Lukman, Ichsan Al Sabah (2023) Optimasi Asp Injection Dalam Meningkatkan Oil Recovery Pada Sandstone Reservoir Menggunakan Metode Super Learner. Other thesis, Universitas Islam Riau.

Text 163210033.pdf - Submitted Version Restricted to Registered users only Download (3MB) | Request a copy

Abstract

Enhanced Oil Recovery (EOR) is a recovery method after primary and secondary recovery is carried out to increase recovery factor (RF) and prevent a drastic decline in production. One of the effective EOR methods to use is Chemical EOR, namely ASP (Alkaline, Surfactant, Polymer) injection. This method has been applied in previous studies. The synergy of this method aims to increase the oil recovery factor (RF) by changing the wettability of the rock, reducing the IFT value and increasing the sweep efficiency. In this research, optimization of several parameters that affect the increase in RF in the ASP Injection method, namely the concentration of surfactant injection (SO4 2- ), alkali (Na+, CO3 2- ), and polymers. With these parameters, of course, it is desirable to have a balance between compounds in order to achieve optimal results. Therefore, if the optimization is only done using conventional simulators such as Computer Modeling Group (CMG) software, making the base model will become more complicated and require longer time just to do RF calculations. To simplify the optimization of ASP injection, the research will be continued using the Super Learner (SL) method. This SL method is an ensemble learner method that uses k-fold cross validation from several base learners with a fairly efficient input algorithm. With a ratio of 80:20 between training and testing the SL model has a fairly good performance with an R2 value of 0.988 and an RMSE of 0.304. Then predictions were made on these parameters and optimization of the RF value was found to be 79.49% with the injection parameters of surfactants, polymers, SO4 2- , CO3 2- , and Na+ where respectively 5483.29 ppm, 2242.61 ppm, .5610.15 ppm, 7053.59 ppm, and 9939.35 ppm.

Actions (login required)

View Item