1 January 2016 in Journal of Cleaner Production
TPThomas Wilis Cândido Pereira
Thomas Wilis Cândido Pereira
FPFábio de Assis Ressel Pereira
Fábio de Assis Ressel Pereira
1 October 2019 in Journal of Petroleum Science and Engineering
FPFabio de A. Ressel Pereira
Fabio de A. Ressel Pereira
1 December 2020 in Journal of Petroleum Science and Engineering
1 November 2020 in Journal of Natural Gas Science and Engineering
SFSamuel Vítor de Almeida Franco
Samuel Vítor de Almeida Franco
11 March 2020 in Oil & Gas Science and Technology – Revue d’IFP Energies nouvelles
FPFabio de Assis Ressel Pereira
Fabio de Assis Ressel Pereira
MRMarcos Pellegrini Ribeiro
Marcos Pellegrini Ribeiro
The recent development of Electrical Submersible Pump (ESP) in the skid, installed in the seabed downstream of the wellhead in an offshore oil production system, is an alternative to the conventional system with the set installed at the bottom of the producing well, facilitating interventions in case of failure. The pump is driven by an electric motor whose cooling must be efficient to ensure the continuity of its operation. The heat withdrawal is performed by the fluid produced. The purpose of this article is to understand the process of electric motor cooling to the single-phase and turbulent flow with convection heat transfer in an annular geometry, which represents the space formed between a capsule and the ESP in the Skid system motor. With this objective it is employed a Computational Fluid Dynamics (CFD) code to solve the governing equations of the turbulent heat transfer single-phase flow. The standard κ-ε model with improved wall function (Enhanced Wall Treatment) is used to closure turbulence problem. This study considered flow rates range of 2200–4200 m3/d (representing Reynolds numbers range of 27 000–133 000 approximately), Prandtl numbers 7–37, three configurations of different annular geometries, one concentric and two eccentric, together with the condition of the constant temperature on the motor surface (130 °C) and capsule (4 °C). The simulations are validated by comparing the Nusselt number in the developed region with the Gnielinski correlation. It is observed that if the constant heat flux condition were used, the motor temperature would have lower values at the beginning and larger at the end of the geometry. Therefore, the higher the Nusselt number, the greater the heat transfer, thus intensifying the cooling of the electric motor. In the eccentric geometry a momentum transfer from the lower to the upper annular region is observed, causing the Nusselt number present an angular variation. In eccentric geometries the flow develops in greater lengths, observing that the greater the eccentricity, the greater this length. Finally, for the ESP in the Skid system the use of an eccentric geometry is not adequate.
30 May 2019 in Journal of Dispersion Science and Technology
Gravity separation of liquid-liquid dispersions is present in many industrial applications, and it involves the complex governing mechanisms of creaming and coalescence in concentrated mixtures. The aim of this work is to model the laminar creaming zone of batch gravitational separation with coalescence using CFD tools. The numerical results were compared to the experimental decay curves of bottle tests. The initial DSD representation by the population balance model was obtained with the Gauss-Legendre Quadrature. The number of velocity groups and the application range of the drag modifier showed to be a function of the initial DSD profile and polydispersity, respectively. The coalescence models evaluated were the constant coalescence efficiency, the critical approach velocity and the film drainage for a deformable partially mobile interface. The coalescence models parameters were calibrated for a determined initial condition (phases volume fraction and droplet size distribution (DSD)) and then applied for different initial conditions. All models were able to model the separation of all initial conditions with the same calibrated parameters, although the drainage model did not need to be calibrated. This suggests that when employed in the modeling of a continuous separator, these can be calibrated by the decay curves of bottle tests.