This paper presents the workflow and learned lessons during the construction of a fully compositional integrated subsurface/surface model for the Santa Barbara and Pirital fields, which are important oil production units located to the east of Venezuela. In this approach, the numerical reservoir simulation models, wells and surface facilities were coupled in order to obtain production profiles considering both changes in the reservoir conditions and surface restrictions, achieving an assertive planning of asset development.
The applied methodology is based on the construction of more than 150 compositional well models, performing sensitivity analysis to define multiphase flow correlations for vertical pipe and chokes. A network model, which comprises more than 900 Km of lines, 3 main flow stations, and 3 separation levels, was also built in compositional mode honoring line sizes, lengths and elevation changes. Two numerical simulation models represent the most reliable characterization of the main reservoirs. Each model was initialized and ran separately, in order to discard internal inconsistencies. Then, the integration was performed considering the sand face on the wells as the coupling point.
The integrated asset modeling allowed predicting the production behavior of the reservoirs taking into account the constraints of the surface facilities, reducing the uncertainty of forecasts and identifying limitations and bottlenecks at surface level. It was also possible to accurately determine the details of the hydrocarbons streams (NGL) at different pressure stages of the network, which reasonably matched with field data (less than 3% of difference). The result is a versatile tool for the integrated asset management, which allows to sensitize all the elements of the production chain and estimate how each one affect the performance of the asset, discarding the division between departments upstream and downstream and establishing a common management strategy for all disciplines.
The novelty of this work is based on the challenge of building fully compositional coupled models considering giants and complex reservoirs with large surface networks. The proposed methodology and learned lessons will certainly serve as reference for similar future works.