[cxfer]

Dear all,

There is a .csv file (12000 lines) for power density distribution discrete data points in cylindrical coordinate and contains four columns, which are r, phi,z and power density (W/cm3), and the whole power is 45kW. My case is water cooling around the solid boundary.

I want to import the data to subdomain as heat source, and have no idea how to treat the data. I try to directly import the data as three dimensions interpolation function in CFX, but the calculation result for CFD post show that the power is 251kW passed from solid to water boundary and this is incorrect I think. I have no idea what the problem is and what is the correct way to load the heat source data for subdomain.

Any help will be appreciated and thanks in advance.

[ghorrocks]

You are using the correct method - a 3D interpolation function and a source term. So something is wrong with your implementation.
* Are you sure it is 45kW total?
* Have you implemented it as source points or as a source term over a subdomain?
* Are you sure your model is converged? Low imbalances? Is it steady state?
* Have you looked how the points map onto your mesh? Is you mesh too coarse or fine for the points?
* Have you linearised the source term at all?

[cxfer]

Thanks for your reply.

1) I am sure the total power is 45kW and I implement as a source term over a subdomain.

2) The model is converged for steady calculation with residual errors are less than 1e-4.

3) I am not sure how to look the points onto my mesh in CFX. Please tell me a procedure, thanks.

4) My source term is calling expressions named heat1bin by using a 3D interpolation function named heatrphiz1bin, and the unit for heatrphiz1bin is [W/cm^3], the argument unit is [cm,rad,cm].

I have attached part of my soure data file at the end, which contains four columns r, phi, z and power density. Phi is equal to -pi. Actually, the data uniformly distribute from -pi to pi and is r-z distribution. I am not sure how to linearise source term at all. Could you please show me an example.

5) The imbalances values are as followes and maybe it seems some domain imbalances are low, but I have no idea the what the threshold value is:
================================================== ====================
Boundary Flow and Total Source Term Summary
================================================== ====================

+--------------------------------------------------------------------+
| U-Mom-water |
+--------------------------------------------------------------------+
Boundary : inlet -4.8452E-09
Boundary : outlet -1.4201E-04
Boundary : tubewaterbj Side 1 5.4048E-04
Boundary : water Default -5.3609E-07
Boundary : waterfstcubj Side 1 -5.5683E-04
-----------
Domain Imbalance : -1.5890E-04

+--------------------------------------------------------------------+
| V-Mom-water |
+--------------------------------------------------------------------+
Boundary : inlet -6.5741E-09
Boundary : outlet -1.4469E-04
Boundary : tubewaterbj Side 1 2.4296E-03
Boundary : water Default -1.4619E-06
Boundary : waterfstcubj Side 1 -2.6287E-03
-----------
Domain Imbalance : -3.4524E-04

+--------------------------------------------------------------------+
| W-Mom-water |
+--------------------------------------------------------------------+
Boundary : inlet -1.1452E+00
Boundary : outlet -4.2095E-01
Boundary : tubewaterbj Side 1 1.5704E+00
Boundary : water Default -4.5409E-04
Boundary : waterfstcubj Side 1 -5.7709E-03
-----------
Domain Imbalance : -1.9336E-03

+--------------------------------------------------------------------+
| P-Mass-water |
+--------------------------------------------------------------------+
Boundary : inlet 5.2960E-01
Boundary : outlet -5.2960E-01
-----------
Domain Imbalance : 1.4722E-07

+--------------------------------------------------------------------+
| H-Energy-water |
+--------------------------------------------------------------------+
Boundary : inlet 2.6575E+04
Boundary : outlet -2.7295E+05
Domain Interface : tubewaterbj (Side 1) -8.5176E+01
Domain Interface : waterfstcubj (Side 1) 2.4820E+05
-----------
Domain Imbalance : 1.7410E+03

+--------------------------------------------------------------------+
| T-Energy-tube |
+--------------------------------------------------------------------+
Boundary : tube Default -8.4258E+01
Domain Interface : tubewaterbj (Side 2) 8.5176E+01
-----------
Domain Imbalance : 9.1820E-01

