A linear unconstrained model predictive control (MPC) scheme has been designed to optimize the operation of dual frequency electrostatic dehydrators in the Colombian Rubiales and Quifa oil fields. This multi-objective controller optimizes operation by maximizing the amount of daily oil production, while maintaining the base sediment and water (BS&W) specification at the exit of the electrostatic dehydrator at or below 1.0%.
The designed controller uses models that describe the dynamics of the dehydrator system to meet the objectives described above. The model of the entire system consists of two empirical sub-models: one for the crude-steam heat exchangers upstream of the electrotreater and one for the treater. Each of these models is used to design an MPC controller for the corresponding subsystem. The electrotreater MPC works as the master controller, dictating crude temperature setpoints to the exchanger MPC. The exchanger MPC then adjusts the vapor flow valve opening to obtain the optimum temperature for the outlet BS&W specification. The electrotreater controller also adjusts the treater inlet/outlet flows and the transformer voltage setpoints to meet the desired objectives.
This control scheme allows for 1) less variability in the output BS&W, 2) maximized daily crude oil production, and 3) tighter control. The decrease in BS&W variability helps ensure that the product quality control and increases the rate of oil production by minimizing the need for further oil dehydration steps. In terms of OPEX, financial benefits are obtained by optimizing the dehydrator operation due to the reduction in the amount of costly emulsion breaking chemicals used in this and other stages of the dehydration process to ensure an effluent BS&W specification of 1.0%. In terms of CAPEX, this control scheme minimizes the need for additional infrastructure necessary to further dehydrate the oil produced to pipeline specifications.
This is the first time such a control scheme is known to have been developed for oil dehydration facilities. The approach proposed in this paper makes further implementation of advanced control systems an intriguing and promising venture that includes benefits such as increased oil production and decreased operating and capital costs.