Abstract
The Ordos Basin is currently one of China's fastest-growing regions for the large-scale deployment of carbon capture, utilization, and storage (CCUS) technology. However, the determination of the minimum miscibility pressure (MMP), a critical parameter for CO(2) flooding reservoir screening and engineering design, has long been constrained by the high costs and long durations associated with experimental and numerical methods. There is an urgent need to develop an efficient and accurate model for predicting MMP rapidly. In this study, a series of slim-tube experiments to refine the MMP data set under representative reservoir conditions in the Ordos Basin were conducted. Subsequently, key parameters influencing MMP were identified through a comprehensive literature review and the Pearson correlation analysis. Based on leave-one-out cross-validation (LOOCV), an optimal MMP prediction model was developed. Its coefficient of determination (R (2)) is 0.84, with a root-mean-square error (RMSE) of 0.75 MPa, mean absolute error (MAE) at 0.48 MPa, and mean absolute percentage error (MAPE) at 2.53%. Compared with existing MMP prediction models, the proposed model significantly improves both prediction accuracy and convenience for MMP estimation in the Ordos Basin reservoirs. It provides a reliable tool that enhances reservoir screening efficiency for CO(2) flooding in basin reservoirs by enabling efficient and economical MMP determination.