+--------------------------------------------------------------------+
| T-Energy-fstcu |
+--------------------------------------------------------------------+
Boundary : fstcu Default -2.9097E+03
Domain Interface : waterfstcubj (Side 2) -2.4820E+05
Domain Interface : smcubj (Side 2) 2.5220E+05
-----------
Domain Imbalance : 1.0922E+03

+--------------------------------------------------------------------+
| T-Energy-sm |
+--------------------------------------------------------------------+
Sub-Domain : smheat 2.5132E+05
Domain Interface : smcubj (Side 1) -2.5220E+05
-----------
Domain Imbalance : -8.7916E+02


+--------------------------------------------------------------------+
| Normalised Imbalance Summary |
+--------------------------------------------------------------------+
| Equation | Maximum Flow | Imbalance (%) |
+--------------------------------------------------------------------+
| U-Mom-water | 1.5704E+00 | -0.0101 |
| V-Mom-water | 1.5704E+00 | -0.0220 |
| W-Mom-water | 1.5704E+00 | -0.1231 |
| P-Mass-water | 5.2960E-01 | 0.0000 |
+----------------------+-----------------------+---------------------+
| H-Energy-water | 2.7295E+05 | 0.6379 |
| T-Energy-tube | 2.7295E+05 | 0.0003 |
| T-Energy-fstcu | 2.7295E+05 | 0.4002 |
| T-Energy-sm | 2.7295E+05 | -0.3221 |
+----------------------+-----------------------+---------------------+

Source data files [cm,rad,cm] for 1st to 3rd columns and power density [W/cm^3] for the 4th column:

