Digital Twin: New Back Allocation Algorithm Based on the Maximum Likelihood Principle

Abstract

Abstract The rapid advancement of digital technologies in the oil and gas sector has led to the development of innovative solutions for optimizing production and enhancing asset management. One such solution is the Digital Twin, a dynamic, real-time digital representation of physical assets. This paper introduces a novel back allocation algorithm for gas and gas condensate fields based on the Maximum Likelihood Estimation (MLE) principle, integrated within the Digital Twin framework. The core thesis of this paper is the application of the MLE principle to the back allocation of fluids to wells and reservoirs, solving this task with the highest possible likelihood. The MLE-based approach allows for precise determination of fluid origins, pinpointing the specific wells and even the intervals within those wells, including the detailed component content. For the MLE algorithm to function correctly, robust fluid characterization is essential. This involves comprehensive compositional grading and the use of Equations of State (EOS) to model fluid properties accurately. Understanding the compositional distribution from the output to the input, along with knowledge of how composition changes with depth and phase behavior at every node in the system, is crucial for achieving accurate back allocation using MLE. Key aspects of the Digital Twin and MLE-based back allocation algorithm include: Accurate Fluid Allocation: Determining the most probable production rates and fluid origins by adjusting parameters such as flow rates, fluid compositions, and equipment dimensions to minimize deviations. Compositional Grading: Utilizing detailed compositional data to understand how fluid properties change with depth, which is vital for accurate modeling and allocation. Equation of State (EOS) Models: Employing EOS models to accurately describe fluid phase behavior under various conditions, providing the necessary data for the MLE algorithm to function effectively. This integration of MLE within the Digital Twin environment enhances the accuracy of production allocation, leading to significant improvements in operational efficiency and economic performance. Case studies from gas and gas condensate fields demonstrate the algorithm's capability to optimize production allocation, reduce operational risks, and enhance overall asset management. This paper will detail the technical aspects of the MLE-based algorithm, the importance of fluid characterization and EOS models, and the resulting benefits in terms of production optimization and economic gains. The findings suggest that adopting this advanced back allocation approach can lead to substantial improvements in field development strategies and operational efficiency.