Directed energy deposition (DED) is a family of additive manufacturing technologies. With these processes, metal parts are built layer by layer, introducing dynamics that propagate in time and layer-domains, which implies additional complexity and consequently, the resulting part quality is hard to predict. Control of the deposit layer thickness and height is a critical issue since it impacts on geometrical accuracy, process stability, and the overall quality of the product. Therefore, online feedback height control for DED processes with proper sensor strategies is required. This work presents a novel vision-based triangulation technique through an off-axis located CCD camera synchronized with a 640 nm wavelength pulsed illumination laser. Image processing and machine vision techniques allow in-line height measurement right after metal solidification. The linearity and the precision of the proposed setup are validated through off-and in-process trials in the laser metal deposition (LMD) process. Besides, the performance of the developed in-line inspection system has also been tested for the Arc based DED process and compared against experimental weld bead characterization data. In this last case, the system additionally allowed for the measurement of weld bead width and contact angles, which are critical in first runs of multilayer buildups.

Understanding the deterioration and predominant damage mechanisms of structures is highly relevant, especially for safety-critical components. Non-Destructive Testing (NDT) plays a crucial role in assessing and monitoring their integrity by evaluating damage evolution. However, when it comes to complex structures, the existing NDT methods face challenges in their application, as is the case with sandwich structures. This study employs two NDT methods to analyze the initiation of damage in such structures during a fatigue test. The Acoustic Emission (AE) technique utilizes membrane-free microphones with a broad bandwidth to capture acoustic events from difficult-to-access areas. A machine learning algorithm is used to classify these events to determine their source and associated damage mechanisms. Additionally, Digital Image Correlation (DIC) is employed to measure strain evolution without contact during the test. This method is particularly relevant due to the complex and thin geometry of sandwich structures, where other methods are not applicable. Strain redistributions are considered relevant damage indicators. The results indicate that Acoustic Emission serves as an early indicator of damage, with the cumulative number of events and peak frequency correlating well with the severity of the damage. In contrast, DIC revealed clear indications of damage or deterioration, albeit at a later stage compared to AE.

This study investigates the feasibility of using Near-Infrared Hyperspectral Imaging (HSI) for monitoring the curing state in Liquid Resin Infusion (LRI) manufacturing. By capturing spectral reflectance data throughout the heating and curing phases of the process, we explore the potential of HSI as a non-destructive technique for assessing resin polymerization. Compared to conventional sensor-based methods, HSI offers two key advantages: it eliminates the need to embed sensors within the manufactured part and provides spatially resolved information across the entire surface, enabling the detection of cure inhomogeneities. First, an in-depth study was conducted on the spectral features captured by NIR-HSI and their temporal evolution throughout the process, comparing them with key absorption bands related to the epoxy resin used in LRI. Afterwards, spectral data was correlated with the curing state inferred from commercial dielectric sensors using regression models, including Partial Least Squares Regression (PLS-R) and Support Vector Regression (SVR). These regression models were trained using the full spectral range of a specific spatial region of interest, as well as wavelength subsets selected through feature selection techniques such as Mutual Information (MI) and Random Forest Regressor (RF). The results indicate a strong correlation between the spectral data and the inferred curing state, with the SVR model trained on the full spectral range achieving the highest performance (R²=0.9983, RMSE = 0.0114, MAE = 0.0086). Moreover, the predicted curing maps display consistent trends across all defined regions of interest (ROIs), further demonstrating the capability of HSI to capture spatial variations in the curing process. These findings reinforce the potential of Near-Infrared Hyperspectral Imaging as a powerful, non-destructive tool for real-time monitoring of resin polymerization in LRI manufacturing.

Aircraft heat exchangers play a crucial role in maintaining thermal balance and ensuring that essential components operate efficiently and safely. In this context, it is highly relevant to gain knowledge about the deterioration and the predominant damage mechanisms of these components. In this study, two different non-destructive methods are employed to analyze the damage initiation of a heat exchanger part during fatigue test. On the one hand, Acoustic Emission technique is employed using membrane-free microphones capable of capturing a broad bandwidth. The acoustic events were classified by a machine learning algorithm to determine their source and damage mechanisms. On the other side, Digital Image Correlation (DIC) allowed the measurement of the strain evolution along the test and more precisely, redistributions were considered as damage indicators. As a result, it was identified that AE served as early damage indicator as the cumulative number of events was in good agreement with the severity of the damage. With respect to classification two clear clusters ascribed to different type of events were identified. In the case of DIC, strain redistributions gave clear indications of damage or deterioration but at a later stage compared to AE

To ensure consistent quality in composite aerostructures, advanced non-invasive monitoring techniques are needed to detect global and local deviations during manufacturing. This study presents a real-time, spatially resolved method for monitoring the curing stage of Liquid Resin Infusion (LRI) using Near-Infrared Hyperspectral Imaging (NIR-HSI). Unlike traditional point-based tools such as disposable dielectric sensors, NIR-HSI enables full-field, non-contact assessment of the chemical evolution of the resin, providing valuable spatial information for detecting inhomogeneities caused by temperature gradients or uneven resin flow, factors known to affect the final mechanical properties of composites. Previous investigations demonstrated that hyperspectral data acquired during LRI correlate with the degree of cure estimated from a dielectric sensor. In the present study, we extend this analysis through a new experimental campaign designed to validate our earlier findings and strengthen the predictive model. To improve robustness and generalizability, the curing temperature, a key driver of cure kinetics, was systematically varied to introduce controlled changes in cure behavior. This increased variability enhances model reliability and supports more accurate prediction of curing progression under realistic manufacturing conditions.