Implementation of a Deep Learning Technique for Fast Predictions of the Oil-CO2 Minimum Miscibility Pressure in Unconventional Reservoirs

Hao Sun


The 1D CNN model developed in this work offers a significant speedup in predicting the confined MMP compared to the numerical method. Sensitivity analysis of the confined MMP against temperature and a pore size is performed to confirm that the prediction results can reasonably interpret the physical behavior of the confined MMP.