Rafsanjani, Reyhan and Dahlia, Agus and Ambia, Fajril and Rita, Novia and Husbani, Ayyi (2025) Analisis Komparatif Model Capacitance-Resistance dan Machine Learning untuk Prediksi Produksi CO2-EOR : Studi Kasus Konektivitas Dinamis pada Reservoir Heterogen. Mechatronics, Electrical Power, and Vehicular Technology.

Text skripsi_213210141_watermark.pdf - Published Version Restricted to Registered users only Download (1MB) | Request a copy

Abstract

study evaluates integrated Capacitance-Resistance Models (CRMP and CRMIP) and machine learning algorithms (Random Forest and XGBoost) for predicting oil production performance in CO₂-Enhanced Oil Recovery (EOR) operations at the Volve Field. Reservoir simulation was performed using tNavigator with CO₂ injection at 941 tons/day (35 MMSCF/day) over a 20-year period. The results demonstrate that machine learning models significantly outperform conventional CRM methods, with XGBoost achieving exceptional accuracy (R² = 0.99-1.00, MAPE = 0.44-2.24%) compared to CRMP/CRMIP (R² = 0.55-0.72, MAPE = 16-23%). The CO₂ injection scenario substantially enhanced oil recovery, achieving cumulative production of 15.73 MMSTB (RF 20.45%) compared to Waterflooding, which yielded 9.38 MMSTB (RF 12.19%), representing an incremental recovery of 6.35 MMSTB. Interwell connectivity analysis revealed heterogeneous reservoir responses with time constants ranging from 916 to 927 days. The integration of physics-based models (CRM) with non-linear machine learning algorithms has been shown to significantly improve prediction accuracy and provide a more comprehensive understanding of reservoir dynamics, thereby supporting the optimization of CCUS implementation in heterogeneous reservoir systems.

Actions (login required)

View Item