Geometric modelling of the volume of investigation of well logs for thin-bed characterization

Masoudi Pedram (1, 2) , Hossein Memarian (1) , Tahar Aifa (2) , Behzad Tokhmechi (3)

1 School of Mining Engineering, University College of Engineering
2 GR - Géosciences Rennes
3 Department of Petroleum Engineering

The fuzzy membership function is used to model the volume of investigation of well logs geometrically. We discuss the fact that the spacing of a transmitter–receiver is not a precise parameter for addressing the vertical resolution of well logs. Instead, the vertical resolution of membership function (VRmf) is developed and estimated by variography analysis. In the five studied wells, the vertical resolution of gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI) and sonic (DT) logs are estimated to be 61, 76, 76 and 61 cm, respectively. The simplest membership function for describing the volume of investigation of the GR, RHOB and NPHI is the triangle. For DT it is a complex shape. Being compatible with volumetric records in the well logs, the volumetric Nyquist frequency is introduced while considering the VRmf. Based on triangular membership functions, a thin-bed geometric simulator is designed. Regression models, i.e. deconvolution relations, are developed between the real thickness and the real petrophysical variation of a thin bed as outputs, and the same log-derived parameters are used as inputs. The shoulder-bed effect in GR, RHOB and NPHI is reduced by two to three times due to the mean squared error (MSE). To check the applicability of the deconvolution relations for the real data, ten thin beds are chosen within a well at the interval of the Sarvak Formation. In all the observations, the shoulder-bed effect is reduced after deconvolution. The thickness of the thin beds is estimated with a standard deviation of 4.4 cm, which is a precise value. The method is applied to the cored interval of the Sarvak Formation in a nearby well to characterize a porous carbonate thin bed sandwiched between dense carbonates. The estimated thin-bed thickness (13 ± 7.5 cm) is close to the in situ thin-bed thickness (<25 cm). Furthermore, the NPHI (total porosity) of the thin bed is estimated to be 11.7%, which is compatible with the core porosity (effective porosity), which is 8%, since the effective porosity should be less than the total porosity.

Keywords: thin-bed characterization; well log deconvolution; shoulder-bed effect; thin-bed simulator; volumetric Nyquist frequency

Domains: Applied geology

Masoudi Pedram, Hossein Memarian, Tahar Aifa, Behzad Tokhmechi. Geometric modelling of the volume of investigation of well logs for thin-bed characterization. Journal of Geophysics and Engineering, 2017, 14 (2), pp.426-444. ⟨10.1088/1742-2140/aa59d4⟩. ⟨insu-01489012⟩