Investigation of deep neural network with drop out for ultrasonic flaw classification in weldments

Munir, Nauman; Kim, Hak-Joon; Song, Sung‐Jin; Kang, Sung-Sik

doi:10.1007/s12206-018-0610-1

Cited by 65 publications

(24 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Data augmentation is a well known technology in the ML literature and is commonly considered to be a key enabling technique when working with limited data sets Chollet [8]. Data augmentation has also previously used for NDT applications of ML [25]. In present study, extensive data augmentation was utilized using the previously developed virtual flaw technology.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Augmented Ultrasonic Data for Machine Learning

et al. 2021

View full text Add to dashboard Cite

Flaw detection in non-destructive testing, especially for complex signals like ultrasonic data, has thus far relied heavily on the expertise and judgement of trained human inspectors. While automated systems have been used for a long time, these have mostly been limited to using simple decision automation, such as signal amplitude threshold. The recent advances in various machine learning algorithms have solved many similarly difficult classification problems, that have previously been considered intractable. For non-destructive testing, encouraging results have already been reported in the open literature, but the use of machine learning is still very limited in NDT applications in the field. Key issue hindering their use, is the limited availability of representative flawed data-sets to be used for training. In the present paper, we develop modern, deep convolutional network to detect flaws from phased-array ultrasonic data. We make extensive use of data augmentation to enhance the initially limited raw data and to aid learning. The data augmentation utilizes virtual flaws—a technique, that has successfully been used in training human inspectors and is soon to be used in nuclear inspection qualification. The results from the machine learning classifier are compared to human performance. We show, that using sophisticated data augmentation, modern deep learning networks can be trained to achieve human-level performance.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Recently Meng et al [23], Zhu et al [39] and Munir et al [25] used deep CNNs for defect classification in ultrasonic and EC-data. Meng et al [23] used deep neural networks with an SVM top layer for enhanced classification capability.…”

Section: Introductionmentioning

confidence: 99%

Augmented Ultrasonic Data for Machine Learning

et al. 2021

View full text Add to dashboard Cite

show abstract

“…As described in Section II, the conventional NDT techniques still face several challenges due to the automation complexity [57], poor quality of the spatial image and blurred defect shape, extensive data interpretation as well as penetration limitation which degrade the inspection reliability. Recently, the development of NDT automation with high reliability has emerged as an active research area [58] because of the needs of avoiding the dependence on the skills and experience of the operators [57]. Therefore, signal processing with intelligent classifier can provide reliable and fast inspection [59] which could increase the sensibility of defect detection and automate the monitoring procedure [60].…”

Section: Artificial Intelligence Applications In Conventional Ndtmentioning

confidence: 99%

“…Practically, this massive sample in NDT is difficult to be acquired for the same sample under the inspection. Moreover, a large number of hidden layers and learnable coefficients are needed for accurate classification which makes DNN suffers from overfitting without regularization to integrate this massive structure of DNN [57]. In addition, in case of high noise of ultrasonic NDT, DNN gives the poor performance of defect detection compared to the conventional neural network [85].…”

Section: ) Deep Neural Network (Dnn)mentioning

confidence: 99%

Prospect of Using Artificial Intelligence for Microwave Nondestructive Testing Technique: A Review

2019

View full text Add to dashboard Cite

The development in materials technology has produced stronger, lighter, stiffer, and more durable electrically insulating composites which are replacing metals in many applications. These composites require alternative inspection techniques because the conventional nondestructive testing (NDT) techniques such as thermography, eddy currents, ultrasonic, X-ray and magnetic particles have limitations of inspecting them. Microwave NDT technique employing open-ended rectangular waveguides (OERW) has emerged as a promising approach to detect the defects in both metal and composite materials. Despite its promising results over conventional NDT techniques, OERW microwave NDT technique has shown numerous limitations in terms of poor spatial resolution due to the stand-off distance variations, inspection area irregularities and quantitative estimation in imaging the size of defects. Microwave NDT employing OERW in conjunction with robust artificial intelligence approaches have tremendous potential and viability for evaluating composite structures for the purpose mentioned here. Artificial intelligence techniques with signal processing techniques are highly possible to enhance the efficiency and resolution of microwave NDT technique because the impact of artificial intelligence approaches is proven in various conventional NDT techniques. This paper provides a comprehensive review of NDT techniques as well as the prospect of using artificial intelligence approaches in microwave NDT technique with regards to other conventional NDT techniques. INDEX TERMS Microwave nondestructive testing, open-ended rectangular waveguides, artificial intelligence.

show abstract

“…This approach of using manually engineered features is termed shallow ML. Ultrasonic measurements have been combined with shallow ML algorithms such as Artificial Neural Networks (ANNs) [22][23][24][25][26][27][28][29] and Support Vector Machines (SVMs) [23,25,30,31], using waveform features from the time domain [23,25,27,31,32] and frequency domain [24,27,31,32] after analyses such as wavelet transforms [22,24]. These have been used for applications such as predicting sugar concentration during fermentation [33], measuring particle concentration in multicomponent suspensions [34], and classification of heat exchanger fouling in the dairy industry [23,25].…”

Section: Introductionmentioning

confidence: 99%

Monitoring Mixing Processes Using Ultrasonic Sensors and Machine Learning

Bowler

Bakalis

Watson

2020

Sensors

View full text Add to dashboard Cite

Mixing is one of the most common processes across food, chemical, and pharmaceutical manufacturing. Real-time, in-line sensors are required for monitoring, and subsequently optimising, essential processes such as mixing. Ultrasonic sensors are low-cost, real-time, in-line, and applicable to characterise opaque systems. In this study, a non-invasive, reflection-mode ultrasonic measurement technique was used to monitor two model mixing systems. The two systems studied were honey-water blending and flour-water batter mixing. Classification machine learning models were developed to predict if materials were mixed or not mixed. Regression machine learning models were developed to predict the time remaining until mixing completion. Artificial neural networks, support vector machines, long short-term memory neural networks, and convolutional neural networks were tested, along with different methods for engineering features from ultrasonic waveforms in both the time and frequency domain. Comparisons between using a single sensor and performing multisensor data fusion between two sensors were made. Classification accuracies of up to 96.3% for honey-water blending and 92.5% for flour-water batter mixing were achieved, along with R2 values for the regression models of up to 0.977 for honey-water blending and 0.968 for flour-water batter mixing. Each prediction task produced optimal performance with different algorithms and feature engineering methods, vindicating the extensive comparison between different machine learning approaches.

show abstract

Investigation of deep neural network with drop out for ultrasonic flaw classification in weldments

Cited by 65 publications

References 11 publications

Augmented Ultrasonic Data for Machine Learning

Augmented Ultrasonic Data for Machine Learning

Prospect of Using Artificial Intelligence for Microwave Nondestructive Testing Technique: A Review

Monitoring Mixing Processes Using Ultrasonic Sensors and Machine Learning

Contact Info

Product

Resources

About