Uncovering Physics-Based Water Production Patterns Through Well and Reservoir Management Process Modeling

Abstract Reservoir management focuses on understanding flow processes within the drainage volume of wells, where high velocities and unfavorable mobility ratios often challenge conventional flow assumptions. Even subtle geological or structural features can significantly affect fluid flow, yet these influences are difficult to represent accurately using traditional numerical models. This paper demonstrates that Well and Reservoir Management (WRM) Process Modeling offers a practical and effective alternative for simulating such complex environments. WRM Process Modeling introduces a streamlined, physics-driven approach to flow analysis, prioritizing critical geo-parameters and augmenting field data with physically consistent synthetic data to enable the use of machine learning (ML) algorithms to identify and rank the key parameters influencing well behavior. A synthetic case study is presented to illustrate the method. The process begins with expert input from WRM practitioners to identify all parameters potentially influencing a well's WOR signature, along with their plausible ranges. Monte Carlo sampling is then applied to these parameters to generate multiple realizations of conceptual reservoir models. These realizations form a catalog of geological scenarios that associate specific combinations of parameters with corresponding WOR and bottomhole pressure signatures. Comparing actual well data with this catalog enables the identification of the most likely flow setup through pattern recognition powered by ML. Once the well setup is established, targeted and optimized remedial actions can be designed to mitigate excessive water production. WRM Process Modeling facilitates rapid and robust decision-making based on a physics-consistent understanding of well behavior, avoiding unnecessary complexity and reducing human bias. Uncertainty is addressed through ensemble modeling of conceptual reservoir scenarios, eliminating the need for full data assimilation or history matching, and thereby accelerating the evaluation cycle. With increasing environmental scrutiny on produced water volumes, reservoir management has a growing role in ensuring sustainable oil and gas operations. This methodology supports that goal by strengthening the technical foundation for well intervention decisions aimed at minimizing water production while maintaining operational performance.