedding the trained neural network model parameters into ultrasonic equipment, and performing real-time diagnosis classification and defect visualization marking on the ultrasonic signals acquired at the current probe position.
Preferably, in step 1, data acquisition is performed on different areas of the sample piece by using an ultrasonic device, so as to obtain ultrasonic signal sampling data and ultrasonic probe picture data with different sensitive waveforms, and establish an ultrasonic database containing different sensitive waveforms.
Preferably, in step 1, the step of establishing the ultrasonic database specifically comprises the following steps,
step S11, placing the tested sample of the composite material on a detection platform;
step S12, starting the ultrasonic equipment and adjusting ultrasonic parameters;
step S13, uniformly spreading a couplant on a composite material sample piece, attaching an ultrasonic receiving and transmitting probe on a sample to be tested to sample data in different areas, and obtaining ultrasonic signal sampling data;
step S14, exporting and dividing the ultrasonic signal sampling data into a training data set and a testing data set;
and step S15, acquiring pictures of the ultrasonic probe at different angles by using the camera to clamp the ultrasonic probe by using a hand-held mode and an automatic scanning machine, acquiring picture data of the ultrasonic probe, and dividing the picture data of the ultrasonic probe into a training data set and a testing data set.
Preferably, in step 1, the data set in the ultrasonic database shares three types of normal composite data, defect composite data and reinforced composite data according to different sampling regions.
Preferably, in step 2, a one-dimensional convolution network model is used for extracting characteristics of ultrasonic time sequence signals in an ultrasonic database, a depth residual error network model is used for extracting visual characteristics of an ultrasonic probe, and the extracted characteristics are connected in parallel by a classifier to construct a deep learning neural network algorithm model.
Further, in step 3, when a one-dimensional convolution network model and a residual error feature extraction backbone network model in the neural network deep learning algorithm model are trained, gradient pass-back is performed, and parameters of the convolution network model are updated.
Further, in step 3, a conjugate gradient method is adopted to predict the motion track of the probe, the output of the one-dimensional convolution network module classifier is connected in parallel to the input of the prediction result, and the position of the probe is recorded to obtain the track of the probe with the defect information of the composite material.
Preferably, in step 4, after real-time diagnosis and classification and visual defect marking, an intelligent detection marking system for automatically scanning the composite material sample piece is set up, and full-automatic scanning and result visual marking verification are performed on the composite material sample piece.
Furthermore, the intelligent detection and marking system for the automatically scanned composite material sample piece comprises a camera, a servo mechanism, a data processor, a probe and a servo mechanism controller;
the probe points to the sample to be measured, the probe is fixed on a servo mechanism, and the servo mechanism drives the probe to move; the camera points to the sample to be measured;
the data processor is used for processing data acquired by the camera and the probe; the servo mechanism controller is used for controlling the servo mechanism;
and the camera and the data processor are in data transmission through a USB (universal serial bus) line.
Further, the data processor is connected with the display, and the GUI window is adopted to display the diagnosis result, the defect information and the motion trail of the probe.
Compared with the prior art, the invention has the following beneficial technical effects:
the invention provides a composite material defect ultrasonic detection method based on a convolutional network and track tracking, which is characterized in that a one-dimensional convolutional neural network model for detecting composite material damage is obtained by a data-driven decision method and a deep learning technology, and a plurality of types of characteristics can be fused by utilizing a neural network, so that the model has the capability of extracting and identifying the plurality of types of characteristics, the defect state of a real-time position is recorded by combining track tracking, and the model is embedded into equipment to obtain ultrasonic equipment with a real-time intelligent diagnosis function.
Drawings
FIG. 1 is a schematic diagram of an ultrasonic intelligent identification and labeling hardware system of the present invention;
FIG. 2 is a diagram of a neural network model architecture constructed in accordance with the present invention;
FIG. 3 is a data flow thread diagram of the ultrasonic intelligent identification and labeling system of the present invention;
in the drawings: 1 is a camera; 2 is a USB line; 3 is a servo mechanism; 4 is a data processor; 5 is a probe; 6 is a sample to be measured; and 7, a servo mechanism controller.
Detailed Description
The present invention will now be described in further detail with reference to specific examples, which are intended to be illustrative, but not limiting, of the invention.
