This work presents the methodology and results of the optimization of the internals (Inlet distributor, oil and water collectors) of a 20,000 BPD (0.037 m3/s) gun-barrel tank starting from an existing design. Computational fluid dynamics (CFD) was applied to simulate and evaluate the performance of various internal configurations. These simulations were performed to determine the best configuration to ensure efficient separation of the oil-water mixture and oil with a low BSW content < 2% at the outlet. The simulations were carried out using the commercial software ANSYS Fluent under the two-phase flow VOF model and k-ε realizable turbulence model. Further CFD simulations were performed to evaluate the behavior of the gun barrel tank under different operating conditions (Different inlet flow rate) and to determine the maximum operation flow which allows obtaining the crude-oil with a maximum BSW content of 0.5%. From the simulation results, an operating curve (operating flow vs retention time) was constructed. This information allows, in practice, to identify the inlet flow rate based on the desired content of BSW in the separated oil.

The heat exchange in stirred tanks occurs mainly due to the impeller rotation and overall circulation of the fluid that promotes a more effective heat transfer between fluid and heat transfer surfaces. The type of impeller used and the size of the vessel (that affects the relation heat transfer area and volume of the vessel) have significant effects on the heat transfer. The heat transfer coefficient is very much dependent on the impeller and the speed of rotation. Empirical correlations are usually used to estimate the process side heat transfer coefficient. However, its dependence on the geometrical parameters restricts the use of those correlations to specific tank configurations. In this respect, the use of computational fluid dynamics (CFD) has recently emerged as an alternative to experimental studies. The procedure proved to be faster and lower cost, and the results proved to be accurate. In this study, CFD was applied to obtain a Nusselt number correlation for a jacketed stirred tank equipped with a six‐blade Rushton turbine impeller. The Nusselt number correlation obtained from the simulated model agrees with experimental data providing a reliable representation of the heat transfer in the tank.