[TRNSYS-users] Time dependency

[Kim Goethals <kim.goethals@ugent.be>]

Dear TRNSYS-users,


I am performing a simulation of a typical office room which is cooled by 
night ventilation. To study the influence of convective heat transfer I 
implement convective heat transfer correlations. Since I got a 
convergence error doing so, I recall the needed parameter values from 
the previous time step to calculate the convective heat transfer 
coefficient. To get a time-independent solution I decreased the time 
step gradually. However, when I use a time step of 0.5 or 1s the results 
deviate seriously. Therefore, I have completely stripped my model to 
eliminate the influence of controls of installations, schedules etc. 
Moreover, I have used constant convective heat transfer coefficients. A 
graph of the indoor temperature shows that the simulation with a time 
step of 0.5s doesn’t start from the initial value while the simulation 
with 1s approaches the temperature profile of 0.5s. Does anyone have an 
explanation?


Best regards,

Kim

begin:vcard
fn:Kim Goethals
n:Goethals;Kim
org:Universiteit Gent;Vakgroep Architectuur & Stedenbouw
adr;quoted-printable:;;Plateaustraat 22;Gent;Oost-Vlaanderen;9000;Belgi=C3=AB
title:ir.-arch.
tel;work:092647861
tel;fax:092644185
url:www.architectuur.ugent.be
version:2.1
end:vcard