Unknown input observers for fault diagnosis in Lipschitz nonlinear systems

Liu, Xiaoxu and Gao, Zhiwei (2015) Unknown input observers for fault diagnosis in Lipschitz nonlinear systems. In: International Conference on Mechatronics and Automation, 2-5 August 2015, Beijing.

Text ICMA2015-271.pdf Restricted to Repository staff only Download (696kB) | Request a copy

Abstract

This paper considers the problem of robust fault detection for Lipschitz nonlinear systems impaired by faults and unknown inputs in both process and sensors. The purpose of the proposed strategy is to estimate both the system states and the considered faults existing in actuators as well as sensors while minimize the influences from disturbances. An innovative robust observer design methodology is developed through an integration of fault estimation approach and unknown input observer (UIO).

In contrast to previous studies, the considered unknown inputs do not only exist in system process but also sensors. Moreover, to meet the practical engineering situations, they are not assumed to be decoupled completely. With the assist of linear matrix
inequality (LMI) method, observer parameters are determined to attenuate the influences of unknown inputs which cannot be decoupled, as well as guarantee the stability of the estimation dynamics. Simulation results are presented to demonstrate that the proposed observer scheme performs reasonably well.

Actions (login required)

View Item

Downloads