Achievement on Automatically Simulation Model Calibration to Provide Fit-For Basin 4D Fea Dynamic Drilling Analysis

Abstract

After years of using 4D simulation for drilling product design and service event root cause analysis, there is a growing demand for an automated workflow that can provide a standardized and efficient model calibration process. This workflow aims to reduce the dependency on manual adjustments, improve the accuracy of simulation inputs, establish a strong correlation between simulation results and field measurements, and enhance drilling performance prediction. To achieve these goals, the auto calibration workflow incorporates various optimization techniques based on the physics and attributes of drilling simulation engines. For instance, grid search is employed to optimize discrete formation descriptions using a library of test rock files, while gradient descent is utilized for the optimization of continuous model parameters. Furthermore, the optimization of certain model parameters is simplified by leveraging special drill state data, thereby reducing the complexity of the overall optimization process. These approaches have significantly improved the performance of optimization, even when dealing with computationally intensive simulations. This paper provides a comprehensive review of the newly developed automated calibration workflow, including the algorithm development and field results. It also discusses the lessons learned from this work and outlines future directions for research and development.