[esp-r] Re: simulation message - non-convergence of air flow networks
Geissler Achim
achim.geissler at fhnw.ch
Wed Oct 30 16:38:27 GMT 2013
Hi Arianna,
interesting network. I have never used the component type "self regulating vent". In case you still have not resolved your problems, I would start out with replacing these vents by cracks or simply deleting them (if each zone then still has a flow path to exterior). Then try the network again.
Depending on how these vents actually work, they may induce flow instabilities in the network. Also, assuming your doors to the hallway are all open, you have huge openings here to a probably limited volume (hallway). "specific air flow door" is, I think, the bi-directional flow component. Again, this configuration may lead to instabilities. Again, you could replace the doors with "cracks beneath closed doors" and try again.
When you write you have tried shorter time steps, what time steps do you mean? Ten minutes? Less? Have you tried really short time steps below a minute?
When you start a simulation via prj in interactive mode you can toggle the parameters for air flow convergence. In general, the residue parameters are not very strict, however, each time step has only 100 iterations max. If you get the warning each and every time step, then it will probably not help to increase this. If the warning pops up only occasionally, then it may help. If running bps via script,
=======
bps -file $CONFIG -mode text > ${RESULTS}.scratch <<XXX
.
.
.
XXX
==
you can append something like
==
# Postprocessing
MONTHLIST="1 2 3 4 5 6 7 8 9 10 11 12"
for MON in $MONTHLIST
do
case $MON in
1|3|5|7|8|10|12)
DAYS=31 ;;
2|4|6|9|11)
DAYS=30 ;;
esac
FAILED=`grep -c "Month $MON" ${RESULTS}.scratch`
FRACTION=`echo "scale = 2 ; 100* $FAILED / (24*$DAYS*$NTSTEP)" | bc`
echo "Timesteps failed in $MON: $FAILED, fraction of simulation time steps: $FRACTION %" >> ${RESULTS}_convergence-failed.lst
echo "===== convergence failure timestep list =======" >> ${RESULTS}_convergence-failed.lst
grep "Month $MON" ${RESULTS}.scratch >> ${RESULTS}_convergence-failed.lst
echo "===== ================================= =======" >> ${RESULTS}_convergence-failed.lst
echo " " >> ${RESULTS}_convergence-failed.lst
#rm ${CONFIG}.scratch
done
=======
to the shell script to generate a report on how often the warning actually occurs.
Hope this helps on a bit further.
Best,
Achim
> -----Ursprüngliche Nachricht-----
> Von: esp-r-bounces at lists.strath.ac.uk [mailto:esp-r-bounces at lists.strath.ac.uk] Im
> Auftrag von Arianna Sdei
> Gesendet: Montag, 28. Oktober 2013 18:29
> An: Jon Hand; Achim Geissler; Aizaz Samuel
> Cc: esp-r at lists.strath.ac.uk
> Betreff: [esp-r] Re: simulation message - non-convergence of air flow networks
>
> OK Jon,
>
> thank you! I have made a clear drawing of my network and I have just printed out
> the flow network file in attachment. At the end on the file there are a series of nul
> values, do you think that could be the problem? How can I solve it? Once I have
> checked that every node and component has a connection and a flow path to a
> boundary pressure how can I check that the iteration convergence control is not
> too strict?
>
> Let me know,
>
> All the best,
>
> Arianna
>
> -----Original Message-----
> From: Jon Hand [mailto:jon at esru.strath.ac.uk]
> Sent: 28 October 2013 16:52
> To: Arianna Sdei; Achim Geissler; Aizaz Samuel
> Cc: esp-r at lists.strath.ac.uk
> Subject: RE: [esp-r] Re: simulation message - non-convergence of air flow
> networks
>
>
> It is ALWAYS GOOD to clearly draw (on paper) your flow network with every name
> of every entity in the network file shown on your sketch. Use coloured pens to
> draw the connections between nodes and components.
>
> If you are controlling components - like a window - if that is the only connection
> between the room and the outside then you should also have a crack defined
> between the room node and the boundary node.
>
> And print the flow network file or the section of the model contents report related
> to network flow and control. Go though each item and each connection and
> ensure that the sketch and the file is consistent.
