TRANSFER FUNCTION FOR HEAVY AND THICK WALLS

2 posts / 0 new
Last post

Dear colleagues,

As you probably know, Conduction Transfer Function (CTF) method does not handle thick and massive walls, because the accuracy of the CTF coefficients is degraded. In particular, in DOE-2, you get the following message:

"Cannot simulate LAYERS .... as given. The wall is too thick or too dense"

I have heard of a possible solution to solve this problem by dividing the wall into thin layers. I have made an attempt in e-quest, but the problem remains, and I do not know how to proceed.

Any insight on this issue would be appreciated.

Sincerely,

LPL

-------------------------------------------------------------
Luis P?rez-Lombard
Dr. Ingeniero Industrial
Profesor Contratado Doctor
Departamento de Ingenier?a Energ?tica
Escuela Superior de Ingenieros
Universidad de Sevilla
Camino de los Descubrimientos s/n
41092 SEVILLA
Tfno.: 95.448.72.56
Fax: 95.446.31.53
e-mail: lpl at us.es
-------------------------------------------------------------

lpl's picture
lpl
Offline
Joined: 2011-10-01
Reputation: 0

I've answered the same question when you posted it on BLDG-SIM, which I'm repeating here
in case there are others who may have different ideas. BTW, what is the wall layer you're
modeling that does not converge within 100 time steps (> 4 days!). I've only run up
against this limit when I was trying to model the foundation layer with lots of dirt.

Joe

------- posted May 16, 2014 on BLDG-SIM ---
Luis,

I'm not sure if I have an answer for your problem, but I want to point out some technical
details that may be being overlooked.

1. DOE-2 does not use CTFs, it uses Response Factors, which shares the same approach as
CTFs but uses a different formulation.

2. When DOE-2 says it cannot simulate a layer because it's too thick, that's not because
the RF methodology fails but because DOE-2 has a maximum of 100 time-steps, i.e., 100
hours, for the three response factors (X,Y,Z) to converge to constant COMMON RATIOs that
are used to compute the residual responses from that point on.

3. There are two situations that could cause such convergence problems - (a) the layer is
simply too massive. I know from trial and error that the thickest single-layer of dirt
allowed in DOE-2 is 5.25 ft (1.6 meters), (b) this thickness will be greatly reduced if
there are multiple layers with contrasting thermal properties; again, from
trial-and-error, if there is a layer of insulation below the dirt, the maximum allowable
thickness of the dirt layer will be reduced by half or more. You've mentioned that
someone has mentioned creating multiple thin layers. This will not help with (a), but
could help with (b), provided that the thin layers are graduated in their thermal
properties; I've never tried this so I can't tell how much benefit would there be, but my
guess is not very much.

3. I don't know EnergyPlus well enough to say whether it has a similar hard limit on
convergence and if so, what is it. If not, you might be able to model the same layer in
EnergyPlus and not encounter any difficulties.

4. If all else fails, as others have already mentioned, you can do an explicit solution
using the Finite Difference Method in EnergyPlus, but expect that to slow down your
simulations considerably.

Joe
------------------------------------------

Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 108D
Moraga CA 94556
yjhuang at whiteboxtechnologies.com
http://weather.whiteboxtechnologies.com for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 406