Uncertainty Estimation in the Volume of Tight Sandstone Gas Reservoir: A Case Study From Sulige Field, Ordos Basin, China
In tight gas production projects, many investment and development plans are based on the volume estimate. Generally, numerous types of uncertainties exit in the volume since the input variables always contain uncertainties to some degree that propagate into reserve estimates. Making decisions in a reservoir management requires a method which is able to accurately quantifying the uncertainties and risks in the volume. This paper carries out a method for uncertainty estimation in the volume of tight sandstone gas reservoir, which has significant meaning for providing the foundation for business decisions on filed development strategy.
The Su6 experimental area which is characterized by low permeability and strong heterogeneity is one of the most prolific gas areas of the Sulige field Ordos Basin, China. So the reserve parameters are difficult to correctly acquire when estimating complex reservoir's reserves in terms of traditional volume method. Stochastic modeling algorithms could be used to provide a measure of the uncertainties inherent to each input data set used to build 3D static reservoir models by means of multiple realizations involving facies and petrophysical parameters.
The workflow is based on the following steps:(1)Object-based modeling algorithm was used to integrate well-log fluvial facies, orientation of the mainstream line, object dimension in building the 3D facies model.(2)Facies model was used to the constrain the spatial distribution of petrophysical parameters, trended by impedance data derived from logging constrained inversion as the second variable, in building petrophysical parameters using SGS algorithm.(3)Through sensitivity analysis, we found that the highest ranked contributors to the uncertainties in the volume were the object dimension, seeds number, range and sill of variogram used to calculate petrophysical parameters in possible-reservoir facies.(4)Three possible volumetric values(P90.50,P10) were provided by the final volumes distribution of one-hundred realizations. Finally, the known volume lies within the estimated proven to probable range.
AAPG Datapages/Search and Discovery Article #90260 © 2016 AAPG/SEG International Conference & Exhibition, Cancun, Mexico, September 6-9, 2016