Dimensionless Data-Driven Model for Cuttings Concentration Prediction in Eccentric Annuli: Statistical and Parametric Approach

Insufficient hole cleaning in deviated eccentric annulus yields many drilling problems due to excessive accumulation of cuttings. The majority of the implemented studies in such issues are computational or experimental making them complicated, expensive, and time-consuming. This paper presents a statistical data-driven model to predict cuttings concentration (CA), as a hole cleaning efficiency indicator, under different presumed eccentricity (ε) values of 0.0, 0.4, and 0.8. The model was constructed on the basis of Buckingham Pi theorem utilizing oil field data collected from six offshore deviated wells for developing reliable CA predictive tool. Least linear regression approach was employed for each ε value considering seven dimensionless groups of drilling parameters, rheological parameters, cuttings density, and hole cleaning efficiency indicators, such as transport velocity ratio (VTR), carrying capacity index (CCI), and equivalent circulating density (ECD). The field data were filtered and divided into 75% for training and 25% for validation. Furthermore, a sensitivity study was conducted by parametric differentiation of independent variables regarding the developed CA correlations. Results showed that, VTR, CCI followed by ECD had the highest significance, especially in eccentric annuli. In contrast, cuttings density had the lowest significance in all cases. The proposed method outperformed Duan and Khaled models with root mean square error of 1.45, 1.77, 2.89, and mean absolute percentage error of 34.6, 41.3, 54.5% for ε = 0.0, 0.4, and 0.8, respectively. This study could provide promising and practical methodology in estimating CA and improving deviated drilling efficiency up to 61.8° from vertical.