Although a systematic evaluation of the uncertainty associated to water saturation profiles calculated from log data could have a significant impact on reservoir exploitation strategies, a probabilistic interpretation methodology of log data has not been developed yet, due to the complexity of the problem. By contrast, metrology presents a world-wide consensus on techniques to quantify uncertainty. In this paper a probabilistic approach based on the Monte Carlo method is presented, aimed at identifying the potential sources of and quantify uncertainties associated to measurements and interpretation of log data. The methodology was applied to the determination of the water saturation profile for a gas bearing, shaly sand formation. Three water saturation models were adopted to perform the probabilistic analysis. In all cases results showed that calculated water saturation values follow a log-normal distribution when proper probability distributions for all the parameters showing in each water saturation model were taken into account. Furthermore, results provided an estimation of the dispersion of the saturation values caused by random errors. The probabilistic analysis also allowed the identification of the most critical parameters in the determination of porosity and water saturation values and, therefore, they made evident that quality of data, appropriate selection of acquisition techniques, and adoption of appropriate effective porosity values and saturation models are relevant to significantly improving the reliability of water saturation calculations from well log data.

This content is only available via PDF.
You can access this article if you purchase or spend a download.