Skip directly to content

Nonlinear Model Predictive Control of an Oil Well with Echo State Networks

on Wed, 02/21/2018 - 16:44
TitleNonlinear Model Predictive Control of an Oil Well with Echo State Networks
Publication TypeJournal Article
Year of Publication2018
AuthorsJordanou JPanaioti, Camponogara E, Antonelo EA, de Aguiar MAurelio S
JournalIFAC-PapersOnLine (3rd IFAC Workshop on Automatic Control in Offshore Oil and Gas Production OOGP 2018)
Pagination13 - 18
KeywordsEcho State Networks, Gas, Model Predictive Control, Oil, System Identification

In oil production platforms, processes are nonlinear and prone to modeling errors, as the flowregime and components are not entirely known and can bring about structural uncertainties,making designing predictive control algorithms for this type of system a challenge. In thiswork, an efficient data-driven framework for Model Predictive Control (MPC) using Echo StateNetworks (ESN) as prediction model is proposed. Differently from previous work, the ESN model for MPC is only linearized partially: while the free response of the system is kept fullynonlinear, only the forced response is linearized. This MPC framework is known in the literatureas the Practical Nonlinear Model Predictive Controller (PNMPC). In this work, by using theanalytically computed gradient from the ESN model, no finite difference method to compute derivatives is needed as in PNMPC. The proposed method, called PNMPC-ESN, is applied tocontrol a simplified model of a gas lifted oil well, managing to successfully control the plant,obeying the established constraints while maintaining setpoint tracking.