Heavy oil is particularly known for the difficulty it presents in obtaining accurate fluid property measurements. The uncertainty of heavy oil fluid property measurements affects the quality of fluid property data which in turn affecting the accuracy of production forecast. In our investigation, we included the impact from certain heavy oil fluid property uncertainties, including dead oil viscosity, live oil viscosity and saturation pressure (or the initial gas-oil ratio, GOR). The reservoir simulations were performed for case studies with an API 9 reservoir fluid and geology similar to those typically found in the Faja region in Venezuela. Two base cases and the underlying process mechanisms were established, one assuming primary production process and one assuming steam-assisted gravity drainage process. We then performed sensitivity studies of the different fluid property variations to determine the effect on the process performances. We found that the process mechanism dictated the magnitude of the sensitivity to different fluid properties. The uncertainties of the recovery prediction obtained from a reservoir simulator confirmed how important it is to select accurate fluid property data on the dynamic simulation prediction.
In upfront reservoir engineering study, the reservoir and fluid properties are very important input data for the dynamic reservoir simulation as they dictate the accuracy of the simulation prediction. The uncertainties of the reservoir properties such as the geologic and petrophysic data are very often given more attention than the uncertainties of the fluid properties such as oil viscosity and initial gas-oil ratio (GOR).
It is well known that there are uncertainties in fluid property data especially for heavy oil as accurate measurements of heavy oil properties are quite difficult to obtain. For example, it has been observed that a large variation in the heavy oil viscosity measured from different laboratories using different measuring techniques. In addition to the inconsistency from the measurements, fluid sampling is another main source contributing to the accuracy of fluid property data. Therefore, it is essential to study the effect of the variation in fluid properties such as oil viscosity and oil saturated pressure (or initial GOR) on the prediction of the heavy oil recovery performance from dynamic reservoir simulation.
The approach implemented in the sensitivity study for the Faja heavy oil is described below: The thermal reservoir simulator, ECLIPSE Thermal, was used to predict the recovery performance of the primary and steam-assisted gravity drainage (SAGD) processes in a typical reservoir in the Faja region in Venezuela. Primary and SAGD processes were selected as the recovery processes of interest for this study and the process mechanism for each of these processes would be investigated. Then, sensitivity studies on different fluid properties including dead oil and live oil viscosities, and oil saturation pressure (or initial GOR) were conducted. The recovery performance from the sensitivity simulation runs were analyzed and compared with that from the base case.
The following reservoir and fluid characterizations were used as input data to initially build a static model.
