TY - JOUR
T1 - Unknown Input Observer Based Robust Fault Estimation for Systems Corrupted by Partially-Decoupled Disturbances
AU - Gao, Zhiwei
AU - Liu, Xiaoxu
AU - Chen, Michael
PY - 2016/4
Y1 - 2016/4
N2 - Robust fault estimation plays an important role in real-time monitoring, diagnosis, and fault-tolerance control. Accordingly, this paper aims to develop an effective fault estimation technique to simultaneously estimate the system states and the concerned faults, while minimizing the influences from process/sensor disturbances. Specifically, an augmented system is constructed by forming an augmented state vector composed of the system states and the concerned faults. Next, an unknown input observer (UIO) is designed for the augmented system by decoupling the partial disturbances and attenuating the disturbances that cannot be decoupled, leading to a simultaneous estimate of the system states and the concerned faults. In order to be close to the practical engineering situations, the process disturbances in this study are assumed not to be completely decoupled. In the first part of this paper, the existence condition of such an UIO is proposed to facilitate the fault estimation for linear systems subjected to process disturbances. In the second part, robust fault estimation techniques are addressed for Lipschitz nonlinear systems subjected to both process and sensor disturbances. The proposed technique is finally illustrated by the simulation studies of a three-shaft gas turbine engine and a single-link flexible joint robot.
AB - Robust fault estimation plays an important role in real-time monitoring, diagnosis, and fault-tolerance control. Accordingly, this paper aims to develop an effective fault estimation technique to simultaneously estimate the system states and the concerned faults, while minimizing the influences from process/sensor disturbances. Specifically, an augmented system is constructed by forming an augmented state vector composed of the system states and the concerned faults. Next, an unknown input observer (UIO) is designed for the augmented system by decoupling the partial disturbances and attenuating the disturbances that cannot be decoupled, leading to a simultaneous estimate of the system states and the concerned faults. In order to be close to the practical engineering situations, the process disturbances in this study are assumed not to be completely decoupled. In the first part of this paper, the existence condition of such an UIO is proposed to facilitate the fault estimation for linear systems subjected to process disturbances. In the second part, robust fault estimation techniques are addressed for Lipschitz nonlinear systems subjected to both process and sensor disturbances. The proposed technique is finally illustrated by the simulation studies of a three-shaft gas turbine engine and a single-link flexible joint robot.
KW - Augmented system approach
KW - Fault diagnosis
KW - augmented system approach
KW - fault diagnosis
KW - fault estimation
KW - linear matrix inequality (LMI)
KW - unknown input observer (UIO)
U2 - 10.1109/TIE.2015.2497201
DO - 10.1109/TIE.2015.2497201
M3 - Article
VL - 63
SP - 2537
EP - 2547
JO - IEEE Transactions on Industrial Electronics
JF - IEEE Transactions on Industrial Electronics
SN - 0278-0046
IS - 4
ER -