UFR 3-10 Evaluation: Difference between revisions

From KBwiki
Jump to navigation Jump to search
(New page: {{UFR|front=UFR 3-10|description=UFR 3-10 Description|references=UFR 3-10 References|testcase=UFR 3-10 Test Case|evaluation=UFR 3-10 Evaluation|qualityreview=UFR 3-10 Quality Review|bestp...)
 
Line 17: Line 17:
== Comparison of CFD calculations with Experiments ==
== Comparison of CFD calculations with Experiments ==


The computations by Andersson et al. (1993) gave the following main results: besides the shortcomings following from the use of wall functions (e.g. the inability to capture the near-wall peak in the distribution of the streamwise turbulence intensity), the RSM closure gives an adequate prediction of the mean flow and turbulence characteristics. The k-ε computations fail to reproduce the experimentally determined growth rate and wall friction, which are overpredicted.
The computations by Andersson et al. (1993) gave the following main results: besides the shortcomings following from the use of wall functions (e.g. the inability to capture the near-wall peak in the distribution of the streamwise turbulence intensity), the RSM closure gives an adequate prediction of the mean flow and turbulence characteristics. The k-ε computations fail to reproduce the experimentally determined growth rate and wall friction, which are overpredicted.


The same general results as indicated above were obtained with the standard k-ε model using the commercial codes PHOENICS, FLUENT, and CFDS-FLOW3D (Hemström, 1995). With the same grid, turbulence model, and boundary conditions, the results were almost identical for the three codes, giving confidence to the numerical solutions of the problem.
The same general results as indicated above were obtained with the standard k-ε model using the commercial codes PHOENICS, FLUENT, and CFDS-FLOW3D (Hemström, 1995). With the same grid, turbulence model, and boundary conditions, the results were almost identical for the three codes, giving confidence to the numerical solutions of the problem.


The Proceedings from the Workshops do not (except for a few exceptions) contain direct comparisons of design or assessment parameters such as the growth rate of the jet. Therefore some of the parameters are difficult to assess. On the other hand, the skin friction coefficient is given and direct comparisons of C<sub>f</sub> are possible.
The Proceedings from the Workshops do not (except for a few exceptions) contain direct comparisons of design or assessment parameters such as the growth rate of the jet. Therefore some of the parameters are difficult to assess. On the other hand, the skin friction coefficient is given and direct comparisons of C<sub>f</sub> are possible.
Line 25: Line 25:
Sample results from some ARSM and RSM models are presented in Figures 11 and 12. These computations show good agreement with experimental data on both mean velocity and skin friction.
Sample results from some ARSM and RSM models are presented in Figures 11 and 12. These computations show good agreement with experimental data on both mean velocity and skin friction.


Computations using k-εwith wall functions do in general fail to predict growth rate and skin friction – overpredictions of the order 20-30 % are common. Some of the near-wall two-equation models appear to predict the growth rate quite well, while C<sub>f</sub> is slightly overpredicted. The unique characteristic of the wall jet that the point of zero shear stress does not coincide with the point of zero mean velocity gradients cannot, of course, be predicted with such models.<br style="page-break-before: always" clear="all" />
Computations using k-&epsilon; with wall functions do in general fail to predict growth rate and skin friction – overpredictions of the order 20-30 % are common. Some of the near-wall two-equation models appear to predict the growth rate quite well, while C<sub>f</sub> is slightly overpredicted. The unique characteristic of the wall jet that the point of zero shear stress does not coincide with the point of zero mean velocity gradients cannot, of course, be predicted with such models.
 
<br style="page-break-before: always" clear="all" />


[[Image:U3-10d32_files_image048.jpg]]
[[Image:U3-10d32_files_image048.jpg]]

Revision as of 16:27, 11 March 2009


Front Page

Description

Test Case Studies

Evaluation

Best Practice Advice

References




The plane wall jet 

Underlying Flow Regime 3-10               © copyright ERCOFTAC 2004


Evaluation

Comparison of CFD calculations with Experiments

The computations by Andersson et al. (1993) gave the following main results: besides the shortcomings following from the use of wall functions (e.g. the inability to capture the near-wall peak in the distribution of the streamwise turbulence intensity), the RSM closure gives an adequate prediction of the mean flow and turbulence characteristics. The k-ε computations fail to reproduce the experimentally determined growth rate and wall friction, which are overpredicted.

The same general results as indicated above were obtained with the standard k-ε model using the commercial codes PHOENICS, FLUENT, and CFDS-FLOW3D (Hemström, 1995). With the same grid, turbulence model, and boundary conditions, the results were almost identical for the three codes, giving confidence to the numerical solutions of the problem.

The Proceedings from the Workshops do not (except for a few exceptions) contain direct comparisons of design or assessment parameters such as the growth rate of the jet. Therefore some of the parameters are difficult to assess. On the other hand, the skin friction coefficient is given and direct comparisons of Cf are possible.

Sample results from some ARSM and RSM models are presented in Figures 11 and 12. These computations show good agreement with experimental data on both mean velocity and skin friction.

Computations using k-ε with wall functions do in general fail to predict growth rate and skin friction – overpredictions of the order 20-30 % are common. Some of the near-wall two-equation models appear to predict the growth rate quite well, while Cf is slightly overpredicted. The unique characteristic of the wall jet that the point of zero shear stress does not coincide with the point of zero mean velocity gradients cannot, of course, be predicted with such models.


U3-10d32 files image048.jpg

Figure 11. Mean velocity profiles at x/b = 70. Second-moment closures.


U3-10d32 files image051.gif

Figure 12. Umax/Uref and local skin friction coefficient at x/b = 70. Second-moment closures.

© copyright ERCOFTAC 2004



Contributors: Jan Eriksson; Rolf Karlsson - Vattenfall Utveckling AB


Front Page

Description

Test Case Studies

Evaluation

Best Practice Advice

References