Authors:
Sumika Arima
;
Ushio Sumita
and
Jun Yoshii
Affiliation:
University of Tsukuba, Japan
Keyword(s):
Semi-conductor manufacturing, Minor-stoppages, Sequential association rules, Preventive maintenance policies.
Related
Ontology
Subjects/Areas/Topics:
Applications
;
Automation of Operations
;
Maintenance
;
Methodologies and Technologies
;
Operational Research
;
Pattern Recognition
;
Risk Management
;
Software Engineering
;
Software Project Management
Abstract:
In semi-conductor manufacturing, the machine downtimes due to minor-stoppages often exceed 40% of the
working hours of a day, and would amount to the huge loss. However, effective methodological tools for
predicting and preventing the minor-stoppages are hard to come by. The purpose of this research is to fill this
gap by establishing effective preventive maintenance policies for controlling minor-stoppages. Our approach
is to develop association rules based on sequential data along the time axis so that the resulting rules could be
used for predicting occurrences of certain minor-stoppages. The proposed methodology is applied to a real
data set and yields two preventive maintenance policies in a concrete form, thereby demonstrating its power
and usefulness. While the paper focuses on the testing process, the methodology proposed in this paper is
valid for other production processes, provided that similar sequential data could be collected.