Advancing Porosity Prediction: Integrating Seismic Inversion and Machine Learning for Enhanced Subsurface Characterization

Abstract The accurate prediction of porosity is complicated due to the intricate and varied nature of geological formations. This research explores the integration of seismic inversion with machine learning to enhance porosity prediction, utilizing chi angle-based and Acoustic Impedance (AI) cube-based techniques. A comprehensive approach involving data collection, preprocessing, and seismic inversion is used to extract necessary information. Advanced machine learning models are then developed and rigorously trained on diverse datasets, allowing for porosity estimation using both techniques. Key findings include: The chi angle-based method demonstrates high accuracy with a mean absolute error (MAE) of 0.05 and a mean square error (MSE) of 0.002, effectively capturing subtle variations in porosity.The AI cube-based method provides dependable and consistent predictions across various geological settings, with an MAE of 0.07 and an MSE of 0.003. A thorough comparative analysis reveals the complementary nature of these techniques, illustrating how their integration enhances porosity characterization. This innovative approach provides valuable insights for subsurface exploration, significantly improving the accuracy of porosity predictions and supporting more informed decision-making in exploration and production strategies.