Decision Layer by Fusion of Diagnostic Algorithms

Jérôme Lacaille and Tsirizo Rabenoro
Submission Type: 
Full Paper
AttachmentSizeTimestamp
phmc_13_010.pdf2.2 MBOctober 9, 2013 - 2:00am

For manufacturers systems monitoring or production equipment optimization solutions are founded on specific algorithms that produce low level local information about risk of degradation or production loss. In either case local results are combined in synthetic reports aimed to help decision taking at higher level. This work is about the description of an automatic fusion mechanism able to build expert output with direct understanding of the system behavior and help to infer causes of efficiency loss. An example application was built and tested in a semiconductor fab. The algorithms diagnosed yield degradation in different subsystems or work-area and were digested in a weekly report that highlighted the main production problems. We deployed the same methodology for condition based maintenance of aircraft engines on a test platform. The first part of this document sketches out some notations, the second part describes the semiconductor application and the conclusion is dedicated to the transfer in the aeronautic domain for the decision level of an engine fleet health monitoring system.

Publication Year: 
2013
Publication Volume: 
4
Publication Control Number: 
010
Page Count: 
9
Submission Keywords: 
diagnostics
fusion
neural network
decision
genetic algorithm
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