Application of Data Mining in Failure Estimation of Cold Forging Machines: An Industrial Research

Buse TURKOGLU¹, Murat KOMESLI^2*, Mehmet Suleyman UNLUTURK³
¹Norm Group, Department of Research and Development,
10007 Sk. 1/1, AOSB, 35620, Cigli, Izmir, Turkey
buse.turkoglu@norm-fasteners.com.tr
²Yasar University, School of Applied Sciences, Department of Management Information Systems,
Universite cad. 37, 35100, Bornova, Izmir, Turkey
murat.komesli@yasar.edu.tr (^*Corresponding author)
³ Yasar University, Faculty of Engineering, Department of Software Engineering,
Universite cad. 37, Bornova, Izmir, 35100, Turkey
mehmet.unluturk@yasar.edu.tr

ABSTRACT: The industrial companies are now reaching out for solutions that would enable them to reduce the number of manufacturing defects in production so that they may be able to compete and maintain their sustainability in the market. All production processes need to be uninterruptible. This study utilizes data mining algorithms to turn the data created by machines into information. These data mining algorithms are effective tools for reducing the cold forging machine downtime. Furthermore, the selected data mining methodology, the J48 model, generates meaningful results for a large real-life data set and predicts the error according to a behavioral model. The J48 model successfully detected 28 failures from this data set which suggests that it can be a promising method for reducing the periods of downtime of the cold machine.

KEYWORDS: Data mining, Industrial systems, Predictive maintenance, Cold forging machines, Failure estimation.

>>FULL TEXT: PDF

CITE THIS PAPER AS:
Buse TURKOGLU, Murat KOMESLI, Mehmet Suleyman UNLUTURK, Application of Data Mining in Failure Estimation of Cold Forging Machines: An Industrial Research, Studies in Informatics and Control, ISSN 1220-1766, vol. 28(1), pp. 87-94, 2019. https://doi.org/10.24846/v28i1y201909