0.0 -3.1416 54.0 34.152
0.1 -3.1416 54.0 37.815
0.2 -3.1416 54.0 45.555
0.3 -3.1416 54.0 58.026
0.4 -3.1416 54.0 76.11
0.5 -3.1416 54.0 100.245
0.6 -3.1416 54.0 130.458
0.7 -3.1416 54.0 165.732
0.8 -3.1416 54.0 203.622
0.9 -3.1416 54.0 240.48
1.0 -3.1416 54.0 271.779
1.1 -3.1416 54.0 293.073
1.2 -3.1416 54.0 300.78
1.3 -3.1416 54.0 293.601
1.4 -3.1416 54.0 272.184
1.5 -3.1416 54.0 239.763
1.6 -3.1416 54.0 200.736
1.7 -3.1416 54.0 159.915
1.8 -3.1416 54.0 121.458
1.9 -3.1416 54.0 88.209
2.0 -3.1416 54.0 61.479
2.1 -3.1416 54.0 41.37
2.2 -3.1416 54.0 27.0609
2.3 -3.1416 54.0 17.3601
2.4 -3.1416 54.0 11.0469
2.5 -3.1416 54.0 7.0605
2.6 -3.1416 54.0 4.587
2.7 -3.1416 54.0 3.0639
2.8 -3.1416 54.0 2.12106
2.9 -3.1416 54.0 1.52862
3.0 -3.1416 54.0 1.14318
3.1 -3.1416 54.0 0.88431
3.2 -3.1416 54.0 0.70083
3.3 -3.1416 54.0 0.5673
3.4 -3.1416 54.0 0.46533
3.5 -3.1416 54.0 0.38712
3.6 -3.1416 54.0 0.32676
3.7 -3.1416 54.0 0.277689
3.8 -3.1416 54.0 0.238224
3.9 -3.1416 54.0 0.206358
4.0 -3.1416 54.0 0.180363
4.1 -3.1416 54.0 0.158133
4.2 -3.1416 54.0 0.139932
4.3 -3.1416 54.0 0.123732
4.4 -3.1416 54.0 0.110757
4.5 -3.1416 54.0 0.099018
4.6 -3.1416 54.0 0.089226
4.7 -3.1416 54.0 0.080601
4.8 -3.1416 54.0 0.073299
4.9 -3.1416 54.0 0.066834
5.0 -3.1416 54.0 0.061329
5.1 -3.1416 54.0 0.056232
5.2 -3.1416 54.0 0.051678
5.3 -3.1416 54.0 0.047727
5.4 -3.1416 54.0 0.044073
5.5 -3.1416 54.0 0.041124
5.6 -3.1416 54.0 0.03831
5.7 -3.1416 54.0 0.035772
5.8 -3.1416 54.0 0.033576
5.9 -3.1416 54.0 0.031536
6.0 -3.1416 54.0 0.0296124
6.1 -3.1416 54.0 0.027813
6.2 -3.1416 54.0 0.0263022
6.3 -3.1416 54.0 0.0250773
6.4 -3.1416 54.0 0.0237564
6.5 -3.1416 54.0 0.022479
6.6 -3.1416 54.0 0.0214698
6.7 -3.1416 54.0 0.0204408
6.8 -3.1416 54.0 0.0196164
6.9 -3.1416 54.0 0.0186885
7.0 -3.1416 54.0 0.0179052
7.1 -3.1416 54.0 0.0171753
7.2 -3.1416 54.0 0.0164331
7.3 -3.1416 54.0 0.0157596
7.4 -3.1416 54.0 0.0153795
7.5 -3.1416 54.0 0.0147024
7.6 -3.1416 54.0 0.0142971
7.7 -3.1416 54.0 0.013845
7.8 -3.1416 54.0 0.0134202
7.9 -3.1416 54.0 0.0129576
8.0 -3.1416 54.0 0.0124875
8.1 -3.1416 54.0 0.0121485
8.2 -3.1416 54.0 0.0118944
8.3 -3.1416 54.0 0.0115503
8.4 -3.1416 54.0 0.0112998
8.5 -3.1416 54.0 0.0109653
8.6 -3.1416 54.0 0.0107013
8.7 -3.1416 54.0 0.0104736
8.8 -3.1416 54.0 0.0102663
8.9 -3.1416 54.0 0.0100479
9.0 -3.1416 54.0 0.0098511
9.1 -3.1416 54.0 0.0096234
9.2 -3.1416 54.0 0.0094086
9.3 -3.1416 54.0 0.0093249
9.4 -3.1416 54.0 0.0091305
9.5 -3.1416 54.0 0.0089124
9.6 -3.1416 54.0 0.0087873
9.7 -3.1416 54.0 0.0087075
9.8 -3.1416 54.0 0.0086748
9.9 -3.1416 54.0 0.0086838
10.0 -3.1416 54.0 0.033066
10.1 -3.1416 54.0 0.0239277
10.2 -3.1416 54.0 0.0162423
10.3 -3.1416 54.0 0.0119715
10.4 -3.1416 54.0 0.0107763
10.5 -3.1416 54.0 0.0103974
10.6 -3.1416 54.0 0.0101484
10.7 -3.1416 54.0 0.0098799
10.8 -3.1416 54.0 0.009678
10.9 -3.1416 54.0 0.0094497
11.0 -3.1416 54.0 0.0092346
11.1 -3.1416 54.0 0.0089976
11.2 -3.1416 54.0 0.008817
11.3 -3.1416 54.0 0.0086718
11.4 -3.1416 54.0 0.0085095
11.5 -3.1416 54.0 0.0083691
11.6 -3.1416 54.0 0.0081603
11.7 -3.1416 54.0 0.0080751
11.8 -3.1416 54.0 0.0078909
11.9 -3.1416 54.0 0.0077763
12.0 -3.1416 54.0 0.0076281
12.1 -3.1416 54.0 0.0074616
12.2 -3.1416 54.0 0.0074406
12.3 -3.1416 54.0 0.0072588
12.4 -3.1416 54.0 0.0071271
12.5 -3.1416 54.0 0.0070035
12.6 -3.1416 54.0 0.0069258
12.7 -3.1416 54.0 0.0068835
12.8 -3.1416 54.0 0.0068601
12.9 -3.1416 54.0 0.0066474
13.0 -3.1416 54.0 0.0065799
13.1 -3.1416 54.0 0.0065484
13.2 -3.1416 54.0 0.0065106
13.3 -3.1416 54.0 0.00642
13.4 -3.1416 54.0 0.0064197
13.5 -3.1416 54.0 0.0062763
13.6 -3.1416 54.0 0.0061893
13.7 -3.1416 54.0 0.0062775
13.8 -3.1416 54.0 0.0062112
13.9 -3.1416 54.0 0.0061224
14.0 -3.1416 54.0 0.0060645
14.1 -3.1416 54.0 0.0060102
14.2 -3.1416 54.0 0.0059397
14.3 -3.1416 54.0 0.0059289
14.4 -3.1416 54.0 0.0059187
14.5 -3.1416 54.0 0.0058935
14.6 -3.1416 54.0 0.0058308
14.7 -3.1416 54.0 0.0058635
14.8 -3.1416 54.0 0.0058044
14.9 -3.1416 54.0 0.0058515
15.0 -3.1416 54.0 0.0057963
15.1 -3.1416 54.0 0.0057507
15.2 -3.1416 54.0 0.0057633
15.3 -3.1416 54.0 0.0057558
15.4 -3.1416 54.0 0.0057837
15.5 -3.1416 54.0 0.0057018
15.6 -3.1416 54.0 0.0056742
15.7 -3.1416 54.0 0.0057153
15.8 -3.1416 54.0 0.0057669
15.9 -3.1416 54.0 0.0055782
......

