Performance Based Anomaly Detection Analysis of a Gas Turbine Engine by Artificial Neural Network Approach

Amar Kumar, Alka Srivastava, Avisekh Banerjee, and Alok Goel
Submission Type: 
Full Paper
AttachmentSizeTimestamp
phmc_12_107.pdf422.61 KBSeptember 20, 2012 - 9:55am

This present work follows our earlier research efforts on fault diagnosis and prognosis solutions considering statistical and physics based approaches. In-service performance analysis and detection of any malfunctioning in an operating small sized gas turbine engine using artificial neural network approach is the central theme of this work. The measured engine operating and performance parameters are used to train two neural network models, namely back propagation and generalized regression. Following the training and validation of the neural network model, simulation results for test data corresponding to various engine usage stages are found to be close by two models. The analysis identifies an anamoly in the simulated and measured data collected 17 months after the engine overhauling which may be attributed to deliberate adjustments in the operating parameters. A threshold for anomaly detection in terms of the probability levels for variation of the rated power capacity of the engine is also studied.

Publication Year: 
2012
Publication Volume: 
3
Publication Control Number: 
107
Page Count: 
8
Submission Keywords: 
artificial neural networks
anomaly detection
gas turbine
overhauling
performance analysis
Submission Topic Areas: 
Data-driven methods for fault detection, diagnosis, and prognosis
Submitted by: 
  
 
 
 

follow us

PHM Society on Facebook Follow PHM Society on Twitter PHM Society on LinkedIn PHM Society RSS News Feed