Active Fault Diagnosis for a Class of Nonlinear Uncertain Systems: A Distributionally Robust Approach

TL;DR

This paper introduces a distributionally robust active fault diagnosis method for nonlinear uncertain systems, accounting for ambiguity in model parameters and using moment-based measures to distinguish fault scenarios effectively.

Contribution

It presents a novel fault diagnosis approach using total variation distance and moment-based measures, applicable to various distributions beyond normal, with proven optimality conditions and practical implementation.

Findings

Effective fault discrimination demonstrated on a three-tank system

Method handles distribution ambiguity in model parameters

Applicable to non-normal distributions

Abstract

This work is devoted to the development of a distributionally robust active fault diagnosis approach for a class of nonlinear systems, which takes into account any ambiguity in distribution information of the uncertain model parameters. More specifically, a new approach is presented using the total variation distance metric as an information constraint, and as a measure for the separation of multiple models based on the similarity of their output probability density functions. A practical aspect of the proposed approach is that different levels of ambiguity may be assigned to the models pertaining to the different fault scenarios. The main feature of the proposed solution is that it is expressed in terms of distribution's first and second moments, and hence, can be applied to alternative distributions other than normal. In addition, necessary and sufficient conditions of optimality are derived, and an offline active fault diagnosis procedure is provided to exemplify the implementation of the proposed scheme. The effectiveness of the proposed distributionally robust approach is demonstrated through an application to a three-tank benchmark system under multiple fault scenarios.