Desenvolvimento de um sistema baseado em aquisição de imagens para a parametrização do processo GMAW-P em manufatura aditiva

Abstract

In recent years, a large number of researches based on rapid prototyping processes in Additive Manufacturing have been developed in the most diverse areas and using different techniques and materials. The application of Wire and Arc Additive Manufacturing by Arc Process presents a high deposition rate and has been shown to be a very promising technology to meet current requirements for a rapid development and production of new products. The final quality of the produced part can be followed by the behavior of the weld pool, which contains various information such as penetration, width and reinforcement, which determine the geometry of the solidified layer. This thesis describes a technique to assist in the prediction and control of process parameters, which will be used for layer deposition based on Additive Manufacturing with the aid of a vision system, to optimize the homogeneity of the wall profile. The geometry and behavior of the weld pool were analyzed from the images obtained by a high-speed camera and through MATLAB© processing. It has been shown that the behavior of the weld pool is strongly affected by wire feed speed, droplet transfer rate and drop volume when it occurs by pulsed current, in the condition where the current and peak time are kept constant. With the results obtained from the images extracted to evaluate the behavior of the molten metal, in order to understand the changes in the geometry of the deposited layer, it is possible to identify the ability to predict the parameters to be used in the process, ensuring a more constant layer geometry. Thus, by changing the heat input values, and the forces that act on the metal transfer and the weld pool during metal deposition, and consequently reducing the heat applied in a region that is still cooling, it is possible to minimize the reflow of a layer already solidified. The analyzes of the predictive models were developed using the multiple linear regression method, and confirmed the strong influence of base time and electrical voltage on drop formation and behavior of the process during the deposition of the layers.