Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
Modeling of an Experimental Setup of an Electromechanical System through Linear and Nonlinear System Identification
Submission Author:
Leonardo Dias Pereira , RJ
Co-Authors:
Leonardo Dias Pereira, Lara Candido Alvim, Helon Vicente Hultmann Ayala, Hans Ingo Weber
Presenter: Leonardo Dias Pereira
doi://10.26678/ABCM.COBEM2021.COB2021-1875
Abstract
The present work investigates the data-driven modeling of an electromechanical system with three degrees of freedom (3-DOF), that represents a test rig setup consisting of a slender structure subjected to torsional oscillations. It is proposed to identify system parameters using benchmark datasets of test rig setup with linear approximations and, additionally, nonlinear approaches. Three model structures are applied to the system for comparative purposes. First, the linear Black-box models are the AutoRegressive with eXogenous input (ARX) and the AutoRegressive Moving Average with eXogenous input (ARMAX), then, the last one is the Nonlinear AutoRegressive Moving Average with eXogenous input (NARMAX). Finally, the modeling is validated based on a comparison made between the results obtained through One-Step-Ahead and Free Run prediction models, and the actual data acquired was considered for this purpose. This work shows the results through correlation tests and presents the best model structure based on parameters variation and R2.
Keywords
System Identification, test-rig setup, ARX, ARMAX, NARMAX
