Distributed output‐feedback fault detection and isolation of cascade process networks

Distributed output‐feedback fault detection and isolation (FDI) of nonlinear cascade process networks that can be divided into subsystems is considered. Based on the assumption that an exponentially convergent estimator exists for each subsystem, a distributed state estimation system is developed. I...

Full description

Saved in:
Bibliographic Details
Published in:AIChE journal 2017-10, Vol.63 (10), p.4329-4342
Main Authors: Yin, Xunyuan, Liu, Jinfeng
Format: Article
Language:English
Subjects:
actuator faults
chemical processes
Convergence
Fault detection
Faults
Feedback
Noise measurement
nonlinear systems
Output feedback
sensor faults
State estimation
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Distributed output‐feedback fault detection and isolation (FDI) of nonlinear cascade process networks that can be divided into subsystems is considered. Based on the assumption that an exponentially convergent estimator exists for each subsystem, a distributed state estimation system is developed. In the distributed state estimation system, a compensator is designed for each subsystem to compensate for subsystem interaction and the estimators for subsystems communicate to exchange information. It is shown that when there is no fault, the estimation error of the distributed estimation system converges to zero in the absence of system disturbances and measurement noise. For each subsystem, a state predictor is also designed to provide subsystem state predictions. A residual generator is designed for each subsystem based on subsystem state estimates given by the distributed state estimation system and subsystem state predictions given by the predictor. A subsystem residual generator generates two residual sequences, which act as references for FDI. A distributed FDI mechanism is proposed based on residuals. The proposed approach is able to handle both actuator faults and sensor faults by evaluating the residual signals. A chemical process example is introduced to demonstrate the effectiveness of the distributed FDI mechanism. © 2017 American Institute of Chemical Engineers AIChE J, 63: 4329–4342, 2017
ISSN:0001-1541
1547-5905
DOI:10.1002/aic.15791