Symbolic Time-Series Analysis for Anomaly Detection in Mechanical Systems

Abstract

This paper examines the efficacy of a novel method for anomaly detection in mechanical systems, which makes use of a hidden Markov model, derived from the time-series data of pertinent measurement(s). The core concept of the anomaly detection method is symbolic time-series analysis that is built upon the principles of Automata Theory, Information Theory, and Pattern Recognition. The performance of this method is compared with that of other existing pattern-recognition techniques from the perspective of early detection of small fatigue cracks in ductile alloy structures. The experimental apparatus, on which the anomaly detection method is tested, is a multi-degree-of-freedom mass-beam structure excited by oscillatory motion of two electromagnetic shakers. The evolution of fatigue crack damage at one or more failure sites are detected from symbolic time-series analysis of displacement sensor signals.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2006
Accession Number
ADA507170

Entities

People

  • Amol Khatkhate
  • Asok Ray
  • Eric Keller
  • Shalabh Gupta
  • Shin C. Chin

Organizations

  • Pennsylvania State University

Tags

DTIC Thesaurus Topics

  • Algorithms
  • Anomaly Detection
  • Change Detection
  • Detection
  • Dynamic Response
  • Hidden Markov Models
  • Mechanical Structure
  • Military Research
  • Neural Networks
  • Pattern Recognition
  • Probabilistic Models
  • Probability
  • Probability Distributions
  • Random Variables
  • Resonant Frequency
  • Stochastic Processes
  • Time Series Analysis

Fields of Study

  • Engineering

Readers

  • Computer Vision.
  • Control Systems Engineering.
  • Theoretical Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference