Robotized laser beam welding is one important process in manufacturing, offering efficient welding while minimizing the heat input. Nonetheless, this method is sensitive to various deviations, including fixture problems, heat-induced distortions, and inaccuracies in tool handling. Such deviations can lead to significant defects like lack of fusion, particularly when welding square butt joints without gaps. Detecting these defects through visual inspection or non-destructive methods is challenging. To address this, real-time monitoring and automatic intervention are necessary. One effective sensor for monitoring laser beam welding is the photodiode, which captures optical emissions from the process. Research has demonstrated correlations between these emissions and process stability. Photodiodes are costeffective and easily integrated into welding tools, making them ideal for industrial applications. However, the challenge lies in analyzing the output signals and defining thresholds for identifying deviations from normal conditions. Thus, there's a need for an automated method to set threshold values based on measured data. Machine learning offers a solution, particularly through supervised, unsupervised, or semi-supervised methods. Supervised machine learning requires labeled data, involving time-consuming experiments with nominal and deviating cases, making it less feasible for industrial setups. This paper suggests using unsupervised learning for anomaly detection, relying solely on data from nominal welding cases for model training. In this approach, a model is trained using photodiode data from a single nominal weld case and subsequently tested on data collected during experiments involving laser beam offsets during welding. The results demonstrate the promise of this method for monitoring closed square butt-joint laser beam welding, even with limited training data from nominal cases. 

Laser Beam Welding (LBW) requires precise control to ensure high-quality welds. Accurate classification of joint gap widths and detection of tack welds are crucial for optimizing the process and enhancing product reliability. This study investigates the application of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) to classify instant joint gap widths and detect the presence of tack welds during welding. The goal is to facilitate adaptive joint gap bridging in robotized and autogenous butt joint welding. Sequences of images resembling a time series were captured during welding of prepared workpieces with varying joint gap widths along the joint line. The results demonstrate that CNNs significantly outperform RNNs, achieving over 99 percent classification accuracy in both validation and test datasets, and 96 percent accuracy under conditions of substantial noise. These findings underscore the potential of CNNs in enhancing the precision and adaptability of welding automation. However, challenges remain in generalizing the CNN model to diverse and noisy operational environments.

Hot-wire directed energy deposition using a laser beam (DED-LB/w) is a method of metal additive manufacturing (AM) that has benefits of high material utilization and deposition rate, but parts manufactured by DED-LB/w suffer from a substantial heat input and undesired surface finish. Hence, regulating the process parameters and monitoring the process signatures to control the final quality during the deposition is crucial to ensure the quality of the final part. This paper explores the dynamic modeling of the DED-LB/w process and introduces a parameter-signature-quality modeling and control approach to enhance the quality of modeling and control of part qualities that cannot be measured in situ. The study investigates different process parameters that influence the melt pool width (signature) and bead width (quality) in single and multi-layer beads. The proposed modeling approach utilizes a parameter-signature model as F1 and a signature-quality model as F2. Linear and nonlinear modeling approaches are compared to describe a dynamic relationship between process parameters and a process signature, the melt pool width (F1). A fully connected artificial neural network is employed to model and predict the final part quality, i.e., bead width, based on melt pool signatures (F2). Finally, the effectiveness and usefulness of the proposed parameter-signature-quality modeling is tested and verified by integrating the parameter-signature (F1) and signature-quality (F2) models in the closed-loop control of the width of the part. Compared with the control loop with only F1, the proposed method shows clear advantages and bears potential to be applied to control other part qualities that cannot be directly measured or monitored in situ.

In autogenous laser butt welding the variability of the joint gap can cause problems in terms of weld seam quality. A suitable strategy to alleviate this is to dynamically shape the laser beam instead of a circular-shaped beam with typical Gaussian or top hat distributions. Currently available systems cannot reach sufficient performance due to both the real time control system for the shape variation and the limited laser power currently manageable. In the present work, the possibility of bridging the joint gap during welding using a deformable mirror to elongate the focused laser beam from circular to transversal elliptical shape was investigated. The effect of the beam shaping on the geometry of the weld pool and of the weld cross sections was analysed, for different values of the gap in comparison with a circular Gaussian beam. It was demonstrated that the adoption of a transversal elliptical laser beam makes the welding process more stable, especially for large gaps (i.e. larger than the circular beam radius). Thanks to the beam shaping, the extension of the fused zone (in terms of the cross section area, height and width) resulted to be less sensitive to the gap's dimension; in addition, the extension of the heat affected zone and the presence of undercuts were evidently reduced.

To be a successful competitor among other technologies, metallic laser-directed energy depositionusing a laser beam would benefit from the support of intelligent automation making the processrobust, repeatable, and cost-efficient. This calls for technology leaps towards robust and accuratedetection and estimation of the conditions during processing and control schemes for robustperformance. This chapter discusses how developments in sensor technology and model-basedsignal processing can contribute to advancements in in-process monitoring of directed energydeposition using a laser beam and how developments in model-based feedforward- and feedbackcontrol can support automation. The focus is on how machine vision, optical emission spectroscopy,thermal sensing, and electrical process signals can support monitoring, control and better processunderstanding. These approaches are industrially relevant and have a high potential to support amore sustainable manufacturing. 

