Eventos Anais de eventos
COBEM 2023
27th International Congress of Mechanical Engineering
Artificial immune system applied in the detection of damage in an inverted pendulum under the effect of a controller.
Submission Author:
Matheus Medeiros Donatoni , SP
Co-Authors:
Matheus Medeiros Donatoni, Fabio Roberto Chavarette
Presenter: Matheus Medeiros Donatoni
doi://10.26678/ABCM.COBEM2023.COB2023-0031
Abstract
Control systems are experienced in everyday life and provide optimization for several tools and systems that human beings rely on. In some cases, the usage of controllers in dynamic systems directly aims at their applicability and safety increment, hence their proper functioning is directly linked to the appropriate target plant modeling. Considering these points and the different possibilities of damage that dynamic systems can suffer due to the nature of their operation, the need for monitoring the structural health of these systems, which may have their physical characteristics modified, is validated. A commonly used technique for this purpose is Structural Health Monitoring (SHM), which is a data-driven system that can help detect failures, therefore allowing appropriate action arising from an early prognosis. Since it's based on the detection of patterns, a suitable tool used for this purpose is the so-called Artificial Immune System (AIS) which, compared to the Natural Immune System (NIS), is based on the differentiation between self and non-self agents for the classification of signals. In this paper, we propose the application of a SHM system, with AIS as a pattern recognition tool applied to a Rotary Inverted Pendulum (RIP) as a reference model for damage detection in controlled dynamic systems. Since this mechanical system is naturally unstable, its operation is directly dependent on the existence of a controller, which must be able to guarantee its stability and the desired dynamic behavior, such as variable reference tracking. The approach taken uses the controller state variables as the data for monitoring and classifying operational conditions. The proposed technique applied to the analysis of four different structural state conditions, with the combination of multiple sensor data, came in a damage detection rate of 99.07%, and a full classification capability.
Keywords
control, Inverted pendulum, Structural Health Monitoring (SHM), Artificial Immune System