The embodiment of the invention provides a method for defect ultrasonic intelligent detection and marking of a composite material based on a convolutional neural network and track tracking, which comprises the following steps of S1-S5:
step S1, establishing an ultrasonic database containing sensitive waveforms with different characteristics, and providing a sample set for training and testing a convolution network model;
the method comprises the steps of designing and manufacturing a stringer-added composite material defect sample piece, wherein a customized composite laminated board is adopted in the sample piece, the delamination and crack defects of the composite material are simulated in two areas by embedding an iron wafer and a polytetrafluoroethylene sheet, three reinforcing ribs are connected below the laminated board to simulate a real airplane skin lower stringer, and the sample piece is divided into a normal composite laminated board area, a defect area and a laminated board with reinforcing ribs.
The method comprises the following steps of respectively acquiring data of different areas of a sample piece by utilizing ultrasonic equipment to obtain characteristic data with different sensitive waveforms, wherein the specific process of acquiring the data by utilizing the ultrasonic equipment comprises the following steps:
step S11, placing the composite material sample piece on a detection platform, and fixing the sample 6 to be detected;
step S12, starting ultrasonic equipment, and adjusting relevant parameters such as material sound velocity, sampling frequency and the like to appropriate values;
step S13, uniformly spreading a couplant on the composite material sample piece, attaching the ultrasonic transceiving probe 5 to the sample 6 to be tested, and sampling ultrasonic signal data in different areas;
step S14, leading out ultrasonic signal sampling data into a time sequence signal training data set and a time sequence signal testing data set, wherein the time sequence signal data set has three types of normal composite material data, defect composite material data and reinforced rib composite material data according to different sampling areas;
and step S15, acquiring pictures of the ultrasonic probe at different angles, which are held by a hand and an automatic scanning machine, by the camera, constructing a data set by the pictures of the ultrasonic probe acquired by the camera, and dividing the data set into a training data set and a testing data set.
Step S2, designing and constructing a deep learning neural network algorithm aiming at the data set, and determining network model parameters;
in the embodiment, the one-dimensional convolution is used for extracting the characteristics of the ultrasonic time sequence signal, the visual characteristics of the ultrasonic probe are abstracted by using the depth residual error network, and the abstracted characteristics are connected in parallel by using two classifiers to construct the neural network structure.
As shown in fig. 2, the architecture of the one-dimensional convolutional neural network for extracting time-series signal features includes 5 convolutional layers, 5 pooling layers, 1 full-link layer, and 1 output layer, the first convolutional layer uses 16 1 × 5 kernel filters, and after convolution, a relu activation function is used to generate feature signals, so as to obtain 16 feature signals, and pooling the feature signals, where pooling is a maximum value pooling method. The number of output channels is twice of the number of input channels, pooling is maximum pooling, the characteristic coefficients are combined nonlinearly after 5 layers of convolution and pooling operation, then the outputs of convolution are combined in a linear connection mode, and finally a probability vector with the same length as the signal category number is output; the visual feature abstract network is formed by connecting a residual error network and a convolution network in parallel, the first layer is convolution of 7 x 64, then 16 building blocks including 3, 4, 6 and 3 are provided, each building block is 3 layers, and finally a full connection layer is provided, the resnet50 extracts a depth feature and an ultrasonic probe center position contained in each frame of picture, the one-dimensional convolution network extracts a signal feature, and the two inputs, one is a picture and the other is a one-dimensional signal, so that the networks with different characteristics are used for extracting features. In the example, the output layer of the one-dimensional convolution network and the output layer of the residual error network are combined together, the result is input into a two-layer convolution network, and the classification result of the currently input picture signal and the currently input ultrasonic time sequence signal is obtained through a classifier.
S3, training the constructed neural network deep learning algorithm by using the acquired and constructed data set;
in this example, the deep learning model constructed in step S2 is subjected to model parameter training, the model is mainly divided into two modules, and the main training parameters are a one-dimensional convolution network module and a residual error feature extraction backbone network. Inputting an ultrasonic time sequence signal data set into a one-dimensional convolution network branch, extracting a signal characteristic waveform through one-dimensional convolution pooling operation, and optimizing parameters of a one-dimensional convolution network module through a loss value fed back by data of a training set in end-to-end training; after the picture training data set is input into the target Classification branching network, a ResNet50 backbone network and additional convolution blocks (Classification specific Features) are used to extract a depth feature map, and then the feature map is input into a model predictor composed of an initializer and a loop optimizer. The model predictor outputs the weights of the convolutional layer, which are used in the target classification operation of the feature map extracted from the test frame, and the brighter the color on the probability score classification colors belonging to the tracked target, the higher the confidence that the position of the target is, and the peak point of the score is the new position of the target in the test frame.