[ghorrocks]

1) How is the 45kW calculated? Are the points isolated point heat sources, or is it integrated across the domain?

2) For simulations where you are looking at global integrals (eg total heat) I would recommend also making imbalances a convergence criteria. The imbalances can be bad despite the residuals being good.

3) I am saying to look at how many points map to each element. If multiple points map to the same element, then how do you work out what heat source to apply? If elements lie in between points, then how do you calculate the heat value?

It appears you are using the 3D interpolation function, and that will do a distance weighted average between points. This will not be conservative, so you will not get your intended total power out of this approach.

4) If the heat applied is not a function of simulation variables then no linearisation is required.

[cxfer]

1) The power density (unit:W/cm3) deposition distribution data calculated by a third-part monte-carlo software and then export to csv format file and total power is 45kW. The data points are isolated points as shown in my attachment at the end of my previous post.

2) I looking at the total heat by using CFD post software, just check the heat flux integration through solid to fluid boundary. It should be equal to 45kW in my case I think, I am not sure if this method is right. Another question is could you please tell me how to make the imbalances a convergence criteria.

3) I apply my soure data to subdomain by 3D interpolation function, the souce data are lie in between points. You said I won't get my intended total power out of the method. I have no idea what method should I take to import source data to subdomain and how to check the imported data right and is equal to 45kW.

I have no idea how to treat the heat source data of r-z uniform distribution before applying to subdomain if directly importing the source data won't get the total power.

Thank you very much in advance for your reply.

[ghorrocks]

1) So I guess the 45kW is the integral across the volume.

2) As I have no idea how you calculated this I cannot comment.

2A) The imbalances are activated on the Solver Control tab of CFX-Pre, select "Conservation Target".

3) If the heat source is a volume integral as you describe then your approach using a source term with an interpolation function should work.

If you could describe how you calculate the total heat applied in the post processor that would be good. please attach some images of the mesh and geometry.

3)

[cxfer]

1) Yes, 45kW is the integral across the volume.

2) I use CFD-Post to calculate the total heat. I calculate the 'Heat Flux' in location of 'heat source domain interface' by using 'areaInt' Function. The Area Integral of Heat Flux on Domain Interface is 252200[W], which is larger than 45000[W] nearly 5.6 times. I have no idea why. I am not sure if the implementation is right.
(The procedure is: click 'Calculators tab', then double click 'Function Calculator', select "areaInt" Function, the location select 'heat source domain interface', variable select 'Heat Flux'; Direction is 'None'. Then click calculate button could get the results.)

3) ps: Although I tried Google Chrome browser, Microsoft Edge browser and IE browser, I cannot attach images by clicking the image insert button. I have no idea why there is not any respond by clicking it. Besides, any buttons do not respond by clicking above my edit box.

