Deep learning methods for roping defect analysis in aluminum alloy sheets: prediction and grading

Abstract

Roping is a severe band-like surface defect that occurs in deformed aluminum alloy sheets. Accurate roping prediction and rating are essential for industrial applications. Recently, the authors introduced an artificial neural network (ANN) model to efficiently forecast roping behavior across the thickness of large regions with texture gradients. In this study, the previously proposed ANN model for roping prediction is briefly reviewed, and a few-shot learning (FSL)-based method is developed for roping grading with limited samples. To consider the directionality of the roping patterns, the roping dataset constructed from experimental observations is transformed into the frequency domain for more compact characterization. A transfer-based FSL method is further presented for grade roping with manifold mixup regularization and the Sinkhorn mapping algorithm. A new component-focused representation is also implemented for data-processing, exploiting the close correlation between roping and power distribution in the frequency domain. The ultimate FSL method achieved an optimal accuracy of 95.65% in roping classification with only five training samples per class, outperforming four typical FSL methods. This FSL approach can be applied to grade the roping morphologies predicted by the ANN model. Consequently, the combination of prediction and grading using deep learning provides a new paradigm for roping analysis and control.

Keywords

Roping / Artificial neural network (ANN) / Aluminum alloys / Few-shot classification / Surface morphology

References

[1.] Bloeck M. Rowe J. Aluminium sheet for automotive applications. Advanced materials in automotive engineering, 2012, Amsterdam: Elsevier 85-108.

[2.] Shi Y, Jin H, Wu PD, et al. Analysis of roping in an AA6111 T4P automotive sheet in 3D deformation states. Acta Mater, 2017, 124: 598-607.

[3.] Engler O, Schäfer C, Brinkman HJ. Crystal-plasticity simulation of the correlation of microtexture and roping in AA 6xxx Al–Mg–Si sheet alloys for automotive applications. Acta Mater, 2012, 60: 5217-5232.

[4.] Guillotin A, Guiglionda G, Maurice C, et al. Quantification of roping intensity on aluminium sheets by areal power spectral density analysis. Mater Charact, 2010, 61: 1119-1125.

[5.] Guillotin A, Guiglionda G, Maurice C, et al. Correlation of surface roping with through-thickness microtextures in an AA6xxx sheet. Metall Mater Trans A, 2011, 42: 1919-1924.

[6.] Wu PD, Lloyd DJ, MacEwen SR. A simple model describing roping in A1 sheet. Scr Mater, 2003, 48: 1243-1248.

[7.] Qin L, Seefeldt M, Van Houtte P. Analysis of roping of aluminum sheet materials based on the meso-scale moving window approach. Acta Mater, 2015, 84: 215-228.

[8.] Hu Y, Zhou G, Yuan X, et al. An artificial neural network-based model for roping prediction in Aluminum alloy sheet. Acta Mater, 2023, 245: 118605.

[9.] Marteau J, Deltombe R, Bigerelle M. Quantification of the morphological signature of roping based on multiscale analysis and autocorrelation function description. Materials, 2020, 13: 3040.

[10.] Schäfer C, Brinkman HJ, Engler O, et al. Quantification of roping in aluminium sheet alloys for car body applications by combining 3D surface measurements with Fourier analysis. Int J Mater Res, 2015, 106: 248-257.

[11.] Hu Y, Zhou G, Liu R, et al. On the correlation between roping, texture, and morphology of aluminium alloy sheets. J Mater Res Technol, 2023, 26: 571-586.

[12.] Wei B, Hao K, Tang X, et al. A new method using the convolutional neural network with compressive sensing for fabric defect classification based on small sample sizes. Text Res J, 2019, 89: 3539-3555.

[13.] Mundt M, Majumder S, Murali S et al (2019) Meta-learning convolutional neural architectures for multi-target concrete defect classification with the concrete defect bridge image dataset. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 11196–11205

[14.] Konovalenko I, Maruschak P, Brezinová J, et al. Steel surface defect classification using deep residual neural network. Metals, 2020, 10: 846.

[15.] Mangla P, Kumari N, Singh M et al (2020) Charting the right manifold: Manifold mixup for few-shot learning. In: Proceedings of the IEEE/CVF winter conference on applications of computer vision, pp 2218–2227

[16.] Hu Y, Vincent G, Stéphane P (2021) Leveraging the feature distribution in transfer-based few-shot learning. In: International conference on artificial neural networks. Springer International Publishing, Cham, pp 487–499

[17.] Chen WY, Liu YC, Kira Z et al (2018) A closer look at few-shot classification. In: International conference on learning representations. arXiv:1904.04232. https://doi.org/10.48550/arXiv.1904.04232

[18.] Lichtenstein M, Sattigeri P, Feris R, et al. TAFSSL: task-adaptive feature sub-space learning for few-shot classification. European conference on computer vision, 2020, Cham: Springer 522-539.

[19.] Zhang H, Cisse M, Dauphin YN et al (2018) Mixup: beyond empirical risk minimization. In: international conference on learning representations, arXiv:1710.09412. https://doi.org/10.48550/arXiv.1710.09412

[20.] Verma V, Lamb A, Beckham C, et al (2019) Manifold mixup: better representations by interpolating hidden states. In: International conference on machine learning, PMLR, pp 6438–6447

[21.] GMW-15420 (2012) Global qualification process for aluminum sheet. General Motors, US

[22.] Vinyals O, Blundell C, Lillicrap T et al (2016) Matching networks for one shot learning. Adv Neural Inf Process. arXiv:1606.04080. https://arxiv.org/abs/1606.04080

[23.] Russakovsky O, Deng J, Su H, et al. ImageNet large scale visual recognition challenge. IJCV, 2015, 115: 211-252.

[24.] Zagoruyko S, Komodakis N (2016) Wide residual networks. In: British machine vision conference. arXiv:1605.07146. https://doi.org/10.48550/arXiv.1605.07146

[25.] Finn C, Abbeel P, Levine S (2017) Model-agnostic meta-learning for fast adaptation of deep networks. In: International conference on machine learning, PMLR, pp 1126–1135

[26.] Munkhdalai T, Yu H (2017) Meta networks. In: International conference on machine learning, PMLR, pp 2554–2563

[27.] Snell J, Swersky K, Zemel R (2017) Prototypical networks for few-shot learning. Adv Neural Inf Process. arXiv:1703.05175. https://arxiv.org/abs/1703.05175

[28.] Zhou B, Khosla A, Lapedriza A, et al (2016) Learning deep features for discriminative localization. In: Proceedings of the IEEE conference on computer vision and pattern recognition, CVPR, pp 2921–2929

[29.] Van der Maaten L, Hinton G. Visualizing data using t-SNE. J Mach Learn Res, 2008, 9: 2579-2605.

[30.] McInnes L, Healy J, Saul N et al (2018) UMAP: uniform manifold approximation and projection. arXiv:1802.03426. https://doi.org/10.48550/arXiv.1802.03426

Funding

National Natural Science Foundation of China http://dx.doi.org/10.13039/501100001809(52105384)