Um Algoritmo de Detecção de Falhas em Sistemas Multivariados

Abstract

This paper presents a new methodology for system design and fault detection in multivariate plants, statistical quality control and reduction of variables techniques were employed. The technique allows the proper handling of joint variables and reducing the number of variables with the least possible loss of information. The results were evaluated through a faults simulator of a DC drive system. New techniques were developed such as: Effective methods of control chart project; new architectures for efficient implementation of variable reduction, optimizing the plant model.