A Non-Redundant Sensor Validation Scheme for Transient and Steady-State Conditioning Monitoring

Abstract

This paper presents a sensor validation scheme capable of detecting failed sensor hardware without sensor redundancy and during non-steady state monitoring conditions. The technical approach utilizes neural networks and fuzzy logic to accomplish the desired goal. Neural networks are used to recognize the nonlinear, inter-relationships between the different types of sensors used in a transient or steady-state measurement environment. Fuzzy logic is used to pre- and post-process the measurement data in order to determine general characteristics about the state of the process being monitored. Different types of neural network architectures were developed and tested to determine their suitability to solving this problem. The feasibility of the method was proven through computer simulation utilizing gas turbine engine data as input to the validation system.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1996
Accession Number
ADP010175

Entities

People

  • Michael J. Roemer

Tags

Communities of Interest

  • Energy and Power Technologies
  • Sensors

DTIC Thesaurus Topics

  • Algorithms
  • Computers
  • Computing System Architectures
  • Confidence Limits
  • Detectors
  • Failure Mode And Effect Analysis
  • Fuzzy Logic
  • Fuzzy Sets
  • Gas Turbines
  • Logic
  • Measurement
  • Network Architecture
  • Neural Networks
  • Pattern Recognition
  • Steady State
  • Supervised Machine Learning
  • Turbines

Fields of Study

  • Engineering

Readers

  • Computational Modeling and Simulation
  • Neural Network Machine Learning.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks