Thesis (M.Eng.)--Chulalongkorn University, 2009

To investigate a probabilistic approach to estimate possible water injection volume into multilayer depleted reservoirs by accounting for uncertainties in OGIP, rate allocation and injection skin. The estimation of injection volume depends on the knowledge of the original and current reservoir pressures and cumulative production; given that the final injection pressure should not be higher than the original undepleted reservoir pressure. Since the well has commingled completion, the data for the original and current reservoir pressures, cumulative production and injection skin for each individual reservoir are not available. Therefore, a conventional method of estimating water injection volume based on history matching and material balance is not practical for multilayer reservoirs. The probabilistic estimation introduced in this study uses integrated production modeling (IPM) to forecast production and injection profiles. This estimation is validated by comparing water injection history vs. simulated water injection volume where simulated water injection volume is generated from applying probabilistic estimation. Openserver is used to generate a large number of realizations to create a cumulative distribution function for cumulative water injection volume. From this study, the probabilistic estimation can provide a reliable estimate for water injection volume, total OGIP and the end of injection period. Injection skin has little effect on cumulative water injection but has important effect on the amount of time needed to inject water until a reservoir pressure reaches its original reservoir pressure.