Estimation of APU Failure Parameters Employing Linear Regression and Neural Networks

Renata M. Pascoal, Wlamir O. L. Vianna, João P. P. Gomes, and Roberto K. H. Galvão
Submission Type: 
Full Paper
Supporting Agencies (optional): 
FAPESP (grant 2011/17610-0) and CNPq
phmc_13_042.pdf591.38 KBOctober 9, 2013 - 12:06am

This study is concerned with the building of an appropriate model to estimate failure parameters of an Auxiliary Power Unit (APU). Linear and nonlinear models were used in order to evaluate which model is more suitable for this application. Data for model building and testing were obtained by simulating a nonlinear dynamic model of APU in Matlab/Simulink for various operating conditions to which it may be subjected to and with different levels of failure parameter degradation. Linear models were obtained by least-squares regression, whereas nonlinear models were obtained by training neural networks. The results obtained with these two models were compared. As a result, the neural network models were found to provide a better estimate of the APU failure parameters.

Publication Year: 
Publication Volume: 
Publication Control Number: 
Page Count: 
Submission Keywords: 
Neural Networks
Linear Regression
failure parameters
Submission Topic Areas: 
Model-based methods for fault detection, diagnostics, and prognosis
Submitted by: 

follow us

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