Scholarlink

<p>This research aims to develop a robust fault detection and isolation system for uncertain nonlinear processes. The proposed solution extends Kernel Principal Component Analysis to accommodate interval data, requiring the formulation of a resolution method for a parametric programming problem to construct the interval kernel model. The monitoring process incorporates interval extensions of residuals and principal component-based detection indices. For fault isolation, an interval-based extension of the unified contribution plots technique is introduced. Simulations on spacecraft reaction wheels data reveal that the proposed approach achieves higher fault detection rates and lower false alarm rates compared to existing nonlinear fault diagnosis methods. These findings highlight the effectiveness and practical potential of the proposed method for improving fault detection and isolation in complex, uncertain nonlinear systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Interval Kernel PCA for Nonlinear Fault Detection and Isolation: Application to a Spacecraft Reaction Wheels

  • 期刊论文

  • 发表日期：

    2025年9月2日

查看全文
  • Imen Gueddi,
  • Othman Nasri

摘要

This research aims to develop a robust fault detection and isolation system for uncertain nonlinear processes. The proposed solution extends Kernel Principal Component Analysis to accommodate interval data, requiring the formulation of a resolution method for a parametric programming problem to construct the interval kernel model. The monitoring process incorporates interval extensions of residuals and principal component-based detection indices. For fault isolation, an interval-based extension of the unified contribution plots technique is introduced. Simulations on spacecraft reaction wheels data reveal that the proposed approach achieves higher fault detection rates and lower false alarm rates compared to existing nonlinear fault diagnosis methods. These findings highlight the effectiveness and practical potential of the proposed method for improving fault detection and isolation in complex, uncertain nonlinear systems.