PENENTUAN WAKTU DEWATERING PADA SUMUR COALBED METHANE (CBM) JENIS VERTICAL WELL SEBAGAI FUNGSI SIFAT-SIFAT RESERVOIR DAN BATUBARA DENGAN MENGGUNAKAN ARTIFICIAL NEURAL NETWORK

• Main Author: Anggi Naluriawan Santoso, Ade
• Format: Final Project
• Language: Indonesia
• Subjects: Pertambangan dan operasi berkaitan
• Online Access: https://digilib.itb.ac.id/gdl/view/64090
• Institution: Institut Teknologi Bandung

Increasing gas demand significantly in Indonesia and its surrounding region in the last decade affects on exploration of gas resources. However, the conventional gas reserves are decreased and more difficult to be found. The alternate gas sources, therefore should be substituted to fullfill the gas demand in the near future. Coalbed methane (CBM) is one of the unconventional gas which main composition is methane which is adsorbed in the matrix of coal and formed during the coalification process. CBM is produced by dewatering process, that is producing the water filled in the cleats inside coal layer to decrease reservoir pressure until desorption pressure is reached. Once desorption pressure is reached, methane released from coal matrix surface and flows through micropore to natural fractures towards the wellbore. This period of water producing is called dewatering time. Dewatering time prediction is very important for developing CBM project because this period of time will affect the value of the project and as the key to efficient gas production from these reservoir. To predict dewatering time, a reservoir engineer occasionally use reservoir simulation. Last research that already conducted by Halim, H (2013) gives an equation to estimate the dewatering time as function of reservoir area, reservoir thickness, fracture porosity, fracture permeability, desorption pressure, reservoir pressure, and bottomhole pressure. This study aims at estimation of dewatering time on Vertical CBM as function of reservoir and coal properties using Artificial Neural Network (ANN) technique. ANN technique is a poweful modelling tool to establish the complex relationship between input parameters and the dewatering time . A total number of data points consist of 335 sets has been used for generating, validating, and testing the ANN model. The best ANN model gives average data errors = 6% and will be compared with the latest equation that gives average data errors = 9%.