Modeling and Interpretation of Transient Near-Wellbore Phenomena for Improved Well Control and Formation Evaluation

Abstract

Transient near-wellbore phenomena strongly influence measured pressures, temperatures, and flow rates in wells during drilling, completion, and production operations. These short-time responses encode information about local formation properties, near-wellbore damage or stimulation, and dynamic interactions between the wellbore and the surrounding porous medium. At the same time, they govern early-time well control behaviour, including detection of influxes, losses, and unstable flow regimes. Modeling and interpretation of these phenomena require rigorous treatment of multiphase flow, thermal and mechanical coupling, and the finite dimensions and storage of the wellbore. This work develops a mathematical and computational framework for describing transient near-wellbore processes and provides strategies for extracting formation and operational parameters from high-frequency measurements. The framework combines partial differential equation models for porous media flow and energy transport with reduced-order descriptions of the wellbore and its control logic. Numerical schemes tailored to the strong gradients and short time scales near the wellbore are discussed alongside data-driven surrogates that approximate the full physics. The analysis explores how different sources of non-ideal behaviour, such as skin, partial penetration, non-Darcy effects, and geomechanical coupling, distort the early-time signatures used in conventional interpretation. Machine-learning-based inversion and statistical analysis are used to map transient responses to uncertain formation properties in a probabilistic manner. The resulting methodology is intended to inform well control strategies and formation evaluation workflows that explicitly account for near-wellbore transients, without overstating their predictive capability, and to clarify the limits of interpretability under realistic levels of noise, operational constraints, and model uncertainty.