Transactions on Machine Learning and Data Mining (ISSN: 1865-6781)


Volume 5 - Number 2 - October 2012 - Pages 87-116


Forecasting Failures in a Data Stream Context Application to Vacuum Pumping System Prognosis

F. Martin2, N. Méger1, S. Galichet1, and N. Bécourt2

1 LISTIC Laboratory, University of Savoie, B.P. 80439, 74944 Annecy-Le-Vieux, France
2 ADIXEN, B.P. 2069, 74009 Annecy Cedex, France


Abstract

This paper presents a local pattern-based method for fore- casting failures in a data stream context. It also details a successful ap- plication to complex vacuum pumping system prognosis. More precisely, using historical data, the behavior of a set of pumping systems is first modeled by extracting a given type of episode rules, namely the First Local Maximum episode rules (FLM-rules). Each rule comes along with its proper temporal information: its optimal temporal window width. The most reliable FLM-rules are then selected to further forecast system failures in a data stream context. A forecast time interval is supplied for each forecasted failure by merging the temporal information of FLM- rules. The results obtained for production data are very encouraging. Failures are predicted with a good temporal accuracy and precision while very few false alarms are generated. The method presented in this paper is patented and it is being deployed for a customer of the semi-conductor market.


PDFDownload Paper (337 KB)


Back to Table of Contents