Fuzzy-Based Prediction of Spatio-Temporal Distribution of Wet Muck in Block Cave Mine of PT Freeport Indonesia

Abstract

Mud rushes, or wet muck spills, are hydro-geotechnical challenges in block cave mines where wet muck spills out of drawpoints formed by the accumulation of fine materials and water in drawbells. The purpose of this paper is to share the results of the developement of an improved predictive tool that can be used to manage wet muck spills. The tool was developed based on the hybrid modeling of wet muck distribution using fuzzy logic and fuzzy number operations. The fuzzy logic operations were applied to model the spatial distribution of wet muck classes, providing the spatial model of drawpoint status based on five contributing factors, i.e. the height of draw, the water content, the grain size of the fine material, rainfall, and no-mucking days. The fuzzy number operations were used in accordance with the mass balance principle to estimate the temporal distribution of wet muck that forms a mud deposit consisting of fine materials and water in a drawbell. The mass balance principle was expressed using the fuzzy ordinary differential equation, including the uncertainty of joining variables. A wet muck spill event at the Deep Ore Zone (DOZ) block cave mine of PT Freeport Indonesia was utilized as a case study as well as to validate the proposed method. The fuzzy-based approach shows promising results in predicting wet muck spill events.