This paper describes the challenge of managing and optimizing the production of a large land based oilfield with hundreds of ESP-boosted wells arranged in widely distributed well clusters which production converges to major trunklines traversing the field. The Rubiales field, located in the eastern plains of Colombia has challenging features, characteristics and layout that demand effective model-based production optimization and control. The field's gathering system feeds the commingled production to two central field processing plants.
The flow of the numerous wells and streams of the network are interdependent as there are no gas separation facilities at the clusters or at any other location in the network between the wellhead sources and the entry to the processing plants. This creates an interdependency of well streams. Thus, any production change at a single well affects the pressure and rate of all other wells in the network and consequently the total field production. The water rate from each individual producing well strongly depends on the drawdown and the stage of depletion of that particular well, and how it is controlled by varying the speed of its ESP. High water cuts of most producing wells and the constraints on water treatment and disposal at the field level dictates a need for frequent readjustment of individual well ESP speed.
Adjusting ESP speeds to maximize the field oil production, subject to field water production constraints, must also take into account a variety of additional constraints related to system limitations, ESP performance, power consumption, production operations and reservoir recovery strategy. One cannot rely solely on operational intuition and empirical field practice for individual ESP control. Rather, a model-based optimization system has been implemented, taking into account all field and well constraints. The implemented system is robust, fast and easy to tune. Furthermore, inflow of heavy and viscous Rubiales oil into the horizontal wellbores is driven by a strong and active aquifer in a highly heterogeneous and permeable reservoir. This results in rapid changes of produced water cut in response to small changes in drawdown, demanding effective tuning of a predictable well inflow function for the purpose of optimization.
This paper describes the model-based optimization system employed in the Rubiales field. The system is customized to the large scale and special features of Rubiales, such as the demanding production performance of its wells, the constraints of facilities, and the objective to maximize profit given by production revenues less OPEX.