Relevance Vector Machine: a new approach for permeability prediction of petroleum reservoirs

Permeability is one of the most important rock properties showing the ability of rocks in the conduct of fluids such as oil, water and gas throughthe pore spaces of reservoir. It is one of the most difficult petrophysicalproperties to determine and predict. The conventional methods for permeability determination are core analysis and well test data. Thesemethods are, however, very expensive and time-consuming. One of thecomparatively inexpensive and readily available sources of inferringpermeability is from well logs. In addition, artificial Intelligent (AI) has many applications in the petroleum engineering and permeability prediction over the past decade. The aim of this paper is to introduce a novel machine learning technology called Relevance Vector Machine(RVM) for predicting the permeability of three gas wells in southern Pars field. Comparing the obtained results of the RVM with that of support vector machine (SVM) has shown that RVM is a better and precious method than SVM in prediction of permeability.

کلیدواژه ها:

Relevance vector machine ، Permeability ، Prediction ، Southern Pars Field ، Support vector machine