# Difference between revisions of "UFR 2-04 Best Practice Advice"

m (Dave.Ellacott moved page SilverP:UFR 2-04 Best Practice Advice to UFR 2-04 Best Practice Advice over redirect) |
|||

(5 intermediate revisions by the same user not shown) | |||

Line 32: | Line 32: | ||

While qualitative patters of the flow around blades are generally independent of the model of turbulence, quantitative agreement, as well as correct prediction of the development of the boundary layer, call for a correct choice of the model in turbulence in use. So | While qualitative patters of the flow around blades are generally independent of the model of turbulence, quantitative agreement, as well as correct prediction of the development of the boundary layer, call for a correct choice of the model in turbulence in use. So | ||

− | • Use the low-Reynolds-number | + | • Use the low-Reynolds-number k–ε for a fair prediction of the pressure coefficient distribution only at the pressure side. |

− | • Use corrections to the low-Reynolds-number | + | • Use corrections to the low-Reynolds-number k–ε modeling of the laminar flow, to improve the prediction of the pressure distribution at the suction side. |

− | • Use a | + | • Use a y<sup>+</sup> maximum spacing for the first nodes off the solid wall. |

• Employ a non-linear model for a much more sensitive response. | • Employ a non-linear model for a much more sensitive response. | ||

− | == <span lang="EN-US">7.4.<span style="font: 7.0pt "Times New Roman""> | + | == <span lang="EN-US">7.4.<span style="font: 7.0pt "Times New Roman""> </span></span>Recommendations for Future Work == |

− | Despite the fact that the test case reviewed herein first appeared in the open literature in the early nineties, most of the relevant computational studies were carried out in before 1996. Therefore, turbulence modeling is restricted to variants of the low-Reynolds-number | + | Despite the fact that the test case reviewed herein first appeared in the open literature in the early nineties, most of the relevant computational studies were carried out in before 1996. Therefore, turbulence modeling is restricted to variants of the low-Reynolds-number k–ε model with several modifications in order to improve modelling, and only in one study a non-linear model has been employed. Since more promising models of turbulence have sprung in the recent years (Menter’s k–ω SST, Durbin’s v2f and Spalart and Almaras for example) it would be of great value to continue the numerical study of this test case with such models of turbulence for closure, so as to update the assessment of the turbulence modelling. |

<font size="-2" color="#888888">© copyright ERCOFTAC 2004</font><br /> | <font size="-2" color="#888888">© copyright ERCOFTAC 2004</font><br /> |

## Latest revision as of 19:39, 11 February 2017

## Contents

# Flow around (airfoils and) blades (subsonic)

Underlying Flow Regime 2-04 © copyright ERCOFTAC 2004

# Best Practice Advice

## Best Practice Advice for the UFR

## 7.1. Key Physics

The key physics behind the (Subsonic) Flow around (Airfoils and) Blades encompass the development of a two-dimensional boundary layer under pressure gradient. Since we are dealing with turbomachinery applications, in most cases free-stream turbulence strongly influences the development of the boundary layer flow too. Downstream of the blade, the wake/core flow interaction is prevailing.

## 7.2. Numerical Modeling Issues

• Use a C- or an O-type grid around the airfoil to generate a regular structured grid in the leading edge region of a blade with a blunt leading edge.

• Use very fine grids (in both the stream-wise and the pitch-wise direction) to capture the boundary layer development on the suction side.

• Employ a high order accuracy discretization scheme to reduce the numerical dissipation and to accurately compute the pressure-losses.

## 7.3. Physical Modeling

While qualitative patters of the flow around blades are generally independent of the model of turbulence, quantitative agreement, as well as correct prediction of the development of the boundary layer, call for a correct choice of the model in turbulence in use. So

• Use the low-Reynolds-number k–ε for a fair prediction of the pressure coefficient distribution only at the pressure side.

• Use corrections to the low-Reynolds-number k–ε modeling of the laminar flow, to improve the prediction of the pressure distribution at the suction side.

• Use a y^{+} maximum spacing for the first nodes off the solid wall.

• Employ a non-linear model for a much more sensitive response.

## 7.4. Recommendations for Future Work

Despite the fact that the test case reviewed herein first appeared in the open literature in the early nineties, most of the relevant computational studies were carried out in before 1996. Therefore, turbulence modeling is restricted to variants of the low-Reynolds-number k–ε model with several modifications in order to improve modelling, and only in one study a non-linear model has been employed. Since more promising models of turbulence have sprung in the recent years (Menter’s k–ω SST, Durbin’s v2f and Spalart and Almaras for example) it would be of great value to continue the numerical study of this test case with such models of turbulence for closure, so as to update the assessment of the turbulence modelling.

© copyright ERCOFTAC 2004

Contributors: E. S. Politis - NTUA