Sequential Pattern Mining Method for Predictive Maintenance of Large Mining Trucks


Kahraman A., Kantardzic M., Kahraman M. M., Kotan M.

1st Mediterranean Forum - Data Science Conference, MeFDATA 2020, Virtual, Online, 24 October 2020, vol.1343 CCIS, pp.126-136 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 1343 CCIS
  • Doi Number: 10.1007/978-3-030-72805-2_9
  • City: Virtual, Online
  • Page Numbers: pp.126-136
  • Keywords: Mine equipment, Mining trucks, Predictive maintenance, Sequential pattern mining
  • Recep Tayyip Erdoğan University Affiliated: Yes

Abstract

In recent decades, various solutions had been sought for reducing operating costs while increasing the production of minerals in mining operations. Equipment health monitoring technologies had been used for monitoring and increasing the availability of machines. However, the data obtained from these technologies had only been used for monitoring the equipment health, and not for the prediction of failures. In this paper, it was relied on alarms and signals collected through real-time health monitoring technologies for predicting crucial mining truck failures. Sequential Pattern Mining (SPM) Method for Predictive Maintenance had been developed and implemented as a methodology to discover which group of alarms and signals might be related to specific truck failures. The results indicate that the SPM method is able to detect machine failures of trucks with high accuracy with an average 96%. The proposed methodology may reduce the maintenance time, and the expenditures caused by truck breakdowns in the mining industry.