Segregation Of Rock Properties Using Machine Learning Algorithm With Euclidean Distance

Abstract

In rock drilling applications, abrasion causes wear in inserts and hostile working conditions cause damage to other bit components. The effects of physico-mechanical properties of rock on the tool wear are investigated by several researchers in the past. So, it becomes imperative to exhibit good scalability of rock properties by segregating rock samples having similar properties for natural homogeneous rock property groupings. The aim of this work is to segregate groups with similar type of rock properties and assign them into a cluster. This work considers a machine learning based hierarchical clustering approach to segregate groups of rock with similar traits. The results obtained from this study initiate a conversation on the proper choice of rock and tool material for doing laboratory studies using wear test apparatus. The analysis's findings map the distinct qualities of the rock for different mining areas by classifying groups of rocks with comparable characteristics.