Prediction of Bottomhole Pressure of Geothermal Wells Using Artificial Neural Network Model Muhammad Taufiq Fathaddin (a*), Rini Setiati (a), Muhammad Burhannudinnur (a), Havidh Pramadika (a), Agus Guntoro (a), Fajar Hendrasto (a), Alvita Kumala Sari (b)
a) Universitas Trisakti, Jakarta 11440, Indonesia
*muh.taufiq[at]trisakti.ac.id
b) PT. Pertamina EP Cepu, Cepu 50275, Indonesia
Abstract
Many geothermal wells produce a two-phase flow of water and steam. These geothermal fluids experience a loss of pressure and temperature as they move from the bottom of the hole to the well head. Loss of fluid pressure is caused by friction with the well wall, acceleration and elevation of the flow. The loss of fluid temperature is caused by changes in geothermal temperature along the well which is a function of depth. In this study, an artificial neural network model is applied to determine well bottom flow pressure. The parameters of well head pressure, well diameter, well length, mass flow rate, enthalpy, heat transfer, well head and well bottom temperature are used as input. Prediction of well bottom flow pressure using this proposed method provides predictions with an average error of 3.28%. Additionally, this method provides more accurate results than the previous method which gave an average error of 6.29%.
