Neural Network Permeability Prediction of Heterogeneous Gas Sands Using NMR Data

Abstract

Analysis of heterogeneous gas sand reservoirs is one of the most difficult problems. These formations are extremely complex, producing from multiple layers with different permeability that is often enhanced by natural fracturing. Therefore, looking for using new well logging techniques like NMR in individual bases or in combination with conventional open hole logs and building new interpretation methodology is essential to well define and obtain the representative reservoir characterizations. Nuclear magnetic resonance (NMR) logs differ from conventional logs because the NMR measurements provide mainly lithology independent detailed porosity and a good evaluation hydrocarbon potential. NMR logs can be used to determine formation permeability and capillary pressure. This paper concentrates on permeability estimation from NMR logging parameters. There are three models to derive permeability from NMR; Kenyon model, Coates-Timer model and Bulk Gas Magnetic Resonance model. These models have their advantages and limitations related mainly to the nature of reservoir properties. This paper determine permeability from BG model and introduces the artificial network model to derive formation permeability using data from NMR and other open hole logs data. The permeability results of artificial neural model and BG model will be validated by core permeability for studied well.