ARX Model, Addictive Manufacturing, Modelling, Directed Energy Deposition, Estimation
Abstract
Additive manufacturing is one of the disruptive technologies associated with Industry 4.0, which enables new ways of processing materials and manufacturing parts. The Laser Directed Energy Deposition technique, a variant of 3D printing, stands out for enabling work with metal alloys. However, to achieve quality in the result, it is necessary to control several variables involving the processing conditions, whose complexity due to nonlinear physical-chemical interactions presents a challenge, thus, becoming the focus of this currently research. While in the literature solutions involving the use of AI techniques or computer simulations can be found to obtain a model of the system, which require more complex computational resources, this work has as its main objective to obtain a model of the system from data collected experimentally together with autoregressive modeling with exogenous inputs (ARX), to contribute to the application of a real-time control loop. To obtain the parameters of the ARX model, the recursive estimation technique with a least squares algorithm with forgetting factor was applied. The results obtained from the simulations were presented, with details of the application of the ARX modelling technique. The model results were compared with the experimental data for validation.
