Multiple‐model sliding mode observer for an artificial gas lift system

Abstract

Artificial gas lift is widely used in the oil industry to enhance oil recovery. Active feedback control of this process would lead to its stabilization, that in some operating modes may otherwise be unstable, and to the increase of oil production. However, the control strategies are normally constrained to the use of the surface-measured process variables. The use of down-hole measurements would improve the performance of the control system but is technically hardly feasible due to the necessity of placing instruments in harsh conditions. The use of state observation might be a feasible alternative to the down-hole measurements. Recent development of a new accurate model of the artificial gas lift process enables us to increase accuracy of observation due to the account of pressure and density distribution along the well depth. Besides, a new approach to a nonlinear model treatment proposed in the present paper leads to a high computational efficiency of the observer for the gas lift. The paper presents an approach to the design of a novel sliding mode observer for the gas lift process.The observer uses multiple linearized models representing deviations from a set of equilibrium points. These models are then incorporated to produce estimates for the gas lift process variables. The approach is supported by simulations.

Observability of probabilistic Boolean multiplex networks

“…The optimal control of Boolean networks was considered in [30,31]. In addition, the design of sliding mode observers for different systems was studied in [32–36].…”