[esp-r] Re: Underrelaxing by using Timestep-controller 4

[Jon Hand]

Several users I talked to who have large openings first change the building
timestep to one minute and only if that does not work do they consider other options.
Since we have not had a notification about controller-4 recently it could be that
no one has tried controller 4 for some time.

For two rooms which would tend to drive lots of air movement between them I
would suggest that you try the one minute timestep without the controller 4. That
should result in much smaller oscillations and might just work.

Later we might have a chance to do a single-step debug to see how and where
the variable itrapas is being set and unset. 

-Jon Hand
________________________________________
From: esp-r-bounces at lists.strath.ac.uk [esp-r-bounces at lists.strath.ac.uk] On Behalf Of Pueltz, Gunter [Gunter.Pueltz at MuellerBBM.de]
Sent: 04 May 2011 12:39
To: esp-r at lists.strath.ac.uk
Subject: [esp-r]  Underrelaxing by using Timestep-controller 4

Dear ESP-specialists and developer,

for a simulation project with 2 big rooms, which are connected by a large opening and thus show an intensive bidirectional air flow (= intensive air exchange between the 2 zones, see attached screenshot Simulation-model.png), I am using Timestep-controller 4 and a small timestep (10 min).
If I do NOT use the underrelaxation-method for the new air-flow-values (the default in source file mfmach.F is, that the do-loop 2222 for underrelaxing the new air-flow-values is commented out; see mFmach.F line 522 - 530 resp. attached screenshot mFmach-line522-530.png), ESP struggles with very intensive oscillations (see attached screenshot results-without-relaxation.png).
If this underrelaxation do-loop 2222 is activated (relaxation-factor = 0.25), these oscillations do not appear anymore (see attached screenshot results-with-relaxation.png).
The first thing I want to point out is, that these results induce NOT to comment out this underrelaxation do-loop 2222 in future versions of ESP-r !

If the do-loop 2222 is activated manually, a second problem occurs unfortunately :
The boolean variable itrpas should be greater zero, if timestep-controller 4 is used (see mfmach.F, line 523). But I found, that itrpas ALWAYS is zero, even while using timestep-controller 4 !?!?! For me it seems, that itrpas is not set to its correct value ...
I would like to recommend the ESP-developers to fix this bug ;-)

Moreover I am wondering that only one, single user of ESP-r seems to face this problem with oscillations, which are caused by large openings between two thermal domains within a building, inducing an intensive air exchange ... ???

-Gunter



--

Dr. Gunter Pültz

Müller-BBM GmbH
Robert-Koch-Straße 11
82152 Planegg
Telefon +49(89)85602-336
Telefax +49(89)85602-111
E-Mail Gunter.Pueltz at MuellerBBM.de
http://www.MuellerBBM.de

HRB München 86143
Geschäftsführer: Horst Christian Gass, Bernd Grözinger,
Dr. Carl-Christian Hantschk, Stefan Schierer, Dr. Edwin Schorer, Norbert Suritsch