When end-to-end training is carried out, test loss is calculated behind a network of a test frame, then gradient return is carried out by using a pytorch deep learning frame, and parameters of a convolutional network are updated, so that network parameters such as feature extraction and online learning are optimized. In the regression process, a conjugate gradient method is used to replace the steepest descent method, an ideal gradient direction is obtained through iteration, the motion track of the probe is predicted, the output of the one-dimensional convolution network module classifier is connected in parallel to the input of the prediction result, the position of the probe is recorded, and the track of the probe with the defect information of the composite material is obtained.
S4, embedding the trained neural network model parameters into the developed ultrasonic equipment to realize real-time diagnosis classification and defect visualization marking of the ultrasonic signals acquired at the current probe position;
in this embodiment, the same feedforward neural network as that in step S2 is constructed in a development kit of the ultrasound apparatus according to convolution operation, pooling operation, and full-link operation by using the neural network model trained in step S3, network parameters of the feedforward neural network are given by the model file with the highest accuracy trained in step S2, that is, a signal classification thread is created in software control of the ultrasound apparatus, the operation of the thread is to perform operations such as convolution, pooling, and full-link on an input one-dimensional time sequence signal and picture data acquired by a camera, and the signal is obtained by using the feedforward neural network, and the confidence that the signal belongs to different regions is obtained by using a final classifier, and the result of the maximum probability is displayed on a visualization window.
The Data input of the signal diagnosis network module is controlled by a Data reading thread, the thread uses a pointer to access a Data cache region at a fixed time interval, when Data exist in the Data cache region, the Data are stored into a Data class, and the time sequence signal input of the model is completed after the Data are called; and the data input of the probe visual tracking module is controlled by a picture reading thread, the thread also accesses the data cache region by a pointer with a fixed frame number, when data exists in the data cache region, the data is stored into an Image class, and the picture data input of the model is completed after the data is called.
After the real-time diagnosis of the ultrasonic signals is finished, the diagnosis result is visually displayed in a GUI window, and the signal state of the current position and the defect information of the previously detected region are displayed at the same time.
S5, setting up an intelligent detection and marking system for automatically scanning the composite material sample piece, and realizing full-automatic scanning of the composite material sample piece and visual marking verification of a result;
the intelligent detection and marking system for the automatic scanning composite material sample piece comprises a camera 1, a servo mechanism 3, a data processor 4, a probe 5 and a servo mechanism controller 7; the probe 5 points to a sample 6 to be measured, the probe 5 is fixed on the servo mechanism 3, and the servo mechanism 3 drives the probe 5 to move; the camera 1 points to the sample 6 to be measured; the data processor 4 is used for processing the data collected by the camera 1 and the probe 5; the servo mechanism controller 7 is used for controlling the servo mechanism 3; the camera 1 and the data processor 4 perform data transmission through the USB cable 2.
The embodiment combines the prior ultrasonic equipment embedded with a convolutional neural network model, a vision-based probe trajectory tracking system and an automatic scanning platform. The automatic scanning platform consists of a scanning platform control system, a power supply and a three-way guide rail, and an ultrasonic probe, a composite material test piece, a camera and other hardware are combined and installed on the scanning platform to obtain a complete system as shown in figure 1. The detection scanning process of the system can be manually operated or automatically operated, the scanning speed and the scanning direction can be controlled by a set program, the pressure of the piezoelectric crystal is sensed by the sensor, power-off protection can be performed when the pressure exceeds a set pressure value, and the integrity of equipment and the safety of experimental operation are guaranteed. The visualization of the probe trajectory is completed in step S4, and by inputting the signal classification data file output in step S3 into the trajectory tracking process, the current trajectory display color is controlled based on the current ultrasound signal classification before the trajectory visualization: green corresponds to the normal skin region category, red corresponds to the defect composite region category, and yellow corresponds to the reinforcing rib region category. Therefore, the software and hardware connection of the whole system is completed, and the automatic scanning of sample hardware, the real-time diagnosis of ultrasonic signals at any position and the marking of the defect condition of the composite material in a scanning moving area are realized.
