Design Optimization of a 3-phase(oil,water,gas) Separator

The objectives of the study were to develop an improved separator design and quantify the reduction in material usage for the improved design.
Normally, the production of the reservoir fluids to be handled by the separator changes over the life of the field in terms of the relative amount of oil, water and gas. The world crude oil reservoir would not get any bigger in years to come. Hence the field area available to build a platform over it would be limited. Therefore the surface production facilities and equipment should be designed as compact as possible to serve the purpose. This study aims for optimum design of the separator and hence the platform supporting it by redesigning the 3-phase separator which would be fit-for-purpose only.
The scope of the study were a) enhance understanding on the current design and applications of a 3-phase separator, b) review the criteria used in the designs, and c) determine the possible areas for improvements without sacrificing the performance of the 3-phase separator. Finally a case study was carried out using typical reservoir data of Sumandak Field. Microsoft Excel Solver was used to calculate the required separator sizing parameter throughout the study.
The design optimization process started off with the analysis of the reservoir crude compositions and properties. The mole fractions and K values of the hydrocarbon components in the crude were determined at a specific operating temperature and pressure. Flow stream characterizations (i.e. gas and liquid process parameters) were then determined from the crude compositions and properties. Using the calculated process parameters, the study on the separator sizing was carried out using the appropriate design guidelines. Results of the study had shown that the slenderness ratio, SR, was the most critical factor in order to achieve an optimum size of a 3-phase separator. The study concluded that with optimum design at SR = 4.5, with d = 7.92 ft and Lss = 35.0 ft would reduce the d by 26.9% and Lss by 19.9%, based on Sumandak Field data.