Comparison of Inlet Observers for a De-Oiling Gravity Separator

Current offshore Oil & Gas installations in the Danish North Sea are unable to consistently comply with the oil-in-water concentration discharge regulations. An increasing water-cut, stricter governmental regulations, and fluctuating flow rates can be problematic for the offshore produced water treatment facilities to deal with, especially since the systems were designed decades ago. In this study, different extended Kalman filter observers are designed to estimate the unknown inlet conditions of a gravity separator and compared in simulation. These observers are evaluated based on their computational performance, estimation precision, and convergence time from initial state estimate. The results show that it is possible to get a good estimate of the gravity separator’s unknown inlet conditions, which is necessary for efficiency-based model predictive control.