Simscape Thermal Liquid Pipe Outlet Temperature Discrepancy

Hello,

I am attempting to learn SimScape and validate my understanding against various book examples. One such example is that in Fundamentals of Heat Transfer 7th Edition by Incropera Dewitt. I found a picture of the problem setup here:

Although the book example has the goal of determining the convection coefficient of the tube and specifies the outlet temperature directly, I am interested in calculating the outlet temperature instead. I wrote a code which does this and I was careful to use the same correlations used by the Pipe (TL) block as stated in its documentation (for example Haaland's equation for friction and Gnielinski's correlation for Nusselt number). The book's outlet temperature is 57°C and my code produces 57.5°C which is agreeable to me.

Using the same script to drive a Simscape model I built for this problem I only achieve a temperature of 48°C on the outlet. I believe that my SimScape model is correctly built to replicate this problem but I cannot find the source of the discrepancy. I feel that an error of 9°C is a bit large for errors in fluid properties, etc.

Can someone please verify I am not incorrect in my approach?

[Phil U]

Dear cmmcnamara,

I found your post while trying to solve a very similar example. Like you, I want to compare the results of the Simscape pipe (from the thermal liquid library) with the results of a textbook. According to VDI Heatatlas (p. 697 – 698), the outlet temperature of a 1 m long pipe with a 10mm diameter should be 52.2 °C assuming a constant wall temperature of 100 °C and an inlet temperature boundary condition of 10 °C in combination with a massflow 0.5 m/s.

Leaving all settings initial, I receive an outlet temperature of 40.12 °C, which is an unacceptable difference compared to the textbook solution. However, I found out that the value of the “internal surface absolute roughness” (in the “Friction and Heat Transfer” setting of the pipe) seems to be of major importance for the result. In other words, if I play around with that value, I can easily fit the Simscpape model results to the analytical solution.

Another interesting result occurs if I make a subsystem with 10 pipe segments (each pipe segment has now a length of 0.1m). With the same setting, the outlet temperature is 46.34 °C. It seems like the number of elements is also of great importance for the numerical solution. It would be great if some of the Mathworks people around that forum can reply to that post. Browsing several forums across the internet, I more and more get the impression that a lot of people out there deal with the same kind of problem. With the results given above, it is difficult for me to “trust” the Simscape solution.

Here you can see the results of the pipe (only one pipe segment 1m long):

Here you can see the results of the pipe (10 segments 0.1 m each):

Answer 1

[Nicolas Schmit]

The Pipe block from the foundation domain in its actual implementation does not model temperature gradients. If you open the help of the Pipe bloc and go to “Assumptions and Limitations”, you will find the following.

Heat transfer is calculated with respect to the temperature of the fluid volume in the pipe. To model temperature gradient due to heat transfer along a long pipe, connect multiple Pipe (TL) blocks in series.

To model temperature gradients, you can either

• Connect several Pipe (TL) blocks from the Foundation domain in series.
• Use the Pipe (TL) block from Simscape Fluids > Thermal Liquid. This block has a parameter “Number of segments” that you can adjust.

[HabenG]

Interesting, this helps a lot. is there a reasonable way of segmenting a pipe to get realistic results or is it just trial and error?