In preparation for the SPE-ATW held in Brugge in June 2008 a unique benchmark project to test the use of flooding optimization and history-matching methods was organized in the form of an interactive competition during the months preceding the ATW. In total nine different groups participated and presented their results during the workshop. Prior to the Brugge workshop, early 2008, a 3D synthetic dataset was made available to the participants by TNO. The dataset consisted of 104 upscaled realizations of a 3D geological model, well-log data from wells with fixed positions, the first 10 years of the production history of the field (including measurement errors), inverted time-lapse seismic data in terms of (uncertain) pressures and saturations, and economic parameters for oil and water (price and discount rate). Participants were asked to provide a history match (either a single matched "best" model or a matched ensemble) based on the available data, and an optimal production strategy (without infill drilling) for the next period (10–20 years). Their strategy was tested on the "real field" to obtain additional production data over the 10-year period. Using these production data, the participants updated their reservoir model and revised their optimal production strategy for the final period of production (20–30 year). The final objective of the exercise was to optimize, within a time constraint, the NPV of a waterflooded oilfield having smart wells that can be controlled by an inflow control valve per completed layer. The results of the nine participants are compared to an optimization of the "real field" as performed by TNO. This paper gives an overview of the results obtained from this benchmark study. After the Brugge workshop the participants that were not able to finalize the exercise in time were given an additional two months time. The results of this additional exercise are reported in this paper as well.

No abstract

Within the research framework of the "Integrated System Approach Petroleum Production" (ISAPP) knowledge center of TNO, TU Delft and Shell, the necessity of taking the interaction between dynamic reservoir and dynamic well behavior into account when optimizing a producing asset is investigated. To simulate dynamic phenomena in the well and in the reservoir, a dynamic multiphase well simulation tool (OLGA) and a dynamic multiphase reservoir simulator (MoReS) have been used. Both simulators have been coupled using an explicit scheme. The dynamic well simulator, the dynamic reservoir simulator and the coupled dynamic well-reservoir simulator have been used to simulate a realistic test case which consists of a horizontal well with three inflow sections located in a thin oil rim. A number of scenarios are investigated that play a crucial role during different stages of the well's lifetime: naturally occurring phenomena, e.g. coning, and production dynamics, e.g. shut-in. The results of dynamic well simulations, dynamic reservoir simulations and coupled well-reservoir simulations are presented and an overview is given of the cases where the results of the coupled simulations are significantly more accurate in comparison to stand-alone well or reservoir simulations. For gas coning it is shown that the coupled simulator has much faster pressure transients after gas breakthrough than the dynamic reservoir simulator. Therefore, the coupled well-reservoir simulator should be used to simulate gas breakthrough and to optimize production using gas coning control. For small time scale phenomena, order of less then one day, the well and reservoir transients overlap. Simulations show that the coupled simulator is essential for an accurate prediction of the well-reservoir interaction during these small time scale phenomena. Introduction Production instabilities are undesirable and play a crucial role in the production lifetime and ultimate recovery of any reservoir. These instabilities can arise from or be governed by the interaction between the well and the reservoir.1 Production instabilities can be subdivided into two groups. Firstly, the naturally occurring dynamical phenomena, such as coning and slugging. Secondly, the production dynamical phenomena, such as shut-in, clean-up and gas lift heading. Figure 1 displays the time and spatial scales for different naturally occurring and production dynamical phenomena. The values of the time and spatial scales are indicative and based on experience. There are several phenomena which have a certain amount of overlap. In these areas it is expected that the well dynamics are strongly influenced by the reservoir dynamics and visa versa. Simulations are widely used to predict oil and gas production. The current status of these simulations is to either use a dynamic well model combined with some analytical reservoir model2 or to use a dynamic reservoir model combined with either lift tables or a steady state well model.3, 4 The disadvantage of these models is the fact that they underestimate the pre-mentioned well-reservoir interactions and therefore give non-realistic production forecast in cases where well-reservoir interactions play a crucial role.

No abstract

No abstract