CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   hi ( February 12, 2014 12:36

hi iam simulating turbulent flow in concentric double tube heat exchanger. the inner tube is hot water and outer is cold water counter flow. when i have got the heat transfer coefficient for annulus and compare with literature it is wrong ?

for nusslet number in literature i use petkhove also ditus-boilter correlation with hydraulic diameter with wetted PERIMETER for Re. and use heat trasfer PERIMETER(i mean just inner tube diameter) for getting heat transfer from nusslet number.

but when i use wetted PREMETER for claculating heat transfer coefficient from nusslet number i got true answer. but mainly i should use just inner tube PREMIETER for calculating heat transfer coefficient from nusslet.

ghorrocks February 12, 2014 17:41


But in this case you also have to be careful the results you are comparing to are correct.

AKHILESH S L February 13, 2014 02:23

can you please show a schematic with dimensions of your problem? which material you are using as heat exchanger wall? February 14, 2014 03:17

thank u for fast reply
1 Attachment(s)
for calculating heat transfer coefficient in the annulus the hydraulic diameter of the annulus is (Dh=D0-Di) and i use this diameter for calculating Re number. then i use petukhov correlation for finding nusslet number. it should be noted that for annulus the nusselt number should be based on the equivalent diameter calculated from (De=(4*Ac)/Ph) which here Ph is the perimeter of heat transfer area.(i mean inner tube perimeter) and then there is different between De and Dh which we use Dh for calculating Re and De for calculating heat transfer coefficient. but when i get the heat transfer coefficient with Dh from petukhov it is the same with fluent .but when i use De for getting heat transfer coefficient(HTC) from nusselt number it is different from fluent HTC. mainly i should use De for calculating HTC from nusselt number.


ghorrocks February 14, 2014 05:20

You seem to be comparing against empirical correlations. You should not expect these to be very accurate. February 15, 2014 03:47

thanks a lot friends,

i found my problem between your guidance. the problem was that fluent solve the momentum and energy equation with Dh and i use another base i mean De for calculating the nusselt number from empirical correlations.


JuPa February 16, 2014 20:25

Be careful. Literature often define heat transfer coefficients and Nusselt numbers relative to a freestream or bulk temperature. When you post process the results make sure you do the same. By default CFX uses the neighbouring node as the relative temperature to determine heat transfer coefficients and Nusselt numbers. This can be changed in the expert parameters section.

All times are GMT -4. The time now is 13:26.