Laser beam welding offers high productivity and relatively low heat input and is one key enabler for efficient manufacturing of sandwich constructions. However, the process is sensitive to how the laser beam is positioned with regards to the joint, and even a small deviation of the laser beam from the correct joint position (beam offset) can cause severe defects in the produced part. With tee joints, the joint is not visible from top side, therefore traditional seam tracking methods are not applicable since they rely on visual information of the joint. Hence, there is a need for a monitoring system that can give early detection of beam offsets and stop the process to avoid defects and reduce scrap. In this paper, a monitoring system using a spectrometer is suggested and the aim is to find correlations between the spectral emissions from the process and beam offsets. The spectrometer produces high dimensional data and it is not obvious how this is related to the beam offsets. A machine learning approach is therefore suggested to find these correlations. A multi-layer perceptron neural network (MLPNN), support vector machine (SVM), learning vector quantization (LVQ), logistic regression (LR), decision tree (DT) and random forest (RF) were evaluated as classifiers. Feature selection by using random forest and non-dominated sorting genetic algorithm II (NSGAII) was applied before feeding the data to the classifiers and the obtained results of the classifiers are compared subsequently. After testing different offsets, an accuracy of 94% was achieved for real-time detection of the laser beam deviations greater than 0.9 mm from the joint center-line.

This paper presents an experimental study where a photodiode integrated into a laser beam welding tool is used to monitor laser beam spot deviations fromthe joint, the beam offset. The photodiode system is cost effective and typically easy to implement in an industrial system. The selected photodiode is a silicondetector sensitive in the spectral range between 340-600nm which corresponds to the spectral emissions from the plasma plume. The welding application is closed-square-butt joint welding where a laser beam offset can cause lack of fusion in the resulting weld. The photodiode signal has been evaluated by two different change point detection methods, one off-line and one on-line method, with respect to their detection performance. Off-line methods can be used to guide post weld inspection and on-line methods have the potential to enable on-line adaptive control or the possibility to stop the process for repair. The performance of the monitoring system and the change point detection methods have been evaluated from data obtained during laser beam welding experiments conducted on plates of stainless steel. The results clearly indicates the possibility to detect beam offsets by photodiode monitoring.

Adding filler wire control to autogenous laser beam welding of squared butt joints offers a means to widen up the tight fit-up tolerances associated with this process. When the gap width varies, the filler wire rate should be controlled to assure a constant geometry of the resulting weld seam. A dual mode sensing system is proposed to estimate the joint gap width and thereby control the filler wire rate. A vision camera integrated into the welding tool together with external LED illumination and a laser line projection enables two sensing modes, one surface feature extraction mode and one laser triangulation-based mode. Data from the both modes are fused in a Kalman filter, and comparisons show that the fusing of the data gives more robust estimation than estimates from each single mode. A feed-forward control system adaptively adjusts the filler wire rate based on the estimations ofthe joint gap width in front of the keyhole. The focus is on keeping the data processing simple and affordable, and the real-time performance of the sensor and control system has been evaluated by welding experiments. It is shown that the proposed system can be used for on-line control of the filler wire rate to achieve a constant weld geometry during varying joint gap widths

This paper presents an experimental study where a photodetector is used in a laser beam welding tool to monitor beam deviations (beam offsets) in stake welding of tee joints. The aim is to obtain an early detection of deviations from the joint centerline in this type of welding where the joint is not visible from the top side. The photodetector used is a GaP diode sensitive in the spectral range 150-550 nm corresponding to the spectral emissions form the plasma plume during keyhole welding. The photodetector signal has been evaluated by change point detection methods with respect to their detection performance. Both an off-line and an on-line method have been evaluated. The off-line method can be used to guide post weld inspection and the on-line method has the potential to enable on-line adaptive position control and/or the possibility to stop the process for repair. The results shows that the proposed method can be used as a go/no go system and to guide post weld inspection.

Robotized laser beam welding of closed-square-butt joints is sensitive to the positioning of the laser beam with respect to the joint since even a small offset may result in a detrimental lack of sidewall fusion. An evaluation of a system using a photodiode aligned coaxial to the processing laser beam confirms the ability to detect variations of the process conditions, such as when there is an evolution of an offset between the laser beam and the joint. Welding with different robot trajectories and with the processing laser operating in both continuous and pulsed mode provided data for this evaluation. The detection method uses wavelet analysis of the photodetector signal that carries information of the process condition revealed by the plasma plume optical emissions during welding. This experimental data have been evaluated offline. The results show the potential of this detection method that is clearly beneficial for the development of a system for welding joint tracking.