According to the method, the sensitive waveform of the composite material is extracted by utilizing one-dimensional convolution, and an ultrasonic signal diagnosis model with high accuracy is obtained; aiming at the difference between a deep learning environment and a software and hardware control environment of ultrasonic equipment, the method provides the method for training network model parameters by using a deep learning neural network, constructs feedforward neural networks with the same network structure and calls model parameters to be embedded into the ultrasonic equipment, removes a complex back propagation operation process, deploys the model in an ultrasonic detection device in a light weight manner, and realizes real-time intelligent diagnosis and defect visualization marking of ultrasonic signals; the invention utilizes the developed intelligent ultrasonic equipment to match with the automatic scanning device system, realizes the visual marking of the defects on the scanned path under the condition of single-point high-accuracy diagnosis without information loss, realizes the automatic nondestructive detection of the composite material, and verifies that the method has better research prospect.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on differences from other embodiments. In particular, as for the system embodiment, since it is substantially similar to the method embodiment, the description is relatively simple, and reference may be made to the partial description of the method embodiment for relevant points.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (10)
1. A composite material defect ultrasonic detection method based on convolution network and trajectory tracking is characterized by comprising the following steps,
step 1, establishing an ultrasonic database;
step 2, acquiring a data set from an ultrasonic database, constructing a deep learning neural network algorithm model, and determining parameters of the neural network model;
step 3, training the constructed neural network deep learning algorithm model by using the acquired and constructed data set;
and 4, embedding the trained neural network model parameters into ultrasonic equipment, and performing real-time diagnosis classification and defect visualization marking on the ultrasonic signals acquired at the current probe position.
2. The ultrasonic composite defect detection method based on the convolutional network and the trajectory tracking as claimed in claim 1, wherein in step 1, an ultrasonic device is used to collect data of different areas of the sample piece, so as to obtain ultrasonic signal sampling data and ultrasonic probe picture data with different sensitive waveforms, and establish an ultrasonic database containing different sensitive waveforms.
3. The ultrasonic composite defect detection method based on the convolution network and the trajectory tracking as claimed in claim 1, wherein the step 1 of establishing the ultrasonic database specifically comprises the following steps,
step S11, placing the composite material sample to be tested (6) on a detection platform;
step S12, starting the ultrasonic equipment and adjusting ultrasonic parameters;
step S13, uniformly spreading a couplant on a composite material sample piece, attaching an ultrasonic transceiving probe (5) to a sample (6) to be tested, and sampling data of different areas to obtain ultrasonic signal sampling data;
step S14, exporting and dividing the ultrasonic signal sampling data into a training data set and a testing data set;
and step S15, acquiring pictures of the ultrasonic probe (5) at different angles by using a camera (1) to hold the ultrasonic probe by a hand and an automatic scanning machine, acquiring picture data of the ultrasonic probe, and dividing the picture data of the ultrasonic probe into a training data set and a testing data set.
4. The ultrasonic composite defect detection method based on the convolutional network and the trajectory tracking as claimed in claim 1, wherein in step 1, the data sets in the ultrasonic database have three types of normal composite data, defect composite data and reinforced composite data according to different sampling regions.
5. The ultrasonic composite defect detection method based on the convolution network and the trajectory tracking as claimed in claim 1, wherein in step 2, a one-dimensional convolution network model is used for feature extraction of ultrasonic time sequence signals in an ultrasonic database, a depth residual error network model is used for extracting visual features of an ultrasonic probe, and the extracted features are connected in parallel by a classifier to construct a deep learning neural network algorithm model.
6. The ultrasonic testing method for composite material defect based on convolution network and trajectory tracking as claimed in claim 5, wherein in step 3, when training the one-dimensional convolution network model and the residual error feature extraction backbone network model in the neural network deep learning algorithm model, gradient pass-back is performed to update the parameters of the convolution network model.
7. The ultrasonic composite defect detection method based on the convolutional network and the trajectory tracking as claimed in claim 5, wherein in step 3, the motion trajectory of the probe (5) is predicted by adopting a conjugate gradient method, the output of the one-dimensional convolutional network module classifier is connected in parallel to the input of the prediction result, and the position of the probe (5) is recorded to obtain the trajectory of the probe with the composite defect information.
8. The ultrasonic detection method for the defects of the composite materials based on the convolutional network and the track tracking as claimed in claim 1, characterized in that in step 4, after real-time diagnosis and classification and defect visualization marking, an intelligent detection marking system for automatically scanning the composite material sample pieces is set up, and full-automatic scanning and result visualization marking verification are performed on the composite material sample pieces.
9. The ultrasonic composite defect detection method based on convolutional network and trajectory tracking as claimed in claim 8, wherein the intelligent detection and labeling system for automatically scanning the composite sample piece comprises a camera (1), a servo mechanism (3), a data processor (4), a probe (5) and a servo mechanism controller (7);
the probe (5) points to a sample (6) to be detected, the probe (5) is fixed on the servo mechanism (3), and the servo mechanism (3) drives the probe (5) to move; the camera (1) points to a sample (6) to be measured;
the data processor (4) is used for processing data acquired by the camera (1) and the probe (5); the servo mechanism controller (7) is used for controlling the servo mechanism (3);
and the camera (1) and the data processor (4) are in data transmission through the USB cable (2).
