Investigation on Separation Efficiency in Supersonic Separator with Gas-Droplet Flow Based on DPM Approach

Abstract

Supersonic separator is a new technology based on the adiabatic expansion of swirling gas flow, and at present it has demonstrated great application potential in separating and processing droplet liquid contained in natural gas. However, its coefficient of performance is still low and there seems to be a large gap in the method that evaluates the separation efficiency in a satisfactory manner. In order to promote the wide application of this technology in the dehydration field, it is necessary to find a new and feasible approach that can be used to predict the flow characteristic and separation performance inside a supersonic separator. In this paper, a comprehensive three-dimensional fluid numerical model to study the flow behavior and separation efficiency in a supersonic separator was established coupled with the discrete particle model (DPM). The mixture of air and water droplets was chosen as working fluid. The gas phase was modeled with compressible Navier–Stokes equations for two-phase flow and the RSM turbulence model was taken into account. The droplet phase was modeled with the discrete particle model (DPM), in which the droplets are assumed to have the same sphere shape and ignore the phase transition and nucleation process. A pilot test facility was carried out to validate the numerical model. The experimental results not only indicate that the new dehydration device can efficiently separate the liquid droplets from wet gas, but also prove that the numerical results were great agreements with the experimental results. Furthermore, based on the proposed numerical approach, the gas-droplet turbulent flow structures were predicted, the effects of different structure parameters and operation conditions on the separation efficiency were also investigated. The current works settle a foundation for further explorations on the supersonic gas–liquid separation flows inside a supersonic separator as well as the possible new applications.