Application of Artificial Intelligence and Machine Learning Technique for Nonlinear Flow Modelling Applicable in Petroleum Exploration and in Porous Media Flow

Understanding and modelling nonlinear filtration through porous media is especially important in the energy sector since it directly controls the discharge in petroleum and gas wells. The efficiency of hydraulic fracturing process completely depends on our understanding of the nonlinear filtration process in porous media. However, the process is extremely complex, and therefore, theoretical and empirical understanding of the subject is very much ambiguous. This is because there are a lot of factors which are involved in the process. Efficient modelling of such process requires a clear understanding about the parameters which are important for the model. Realising the potential of machine learning and artificial intelligence in the field, we have attempted to arrange the parameters according to their importance using a random forest-based feature selection technique. Understanding the effect of these characteristics on the modelling of the porous media flow in the post-laminar regime may be of great use to academics, planners, and designers especially associated with energy sector.