Gold:UFR 2-02 Quality Review

Front Page

Description

Test Case Studies

Evaluation

Best Practice Advice

References

Flow past cylinder 

Underlying Flow Regime 2-02               © copyright ERCOFTAC 2004

Quality Review

INSTRUCTIONS TO THE REVIEWER

Please indicate your agreement or disagreement with the comments below, by ticking either the YES or NO box (using symbol þ). If you would like to comment on any of the questions, please also tick the CO (comment) box, and add your comments in the box provided at the end of each section. Please make sure that all questions are answered.

 

When you have completed the review, please indicate below your overall judgment of the UFR and its documentation:

 

 

Accept þ

 

 

Reject

 

 

 

Accept provided the following conditions are met

 

	The report is of very high quality. It gives a detailed overview of the experiments and a large number of CFD studies. It provides a discussion of the state of the art concerning the simulation of the testcase.   Some smaller spelling corrections have been added using the “track changes” command of MS-word.

 

 

Underlying Flow Regime Title: Flow Past Cylinder

UFR Author and UFR number: W. Rodi / D32-V1-TA4-P04

Reviewer (Name/Organisation) : F. Menter ANSYS-CFX

 

 

1	TOP LEVEL CHECK	YES	NO	CO
1. 1	Is the selected test-case study a good representation of the assigned UFR?	þ	q	q
1. 2	Does the test-case study include both flow measurements and CFD calculations?	þ	q	q
1. 3	Does the document under review comply with the D32 template	þ	q	q
1. 4	Should any parts be expanded, condensed or deleted?	q	þ	q
1. 5	Are the illustrations and their captions clear and informative?	þ	q	q
1. 6	Are the references adequate and complete?	þ	q	q
1. 7	If any hyperlinks are used, do these function correctly?	q	q	þ
Comments:
I replaced the link to flownet, as there was a strange error – it looked correct but gave the message: Cannot locate server: http://dataserv.inriu.fr/flownet/), although the hyperlink text had it correct “inria” not “inriu”

 

 

 

 

 

DETAILED CHECK

 

2	REVIEW OF UFR STUDIES AND CHOICE OF TEST CASE	YES	NO	CO
2. 1	Have past studies of the UFR been reviewed adequately?	þ	q	q
2. 2	Is the chosen test-case study selected from an established database or comparison exercise?	þ	q	q
2. 3	Have the test-case experiments been devised for CFD validation?	þ	q	q
Comments:

 

 

3	DESCRIPTION OF THE STUDY TEST CASE	YES	NO	CO
3. 1	Is the geometry described adequately, including an appropriate sketch?	þ	q	q
3. 2	Are the flow parameters defining the flow regime specified?	þ	q	q
3. 3	Are the principal measured quantities (i.e. those by which success or failure of CFD is to be judged) specified?	þ	q	q
3. 4	Is the description fully self-contained and sufficiently detailed ? (the level of detail required depends on whether a hyperlink to a detailed database is provided)	þ	q	q
Comments:
The report gives a very detailed description of the testcase

 

4	Test CASE EXPERIMENTS	YES	NO	CO
4. 1	Is the test-case facility described adequately?	þ	q	q
4. 2	Are the measurement techniques explained?	þ	q	q
4. 3	Is the quality/accuracy of the measured data discussed?	þ	q	q
4. 4	Are the following quality aspects addressed in this discussion :-
a)	Closeness of flow to target/design conditions?	q	q	q
b)	Accuracy estimation of measured quantities?	þ	q	q
c)	Checks on global conservation of conserved quantities?	q	þ	q
d)	Consistency in the measurements of different quantities?	q	þ	q
e)	Other (briefly describe)			q
4. 5	Is the evidence of data quality judged to be sufficient?	þ	q	q
4. 6	Is the information provided at the flow boundaries sufficient to specify or estimate reasonably well the boundary conditions required for a CFD calculation?	þ	q	q
4. 7	Is the overall discussion self-contained and sufficiently detailed? (the level of detail required depends on whether a hyperlink to a detailed database is provided)	þ	q	q
Comments:

 

 

5	CFD METHODS	YES	NO	CO
5. 1	Is an overview given of the methods used?	þ	q	q
5. 2	Have the following aspects of the methods used been explained adequately:-
a)	The codes employed?	þ	q	q
b)	The turbulence/physical models used?	þ	q	q
c)	The wall treatments applied?	þ	q	q
d)	The numerical boundary conditions?	þ	q	q
5. 3	Are comments made on how well the boundary conditions replicate conditions in the test rig?	þ	q	q
5. 4	Is the quality of the calculations discussed?	þ	q	q
5. 5	Are the following quality aspects addressed in this discussion?			q
a)	The discretisation scheme(s) and solver(s)?	þ	q	q
b)	The sufficiency of grid resolution(s) ?	q	q	q
c)	Sensitivities to uncertainties in the boundary conditions	þ	q	q
d)	Comparisons between separate calculations using the same physical model	þ	q	q
e)	Other (briefly describe)			q
5. 6	Is the evidence of CFD quality judged to be sufficient in all cases?	q	q	þ
Comments:
The issue of sufficient grid resolution is discussed – however no definite conclusions are possible, as there was no systematic grid refinement with one given turbulence model and one code. For LES, the question is even more difficult, as no grid independent results can be obtained I principle. An open question seems to be the spanwise extent of the domain for LES/DES simulations.

 

 

6	COMPARISON OF CFD CALCULATIONS WITH EXPERIMENT	YES	NO	CO
6. 1	Are key comparisons of CFD results with experiment presented in the form of tables or plots?	þ	q	q
6. 2	Do these comparisons include the assessment quantities?	þ	q	q
6. 3	Are further comparisons available via hyperlinks to a results database?	þ	q	q
6. 4	Is the performance of the CFD calculations judged by comparison with experiments discussed and analysed in all cases?	þ	q	q
Comments:

 

 

7	BEST PRACTICE ADVICE FOR THE UFR	YES	NO	CO
7. 1	Are model abilities for this test case discussed and analysed in sufficient detail?	þ	q	q
7. 2	Are recommendations provided on which models should be used for this UFR?	þ	q	q
7. 3	Are these recommendations supported by the evidence?	q	q	þ
Comments:
It is not clear if the differences between the wall function and the two-layer solutions are a result of the near wall treatment or of the finer resolution required by the two-layer formulation. Unfortunately no consistent low-Re RANS model has been used, as it is known that the two-layer model results depend on the switching location used in the code.

© copyright ERCOFTAC 2004

---

Contributors: Wolfgang Rodi - Universität Karlsruhe

Front Page

Description

Test Case Studies

Evaluation

Best Practice Advice

References