MINERAL DEPOSIT PROSPECTING MODEL DEVELOPMENT, BASED ON THE FUZZY LOGIC AND WEIGHT OF EVIDENCE (WOFE) METHOD: CASE STUDY OF HIGH SULPHIDATION EPITHERMAL (HSE)

• Main Author: Setyadi, Harman
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/55165
• Institution: Institut Teknologi Bandung

Prospecting is an assessment to evaluate the potential economic mineral deposits occurrence based on the existing exploration data collected. Conventionally, prospect evaluation was done by integrating multiple spatial data exploration on the light table. At this time, integration commonly was done by compiling digital map using geographic information systems (GIS). Empirical analytical methods of prospecting have been developed to obtain better result and to reduce the subjective factor in the interpretation. Fuzzy logic and weight of evidence (WOFE) method was widely implemented by the previous researchers by some modification and combination. Commonly their research was done on the regional scale (1:250,000 – 1:100,000), training data was determined based on the mineral deposits/occurrence which was considered as the homogenous point. In the more detailed deposit scale (1:5,000), the ore body should be consisted by several different alteration-mineralization zone with different physical properties. This feature should be impacted to the accuracy of the mineral prospect prediction. This study was proposed to improve the accuracy of gold deposit prediction models on the deposit scale (1: 5,000) with in 1.5x1.5 km2 area. Training data was determined and controlled based on the nature of the mineralization zone physical properties. Model was developed using input of the geophysical data which was consist of reduce to the pole (RTP), analytical signal (ANS), Resistivity and IPchargeability. Model experiment was carried out to evaluate the prediction accuracy in the different level using different horizontal slice of IP-chargeability and resistivity data at -30m, -50m and -100m. Based on this study, prediction accuracy should be improved by implemented the geological control of rock physical properties, because training data assignment from the different zone could be avoided. Model calculation and data flow was simplified using modified WofE and MS Excel programming. Training data fuzzyiv model is represent the combination of physical properties which was proved by the geological out crop and drill core observation. Model verification using the resource model with 0.15 ppm Au outline resulted good correlation, with accuracy 67%.