| Author | Message | |||
     Sander Meijering (sander) New member Username: sander Post Number: 1 Registered: 06-2008 |
In the sample float_zone a external heatsource is used. I was wondering if its possible to define multiple steady sources within the region with cartesian coordinates. | |||
| Marek Nelson (mgnelson) Moderator Username: mgnelson Post Number: 43 Registered: 07-2007 |
The sample 'float_zone' actually uses an internal heat source in the PDE. It is modeling the internal absorption of RF energy and the dissipation as heat through the boundary. An external heat source would be applied as a boundary condition, not a volume source in the PDE. In this sample we made the source term in the PDE time-dependent in order to simulate the moving RF coil. You can define it to be stationary. You can have as many sources as you like. They can be defined as a source in the PDE or as a boundary condition. Source terms in the PDE can also be redefined by REGION if that is desirable. See "Regional Parameter Values" in the Help index. There is nothing magic about cylindrical coordinates here. This was chosen as the coordinate system because it most easily represented the domain of the rod. You can use Cartesian coordinates. | |||
| Sander Meijering (sander) New member Username: sander Post Number: 2 Registered: 06-2008 |
Thank you for your answer, but its not totaly clear yet. In my case I want to model concrete core activation, so I want pipes with hot or cold water (the sources) in a concrete floor. I tried it with regions, but it doesn't give a sattisfying result. Say I a piece of concrete and make it like: start(0,0) line to (4,0) line to (4,2) line to (0,2) line to close and I would like to define 2 stationary. sources, one on (1,1) and one on (3,1). I can imagine my PDE looks like: div(k*grad(temp)) + source_1 + source_2 = cp*dt(temp) But what would my sources look like? How can I get those cordinates in the defenitions? Thanks again for your quick response!! | |||
| Sander Meijering (sander) Junior Member Username: sander Post Number: 3 Registered: 06-2008 |
This is the original code. It works fine until t = 20 000. The temperatuur goes to far up.
| |||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1131 Registered: 06-2003 |
It's a little hard to determine what you want in this problem. You appear to have an environmental temperature heating the slab from top and bottom (T_ruimte=24), with cooling pipes embedded in the concrete (source=19.5). There is a heat flux at the top boundary of alpha_beton_plafond * (t_ruimte-temp), with the units of energy per unit area per unit time. A similar heat flux applies to the bottom surface. Your "natural" BC on the small pipes creates a flux on the pipe wall that attempts to drive the wall temperature to "source". This flux may or may not be able to actually achieve the target temperature, depending on other sources and sinks in the problem. It also allows the temperature to vary around the circumference of the pipe. I don't know what kind of physical mechanism could generate heat on the pipe surface. I suspect that what you really want is to specify the temperature on the wall of the pipes. This would be Value(temp)=source. Also: 1) Your Threshold statement is wildly wrong. The variation of temperatures within the problem is 4.5 degrees, and you have said that variations less than 1100 need not be modeled accurately! See Threshold in the Help Index. 2) With the parameters you have supplied, the time evolution of the system is completed in 100 time units. The results are unchanging for the remaining 71900 time units. 3) Since the entire evolution is over in 100 time units, an initial timestep of 60 is much too large. It simply jumps over most of the action. Use 0.01 instead. 3) If the walls of the pipes are to be held at a constant temperature (as I infer they are), you can simply apply the Value BC and eliminate the interior of the pipes from the domain. This is done simply by tracing the boundaries as part of the region 1 bounding paths. 4) You don't need separate regions for multiple patches (bubbles) with the same material properties. You can combine them into a single region. I have made these modifications in the attached script.
| |||
| Sander Meijering (sander) Member Username: sander Post Number: 4 Registered: 06-2008 |
Thank you, you've helped me a lot already. My main problem isn's solved yet. What I want to simulate is shutting of the cooling system and watch what the concrete does. So the coolingenergy into the system must be stopped. It can't be done by changing my value(temp) = 19.5 into value(temp) = 0, because it would mean that the temperature drops to zero. Do you have any suggestions for that problem? Could it be done by putting energy into the system? I would like it to be a source like the one in float_zone. Thanks for your help. | |||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1132 Registered: 06-2003 |
1. There is an error in your original script that we didn't catch last time: The heat flow equation is cp*dt(Temp)+Div(flux)=source, with flux=-k*grad(temp). The K does not belong inside the grad as you have presented it, but outside, as in div(k* grad(temp)) + source = rho * cp * dt(temp) 2. There are several ways to implement a source that you can turn off. (The only one you can't do is to switch explicitly from a VALUE to a NATURAL condition at some problem time. FlexPDE won't let you change the category of boundary condition.) a) in the framework of the bubledeck_1 file I posted earlier, you can use a Natural BC throughout the time range, but switch the flux at the required time. In your problem, NATURAL(temp) means the outward normal component of k*grad(temp), or inward flux. The statement NATURAL(temp)=Large*(T0-temp) will drive the boundary temperature toward T0 with a flux proportional to Large. It can therefore be used to simulate a VALUE BC. NATURAL(temp) = IF(t<tshutoff) THEN Large*(T0-temp) ELSE 0 will hold the boundary temperature at T0 until time tshutoff, then insulate the wall. b) you could include the material inside the pipes in the domain and program a volume source inside the pipe ("source" in the equation above). Source = IF(t<tshutoff) THEN Large*(T0-temp) ELSE 0 will hold the temperature inside the pipe at T0 until time tshutoff, then let the pipe material cool due to heat loss to the environment. |
