Nupus, Putri Hayatun (2023) Optimasi Implementasi Steamflood Menggunakan Algoritma Random Forest Pada Coarsening Sand Dan Fining Sand. Other thesis, Universitas Islam Riau.

Text 163210159.pdf - Submitted Version Restricted to Registered users only Download (1MB) | Request a copy

Abstract

Steamflood is an Enhanced Oil Recovery process which is a thermal method that is useful in overcoming heavy oil problems. Heavy oil has an API of 10° - 20°. To overcome this problem, this heavy oil will go through the steam stages which are carried out in steamflood activities so as to reduce the viscosity of the oil. In this study, a steamflood study was carried out on coarsening sand and fining sand using the Computer Modeling Group (CMG) to create a base case scenario and optimizing the steamflood with a steam quality of 50%-90%, a steam rate of 1350 bbl/day-1850 bbl/day, and steam temperature 300?-500?. Optimization analysis will be carried out using several algorithms, XGBoost, ANN, SVM, KNN and Random Forest. The accuracy obtained with some of these algorithm models has different levels of confidence. The highest accuracy is obtained by using the Random Forest model with R2 values for training data and testing data of 0.988 and 0.985 for coarsening sand and 0.995 and 0.993 for fine sand. The steamflood performance optimization results provide steam quality values of 0.88, steam rate of 0.66, and steam temperature of 0.58 in coarsening sand. Whereas in fining sand the value of steam quality is 0.86, steam rate is 0.75, and steam temperature is 0.58.

Actions (login required)

View Item