Classification of reservoir sand-facies distribution using multiattribute Probabilistic Neural Network transform in “Bigola” field, Niger Delta, Nigeria

Abstract

This study estimated reservoir properties, classified the reservoir sand-facies distribution and identified  potential hydrocarbon pay zones. This was with a view to optimizing placements of wells in “Bigola” Field.  Conventional interpretation of seismic and well datasets was carried out to provide the sub-surface  structures  and general stratigraphy of the study area. The 3-D seismic data volume along with well logs were analysed using multiattribute Probabilistic Neural Network (PNN) transform to generate target reservoir  properties (V_Shale , Porosity shale and Resistivity) and consequently horizon based maps. The  generated  structure maps of target horizons TMTS4 and TMTS5 showed closures at the north central and north eastern  parts. The generated reservoir property maps showed the distribution of reservoir sand-facies in the wells and  inter-well regions. Average porosity values ranged from 22-30%, V_Shale values ranged between 8 and 12 %  and resistivity values ranged between 112 and 199ohm-m for hydrocarbon reservoir sands. The study  concluded that the reservoir sand-facies of the north eastern and north central parts of the study area were  more indicative of hydrocarbon presence than those of other parts.

Keywords: Porosity, Resistivity, V_shale , Multi-Attribute, Neural Networks, Well and Niger Delta.