Determinação da região robusta de estabilidade e de desempenho inspirada nos princípios da estatística clássica.

Abstract

This study discusses the development of a methodology based on classical concepts of statistics and probability to analyze and evaluate the robustness of the stability and performance of the control system, particularly those that use the PID as control law. To establish the conditions for the application of the methodology, a recursive system identification method process was developed, in which convolution and phenomenological models were used to represent model and process, together with a self-tuning procedure that is necessary to consider tuning parameters as random variables, and hence the roots of the characteristic equation of the closed loop system. The mapping of the region of robustness has been achieved from the roots of the characteristic equation, considering the statistical distance as the metric represented to the robustness of stability which allows the region to establish a degree of significance. The results obtained demonstrate the potential analytical and evaluation required by the methodology, allowing such analysis also "online" with low computational effort and operational proving to be a powerful tool in the analysis of control system.