Flow behind a blunt trailing edge

Underlying Flow Regime 2-01 © copyright ERCOFTAC 2004

Description

Preface

The flow behind a blunt trailing edge is of interest to turbomachinery research due to its impact in the greater understanding of the unsteady flow phenomena, which appears behind the turbine blades. The main characteristic of the proposed UFR is the vortex shedding, which is generated from the blades and causes the acoustics resonance and the structural vibrations.

This type of flow can be view as a special class of the flows over the bluff bodies that have significant separation regions, as opposed to the streamlined type of fluid flows. Compared with the flow over cylinders or flat plates, the vortex shedding from turbine blades is more complex because the boundary layers on the pressure and suction sides are not identical.

It is well known that the flow over cylinders or flat plates is relatively well documented in the literature, but in our knowledge it seems that for the suggested UFR we have only a few detailed studies, which we have referenced, at the end of this document. We also underline that the experimental data of this proposed UFR are of great importance in the validation of the modern CFD codes and particularly relevant for the turbulence modeling studies and research.

Introduction

In the QNET-CFD project, this UFR is relevant for several Application Challenges mainly documented in the QNET-CFD thematic are six on Turbomachinery Internal flows.

According to Sieverding et al. (2002) the main characteristics of such flows are:

the unsteadiness related to the large coherent structures, known as von Karman vortices,

the unsteadiness related to the interference of wakes with downstream blade rows.

The interest in increasing the knowledge regarding this UFR is marked by one of the most consistent research effort contained in the European Projects BRITE/EURAM Time Varying Wake Characteristics behind Flat Plates and Turbine Blades and Turbulence Modeling of Unsteady Flows in Axial Turbines. The aim of such studies was to gain insight by the combined experimental and computational work.

This research conducted at the end of 90's at von Karman Institute (VKI) showed that a comprehensive investigation requires for analysis a large turbine cascade. Consequently, the requirements for high spatial and temporal resolutions were fulfilled and the confidence in experimental data was drastically increased.

Therefore, the configuration under review will be a very large-scale nozzle guide vane with a relatively thick trailing edge. This blade cascade is sketched in Figure 1.

Figure 1: Blade cascade (Figure 1 in Manna et al. (1997))

Review of UFR studies and choice of test case

As it was already mentioned, the literature on this subject is relatively scarce in offering quantitative information on the phenomena behind the turbine blades. In order to analyze such phenomena, it was concluded that the main quantity responsible for the unsteady flow characteristics, is the pressure distribution around the trailing edge and its intimate link with the vortex-shedding phenomenon.

Looking from the experimental side, the von Karman vortices have been emphasized by different authors and with different techniques: smoke visualization (Han and Cox (1983)), ultra-short Schlieren photographs (Lawaczeck and Heineman (1975), Heinemann and Butefisch (1997)) or interferometric density measurements (Cicatelli and Sieverding (1996)).

In order to linking the vortex shedding to the unsteady pressure at the trailing edge a combined pressure measurement and flow field visualization has been performed. First, this link was established by Carscallen et al. (1998) for a transonic turbine and by Sieverging and his group for low subsonic and high subsonic/transonic range (Cicatelli and Sieverding (1996), Sieverding et al. (2002)).

It should be mentioned that the work of Desse (1998), which combine interferometry and shadow visualization has proven the strong coupling between the unsteady pressure and vortex shedding behind the profiled flat plates. The main finding was that the thickness of the turbulent boundary layer is playing a more important role than the vortex shedding frequency.

On the computational side, it seems that the VKI blade cascade investigated by Sieverdings group (see Figure 1) is preferred for computational tests. This test case it preferred for computational tests because is very well experimentally documented. VKI blade cascade among others in literature of this subject is enough large, a especially the trailing edge dimensions what useful from experimental point to view and allows to performed detailed and accurate experimental investigations. Most available literature concerning unsteady trailing edge flow characteristics refers to tests carried with relatively small size turbine blades whose trailing edge dimensions are too small for comprehensive investigation. We also mention the computational studies of Magagnato (1999), Manna et al. (1997) and Arnone and Pacciani (1997).

Except the first study, which focused on the steady state flow, the other studies are computed with good accuracy of the time averaged edge pressure distribution and the vortex shedding frequency.

Magagnato (1999) represents the only study, which applies non-linear turbulence models, while the other studies preferred to apply the linear model or Baldwin-Lomax turbulence model.

It must be underlined, that Manna et al. (1997) have linked "the accuracy of computing the unsteady flow" to the appropriate resolution of the separating turbulent shear layers.

Consequently, due to the consistency between the experimental and computational developments, our UFR documentation will focus on the experimental work of Cicatell and Sieverding (1997) and its numerical counterpart, namely Manna et al. (1997).

Contributors: Charles Hirsch - Vrije Universiteit Brussel