Fault detection using state observers with unknown input, identified by orthogonal functions

Abstract

In this work a methodology of fault analysis in mechanical systems was developed using Kalman Filter state observes, in which, the input to the observers are identified by Fourier, Legendre and Chebyshev orthogonal functions. The proportional-integral observer is presented to the unknown input identification, this observer can be able to find the unknown inputs of the system and these inputs are used to the fault detection by way Kalman Filter Observer. Here can be seen the methodology of parameters and force identification using only the response of the system thought orthogonal functions. The methodology developed is applied in a composed structure of shake tables from Mechanical Vibrations Laboratory.