Search for collections on Repository Universitas Islam Riau

Prediksi Water Coning Pada Natural Fractured Carbonate Reservoir Menggunakan Metode Random Forest

Ardiansyah, Billy Fergi (2023) Prediksi Water Coning Pada Natural Fractured Carbonate Reservoir Menggunakan Metode Random Forest. Other thesis, Universitas Islam Riau.

[thumbnail of 173210602.pdf] Text
173210602.pdf - Submitted Version
Restricted to Registered users only

Download (5MB) | Request a copy

Abstract

The existence of fractured permeability can create a permeable path for water to the perforation, so that it can cause water breakthrough to occur earlier. The phenomenon of Water Coning is one of the causes of high water production, especially in the Natural Fractured Carbonate Reservoir (NFCR). To build an ideal production scheme and extend the life of production wells in the NFCR reservoir, this study predicts water conning through the water breakthrough time in the NFCR reservoir. This study uses the Random Forest (RF) approach, this method has been widely implemented in the petroleum industry and is proven to be a method that has the lowest level of generalization error and is proven to have good accuracy and performance in predicting modeling and feature importance. This study was initiated by constructing 1000 experiments of 13 parameters, namely from the horizontal permeability of the matrix and fracture, the vertical permeability of the matrix and fracture and to the porosity of the matrix and fracture as the most dominant parameters for the water coning phenomenon in NFCR. Making 1000 experiments using Design of Experiment (DoE) on Computer Modeling Group (CMG - CMOST) software with Water Cut as parameter response. Then proceed with an evaluation using the RF algorithm based on validating the accuracy of the Coefficient of Determination (R2) and MSE using the Python Programming Language. Based on the results of the research that has been done, RF produces fairly accurate water cut predictions with a relatively shorter time, obtained an accuracy model of 0.980 for training data and 0.850 for testing data. With Feature Importance the horizontal fractured permeability parameter has the most influence on the water cut value. Meanwhile, the water breakthrough time occurred on the 48th day after the first well was produced in 2020 with a Water Cut value of 22.46%.

Item Type: Thesis (Other)
Contributors:
Contribution
Contributors
NIDN/NIDK
Thesis advisor
FERIZAL, FIKI HIDAYAT
1024078902
Uncontrolled Keywords: Water Coning, Feature Importance, Random Forest (RF), CMG.
Subjects: T Technology > T Technology (General)
Divisions: > Teknik Perminyakan
Depositing User: Erza Pebriani S.Pd
Date Deposited: 25 Nov 2025 07:09
Last Modified: 25 Nov 2025 07:09
URI: https://repository.uir.ac.id/id/eprint/31931

Actions (login required)

View Item View Item