#TITLE_ALTERNATIVE#

• Main Author: ILHAM MAULANA (NIM : 10212098), FARIDZ
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/22082
• Institution: Institut Teknologi Bandung

In oil and gas exploitation, it cannot be denied that there is a risk of getting a dry well. In order to reduce it, a reservoir has to be described in terms of lithology and <br /> <br /> pore fluid content. Hence, it is important to use a method that can classify well data by its lithology and pore fluid content. One tool that can be used to classify data <br /> <br /> is artificial neural network, which is a method that utilize statistic process to create complex function between an input and target output. After the classification in terms of lithology and pore fluid content are made, cross plotting is applied to have a better understanding in determination of rock physics analysis to discriminate a reservoir. Based on research conducted, the error value is proportional to a number of inputs used in artificial neural network, to note that input variables are originated from direct well log data. Furthermore, derived variables produce smaller error on classification, but the change are smaller than using primary variables from well <br /> <br /> log data. Then, the discrimination between the classified lithology and pore fluid content are made using IP PR and SQp SQs cross plot. In shallow zone, all of lithology and pore fluid content can be discriminate, but in the deep zone can only discriminate carbonate and sandstone, but different pore fluid contents are overlapping each other.