An Analysis and Design of Predictive Maintenance System Using Association Rule Mining in Moulding Machine PT X
Abstract
One of the challenges in the implementation of Total Productive Maintenance (TPM)
in the manufacturing industry is a slow managerial decision-making to respond the
condition in the factory. This research investigates the answers of these challenges by
analyzing and modeling the equipment condition and the response of actions required in a
wooden door manufacturing industry. TPM implementation in this company has deployed
the Overall Equipment Effectiveness (OEE) measurement as an indicator of the equipment
@tilization and condition. Through an analysis and modeling of the OEE value obtained
from the factory, the formulation of Association Rule Mining (ARM) aims to find a rule
that shows the well computed relationship between measurable indicators of OEE with the
response of action required to take in certain condition of machine utilization. Results
obtained from ARM accelerate the decision to establish an appropriate TPM management
Strategy based on the rules. The generated dynamic rules form and facilitate the process
of decision-making by related stakeholders. Furthermore, relying on these rules the action
taken by the company induced to a higher reliable and increasing the effectiveness of
response and efficiency of time and costs.