[Opaque]

Originally Posted by cxfer

Dear all,

There is a .csv file (12000 lines) for power density distribution discrete data points in cylindrical coordinate and contains four columns, which are r, phi,z and power density (W/cm3), and the whole power is 45kW. My case is water cooling around the solid boundary.

I want to import the data to subdomain as heat source, and have no idea how to treat the data. I try to directly import the data as three dimensions interpolation function in CFX, but the calculation result for CFD post show that the power is 251kW passed from solid to water boundary and this is incorrect I think. I have no idea what the problem is and what is the correct way to load the heat source data for subdomain.

Any help will be appreciated and thanks in advance.

Not sure where to start.
1 - Do you understand how the sub-domain source approach works?
2 - Any questions after reading the documentation about sub-domains in the theory/modeling guides documentation?
3 - Do you understand that by imposing a source per unit volume at discrete points in space you are creating a solution that is mesh dependent? Are you using the same mesh used to calculate those power densities, and the associated volumes are identical?
4 - Have you tried a simple problem to understand by trial and error how it seems to be working? Say a single point.

[cxfer]

Thanks for your suggestions, I will try. Maybe the mesh are not identical is the problem.

[ghorrocks]

Attaching images and attachments to the forum is an FAQ: https://www.cfd-online.com/Wiki/Ansy...n_the_forum.3F

[cxfer]

Originally Posted by Opaque

Not sure where to start.
1 - Do you understand how the sub-domain source approach works?
2 - Any questions after reading the documentation about sub-domains in the theory/modeling guides documentation?
3 - Do you understand that by imposing a source per unit volume at discrete points in space you are creating a solution that is mesh dependent? Are you using the same mesh used to calculate those power densities, and the associated volumes are identical?
4 - Have you tried a simple problem to understand by trial and error how it seems to be working? Say a single point.

Hi, Opaque and ghorrocks:

Thanks for your reply.

I have two more questions to consult you:

1)for subdomain:
I tried to change the mesh to be nearly identifical follow the instructions, however nothing change occured. The power is not equal to the input power, and is 225kW calculated by CFD post software, which is 5 times larger than input power densities (W/cm^3) over volumn (45kW). I have attached the output file and the mesh pictures in the attachment, could you please do me a favour to check.

The whole geometry range: z:-20~55cm; r:0~16cm;
Subdomain in sm: heatsm (the highlight part in picture), range: z:0~55cm;r:0~10cm;

If I should also add subdomain in fstcu domain to include the whole geometry. I have tried many times and have no idea how to do next, thanks in advance.

2)for source point:
I tried source point (one point location: x=0,y=0,z=0), the input power of the source point is equal to the power calculated in CFD post software. My question is if source point can only set power [unit:W] for only one point each time. How to cope with many discrete points for 3D in a file? Is it possible? Thanks.

[ghorrocks]

Make sure you understand the questions Opaque has quoted. The sub-domain approach is suitable for a continuous heat source applied over an entire volume, while the source points approach is suitable for when you have heat applied at one (or many) points. Do a test simulation using both of these approaches so you understand what they do.

Once you have decided what approach you want to use then we can look at any total power errors, and how to calculate it.

[cxfer]

Thanks for your reply.

Sorry, I don't express clearly.
Yes, the heat source in my data file is continious heat source distributions (heat generations) over a volumn and therefore the unit is W/cm3. The picture I uploaded is the result of continious heat source assign to a subdomain named "heatsm". The subdomain is highlighted in the uploaded picture by mesh.

Although adjusting the mesh to be nearly identical to the data distributions in input file, the calculated integrated power by cfd post software is not equal to the input, which is 5 times larger than 45kW. I have no idea the reason. I uploaded the output file and the result pictures in my previous post. Please help to check, thanks in advance.

[ghorrocks]

Your output file is full of clipping warnings. You are not going to get an accurate simulation with those in there, so you will have to fix that.

I recommend you do a simple model with a heat source to test things and work out how to accurately model this. Until you can get a simple model to work you have no hope on a complex model like this.