Computational and experimental analysis of supersonic air ejector: Turbulence modeling and assessment of 3D effects

https://doi.org/10.1016/j.ijheatfluidflow.2015.08.003Get rights and content

Highlights

  • Computational and experimental assessment of computational techniques for ejector flows.

  • Comparisons to 2D/3D (kε, kε realizable, kω SST, and stress–ω RSM) turbulence models.

  • kω SST model performs best while ε-based models more accurate at low motive pressures.

  • Good on-design agreement across 2D and 3D models; off-design needs 3D simulations.

Abstract

Numerical and experimental analyses are performed on a supersonic air ejector to evaluate the effectiveness of commonly-used computational techniques when predicting ejector flow characteristics. Three series of experimental curves at different operating conditions are compared with 2D and 3D simulations using RANS, steady, wall-resolved models. Four different turbulence models are tested: kε, kε realizable, kω SST, and the stress–ω Reynolds Stress Model. An extensive analysis is performed to interpret the differences between numerical and experimental results. The results show that while differences between turbulence models are typically small with respect to the prediction of global parameters such as ejector inlet mass flow rates and Mass Entrainment Ratio (MER), the kω SST model generally performs best whereas ε-based models are more accurate at low motive pressures. Good agreement is found across all 2D and 3D models at on-design conditions. However, prediction at off-design conditions is only acceptable with 3D models, making 3D simulations mandatory to correctly predict the critical pressure and achieve reasonable results at off-design conditions. This may partly depend on the specific geometry under consideration, which in the present study has a rectangular cross section with low aspect ratio.

Introduction

Supersonic ejectors have long been used as passive pumping devices for a range of applications such as nuclear reactor cooling, pumping of volatile fluids, and compression of refrigerants in energy systems. Superior ejector performance is typically achieved when maximizing the entrainment of a low pressure stream (suction flow) with respect to a certain amount of high pressure flow (motive flow), or in other words, by maximizing the Mass Entrainment Ratio (MER) defined asMER=ṁsuctionṁmotiveThis entrainment effect is the result of momentum transfer between two fluids through a shear-mixing layer inside the ejector, depicted qualitatively in Fig. 1. A high pressure motive flow enters a converging–diverging nozzle where it chokes at the throat and then accelerates to supersonic velocities in the divergent section. The low pressure stream enters the suction nozzle, accelerates slightly, and then reaches the mixing chamber. At this point, mechanical energy is transferred from the supersonic motive stream to the subsonic suction stream through the development of a turbulent mixing layer. Depending on the geometric design and operating conditions, the resulting mixed stream may reach supersonic conditions before exiting the mixing section. The supersonic mixed flow will then adjust to pressure conditions in a succession of oblique shocks (Matsuo et al., 1999). The position of the shock train depends on the back pressure, where the higher the outlet pressure, the earlier the mixed flow will shock (for very low back pressure, the shock train enters the subsonic diffuser). Downstream of this point, the motive jet becomes subsonic and the pressure increases gradually to the outlet pressure. Under these circumstances (where the suction flow reaches or exceeds sonic velocity), ejector operation is said to be “on-design,” and the suction flow rate is independent of the outlet pressure. On the contrary, if the mixed flow remains subsonic, the amount of suction flow drawn into the ejector depends on the outlet pressure and the operation is said to be “off-design.” The threshold value of outlet pressure between these two operational modes is called the “critical pressure.” Fig. 2 shows a characteristic curve of the ejector, generated at a constant motive and suction inlet pressure, and varying outlet pressure. The critical point is indicated at the critical pressure where the transition between on- and off-design modes occurs.

Section snippets

Prior work

The global behavior described above is the result of a combination of complex flow features inside the ejector including boundary layers subject to adverse pressure gradients, turbulent mixing layers bounded by near-wall regions, compressibility effects like shock-induced separations, vortex shedding, and recirculating regions. It is because of this complexity that ejector designs and performances have thus far been difficult to characterize and optimize. With the advent of modern computational

Experimental setup

A schematic of the experimental apparatus is provided in Fig. 3. The air supply to the motive nozzle is provided by an industrial Ateliers François compressor (Model CE46B with a capacity of 1320 m3/h FAD and motor power of 250 kW). Before entering the ejector, the air accumulates inside a reservoir at ambient temperature and a set pressure of 16.0 bar. The motive stream pressure is then regulated down to the desired inlet pressure with a Bellofram T-2000 pneumatic valve. Motive pressures at 2.0,

Numerical modeling

Simulations are performed using a finite volume approach available in the commercial CFD package ANSYS FLUENT v14.5. For each condition, 2D and 3D steady-state simulations are conducted using four different turbulence models: kε, kε realizable, kω SST, and the stress–ω Reynolds Stress Model (RSM). Dry air is used as the working fluid and is assumed to have the properties of an ideal gas. Spatial discretization of both the conservation and turbulence equations are set to be second-order

Turbulence modeling

Although the working principle of an ejector is simple, correct numerical prediction of the internal dynamics is dependent on the turbulence model used. At present, no universal turbulence model that can accurately predict all possible flow features is available in the literature. Most models achieve a good level of agreement only for the type of flow for which they were previously calibrated. Wilcox (2006) provides an excellent overview of past and current advances in turbulence modeling.

Overview

To best assess the accuracy of 2D and 3D simulations, three experimental characteristic curves at motive pressures of 2.0, 3.5, and 5.0 bar are compared with numerical results. Fig. 9, Fig. 10, Fig. 11, Fig. 12 show this comparison for the four different turbulence models described in the previous section: kε, kε realizable, kω SST, and the RSM. Table 1, Table 2 show the differences between the predictions of each of these models and the data. In Table 1, differences are reported as mean

Conclusion

In the present study, numerical and experimental analyses are conducted to evaluate the effectiveness of common numerical techniques when predicting the flow field inside a rectangular air ejector. Three series of experimental curves at different operating conditions are compared with 2D and 3D, RANS, steady, wall resolved, CFD simulations, using four different turbulence models: kε, kε realizable, kω SST, and the stress–ω Reynolds Stress Model.

A comparison of global parameters finds that

References (30)

  • M. Wolfshtein

    The velocity and temperature distribution in one-dimensional flow with turbulence augmentation and pressure gradient

    Int. J. Heat Mass Transfer

    (1969)
  • H. Al-Ansary et al.

    Numerical and experimental analysis of single-phase and two-phase flow in ejectors

    HVAC&R Res.

    (2004)
  • Bouhanguel, A., Desevaux, P., Gavignet, E., 2009. 3D CFD simulation of a supersonic air ejector. In: International...
  • T.B. Gatsky et al.

    Compressibility, Turbulence and High Speed Flow

    (2013)
  • I.E. Idelchik

    Handbook of Hydraulic Resistance

    (2008)
  • Cited by (117)

    View all citing articles on Scopus
    View full text