Steel crack depth estimation based on 2D images using artificial neural networks

Mohamed, Yasser S.; Shehata, Hesham M.; Abdellatif, Mohamed; Awad, Taher

doi:10.1016/j.aej.2019.10.001

Cited by 24 publications

(15 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Machine learning and, more specifically, deep learning are becoming promising tools for structural health monitoring in civil engineering. This involves in particular surface inspection to detect cracks in buildings, 24,25 roads, 26,27 or metallic structures 28,29 . In these methods, the inspection is done using 2D images of the damaged structure.…”

Section: Introductionmentioning

confidence: 99%

Automatic detection of fatigue crack paths using digital image correlation and convolutional neural networks

Strohmann

Starostin‐Penner

Breitbarth

et al. 2021

Fatigue Fract Eng Mat Struct

View full text Add to dashboard Cite

The occurrence of fatigue cracks is an inherent part of the design of engineering structures subjected to nonconstant loads. Thus, the accurate description of cracks in terms of location and evolution during service conditions is mandatory to fulfill safety‐relevant criteria. In the present work, we implement a deep convolutional neural network to detect crack paths together with their crack tips based on displacement fields obtained using digital image correlation. To this purpose, fatigue crack propagation experiments were performed for AA2024‐T3 rolled sheets using specimens with different geometries. Several hundred datasets were acquired by digital image correlation during the experiments. A part of the displacement data from one of the specimens was then used to train the neural network. The results show that the method can accurately detect the shape and evolution of the cracks in all specimens. Adding synthetic data generated by finite element analyses to the training step improved the accuracy for cracks with stress intensity factors that exceeded the range of the original training data.

show abstract

Section: Introductionmentioning

confidence: 99%

Automatic detection of fatigue crack paths using digital image correlation and convolutional neural networks

Strohmann

Starostin‐Penner

Breitbarth

et al. 2021

Fatigue Fract Eng Mat Struct

View full text Add to dashboard Cite

show abstract

“…However, the accuracy of their networks was not sufficient for practical health monitoring projects. 30 Convolutional neural network (CNN) is one of the most powerful networks for crack detection. CNN is capable of updating the feature extracting weights in the training process to minimize errors.…”

Section: Introductionmentioning

confidence: 99%

“…To facilitate automatic inspection of crack monitoring, Mohamed et al trained several neural networks for estimating crack depth on steel surfaces. However, the accuracy of their networks was not sufficient for practical health monitoring projects 30 . Convolutional neural network (CNN) is one of the most powerful networks for crack detection.…”

Section: Introductionmentioning

confidence: 99%

Arc Length method for extracting crack pattern characteristics

Asjodi

Daeizadeh

Hamidia

et al. 2020

Struct Control Health Monit

View full text Add to dashboard Cite

Although manual crack inspection has been widely used for structural health monitoring over the last decades, the development of computer vision methods allows continuous monitoring and compensates the human judgment inaccuracy. In this study, an image-based method entitled Arc Length method is introduced for extracting crack pattern characteristics, including crack width and crack length. The method contains two major steps; in the first step, the crack zones are estimated in the whole image. Afterwards, the algorithm finds the start point, follows the crack pattern, and measures the crack features, such as crack width, crack length, and crack pattern angle. The efficiency of the method is validated using a few case studies from cracked structural concrete shear walls tested in the laboratory under quasi-static cyclic loadings. The case studies show high efficiency of the proposed method in following the crack patterns even when the crack propagates in two or more branches. The application of this approach plays a significant role in crack monitoring of infrastructures, such as concrete bridges and tunnels.

show abstract

“…[6][7][8][9][10][11][12][13][14][15][16][17][18] Nowadays, for a quantitative description of complex systems, artificial neural networks (ANN) are used. [19][20][21][22][23][24][25][26][27][28] For example in Mahjani et al [23] three-layered, feed-forward ANNs have been constructed to predict the corrosion current of 316 stainless steel as a function of CuSO 4 concentration, solution pH, and electrode surface area. In Kakooei et al [24] corrosion rate of the heat-affected zone region in welded X65 pipeline steel was studied.…”

Section: Introductionmentioning

confidence: 99%

Corrosion damage of 316L steel surface examined using statistical methods and artificial neural network

Kubisztal

Haneczok

2020

Materials & Corrosion

View full text Add to dashboard Cite

Detailed examination of corrosion-induced changes of the 316L steel surface (immersed in 5 wt% NaCl solution) is presented and discussed. The evolution of the stable pit depth (h av) with the immersion time (t) was established using 3D maps and statistic techniques. It was found that h t n av ∝ with n ≈ 0.5. Moreover, determination of the pit area allows estimating the curve current density (j) versus the immersion time and it was found that j t m − ∝ with m ≈ 1. A novel technique for surface corrosion degree determination is based on analysis of 2D grayscale images instead of black and white images showing that corrosion morphology was elaborated. For this purpose a three-layered, feed-forward neural network with the Levenberg-Marquardt backpropagation training algorithm was used. It was shown that a dependence corrosion degree versus immersion time (S-type curve) can be fully described by the proposed procedure.

show abstract

Steel crack depth estimation based on 2D images using artificial neural networks

Cited by 24 publications

References 22 publications

Automatic detection of fatigue crack paths using digital image correlation and convolutional neural networks

Automatic detection of fatigue crack paths using digital image correlation and convolutional neural networks

Arc Length method for extracting crack pattern characteristics

Corrosion damage of 316L steel surface examined using statistical methods and artificial neural network

Contact Info

Product

Resources

About