A New Comprehensive, Mechanistic Model for Underbalanced Drilling Improves Wellbore Pressure Predictions

Abstract

Abstract It is generally accepted that the success of underbalanced drilling (UBD) operations is dependent on maintaining the wellbore pressure between boundaries determined by the formation pressure, wellbore stability, and the flow capacity of the surface equipment. Therefore, the ability to accurately predict wellbore pressure is critically important for both designing the UBD operation and predicting the effect of changes in the actual operation. Most of the pressure prediction approaches used in current practice for UBD are based on empirical correlations, which frequently fail to accurately predict the wellbore pressure. Consequently, the current trend is toward increasing use of prediction methods based on phenomenological or mechanistic models. This paper presents an improved, comprehensive, mechanistic model for pressure predictions throughout a well during UBD operations. The comprehensive model is composed of a set of state-of-the-art mechanistic steady-state models for predicting flow patterns and calculating pressure and two-phase flow parameters in bubble, dispersed bubble, and slug flow. In contrast to other mechanistic methods developed for UBD operations, the present model takes into account the entire flowpath including downward two-phase flow through the drill string, two-phase flow through the bit nozzles, and upward two-phase flow through the annulus. Additionally, more rigorous, analytical modifications to the previous mechanistic models for UBD give improved wellbore pressure predictions for steady state flow conditions. The results of using the new, comprehensive model were validated against two real wellbore configurations with different flow areas. Field data from a Mexican well, drilled with the simultaneous injection of nitrogen and a non-Newtonian fluid and full-scale experimental data from the literature validate the improved model predictions. Additionally, a comparison of the model results against two commercial UBD simulators, which rely on empirical correlations, confirm the expectation that mechanistic models perform better in predicting two phase flow parameters in UBD operations.