>
> Regards, Jon Hand
>
> ________________________________________
> From: esp-r-bounces at lists.strath.ac.uk [esp-r-bounces at lists.strath.ac.uk] on
> behalf of Arianna Sdei [A.Sdei at brighton.ac.uk]
> Sent: 28 October 2013 16:39
> To: Achim Geissler; Aizaz Samuel
> Cc: esp-r at lists.strath.ac.uk
> Subject: [esp-r] Re: simulation message - non-convergence of air flow
> networks
>
> Hi Achim,
>
> thank you for that! How can I do convergence checking? Is there any way ESP-r
> can automatically check if there is a problem with the airflow network? I have tried
> with smaller time steps but no change. Let me know,
>
> Regards,
> Arianna
>
> From: esp-r-bounces at lists.strath.ac.uk [mailto:esp-r-bounces at lists.strath.ac.uk]
> On Behalf Of Achim Geissler
> Sent: 27 October 2013 16:47
> To: Aizaz Samuel
> Cc: esp-r at lists.strath.ac.uk
> Subject: [esp-r] Re: simulation message - non-convergence of air flow networks
>
> Dear Arianna,
>
> in addition to the below, it sometimes helps to reduce simulation time steps
> (increase number of steps per hour). Also, you can change the parameters used
> for convergence checking (either every run manually or in the code and recompile).
>
> Best
> Achim
>
>
>
> On Oct 22, 2013, at 9:04 AM, Aizaz Samuel
> <aizaz.a.samuel at strath.ac.uk<mailto:aizaz.a.samuel at strath.ac.uk>> wrote:
>
>
> Arianna,
>
> Somewhere in your model there is mass imbalance that is more than the tolerance
> allowed within the solver. Usually happens if you close a flow path using control
> (e.g. window closes but air is still being blown into the room from a fan). I would
> recommend to check this by removing control and then running the model. Another
> way forward is to make a sketch of the flow network and visually check if a bad
> flow path occurs given the constraints of the control you are trying to impose. A
> usual solution is to always have a flow path from internal nodes to external ones
> via small cracks (which is correct for most situations).
>
> Another (rarer) reason is that you may have two components in series that have
> hugely different flow resistances.
>
> Regards,
> Aizaz
>
> On 22/10/13 12:22, Arianna Sdei wrote:
> Dear All,
>
> I am simulating using an air flow network and while running the simulation I get the
> following message:
>
> Please check (in the following order) that:
> -each node has some flow path to a boundary pressure and -iteration convergence
> control is not too strict
>
> Can someone help me to interpret what that means and what should I look at in the
> model? Thank you,
>
> Arianna
>
> Arianna Sdei
>
> Research Fellow IFORE (Innovation For Renewal)
>
> UNIVERSITY OF BRIGHTON
>
> Moulsecoomb, Lewes Road, Brighton, BN2 4AT
>
> direct line: 0044 (0)2036747268
>
> mobile: 0044 (0)7981739024
>
>
>
> ___________________________________________________________
> This email has been scanned by MessageLabs' Email Security System on behalf
> of the University of Brighton.
> For more information see http://www.brighton.ac.uk/is/spam/
> ___________________________________________________________
>
>
>
>
> _______________________________________________
>
> esp-r mailing list
>
> esp-r at lists.strath.ac.uk<mailto:esp-r at lists.strath.ac.uk>
>
> http://lists.strath.ac.uk/mailman/listinfo/esp-r
>
> _______________________________________________
> esp-r mailing list
> esp-r at lists.strath.ac.uk<mailto:esp-r at lists.strath.ac.uk>
> http://lists.strath.ac.uk/mailman/listinfo/esp-r
>
> achim.geissler at intergga.ch<mailto:achim.geissler at intergga.ch>
>
>
>
>
>
> ___________________________________________________________
> This email has been scanned by MessageLabs' Email Security System on behalf
> of the University of Brighton.
> For more information see http://www.brighton.ac.uk/is/spam/
> ___________________________________________________________
>
> ___________________________________________________________
> This email has been scanned by MessageLabs' Email Security System on behalf
> of the University of Brighton.
> For more information see http://www.brighton.ac.uk/is/spam/
> ___________________________________________________________
>
> ___________________________________________________________
> This email has been scanned by MessageLabs' Email Security System on behalf
> of the University of Brighton.
> For more information see http://www.brighton.ac.uk/is/spam/
> ___________________________________________________________
>
> ___________________________________________________________
> This email has been scanned by MessageLabs' Email Security
> System on behalf of the University of Brighton.
> For more information see http://www.brighton.ac.uk/is/spam/
> ___________________________________________________________
More information about the esp-r
mailing list