Laser welding is a rapidly developing technology that is of utmost importance in a number of industrial processes. The physics of the process has been investigated over the past 50 years and is mostly well understood. Nevertheless, online laser-quality monitoring remains an open issue until today due to its dynamic complexity. This paper is a supplement to existing approaches in the field of in situ and real-time laser-quality monitoring that presents a novel combination of state-of-the-art sensors and machine learning for data processing. The investigations were carried out using laser welding of titanium workpieces. The quality was estimated a posteriori by the visual inspection of cross-sections of the welded joints. Four quality categories were defined to cover the two main laser welding regimes: conduction and keyhole. The signals from the laser back reflection and optical and acoustic emissions were recorded during the laser welding process and were decomposed with the M-band wavelets. The relative energies of narrow frequency bands were taken as descriptive features. The correlation of the extracted features with the laser welding quality was carried out using the Laplacian graph support vector machine classifier. Also, an adaptive kernel for the classifier was developed to improve the analysis of the distributions of the complex features and was constructed from Gaussian mixtures. The presented laser welding setup and the developed adaptive kernel algorithm were able to classify the quality for every 2 µm of the welded joint with an accuracy ranged between 85.9% and 99.9%. Finally, the results of the developed adaptive kernel were compared with stateof-the-art machine learning methods.

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 reallife 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. The introduction of laser technology in metal welding of metals is dated back to the late 1960s 1,2 when it immediately showed advantages as compared to traditional arc welding 3. The attractions of this technique are in the non-contact processing, the absence of tool wear, high aspect ratio of the melt pool, better material fusion, possibility to process refractory materials, low running costs and high processing speed 3,4. Today, laser welding is a key technology in many fields e.g. automotive 5 and aerospace 3,6 industries, naval and heavy machinery production 7 , medicine and micromechanics 3. Unfortunately, the potential of this technology is not fully exploited, particularly in applications that require the guarantee of high weld quality. The reason is the non-linear nature of light-matter interactions, which complicates the reproducibility of the weld quality in mass production 8-10. The complex dynamics of the process, especially in keyhole welding regime, and its instabilities can cause various defects at the joint 3,10-12. A defect type of particular interest is porosity, which is a hidden threat for the mechanical properties of the workpieces 3,9-11. Obviously, an adequate, robust and low cost quality monitoring system is of great desire. The major challenge in developing such technique is in the difficulties to inspect directly the sub-surface behavior of the process zone in real-life conditions 13. Multiple approaches have been proposed, which are mostly based on mathematical modeling aiming to reconstruct the under surface dynamics using inspections of the surface via measurements of temperature 11,12,14,15 , optical 16,17 and/or acoustic 18,19 emissions (AE). However, those approaches face three main problems. Firstly, modeling often suffers inaccuracies originating from the deviations of the model assumptions from the real parameters' values. More complicated assumptions can be used to imp...

No abstract

No abstract

No abstract

No abstract

Metal-based Laser Powder Bed Fusion (LPBF) suffers from a lack of repeatability and is challenging to model, making their quality monitoring essential and demanding. The reason lies in the high dynamics taking place during the interaction of the laser with metallic powders. To bring this technology to mass production, industries are only interested in the process regime where the built layer's quality meets their standards. All other process regimes leading to poor mechanical properties and/or defect formation such as balling, Lack of Fusion (LoF) pores, keyhole pores, delamination, and crack propagation irrespective of their different regimes are considered anomalies. Today, the common methodology for monitoring uses conventional/supervised Machine Learning (ML) algorithms for the classification task requires collecting a balanced dataset corresponding to each investigated regime from the sensors, which is very expensive and time-consuming. As an alternative, the article proposes a semi-supervised approach where the defect-free regime can be differentiated from the anomalies by familiarising the ML algorithms only with the distribution of acoustic signatures corresponding to the defect-free regime. This work presents two generative Convolutional Neural Network architectures based on Variational Auto-Encoder and General Adversarial Network. As a result, we could classify the anomaly regimes with 96 and 97% accuracy, respectively.

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.