Integrasi Metode Eaton Dan Machine Learning Untuk Prediksi Pore Pressure, In-situ Stress, Dan Optimasi Safe Mud Window Di Lapangan "N", Cekungan Sumatera Utara

Khairunnisa, Nabila Thufail (2026) Integrasi Metode Eaton Dan Machine Learning Untuk Prediksi Pore Pressure, In-situ Stress, Dan Optimasi Safe Mud Window Di Lapangan "N", Cekungan Sumatera Utara. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5017221026-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (11MB) | Request a copy

Abstract

Prediksi tekanan pori menggunakan metode empiris sangat bergantung pada ketersediaan data kalibrasi dan kesesuaiannya dengan kondisi geologi daerah penelitian. Penelitian ini mengintegrasikan metode Eaton dan machine learning untuk memprediksi tekanan pori, mengestimasi in-situ stress, dan menentukan safe mud window pada empat sumur di Lapangan "N", Cekungan Sumatera Utara. Metode Eaton menggunakan eksponen empiris hasil kalibrasi regional, sedangkan model machine learning dibangun dari log DT, RHOB, GR, dan resistivitas menggunakan algoritma Random Forest, XGBoost, dan Neural Network. Overpressure teridentifikasi pada Sumur NBL 26 dan NBL 29 mulai kedalaman sekitar 800 m dengan magnitude masing-masing 311,6 psi dan 112,1 psi di atas tekanan hidrostatik, sedangkan Sumur NBL 60 dan NBL 21 tetap menunjukkan kondisi hidrostatik normal. Shoulder effect, pembalikan tren log DT, dan crossplot Dutta & Katahara mengindikasikan mekanisme lateral drainage sebagai penyebab minor unloading. Seluruh sumur berada pada rezim sesar normal (Sv > 〖Sh〗_max> 〖Sh〗_min) dengan rentang safe mud window 1,03–1,33 SG. Random Forest memberikan performa terbaik (R^2 ≥ 0,92) dan menghasilkan estimasi berat lumpur dengan deviasi 0,01–0,06 SG terhadap metode Eaton. Hasil tersebut menunjukkan bahwa integrasi metode Eaton dan machine learning meningkatkan keandalan prediksi tekanan pori sekaligus mendukung evaluasi geomekanika dan penentuan safe mud window pada formasi shalysand.
===============================================================================================================================
Pore pressure prediction using empirical methods strongly depends on the availability of calibration data and their compatibility with the geological conditions of the study area. This study integrates Eaton's method and machine learning to predict pore pressure, estimate in-situ stress, and determine the safe mud window for four wells in Field "N", North Sumatra Basin. Eaton's method was applied using a regionally calibrated empirical exponent, while the machine learning models were developed from DT, RHOB, GR, and resistivity logs using the Random Forest, XGBoost, and Neural Network algorithms. Overpressure was identified in Wells NBL 26 and NBL 29 from a depth of approximately 800 m, with pressure magnitudes of 311,6 psi and 112,1 psi above the hydrostatic pressure, respectively, whereas Wells NBL 60 and NBL 21 remained under normal hydrostatic conditions. The shoulder effect, reversal of the DT log trend, and Dutta & Katahara crossplot analysis indicate that lateral drainage is the main mechanism responsible for the observed minor unloading. All wells are characterized by a normal faulting stress regime (Sv > 〖Sh〗_max> 〖Sh〗_min), with a safe mud window ranging from 1.03 to 1.33 SG. Among the evaluated models, Random Forest achieved the best performance (R^2 ≥ 0,92) and produced mud weight estimates that differed by only 0.01–0.06 SG from those obtained using Eaton's method. These results demonstrate that integrating Eaton's method with machine learning provides reliable pore pressure predictions while supporting geomechanical

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Metode Eaton, Machine Learning, Safe Mud Window, Tekanan Pori, Cekungan Sumatera Utara Eaton’s Method, Machine Learning, Safe Mud Window, Pore Pressure, North Sumatera Basin
Subjects:	Q Science > QE Geology > QE601 Geology, Structural T Technology > TJ Mechanical engineering and machinery > TJ1260 Drilling and boring.
Divisions:	Faculty of Civil Engineering and Planning > Geophysics Engineering > 33201-(S1) Undergraduate Thesis
Depositing User:	Nabila Thufail Khairunnisa
Date Deposited:	15 Jul 2026 00:55
Last Modified:	15 Jul 2026 00:55
URI:	http://repository.its.ac.id/id/eprint/134687

Actions (login required)

View Item