Improved quantification of fit-for-purpose saturation exponents

Saturation exponent is a key input parameter to those petrophysical algorithms that are used for the evaluation of water saturation S-w. A longstanding problem is that different interpretative algorithms often deliver different predictions of S-w in the same reservoir rock, especially where the reservoir shows shale effects. This problem is overcome by determining a fit-for-purpose saturation exponent, which is specific not only to a given core sample but also to a particular interpretative algorithm for the evaluation of S-w. By using fit-for-purpose saturation exponents and published type algorithms for the petrophysical evaluation of S-w, it is shown that the discrepancies between predicted water saturations obtained using different interpretative equations can be confined to a highly restricted range. Through this process, the sensitivity associated with the choice of a petrophysical model for the evaluation of S-w is demonstrably contained. This concept is extended to take account of the degree of maturity of the core database. Four expanding data scenarios are enacted to derive data-specific values of the fit-for-purpose saturation exponent using the best-performing type algorithms. The first scenario assumes no core data beyond measurements of S-w and resistivity index using simulated formation water; results are erratic. Scenario 2 draws additionally on a measurement of porosity for each desaturated core plug; it sets the intrinsic porosity exponent m* = 2. Values of S-w predicted using saturation exponents established through Scenario 2 compare closely with those evaluations generated through more-data-intensive scenarios. Therefore, Scenario 2 is proposed as the most efficient way of deriving a fit-for-purpose saturation exponent and thence of obtaining a meaningful evaluation of S-w. A specific type algorithm has been identified for doing this most effectively. The method has been validated against benchmarks from a public database. The predictive performance of the preferred type algorithm for evaluating S-w through Scenario 2 matches even a fully comprehensive shaly-sand approach with all its additional characterizing data. Therefore, the use of fit-for-purpose saturation exponents allows data needs to be optimized alongside the further containment of uncertainty in integrated reservoir description. Thus, both costs and risk can be reduced simultaneously.