Having garnered significant attention in the scientific community over the past decade, wire-arc directed energy deposition (arc-DED) technology is at the heart of this investigation into additive manufacturing parameters. Singularly focused on Invar as the selected material, the primary objective revolves around devising a virtual sensor for the indirect size measurement of the bead. This innovative methodology involves the seamless integration of internal signals and sensors, enabling the derivation of crucial measurements sans the requirement for direct physical interaction or conventional measurement methodologies. The internal signals recorded, the comprising voltage, the current, the energy from the welding heat source generator, the wire feed speed from the feeding system, the traverse speed from the machine axes, and the temperature from a pyrometer located in the head were all captured through the control of the machine specially dedicated to the arc-DED process during a phase of optimizing and modeling the bead geometry. Finally, a feedforward neural network (FNN), also known as a multi-layer perceptron (MLP), is designed, with the internal signals serving as the input and the height and width of the bead constituting the output. Remarkably cost-effective, this solution circumvents the need for intricate measurements and significantly contributes to the proper layer-by-layer growth process. Furthermore, a neural network model is implemented with a test loss of 0.144 and a test accuracy of 1.0 in order to predict weld bead geometry based on process parameters, thus offering a promising approach for real-time monitoring and defect detection.

With the increasing number of applications employing additive manufacturing solutions, these deposition processes must become more autonomous, which can be helped by the application of machine learning monitoring. This study presents a fully online, low-cost framework for real-time quality control in Invar wire-arc additive manufacturing (WAAM). Synchronized current and voltage signals are transformed into spatial heatmaps and temporal Markov transition images, which are processed by an optimized ResNet-18 to classify the quality of each layer on-the-fly. Validation using cross-validation on an internal Invar dataset yields an accuracy of up to 94% under clean conditions, with inference times below 20 ms per layer, enabling deployment during natural cooling between layers. These results demonstrate the feasibility of non-intrusive signal-based anomaly detection, enabling rapid identification of weld spalls and useful for scalable and automated WAAM monitoring in industrial environments.

The accurate prediction of weld bead geometry is crucial for ensuring the quality and consistency of wire and arc additive manufacturing (WAAM), a specific form of directed energy deposition (DED) that utilizes arc welding. Despite advancements in process control, predicting the shape and dimensions of weld beads remains challenging due to the complex interactions between process parameters and material behavior. This paper addresses this challenge by exploring the application of symmetrical neural networks to enhance the accuracy and reliability of geometric predictions in WAAM. By leveraging advanced machine learning techniques and incorporating the inherent symmetry of the welding process, the proposed models aim to precisely forecast weld bead geometry. The use of neuronal networks and experimental validation demonstrate the potential of symmetrical neural networks to improve prediction precision, contributing to more consistent and optimized WAAM outcomes.

The combination of safety belts and rollover protective structures (ROPSs) is key in improving the safety of agricultural tractors in the event of rollover. However, we also have the opportunity to enhance the security provided by each ROPS; one such example is the combination of this safety device with adequate mechanical energy absorbers (MEAs). Inexpensive disc-shaped MEAs can be included in the anchoring points of a ROPS onto the chassis of a tractor. Three configurations of ROPS combined with MEAs were tested during the application of loads that simulated the effects of side rollover in the vehicle. The tested configurations included a blank MEA as a reference case alongside a single MEA and a stack assembly containing both elements. The results of the tests show that both the deformation of the ROPS itself and the strain energy are larger in the case of blank MEAs; thus, there is also a risk that the clearance zone will be infringed upon and that the protective structure will collapse. We can conclude that the implementation of an appropriate MEA in ROPS reduces the deformation of the ROPS itself and its strain energy in cases of vehicle rollover; hence, the safety provided by such protection systems may be improved at a low cost.

Disc-shaped mechanical energy absorbers (MEAs), in combination with rollover protection structures (ROPSs), may contribute to the prevention of the infringement of the safety zone and the collapse of the ROPS in case of the overturn of an agricultural tractor. An MEA can absorb a significant amount of potential energy of an overturning tractor and its deformation produces a rotation of the ROPS around the safety zone. In this research, MEAs with two different geometries have been developed. Both geometries present common features, such as disc dimensions, number of rings, and number of arms, but the distribution of the arms differs. Additionally, these MEA were manufactured in steel discs of four different thicknesses, ranging from 2 to 6 mm. The manufactured MEAs were tested in a universal testing machine, and their behavior characterized. From this data, linear models of the MEAs were developed. As a consequence, a number of characteristic parameters were selected and calculated, such as the activation load and the strain energy absorbed in a safe range of applied loads. Some patterns and trends were analyzed from the tested MEAs, which enables a better description of their behavior and the extrapolation of this behavior to other non-tested thicknesses and geometries.