[anon_l]

Hi all,

I am trying to enter the density and specific heat of H2O as a polynomial fumction in star ccm+ since I want to make a simulation for super critical conditions.
I use the data from NIST(http://webbook.nist.gov/cgi/fluid.cg...s&STUnit=N%2Fm)
AND then , I use EXCEL to generate a polynomial function of order 6 for my data .

However , after defining the polynomial function in star ccm+ , I get a weird curve . I am including a picture of that in this post .
Can someone let me know how I can fix this.

Thanks

[kirrer]

I'm not sure you're taking the best approach to simulate super critical fluid, though I don't have the answer how to do it. But, I will share my experience in Excel. Whenever doing any curve fit, ensure that your coefficients don't fall into the Coefficient Round Off Error. I just made that term up, but if your X axis values are large, the coefficients like in your plot are shown rounded off to 1 significant figure with a rather large exponent - like 3E-12 instead of 2.875e-12. Depending on the polynomial order, that rounding might be OK but usually it is not. In your case, you get extreme negative values which make no sense. To make more sense of the error, try to plot the data in Excel (one column is temp, the other has the equation from your fit curve programmed in). You'll likely see similar behavior for the super critical part.

I believe CCM+ has other models which would better handle density in the super critical state than the polynomial function of temperature; I'll let others chime in on that (or check the manual / check with your support engineer). But in general, I would avoid using a 6th order (or even more than 3rd order) polynomial due to the unexpected behavior you've seen, as well as the inherent waviness you show in your 2nd image for the super critical part. Your 6th order function doesn't even follow the given data, so I would imagine it would create some simulation errors, even if you could get it to load in CCM+.

[anon_l]

Hi Kevin @kirrer ,
Thank you very much for your suggestions . I tried making the polynomials to 2nd degree and now my simulation doesnt give any errors .
The only problem that I am now left with is , that my temperature doesnt converge and my residuals show some weird oscillating behaviour , which I am not sure if it's correct ?

[kirrer]

I'm not familiar with super critical fluid simulation, so I can't give you pointers there. But, you would definitely want to provide more information for anyone that can help. For example, what are you trying to simulate? (steady flow, transient motion, turbulent mixing, etc - show a picture of possible). Give some idea of the boundary conditions. Explain the oscillating behavior - at some point, all residuals oscillate, but depending on the value of the oscillations (or the oscillations of engineering quantities), it may or may not be an issue.

There are some real gas models which are more applicable to what you're studying, though they are significantly more in-depth and would require a much greater understanding of the physics and the default model settings to produce an accurate answer - check the documentation for Real Gas models.

[anon_l]

Thanks again @kirrer kevin ,

I am trying to simulate supercritical water , flowing through a simple tube geometry .
The conditions are -
1)inlet h20 temp = 35.4 degree C , outlet temp = 356 degree C
2)mass flow= 587.6 gm/m^2-s
3) heat power = heat flux at inner surface of tube = 0.623 MW/m^2
4)I also use heat flux at the outer wall with a little higher value than the inner wall of the tube
5)pressure = 25 MPA
6)K- epsilon turbulence
7)steady

I have attached some pictures of my model, mesh and residuals for your reference .

Thank you very much

[kirrer]

If you are specifying heat flux at the wall (inner surface) of the tube, why do you need the solid model of the tube at all? Also, if you specify heat flux at the ID of the tube, then also specifying it at the OD could be numerically unstable. I recommend you simulate the interface as an interface, and only apply the heat flux at the outside of the tube. As for the physics models themselves, you're venturing into territory where the default values of H2O vapor likely don't apply (viscosity, specific heat, conductivity), and whatever answer you are looking for (pressure drop, heat transfer, temperature, etc) will change depending on those values. I don't have further advice, except to carefully specify the properties which would affect your answer and to reconsider using real gas models or talking with your support specialist. Of course, others on the board may have better advice than I, so you can wait for a response there as well.

[wassli]

Piecewise polynomials can be fitted because i am also doing work on supercritical freon. you just have to increase the precision of coefficients of polynomials.
1.png

2.png

3.jpg

4.jpg

[anon_l]

@wassli , thanks very much for sharing this , my work with star ccm is almost done now .
But, I will keep your suggestion in mind if I work again with this software in the future .
Thanks

[NeedHelpPls]

Excuse me, I just started to learn STAR CCM+, can anybody explain me how to enter polinomial density? I need to enter this polynom: y =-0,0000325783*x3-0,0000933926*x2-1,2194573442*x+872,4591125572.

Thanks!