Artificial Intelligence Models for Flowing Bottomhole Pressure Estimation: State-of-the-Art and Proposed Future Research Directions

Flowing bottom hole pressure (FBHP) is an important parameter during evaluation of conventional and unconventional oil and gas resources and is mainly used for production optimization, calculation of productivity index and assessment of well performance. FBHP in an oil well is a multifaceted parameter that has a number of factors affecting it. It is characterized by high stochasticity, non-linearity and non-stationarity. Traditionally, production engineers rely on physics-based models and empirical correlations calibrated with measurements from offset wells to calculate FBHP. However, in recent times, there has been a significant shift towards the use of cutting-edge artificial intelligence (AI) algorithms. The present study is designed to provide a historical account of past and present models developed with AI algorithms for estimating FBHP. To achieve this, a deep bibliographic survey was conducted using various peer reviewed journals and relevant oil and gas conference papers. The results of the review have been presented in tables to avoid ambiguity. To make the review novel, the merits and demerits of each of the AI models for FBHP prediction are highlighted and discussed in detail. In this direction 54 models were isolated from the literature. The findings indicate that artificial neural network is the preferred algorithm by several researchers. However, the transparency and interpretability issues associated with the neural network algorithm has propelled researchers to explore the possibility of deploying physics informed machine learning techniques to model FBHP. This review would serve as a valuable reference for production engineers seeking information on AI models for FBHP estimation.