Sensitivity-based Parameter Uncertainty Reduction for Structural Reservoir Models.

@inproceedings{SatijaGM2013,
 abstract = { Uncertainty in the geological structure significantly influences the overall uncertainty in a reservoir. However, this structural uncertainty is currently still not widely incorporated in actual reservoir forecasting. Structural uncertainty has many sources of uncertainty. These sources of uncertainty can be parameterized with uncertain parameters. These uncertain parameters can be categorical, discrete or continuous. Integrating all these uncertain parameters effectively requires generating a large set of structural reservoir models. This is computationally very complex.
In our work, instead of considering all the parameters, we only consider those that have a significant impact on the targeted response variable. These can be identified using sensitivity analysis. In considering only significant parameters, we reduce the uncertainty in the less significant parameters. We do this by considering a narrow range of the values of the less significant parameters. This narrowed range is selected so as to not interfere with the effects of interactions with more significant parameters. In a realistic reservoir with uncertainty in the number of faults, fault throws and fault transmissibility; we illustrate how our approach successfully reduces parameter uncertainty for a given water-flooding case. The reduced reservoir model (with reduced parameter uncertainty) is demonstrated to lead to the same water production response uncertainty as Monte Carlo method that spans the entirety of parameter uncertainty. },
 author = { Satija, Addy AND Caers, Jef },
 booktitle = { Proc. 33rd Gocad Meeting, Nancy },
 title = { Sensitivity-based Parameter Uncertainty Reduction for Structural Reservoir Models. },
 year = { 2013 }
}