Abstract:This paper proposes a nonlinear estimator for the robust reconstruction of process and sensor faults for a class of uncertain nonlinear systems. The proposed fault estimation method augments the system dynamics with an ultra-local (in time) internal state-space representation (a finite chain of integrators) of the fault vector. Next, a nonlinear state observer is designed based on the known parts of the augmented dynamics. This nonlinear filter (observer) reconstructs the fault signal as well as the states of the augmented system. We provide sufficient conditions that guarantee stability of the estimation error dynamics: firstly, asymptotic stability (i.e., exact fault estimation) in the absence of perturbations induced by the fault model mismatch (mismatch between internal ultra-local model for the fault and the actual fault dynamics), uncertainty, external disturbances, and measurement noise and, secondly, Input-to-State Stability (ISS) of the estimation error dynamics is guaranteed in the presence of these perturbations. In addition, to support performance-based estimator design, we provide Linear Matrix Inequality (LMI) conditions for $\mathcal{L}_2$-gain and $\mathcal{L}_2-\mathcal{L}_\infty$ induced norm and cast the synthesis of the estimator gains as a semi-definite program where the effect of model mismatch and external disturbances on the fault estimation error is minimized in the sense of $\mathcal{L}_2$-gain, for an acceptable $\mathcal{L}_2-\mathcal{L}_\infty$ induced norm with respect to measurement noise. The latter result facilitates a design that explicitly addresses the performance trade-off between noise sensitivity and robustness against model mismatch and external disturbances. Finally, numerical results for a benchmark system illustrate the performance of the proposed methodologies.

Rapid Fault Isolation For A Class Of Nonlinear Lipschitz Systems Via Deterministic Learning

Rapid Sensor Fault Diagnosis for a Class of Nonlinear Systems via Deterministic Learning

Performance-Guaranteed Fault-Tolerant Control for Uncertain Nonlinear Systems via Learning-Based Switching Scheme

Fault Isolation of Linear Stochastic Time-Varying Systems with Strong Noise

Small Fault Detection of Discrete-Time Nonlinear Uncertain Systems

ROBUST FAULT DETECTION AND ISOLATION FOR UNCERTAIN LINEAR RETARDED SYSTEMS

Fault diagnosis for a class of nonlinear uncertain systems using deterministic learning approach

Sensor Fault Detection and Diagnostics of Nonlinear Systems with Application to Paper Machines

Automated Fault Accommodation for Discrete-Time Systems Using Online Approximators

Adaptive Learning Based Fault Tolerant Control for Uncertain Nonlinear Systems

Iterative learning based fault estimation for nonlinear discrete-time systems

An Accelerated Iterative Learning Algorithm Based on Feedback Terms for Fault Estimation of Nonlinear Parabolic Distributed Parameter Systems

Integrating reinforcement learning with deterministic learning for fault diagnosis of nonlinear systems

A Probabilistic Approach to Robust Fault Detection for a Class of Nonlinear Systems

Fault Detection and Isolation for Nonlinear Systems with Full State Information

A data-driven fault isolation and estimation approach for unknown linear systems

Fault Identification for a Class of Nonlinear Systems of Canonical Form via Deterministic Learning

Robust Fault Estimators for Nonlinear Systems: An Ultra-Local Model Design

Just-in-time Learning-aided Nonlinear Fault Detection for Traction Systems of High-speed Trains

Iterative Learning Based Estimation of Periodically Occurring Faults

A Robust Fault Detection Approach for Nonlinear Systems