# Flow Simulation with Temperature dependent transport properties

• 48 Views
• Last Post 10 January 2019
ashiadarsh posted this 08 January 2019

Hi All,

I was trying to simulate the flow of water through a curved rectangular duct at a pressure of 6.9 Mpa subjected to variable heat flux boundary condition..here all the physical properties -density,thermal conductivity,viscosity and specific heat are functions of temperature and I have used NIST- real gas model for water in the liquid state.While trying to run the simulation i am getting error-pls refer to the image attached. What could be the problem?

I got a converged solution with the same BCs with constant physical properties.

rwoolhou posted this 08 January 2019

I doubt you need the Real Gas Model, just alter the water density with temperature using piecewise unless the pressure also varies significantly. Please can you copy the error text into a post: I can't read it.

• Liked by
abenhadj posted this 08 January 2019

If you do only have to account temperature dependency then follow what my colleague rwoolhou suggested.

The error message is telling that some properties are not defined within the prescribed range. For each property there is a certain range where properties might be deduced.

Best regards,

Amine

• Liked by
ashiadarsh posted this 10 January 2019

Sir, you mean piece wise linear for density?I tried with that option and the solution is proceeding...what i noticed is that if i am approximating density to higher order polynomial,for better accuracy(say 2nd degree polynomial) i am not able to proceed with the solution..solver is showing error...it worked fine for piece wise linear,inputting values of density at different temp,data taken from NIST.

Can you suggest how many data points we should taken,say for example for a temperature range of 100?

Also for other physical properties like Viscosity,Specific heat and thermal conductivity,can i approx. with higher order polynomial(utpto 4 degree if i m fitting i m getting close to the actual values) ?

rwoolhou posted this 10 January 2019

I suspect you've found the problem that many face using polynomials. If your coefficients are very high/low then slight rounding errors make the value of the equation change dramatically for small changes in the values used.  It's not necessarily a case of the number of points, but using enough significant figures in the polynomial.