Estimation of reservoir quality from multi-attribute analysis by using a probabilistic neural network: case of Sarvak Formation in an offshore oil field

The objective of this research is to study the relationship between reservoir parameters and seismic attributes in order to determine the reservoir quality, estimate the porosity model, and plan for infill drilling in the oil field under study, based on seismic and well log data. Sarvak Formation in this oil field is characterised, using a combination of the seismic data, porosity logs, and seismic attributes. By estimating the porosity model through various methods including single-attribute regression, multi-attribute regression, and artificial neural network, the probabilistic neural network has shown to represent reliable results. Interpretation of data shows that the estimation of porosity model for the Sarvak Formation indicates high reservoir quality. Besides, comparing the porosity values, the well BS-06 has higher porosity than the well BS-01, which indicates that, the higher reservoir quality at the BS-06 well location. The obtained porosity model, showed that the highest porosity values are found around the seismic CDP location No. 30850. Therefore, this area can be considered for prospective infill drilling in the field development plans.