10. The ultrasonic composite defect detection method based on the convolutional network and the trajectory tracking as claimed in claim 9, wherein the data processor (4) is connected with a display, and a GUI window is adopted to display the diagnosis result, the defect information and the motion trajectory of the probe (5).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210810586.2A CN115015394A (en) | 2022-07-11 | 2022-07-11 | Composite material defect ultrasonic detection method based on convolution network and trajectory tracking |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210810586.2A CN115015394A (en) | 2022-07-11 | 2022-07-11 | Composite material defect ultrasonic detection method based on convolution network and trajectory tracking |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115015394A true CN115015394A (en) | 2022-09-06 |
Family
ID=83082502
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210810586.2A Pending CN115015394A (en) | 2022-07-11 | 2022-07-11 | Composite material defect ultrasonic detection method based on convolution network and trajectory tracking |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115015394A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116858943A (en) * | 2023-02-03 | 2023-10-10 | 台州五标机械股份有限公司 | Hollow shaft intelligent preparation method and system for new energy automobile |
| CN117299596A (en) * | 2023-08-14 | 2023-12-29 | 江苏秦郡机械科技有限公司 | Material screening system and method for automatic detection |
-
2022
- 2022-07-11 CN CN202210810586.2A patent/CN115015394A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116858943A (en) * | 2023-02-03 | 2023-10-10 | 台州五标机械股份有限公司 | Hollow shaft intelligent preparation method and system for new energy automobile |
| CN117299596A (en) * | 2023-08-14 | 2023-12-29 | 江苏秦郡机械科技有限公司 | Material screening system and method for automatic detection |
| CN117299596B (en) * | 2023-08-14 | 2024-05-24 | 江苏秦郡机械科技有限公司 | Material screening system and method for automatic detection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11514567B2 (en) | On-line real-time diagnosis system and method for wind turbine blade (WTB) damage | |
| RU2594368C2 (en) | Method of simulating non-destructive testing operations in real conditions using synthetic signals | |
| CN115015394A (en) | Composite material defect ultrasonic detection method based on convolution network and trajectory tracking | |
| US7333898B2 (en) | Passive structural assessment and monitoring system and associated method | |
| CN110378252A (en) | A kind of distress in concrete recognition methods based on depth migration study | |
| EP1873519A2 (en) | Ultrasonic inspection and repair mode selection | |
| CN111553877A (en) | Damage identification and service life evaluation method based on terahertz ceramic matrix composite blade | |
| KR20190130257A (en) | Prediction method for compression strength of concrete structure based on deep convolutional neural network algorithm and prediction system using the method | |
| JP5156707B2 (en) | Ultrasonic inspection method and apparatus | |
| Hu et al. | Study on intelligent and visualization method of ultrasonic testing of composite materials based on deep learning | |
| CN114324580A (en) | Intelligent knocking detection method and system for structural defects | |
| CN117686588B (en) | Nondestructive monitoring system and method for welding quality of gas pipeline | |
| Smith et al. | Model‐Assisted Probability of Detection Validation for Immersion Ultrasonic Application | |
| CN117054517A (en) | Ultrasonic line detection method for display device | |
| Guibert et al. | Smart NDT tools: Connection and automation for efficient and reliable NDT operations | |
| KR102385850B1 (en) | Multichannel mobile non-destructive ultrasonic testing system with artificial neural networks | |
| Liu et al. | Automatic terahertz recognition of hidden defects in layered polymer composites based on a deep residual network with transfer learning | |
| Shang et al. | Classification of Multistates of Internal Damage in Concrete Based on Convolutional Neural Network Analysis of Time-Frequency Images | |
| Alcaide et al. | Damage detection on Aerospace structures using PAMELA SHM System | |
| Smagulova et al. | Machine learning based approach for automatic defect detection and classification in adhesive joints | |
| CN112419823B (en) | Magnetic memory nondestructive testing virtual experiment teaching system and method for stress and defect | |
| CN206038003U (en) | Positioning detecting system is separated to existing building curtain | |
| Manzano et al. | Towards Automation of Subsurface Defects Detection in Composites Through Analysis of Thermal Images Using a Hybrid Approach | |
| ZHANG et al. | Defect identification in concrete using deep learning technique on nonlinear ultrasonic wave | |
| Dragan et al. | Holistic approach for structural integrity evaluation of composite main rotor blades |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |