Heat exchanger in ANSYS

Heat exchangers are a big part of what I dealt with during my time as an engineer. In the world heat exchangers are a big part of many things we use. Such as engines, fridges, computers and even our tablets. As their name suggests, these components are in charge of exchanging heat from one source to another. We know that heat flows in the direction of hot to cold and not the other way. The theory introduced in the heat transfer course stands true and I will calculate the error between the theory and ANSYS at the end of this analysis.

The points I will cover are as follows:

  1. Geometry
  2. Mesh
  3. Setup & calculation
  4. Results


1. Geometry

Let’s talk about the geometry…

Geometry iso
Figure 1 – Isometric view of the heat exchanger

The 3D model was created in SolidWorks since it is much easier for me to sketch in SolidWorks. At first, you would be tempted to say there are 2 parts in the assembly we see in figure 1. However, as preparation for the ANSYS analysis, there are 4 parts here. a) The heat exchanger casing, b) the fluid outside of the pipe and inside the casing, c) the pipe itself and d) the fluid inside the pipe. The reason for this is the way ANSYS recognizes domains and how we define the type of materials and where they dwell.

pipe iso
Figure 2 – The solid shell of the pipe

In figure 2 the highlighted geometry, which will be assigned a hollow pipe made of solid material, running through the heat exchanger shell. The interference checker was used to calculate whether or not there are any clashes between solid entities. In ANSYS we will also insert a contact region between the fluid inside the pipe and the solid pipe. As well as between the pipe outer wall and the fluid inside the heat exchanger shell.

pipe fluid iso
Figure 3 – Fluid domain inside the pipe

Not much to add, This fluid domain serves to simulate the fluid found inside the pipe, the hot fluid is usually put through the inner pipe, which is what I did later on as well.

Shell fluid
Figure 4 – Heat exchanger shell fluid domain

This fluid subdomain represents the cold water that surrounds the smaller pipe, we can easily see the inlet and outlet (arbitrary) and the rest which is not the hot water pipe nor the fluid domain is the outer shell.

Figure 5 – engineering drawing of the geometry


I’d like to close the geometry section with measurements. The cold water outlet\inlet diameter is 2.5 [mm]. The hot water inlet\outlet is 5 [mm]. The shell’s outer diameter is 34 [mm], and a thickness of 2 [mm]. One thing to note is that this drawing is not complete, it is missing the inner part of the tube, which is the cold water pipe.

2. Mesh

From here on I continue my work in ANSYS. The meshing tool is many folds better and the main reason is control. I can control the mesh boundaries and behavior, I can control how fast it expands, I can make the shape I want it to take. Thus make sure the spots that are important for the analysis are well defined and have a dense enough mesh. Or at least, the closest to what I would want to have. Whereas in SolidWorks the control is a scale based controller and the manual controls are very arbitrary and do not allow true FEA-esque analysis.

After switching to ANSYS I realized that there’s no need for the big shell of the heat exchanger because the nature of the analysis is of internal flow rather than external flow. Something interesting to consider for the future is to perform the internal flow analysis and get the temperatures for the outer shell. Then take those temperatures and use them on an external flow analysis for the shell. But I digress.

Figure 6 – Mesh closeup on the interaction region

The areas that are important for me in this analysis is the interaction between the hot water fluid domain and the pipe around it, and between the pipe and the fluid domain of the heat exchanger shell. That doesn’t mean, however, that other areas are not important, they are just less important, which means they get less of a dense mesh as their values are less significant. As you see the protruding, greenish-coloured part has a very nice and orderly outer mesh. This part will be designated as a solid and will be interfaced with fluid around it. In figure 6 we also see how dense the mesh is around the top nozzle and the fluid part of the heat exchanger shell (in rectangles) around the interface with the solid pipe.

Figure 7 – Front view of the cold water pipe and fluid sub-domain

What’s important to note here is the density of the mesh around the cold water pipe, a very fine and appealing mesh exists there. And the fact that if you cross a line between the outer diameter of the pipe and the inner diameter of the pipe you are going to cross at least 3 elements, sometimes more. This is important because of the way the results are calculated, which is through averages. And the inner part which is 2.5 [mm] in diameter, I get good enough gradient for the radial conduction, which will result in a more precise solution.


Figure 8 – Top view of the heat exchanger


This showcases more on a quantitive plane than a quality characteristic. Specifically, you notice the ratio between the elements on the far side of the shell_domain and the nozzle at the top, for example.


Figure 9 – Top view close-up on hot water inlet\outlet


By now, as usual, notice the density of the mesh around the edges. This would be imperative were I to add the outer shell, but as of now, it is present for the good manner of meshing more than a contribution to a more precise solution.


Figure 10 – Top view of the Copper pipe


Figure 11 – cross section view, fluid – solid – fluid mesh interaction

I don’t have much to say about figure 10. Figure 11 however, is something else, first I made a cross section with section plane function. The picture I got was a bit confusing, a lot of partial elements and a weird Image. I learned today that there’s a complete elements option which gives me this cool topography. What it does is it either subtracts or adds the elements I cut to allow the closest representation to what I asked the cross section to be.


section plane
The pyramid is the whole element showcase



Figure 12 – Cross section view of the mesh, U pipe area




The same trend applies here, the further away we are from the interaction zone the less dense we need the mesh to be, which is represented here by the big “rocks” in the middle.

This concludes the Mesh section, I will be moving to the solution setup shortly.


3. problem setup in ANSYS


I imported Water, Copper, and Aluminium to my available materials
Copper has the thermal conductivity coefficient of 387 [W\m*K]
Aluminium has the thermal conductivity coefficient of about 202 [W\m*K]
and water is at a density of 997 [kg\m³]

Boundary conditions:

  1. Cold inlet at 1 [m/s] and 283 [K]
  2. Cold outlet has environment pressure [1 atm] = 0 gauge
  3. Hot inlet at 1 [m/s] and 353 [K]
  4. Hot outlet has environment pressure [1 atm] = 0 gauge
  5. Defined contact region between pipe_fluid_domain and pipe_shell
  6. Define contact region between Pipe_shell and Heat_exchanger_fluid_domain

Residual monitoring

  • The criteria for convergence is 0.001 for all parameters except 1*10^(-6) for energy
  • The solution will run for 1000 iterations



4. Results


Figure 13 – Temperature distribution


Absolutely brilliant.

This heat exchanger could represent, for example, a pre-process for something being done in a factory that requires hot water, let’s say they use steam to move a turbine, and after the work on the system is done, the water leaves the system. The water can be run through a heat exchanger to make the incoming water hotter, before being boiled and superheated, which makes the process more efficient.

As I stated before, the cold water enters the system at 282 [K] and the hot water enters the top nozzle at 353 [K]. What you can see in figure 13 is the interaction and gradient of temperatures in a steady state heat transfer process. The lining that is present in the cold water domain is, of course, the Copper pipe. Notice how the temperature of the pipe at the cold inlet is around 300[K] while the temperature of the pipe changes up to about 334[K] at the outlet.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s