Evolving simple-to-use method to determine water–oil relative permeability in petroleum reservoirs

Abstract

In the current research, a new approach constructed based on artificial intelligence concept is introduced to determine water/oil relative permeability at various conditions. To attain an effective tool, various artificial intelligence approaches such as artificial neural network (ANN), hybrid of genetic algorithm and particle swarm optimization (HGAPSO) are examined. Intrinsic potential of feed-forward artificial neural network (ANN) optimized by different optimization algorithms are composed to estimate water/oil relative permeability. The optimization methods such as genetic algorithm, particle swarm optimization and hybrid approach of them are implemented to obtain optimal connection weights involved in the developed smart technique. The constructed intelligent models are evaluated by utilizing extensive experimental data reported in open literature. Results obtained from the proposed intelligent tools were compared with the corresponding experimental relative permeability data. The average absolute deviation between the model predictions and the relevant experimental data was found to be less than 0.1% for hybrid genetic algorithm and particle swarm optimization technique. It is expected that implication of HGAPSO-ANN in relative permeability of water/oil estimation leads to more reliable water/oil relative permeability predictions, resulting in design of more comprehensive simulation and further plans for reservoir production and management.

Keywords

• Crude oil
• Water
• Optimization
• Relative permeability
• Neural network
• Porous media