Controle robusto de um motor CC usando realimentação de estados via projeto LMIs com escalares otimizados por evolução diferencial

Abstract

When employed in industrial systems, Direct Current (DC) machines can do work that human hands do not have the necessary physical precision or strength. Problems arise, for example, from the wear and tear of the machine, which often works every day for years until it is replaced, or even during its routine operation, since its internal parameters can vary by factors such as temperature and vibrations. However, controllers can be designed to achieve acceptable performance even under the conditions presented, the so-called robust controllers. One of the methods for the synthesis of these controllers is the Linear Matrix Inequalities (LMIs). This work proposes to solve the problem that arises in projects based on LMIs dependent on scalars for the synthesis of robust controllers. These scalars are usually chosen empirically or by trial and error and usually result in inefficient drivers. In order to optimize these scalars, we will use the Differential Evolution (DE), which searches through various random combinations to find optimal solutions within their limits of search. The employed algorithm demonstrated the ability to converge to an optimal minimum within the selected search area, with some limitations. The gains obtained by the algorithm demonstrated good performance in controlling a didactic module under several tests. And, in simulations, demonstrated the ability to stabilize uncertain systems as proposed in the paper.