Skip to content
Longterm Wiki
Back

TEE for ML (Chen et al.)

paper

Authors

Fatemeh Hashemniya·Benoït Caillaud·Erik Frisk·Mattias Krysander·Mathias Malandain

Credibility Rating

3/5
Good(3)

Good quality. Reputable source with community review or editorial standards, but less rigorous than peer-reviewed venues.

Rating inherited from publication venue: arXiv

Technical paper on multi-mode diagnostics algorithms for complex systems, relevant to AI safety through fault detection and system reliability in multi-mode autonomous systems.

Paper Details

Citations
0

Metadata

arxiv preprintprimary source

Abstract

Multi-mode systems can operate in different modes, leading to large numbers of different dynamics. Consequently, applying traditional structural diagnostics to such systems is often untractable. To address this challenge, we present a multi-mode diagnostics algorithm that relies on a multi-mode extension of the Dulmage-Mendelsohn decomposition. We introduce two methodologies for modeling faults, either as signals or as Boolean variables, and apply them to a modular switched battery system in order to demonstrate their effectiveness and discuss their respective advantages.

Summary

This paper addresses the challenge of diagnosing faults in multi-mode systems, which operate across different dynamic configurations and are difficult to analyze using traditional structural diagnostics. The authors propose a multi-mode diagnostics algorithm based on a multi-mode extension of the Dulmage-Mendelsohn decomposition and introduce two fault modeling approaches: signal-based and Boolean variable-based representations. The methodologies are demonstrated on a modular switched battery system, with discussion of their respective strengths and limitations.

Cited by 1 page

PageTypeQuality
AI Governance Coordination TechnologiesApproach91.0

Cached Content Preview

HTTP 200Fetched Mar 20, 202660 KB
# Fault Diagnosability Analysis of Multi-Mode Systems

Fatemeh Hashemniya
Benoît Caillaud
Erik Frisk
Mattias Krysander
Mathias Malandain
Department of Electrical Engineering

Linköping University, SE 581-83, Linköping, Sweden

e-mail: {fatemeh.hashemniya, erik.frisk, mattias.krysander}@ liu.se
National Institute for Research in Digital Science and Technology (Inria),
Inria centre at Rennes University, Rennes, France.

e-mail: {benoit.caillaud, mathias.malandain}@inria.fr

###### Abstract

Multi-mode systems can operate in different modes, leading to large numbers of different dynamics. Consequently, applying traditional structural diagnostics to such systems is often untractable. To address this challenge, we present a multi-mode diagnostics algorithm that relies on a multi-mode extension of the Dulmage-Mendelsohn decomposition. We introduce two methodologies for modeling faults, either as signals or as Boolean variables, and apply them to a modular switched battery system in order to demonstrate their effectiveness and discuss their respective advantages.

###### keywords:

Multi-mode systems, Diagnostics, Dulmage-Mendelsohn decomposition.

,
,
,
, and

## 1 Introduction

Fault detection and diagnosis are important for the health monitoring of physical systems.
Model-based approaches for single-mode, smooth, systems is a well-established field, supported by a large body of literature covering various approaches like structural methods Blanke et al. ( [2006](https://ar5iv.labs.arxiv.org/html/2312.14030#bib.bib4 "")), parity space techniques, and observer-based methods Isermann ( [2006](https://ar5iv.labs.arxiv.org/html/2312.14030#bib.bib11 "")).

While single-mode systems are often described using differential algebraic equations (DAEs), the modeling of non-smooth physical systems yields switched DAEs, also known as multi-mode DAEs (mmDAEs), which combine continuous behaviors, defined as solutions of a set of DAE systems,
with discrete mode changes Trenn ( [2012](https://ar5iv.labs.arxiv.org/html/2312.14030#bib.bib14 "")); Benveniste et al. ( [2020](https://ar5iv.labs.arxiv.org/html/2312.14030#bib.bib2 "")). Direct application of
traditional fault diagnosis methods to all possible configurations of multi-mode systems quickly
becomes intractable, as the number of modes tends to be exponential in the size of the system.
The method proposed by Khorasgani and Biswas ( [2017](https://ar5iv.labs.arxiv.org/html/2312.14030#bib.bib12 "")) works around this issue by coupling a mode estimation algorithm with a single-mode diagnosis methodology, akin to just-in-time compilation in computer science. This approach unfortunately puts the burden on solving mode estimation problems, which often turn out to be intractable for the same reason.

Structural fault detectability and isolability is a graph-based method to evaluate diagnosability properties on
DAEs Frisk et al. ( [2012](https://ar5iv.labs.arxiv.org/html/2312.14030#bib.bib8 "")). It is based on the Dulmag

... (truncated, 60 KB total)
Resource ID: 5a5c934f8df343c9 | Stable ID: YTdlM2M4Nm