CFD Online Discussion Forums

CFD Online Discussion Forums (
-   STAR-CCM+ (
-   -   Simulation of air condition in room (

Love May 9, 2012 14:25

Simulation of air condition in room

we are two students working on a project regarding thermal comfort in a classroom at our school. Since we are new at this kind of simulations we have encountered a few problems.

Some background about the model:

A 3D model of the classroom was created. Due to lack of virtual memory on the school's computers we could only manage to mesh the model with five persons inside. Therefore we use the back of the seats as heat sources since they have almost the same surface area as a human.

Initially we set all human (or seats) to a constant temperature of 28 C assuming that the temperature of a human is relatively constant. The inlets were set as velocity inlets with mesured speeds and temperature (1 m/s, 18 C). The outlet to flow split with a ratio of 1. The surrounding walls, floor and cieling were defined as adiabatic. The window, projector and the lamps were also set to constant temperature.

The physical model used was
steady state
gas (air)
constant density
turbulence model; k-epsilon
coupled energy

The simultation converged and the result was fairly good. The only problem is that each person (seat) does not create enough heat to make the airtemperature close to the one mesured.

Therefore we tried to set each constat temperature to a heatflux source instead (which was actually what we tried in the very beginning since a human doing classroom activities generate around 70 w/m2). However this simulation gives impossible big temperatures, in sizes of 500 C, no matter what we do the rest of the boundaries in terms of setting them to constant temperatures. The high temperatures are located in the cells closest to the sources and the outlet.

Is there anyone who have done similar simulations and could give us hint on how to proceed from here. Should we choose another physical model?

Best regards,

(You might find my name quite funny but it's quite common in Sweden)

abdul099 May 9, 2012 15:14

You didn't take natural convection into account, didn't you? Not sure if that will solve all your issues, but I think that might have an impact on the solution.

Love May 10, 2012 13:33

Hi, thank you for your quick respons. Unfortunately using the natural convection model does not change the problem with high temperatures when applying heat flux into the model.

Best regards

Guava Wang May 15, 2012 01:12

hi,check heat flux 70 w/m2 and the related heat exchange area. Confirm whether the magnitude is suitable.

Felggv May 23, 2012 13:40

I'm also new to this kind of simulation, but as we're all here to learn, I'll risk to suggest you: maybe you could set a simmetry plane on the "humans" and put as many students as you need to make it more realistic, I think this will require less computer effort.

Sorry if it sounds dumb.

abdul099 May 23, 2012 17:02

A symmetry plane on the human bodies doesn't allow to take the energy put from the humans into the room into account. It is way more realistic to set the human bodies to a fixed temperature or specify the heat flux.

All times are GMT -4. The time now is 18:27.