| dc.description.abstract | A newly discovered reservoir needs to undergo a development study phase. Many development scenarios have to be evaluated to determine the scenario, which yields the maximum profit. A research has been carried out by a PhD student under the research center Subsea production and processing (SUBPRO) to develop an automated decision support methodology. The methodology includes integrating the production and economic elements of the field, mathematical optimization, and uncertainty analysis to decide upon the best strategy. \citet{diana2020} and \citet{angga2019automated} had previously implemented this methodology on synthetic reservoirs consisting of simplified reservoir models having identical well performances. This thesis aims to develop the automated decision support methodology by taking into account the separate performances of each well in the reservoirs: oil production rates, Gas Oil Ratio (GOR), and Water Cut (WC) profiles.
There are three main stages of this work. The first stage was the problem formulation of Net Present Value (NPV) optimization as a Mixed Integer Linear Programming (MILP). The decision variables of the optimization were the production potential and drilling schedule. The optimization used production potential as the proxy model of the entire production system and cost proxy model to estimate the development cost. The cost proxy model was modeled as a linear function based on the available data. Since every well had distinct performance, the potentials oil rates were modeled as a non-linear function of cumulative oil productions and well status: active or inactive. Multidimensional Piecewise Linear (PWL) approximation was implemented to represent the non-linear behavior of the production potential imposing the Special Order Set (SOS)1/SOS2 constraints. The second stage was to develop a method to quantify the uncertainty of GOR and WC of the reservoir based on the individual well GOR and WC. Several optimization routines were formulated and performed to generate the new cumulative water and gas production curves by adjusting the well production schedule. Afterward, the NPV optimization routine was performed by employing these new curves to determine which curves have the highest and lowest NPV. The last stage is the uncertainty analysis study to quantify the uncertainties of the optimization result. Five parameters were considered: Initial Oil In Place (IOIP), water-gas profiles, cost figures, production potential, and oil price. The technique used in the study was a probability tree.
Based on the NPV optimization, changing the well status and drilling schedule throughout the production years had increased the NPV compared to the optimization with a fixed drilling schedule. Six extreme water and gas profiles were successfully generated to capture all the possible highest or lowest production of water and gas. From the NPV optimization, considering high production, both water and gas productions had lowered the NPV significantly from the base case. From the uncertainty analysis, the effect of the uncertainty was quantified in NPV, the optimal oil rates, and the optimal number of well. | |
