Deep Learning-Based Ultrasonic Testing to Evaluate the Porosity of Additively Manufactured Parts with Rough Surfaces

Abstract

Ultrasonic testing (UT) has been actively studied to evaluate the porosity of additively manufactured parts. Currently, ultrasonic measurements of as-deposited parts with a rough surface remain problematic because the surface lowers the signal-to-noise ratio (SNR) of ultrasonic signals, which degrades the UT performance. In this study, various deep learning (DL) techniques that can effectively extract the features of defects, even from signals with a low SNR, were applied to UT, and their performance in terms of the porosity evaluation of additively manufactured parts with rough surfaces was investigated. Experimentally, the effects of the processing conditions of additive manufacturing on the resulting porosity were first analyzed using both optical and scanning acoustic microscopy. Second, convolutional neural network (CNN), deep neural network, and multi-layer perceptron models were trained using time-domain ultrasonic signals obtained from additively manufactured specimens with various levels of porosity and surface roughness. The experimental results showed that all the models could evaluate porosity accurately, even that of the as-deposited specimens. In particular, the CNN delivered the best performance at 94.5%. However, conventional UT could not be applied because of the low SNR. The generalization performance when using newly manufactured as-deposited specimens was high at 90%.

Additive Manufacturing Research and Applications

“…This Special Issue offers a wide scope in the research field around 3D printing, including the following [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18]: the use of 3D printing in system design, AM with binding jetting, powder manufacturing technologies in 3D printing, fatigue performance of additively manufactured metals such as the Ti-6Al-4V alloy, 3D-printing method with metallic powder and a laser-based 3D printer, 3D-printed custom-made implants, laser-directed energy deposition (LDED) process of TiC-TMC coatings, Wire Arc Additive Manufacturing, cranial implant fabrication without supports in electron beam melting (EBM) additive manufacturing, the influence of material properties and characteristics in laser powder bed fusion, Design For Additive Manufacturing (DFAM), porosity evaluation of additively manufactured parts, fabrication of coatings by laser additive manufacturing, laser powder bed fusion additive manufacturing, plasma metal deposition (PMD), as-metal-arc (GMA) additive manufacturing process, and spreading process maps for powder-bed additive manufacturing derived from physics model-based machine learning.…”

Measurement of Absolute Acoustic Nonlinearity Parameter Using Laser-Ultrasonic Detection

“…Because these nonlinear elastic constants are closely related to the microstructural features and micro-damage of the components, the β parameter measured using the NUT has been used as an indicator for quantitative NDE [9]. Assuming that the planar longitudinal wave propagates in lossless isotropic media, β can be defined as follows based on the nonlinear wave equation solution [28]:…”

Deep learning in automated ultrasonic NDE -- developments, axioms and opportunities

“…Deep learning has also been used for material characterisation (other than defects), which include the direct or indirect measurement of material properties. For instance, the inference of porosity level (obtained by processing Cscan images) in additive manufacturing parts have been investigated by Park et al [62,63]. Several DL models based on CNNs, DNNs, and ANNs, were designed and tested against real samples by taking as input the ultrasonic signal and as output the porosity level (through multiple classes).…”