A Novel Workflow for Estimating Anisotropic Mechanical and Stress Profile in Unconventional Reservoirs: Utilizing Stoneley Wave Inversion in Data-Rich Wells and Integrating Machine Learning in Data-Limited Wells

This paper introduces an innovative integrated workflow for estimating anisotropic mechanical and stress profiles from acoustic log and geomechanics core measurements, obviating the need for direct model-based full frequency Stoneley waveform inversion. For wells with Stoneley wave data, the workflow leverages synthetic isotropic and anisotropic Stoneley wave slownesses and their correlation with Thomsen parameters to facilitate rapid and stable estimation of anisotropic elastic moduli. For wells lacking Stoneley wave data, the workflow employs machine learning and clustering techniques to effectively estimate Thomsen parameters. This methodology has been successfully applied to the Wolfcamp Shale Formation of the Permian Basin, demonstrating its efficacy in predicting anisotropic elastic and mechanical properties.