A systematic computational method to simulate and detect sub-surface flaws, through non-destructive transient thermography, in aluminium sheets and friction stir welded sheets is proposed. The proposed method relies on feature extraction methods and a data-driven machine learning modelling structure. In this work, we propose the use of a multi-layer perceptron feed-forward neural-network with feature extraction methods to improve the flaw-probing depth of transient thermography inspection. Furthermore, for the first time, we propose Thermographic Signal Linear Modelling (TSLM), a hyper-parameterfree feature extraction technique for transient thermography. The new feature extraction and modelling framework was tested with out-of-sample experimental transient thermography data and results show effectiveness in sub-surface flaw detection of up to 2.3 mm deep in aluminium sheets (99.8 % true positive rate, 92.1 % true negative rate) and up to 2.2 mm deep in friction stir welds (97.2 % true positive rate, 87.8 % true negative rate).
Numerous efforts in the additive manufacturing literature have been made toward in-situ defect prediction for process control and optimization. However, the current work in the literature is limited by the need for multi-sensory data in appropriate resolution and scale to capture defects reliably and the need for systematic experimental and data-driven modeling validation to prove utility. For the first time in literature, we propose a data-driven neural network framework capable of in-situ micro-porosity localization for laser powder bed fusion via exclusively within hatch strip of sensory data, as opposed to a three-dimensional neighborhood of sensory data. We further propose using prior-guided neural networks to utilize the often-abundant nominal data in the form of a prior loss, enabling the machine learning structure to comply more with process physics. The proposed methods are validated via rigorous experimental data sets of high-strength aluminum A205 parts, repeated k-fold cross-validation, and prior-guided validation. Using exclusively within hatch stripe data, we detect and localize porosity with a spherical equivalent diameter (SED) smaller than $$50.00\,\upmu $$
50.00
μ
m with a classification accuracy of $$73.13\pm 1.57\%$$
73.13
±
1.57
%
This is the first work in the literature demonstrating in-situ localization of porosities as small as $$38.12\,\upmu m$$
38.12
μ
m
SED and is more than a five-fold improvement on the smallest SED porosity localization via spectral emissions sensory data in the literature. In-situ localizing micro-porosity using exclusively within hatch-stripe data is a significant step towards within-layer defect mitigation, advanced process feedback control, and compliance with the reliability certification requirements of industries such as the aerospace industry.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.