Adaptive Approximation for Multiple Sensor Fault Detection and Isolation of Nonlinear Uncertain Systems

Abstract

This paper presents an adaptive approximation-based design methodology and analytical results for distributed detection and isolation of multiple sensor faults in a class of nonlinear uncertain systems. During the initial stage of the nonlinear system operation, adaptive approximation is used for online learning of the modeling uncertainty. Then, local sensor fault detection and isolation (SFDI) modules are designed using a dedicated nonlinear observer scheme. The multiple sensor fault isolation process is enhanced by deriving a combinatorial decision logic that integrates information from local SFDI modules. The performance of the proposed diagnostic scheme is analyzed in terms of conditions for ensuring fault detectability and isolability. A simulation example of a single-link robotic arm is used to illustrate the application of the adaptive approximation-based SFDI methodology and its effectiveness in detecting and isolating multiple sensor faults.

A distributed fault diagnosis approach utilizing adaptive approximation for a class of interconnected continuous-time nonlinear systems

“…This objective is achieved in two ways; i) adaptive approximation methods are used for learning the modeling uncertainty (so that the learned modeling uncertainty function is used in the design of the residual signals) and then ii) filtering is used to dampen the effect of measurement noise on the diagnosis thresholds. Adaptive approximation methods have been used in the area of fault diagnosis for learning the modeling uncertainties and the fault function and then exploiting the information for fault isolation purposes [8]- [11]. This work extends these results by taking advantage of the learned function in order to obtain tighter thresholds and enhance fault detection.…”

“…As indicated in [11] and the references therein, the bound η I (x I ,x I , u I )| that would be used instead if no learning was used.…”

An Integrated Learning and Filtering Approach for Fault Diagnosis of a Class of Nonlinear Dynamical Systems

“…L can be computed by applying a set membership technique, as indicated in [40] and the references therein. In addition, the boundη…”

“…The first objective is achieved in two ways: 1) adaptive approximation methods are used for learning the modeling uncertainty (so that the learned modeling uncertainty function is used in the design of the residual signals) and 2) by using filtering to attenuate the effect of measurement noise on the diagnosis thresholds. Adaptive approximation methods have been used in the area of fault diagnosis for learning the modeling uncertainties and the fault function for fault isolation purposes [16], [17], [39], [40]. Therefore, the novelty in this paper is that, both tasks, learning and filtering, are integrated in a unified framework and intertwined through the recent filtering approach in [27], which is decomposed in a two stage filtering process in order to derive the required signals for the adaptive approximation and for the residual derivation.…”

Distributed adaptive sensor fault tolerant control for smart buildings

“…The termsf s ,f (i) are the estimations of the faults f s , f (i) , respectively withf s (t s I ) = 0 andf (i) (t (i) I ) = 0, and Ω s , Ω (i) are filtering terms necessary for ensuring the stability property of the adaptive nonlinear estimation schemes [11] with Ω s (t s I ) = 0 and Ω (i) (t (i) I ) = 0, where t s I and t (i)…”