K-Means Clustering Approach to Determine Ore Type in Laterite Nickel Deposit

Abstract

Nickel demand will increase as the electric vehicle production plan increases in the years ahead. Nickel matte is an important raw material in producing the producing electric cars. The precision and speed with which the kind of laterite nickel ore is determined based on the olivine concentration are critical because it affects energy requirements and nickel recovery. The present challenge is that it takes a long time to obtain the Loss on Ignition (LoI) number, which will be used to calculate the olivine concentration. As a result, an alternate approach for determining the type of ore without going through the phases of assessing the LoI value is required. The k-means clustering method examined data from one of the blocks in the PT Vale Indonesia Tbk mining region. The K-Means method is used to cluster data from 9 chemical elements. The olivine group that will be produced as a result of this analysis is characterized by three scenarios: the first one is the olivine group using two clusters according to the present split of olivine groups, namely high olivine and low olivine. The second scenario employs three clusters: high olivine, medium olivine, and low olivine. The third scenario uses four clusters based on the elbow method and silhouette guidelines. According to the findings, the best number of clusters was two clusters. The level of accuracy for 2 clusters, when compared with the conventional method, achieved 97.9% for cluster 1 and 85.0% for cluster 2.