Additive manufacturing (AM) is considered as a revolution in manufacturing. However, the high expectations face technical difficulties that prevent further penetration into wider industries. The main reason is the lack of process reproducibility and the absence of a reliable and cost-effective process monitoring. This paper is a supplement to existing studies in this field and proposes a unique combination of highly sensitive acoustic sensor and machine learning for process monitoring. The acoustic signals from a real powder-bed fusion AM process were collected using a fiber Bragg grating. The process parameters are intentionally tuned to achieve three levels of quality categories, which are related to the porosity contents inside the workpiece. The quality categories are defined as high, medium, and poor quality and their corresponding porosity contents are 0.07%, 0.30%, and 1.42%, respectively. Wavelet spectrograms of the signals and their encoded label representations, obtained from spectral clustering, are taken as features. A deep convolutional neural network is used to classify the features from each category and the classification accuracy ranges between 78% and 91%. Hence, the proposed method has significant industrial potentials for in situ and real-time quality monitoring of AM processes since it requires minimum modifications of commercially available industrial machines.

This study investigates the use of a low cost microphone combined with state-of-the-art machine learning (ML) algorithms as online process monitoring to differentiate various materials and process regimes of Laser-Powder Bed Fusion (LPBF). Three processing regimes (lack of fusion pores, conduction mode and keyhole pores) and three alloys (316L stainless steel, bronze (CuSn₈), and Inconel 718) were selected. Three conventional ML algorithms and a Convolutional Neural Network (CNN) were chosen to perform the classification tasks resulting in five main findings. First, we proved that the AE features are related to the laser-material interaction and not from undesired machine or environmental noise. Second, the process regimes are classified with high accuracy (> 87%) regardless of the algorithms and materials. Third, it is possible to build a single model from the three materials and still reach high classification accuracy (>86%) of the different regimes. Forth, the AE features used for the classifications are material and regime dependent. Finally, with LPBF processing of multi-materials on the rise, a strategy for classifying the material and the process regimes simultaneously using a CNN multi-label architecture reached a very high classification accuracy (≈ 93%). The results demonstrate the potential of our approaches for online LPBF process monitoring of different materials and regimes.

Laser welding is a key technology for many industrial applications. However, its online quality monitoring is an open issue due to the highly complex nature of the process. This work aims at enriching existing approaches in this field. We propose a method for real-time detection of process instabilities that can lead to defects. Hard X-ray radiography is used for the ground truth observations of the sub-surface events that are critical for the quality. A deep artificial neural network is applied to reveal the unique signatures of those events in wavelet spectrograms from the laser back-reflection and acoustic emission signals. The autonomous classification of the revealed signatures is tested on real-life data, while the real-time performance is reached by means of parallel computing. The confidence of the quality classification ranges between 71% and 99%, with a temporal resolution down to 2 ms and a computation time per classification task as low as 2 ms. This approach is a new paradigm in the digitization of industrial processes and can be exploited to provide feedbacks in a closed-loop quality control system.

Laser powder bed fusion (LPBF) is a metal additive manufacturing technique involving complex interplays between vapor, liquid, and solid phases. Despite LPBF’s advantageous capabilities compared to conventional manufacturing methods, the underlying physical phenomena can result in inter-regime instabilities followed by transitions between conduction and keyhole melting regimes — leading to defects. We investigate these issues through operando synchrotron X-ray imaging synchronized with acoustic emission recording, during the remelting processes of LPBF-produced thin walls, monitoring regime changes occurring under constant laser processing parameters. The collected data show an increment in acoustic signal amplitude when switching from conduction to keyhole regime, which we correlate to changes in laser absorptivity. Moreover, a full correlation between X-ray imaging and the acoustic signals permits the design of a simple filtering algorithm to predict the melting regimes. As a result, conduction, stable keyhole, and unstable keyhole regimes are identified with a time resolution of 100 µs, even under rapid transitions, providing a straightforward method to accurately detect undesired processing regimes without the use of artificial intelligence.

Despite extensive research efforts in the field of laser welding, the imperfect repeatability of the weld quality still represents an open topic. Indeed, the inherent complexity of the underlying physical phenomena prevents the implementation of an effective controller using conventional regulators. To close this gap, we propose the application of Reinforcement Learning for closed-loop adaptive control of welding processes. The presented system is able to autonomously learn a control law that achieves a predefined weld quality independently from the starting conditions and without prior knowledge of the process dynamics. Specifically, our control unit influences the welding process by modulating the laser power and uses optical and acoustic emission signals as sensory input. The algorithm consists of three elements: a smart agent interacting with the process, a feedback network for quality monitoring, and an encoder that retains only the quality critic events from the sensory input. Based on the data representation provided by the encoder, the smart agent decides the output laser power accordingly. The corresponding input signals are then analyzed by the feedback network to determine the resulting process quality. Depending on the distance to the targeted quality, a reward is given to the agent. The latter is designed to learn from its experience by taking the actions that maximize not just its immediate reward, but the sum of all the rewards that it will receive from that moment on. Two learning schemes were tested for the agent, namely Q-Learning and Policy Gradient. The required training time to reach the targeted quality was 20 min for the former technique and 33 min for the latter.

Laser directed energy deposition (L-DED) allows the production of multi-materials and functionally graded material (FGM) parts. But for commercialisation, process and quality monitoring of parts is required. For the first time, a novel monitoring method for chemical composition and process regimes of FGMs is proposed using a cost-effective acoustic emission (AE) (microphone) and optical emission spectroscopy (OES) sensors. Four chemical compositions (100%Ti, 58%Ti42%Nb, 37%Ti63%Nb, and 100% Nb) and two process parameters (475 W – 1'400 mm/min and 175 W – 2'000 mm/min) were selected, leading to four regimes/quality (conduction mode, partial, minor, medium, and severe lack of fusion pores). The signals were classified using seven mainstream artificial intelligence algorithms. The main conclusions are twofold. First, microphones are unsuitable candidates for monitoring the laser-material interaction during L-DED. The acoustic waves generated by the laser-material interaction are shielded by high gas flow surrounding it and so are either highly disturbed or does not reach the microphone. Conversely, OES are suitable candidates as the classification accuracies are higher than 90% for most category and machine learning algorithms, even after drastic feature reduction. Considering the wide range of chemical composition and quality, our proposed methods using OES have high industrialised potentials for them during L-DED FGM.