Real-Time Completion Cost Optimization Using Model Predictive Control

Abstract

Abstract Completion cost structure varies from company to company and is primarily driven by hydraulic horsepower consumption, represented by the average wellhead pressure and slurry rate, given a total amount of proppant and total volume of fluids. In this paper, we use model predictive control (MPC) with system identification to optimize completion cost subject to real-time operational constraints. This can help completion engineers adjust the pumping schedule to optimize completion costs on the fly. We examined a typical cost structure of hydraulic fracturing service companies and illustrated how well completion costs could have been reduced if we adjusted their pumping schedules. We first developed a data-driven model with system identification that characterized how wellhead pressure responded to changes in slurry rate, proppant concentration, and friction reducer (FR) concentration. A linear transfer function model was shown to give acceptable accuracies if the model was updated every 5 to 10 minutes as real-time data were received. MPC was then used to design an optimal schedule subject to a wellhead pressure setpoint. In multiple scenarios with different kinds of constraints, the target wellhead pressure can be achieved by adjusting friction reducer, slurry rate, and proppant concentration. One can deploy such a model at the edge or in a cloud environment to offer suggestions to engineers in the field or in the office. Completion cost savings can potentially be achieved by following the suggested changes.