[esp-r] Re: nodal temperatures in coupled surfaces

[Jon Hand]

A recent question about temperatures within a partition....

::I am working on a model containing two zones which are coupled by an  
::opaque construction with boundary condition "surface in other zone".  
::Since the construction is non-symmetrical, I've inverted it so that  
::the layers match. One of the zones has a large window.

::I am trying to understand how ESP-r handles coupled surfaces. Ideally,  
::the resulting nodal temperatures for both original and inverted  
::construction separating the two zones should be identical. But since  
::ESP-r solves the difference equations one zone at a time, this doesn't  
::seem to be the case and I am observing large differences between the  
::nodal temperatures as viewed from zone1 and zone2. Is this expected?

The solution technique does solve zone-by-zone.  When the boundary
condition on the 'other-side' is needed it looks for such information in
the common blocks.  The current value in the common block might be
temperatures from the current timestep or the previous timestep depending
on the order of the zones.  Normally the values will be close.

IF YOU HAVE a one hour timestep then the information being used could be 
somewhat out of date, especially if there is a rapid change in surface 
temperatures or the radiant environment.  For example, if a room with a large
window is the 4th zone and the room on the other size is the 3rd zone
the boundary condition at the other side of the 3rd zone will lag reality.
If the zones are reversed then the boundary condition on the other side
will have just been solved and the information lag is minimal.

This suggests that the model might be re-run at shorter timesteps and
this would reduce the temporal impact.  It might be that a 15 minute
timestep would be sufficient.

ALSO: be sure to use save level 4 when running the simulation so as to
record the explicit fluxes on both sides of the surface. 

Regards, Jon Hand