I have no experience with chilled ceilings,
but as I understand the manual (chapter 6, page 47), the chilled ceiling
layer has to be the first layer (inside surface) in a wall. Or, if you
have an adjacent wall, it can also be the last one (inside surface of the
adjacent zone).
I was hoping that someone else with
more experience with chilled ceilings would see your posting and give a
more helpful answer.
Best regards,
Knut Erik
From:
Mavi Cambronero Vázquez <maviqnk@gmail.com>
To:
knut.erik.enerstvedt@niva.no
Cc:
trnsys-users@cae.wisc.edu
Date:
05/13/2010 05:38 PM
Subject:
Re: [TRNSYS-users] A lot of convergence
warnings
Thanks for your answers.....
I´ve noticed that the problem is the chilled ceiling used. It´s the first
time that I try to use it and I supose I didn´t succeed. I know how to
create a chilled ceiling layer, but I´m not sure what are the following
steps. What I did was integrate the new layer as the last one in a new
wall. TRNBuild said to me that a wall with chilled ceiling can´t be external,
boundary or internal, so the only option is that it is adjacent....
adjacent to the roof???So I created a new zone called roof! I´m a
bit confused.
In addition to reducing time step length, I would try the following:
- Check if there is sufficient thermal mass in your system (add pipes in
fluid loops, for example).
- Use controllers (such as Type 2) instead of equations (LT, GT and similar)
when variables that can change from one iteration to the next are involved.
- Place components in a "natural" calling order (controllers
before controlled equipment, for example): Control cards->Component
order: Click and drag the different components to order them.
- Use an online plotter to check if there are any unrealistic output values
(very high temperatures, for example).
Best regards,
Knut Erik Enerstvedt
>You could change the simulation
time step (in steps of 1/(2^n) so 1 hour, 0,5 hour, 0,25 hour etc). You
can find this by assembly -> component >order.
>Good luck,
>Mariëlle Nuchelmans
>Hi,
>What can I do when there are lots of convergence problems??? The same
project has different warnigs because of that if the simulation >starts
at 3896 than if it starts at 0.... Does the option "Optimaze component
order" work?? If I select that option other errors appear... :(
------------------------------------------------------------------------------------------------------------
Meld deg på vårt nyhetsbrev på www.niva.no
------------------------------------------------------------------------------------------------------------
Meld deg på vårt nyhetsbrev på www.niva.no