An analytical model for droplet separation in vane separators and measurements of grade efficiency and pressure drop

• An analytical model for efficiency is extended with additional geometrical features.
• A simplified and a novel vane separator design are investigated experimentally.
• Experimental results are significantly affected by re-entrainment effects.
• Outlet droplet size spectra are accurately predicted by the model.
• The improved grade efficiency doubles the pressure drop.

This study investigates the predictive power of analytical models for the droplet separation efficiency of vane separators and compares experimental results of two different vane separator geometries. The ability to predict the separation efficiency of vane separators simplifies their design process, especially when analytical research allows the identification of the most important physical and geometrical parameters and can quantify their contribution. In this paper, an extension of a classical analytical model for separation efficiency is proposed that accounts for the contributions provided by straight wall sections. The extension of the analytical model is benchmarked against experiments performed by Leber (2003) on a single stage straight vane separator. The model is in very reasonable agreement with the experimental values. Results from the analytical model are also compared with experiments performed on a vane separator of simplified geometry (VS-1). The experimental separation efficiencies, computed from the measured liquid mass balances, are significantly below the model predictions, which lie arbitrarily close to unity. This difference is attributed to re-entrainment through film detachment from the last stage of the vane separators. After adjustment for re-entrainment effects, by applying a cut-off filter to the outlet droplet size spectra, the experimental and theoretical outlet Sauter mean diameters show very good agreement. A novel vane separator geometry of patented design (VS-2) is also investigated, comparing experimental results with VS-1. Experimental separation efficiencies based on recorded mass flow balances are very similar for both geometries. However, based on an analysis of the outlet droplet size spectra, VS-2 outperforms VS-1. It is concluded that re-entrainment effects cloud the performance of VS-2 and that the perforations in the wall sections of VS-2 are not very effective. Pressure drop measurements performed on both geometries reveal that the pressure drop for VS-2 is higher than for VS-1 by a factor of almost two for gas velocities up to 10 m/s.