Measurement and inspection of electrical discharge machined steel surfaces using deep neural networks
- PDF / 3,837,586 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 43 Downloads / 181 Views
ORIGINAL PAPER
Measurement and inspection of electrical discharge machined steel surfaces using deep neural networks Jamal Saeedi1 · Matteo Dotta2 · Andrea Galli2 · Adriano Nasciuti2 · Umang Maradia3 · Marco Boccadoro3 · Luca Maria Gambardella1 · Alessandro Giusti1 Received: 20 April 2020 / Revised: 17 August 2020 / Accepted: 12 October 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract We propose an industrial measurement and inspection system for steel workpieces eroded by electrical discharge machining, which uses deep neural networks for surface roughness estimation and defect detection. Specifically, a convolutional neural network (CNN) is used as a regressor in order to obtain steel surface roughness and a CNN based on spatial pooling pyramid is applied for defect classification. In addition, a new method for the region of interest selection based on morphological reconstruction and mean shift filtering is proposed for defect detection and localization. The regressor and classifier based on deep neural networks proposed here outperform state-of-the-art methods using handcrafted feature extraction. We achieve a mean absolute percentage error of 7.32% on roughness estimation; on defect detection, our approach yields an accuracy of 97.26% and an area under the ROC curve metric of 99.09%. Keywords Electrical discharge machining · Convolutional neural networks · Spatial pooling pyramid · Morphological reconstruction · Mean shift filtering
1 Introduction The research presented here uses machine learning for measurement and inspection of steel products processed by a diesinking electrical discharge machining (EDM) device [1]. Specifically, deep learning and image-processing techniques have been investigated, allowing for the cost-efficient integration of machine vision equipment in the EDM machine tools in order to measure steel surface roughness and inspect it for possible defects. Die-sinking EDM comprises an electrode and a workpiece which is immersed in an insulating liquid such as oil or similar dielectric fluids. The electrode and workpiece are then connected to an appropriate pulsed voltage power supply,
B B
Jamal Saeedi [email protected] Alessandro Giusti [email protected]
1
Dalle Molle Institute for Artificial Intelligence (IDSIA USI-SUPSI), Lugano, Switzerland
2
Institute for Mechanical Engineering and Material Technology (MEMTi SUPSI), Lugano, Switzerland
3
Agie Charmilles SA, Losone, Switzerland
which generates an electrical potential between these two parts. When the electrode moves toward the workpiece, a dielectric breakdown happens between the electrodes (in the insulating liquid), which results in a plasma channel formation. EDM has been applied by numerous industries in their manufacturing developments because of its capability to easily erode hard materials, like hardened steel and tungsten carbide, and create accurate and unique shapes. The most common applications for EDM are die making, mold making, and small hole drilling [1]. Roughness measurement fo
Data Loading...
