DEVELOPMENT OF OPTIMIZATION MODEL OF MINE DRAINAGE DRAINHOLE DESIGN CASE STUDY GRASBERG OPEN PIT MINE PT FREEPORT INDONESIA

• Main Author: AGUNG CAHYADI, TEDY
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/31253
• Institution: Institut Teknologi Bandung

To solve groundwater problems on slopes, it is necessary to implant drainage, for example by installing a horizontal drainhole. Installation of drainhole with a certain spaced uniform pattern is widely used in the installation of drainhole in fractured rocks on open pits. However, the condition is less effective because of the excessive ammounts of drainholes installed. Drainholes mounted with limited numbers and are precisely in accordance with the conceptual model will be more optimal than drainholes that are installed in excessive numbers and uniformly spaced. The inaccuracy of the target location in drainhole drilling results in the flow rate and drawdown to be unmaximized, and unused in the Grasberg Mine. Integrated modeling is required starting from hydraulic conductivity model (K), groundwater flow system model, and optimization model in determining optimal drainhole design in the form of Simulation and Optimization (SO). <br /> <br /> <br /> <br /> <br /> <br /> The purpose of this research is to determine the distribution of K in block model using HC-System and supported by the Ordinary Kriging (OK) method, Artificial Neural Network (ANN) with limited data; and determining the optimal drainhole design on the fractured media with integration of Simulation and Optimization model (Multi Stage SO GWsim-GA) through a short time. <br /> <br /> <br /> <br /> <br /> <br /> Hydraulic conductivity distribution (K) modeling is built through the HC-System equation. The equations and variables are used for the distribution of K values by using the Ordinary Kriging (OK), Artificial Neural Network Feed Forward Segmentation (ANNFFS) and Artificial Neural Network Back Propagation (ANNBP) approach. The result of the distribution is used to model the groundwater flow by using the finite difference method found in the Groundwater Simulator (GWsim) on Visual ModFlow software. Groundwater Simulator and Genetic Algorithm (GWsim-GA) are used for optimizing the drainhole design. The verification process is performed to compare between the prediction model and field data. <br /> <br /> <br /> <br /> <br /> <br /> The results showed that the Numerical analysis, using the HC-System, can be used to predict the K value in locations where hydraulic tests aren’t done based on hydrogeological, geotechnical, and geological datas based on rock quality designation (RQD), lithology permeability index (LPI), gouge content designation (GCD) and depth index (DI). The equation primarily used to predict the hydraulic conductivity value (K) along the bore hole length is = 2 x 10-6 x (HC)0.5571 with the correlation coefficient (R) = 0,85. The K model distribution using the ANNBP produces a better distribution compared to the OK & ANNFFS method, with a Standard Error of Estimate (SEE) = 0,14, Root Mean Square (RMS) = 0,41, Normal Root Mean Square (NRMS) = 0,14, R = 0,93 and the % error = 4,98%. The RQD and LPI variables are used as primary parameters in the study process using ANNBP. The K distribution data that’s been verified is used as input data in the ground water flow system model of Grasberg’s soil. The calibration in steady state conditions produced these data R = 0,94, SEE = 11,6 m, RMS = 66,9 m, and NRMS = 9,2%. This condition is used as a basis in modeling the SO GWsim-GA’s verification process in the field. <br /> <br /> <br /> <br /> <br /> Determining the stochastic optimal position, direction, number and length of the drainhole can be done by combining both the Visual ModFlow and GA method (Multi Stage SO GWsim-GA), so that the optimal drainhole search simulation time can be shortened from the previous 2,500 hours to 19.5 hours for the synthetic model, whilst the field model is only performed on average 25 simulations or 25 hours of 180 simulations or 180 hours. <br /> <br /> <br /> <br /> <br /> <br /> The optimal drainhole design of fractured media can be carried out with the integration of a heterogeneous K model, groundwater flow model and optimization model with a function criteria of maximizing the flow rate (Q) discharge and drawdown (?H), and minimizing the drainhole number (N) and drainhole length (L), with constraint limitation factors such as the drawdown (?H), flow rate (Q), drainhole number (N), drainhole length (L), co-ordinate drainhole (X, Y), spacing (S), and drilling direction (?) becoming important factors in optimal drainhole design. <br />