[TRNSYS-users] Problems with convergence - Problems with input/output

[knut.erik.enerstvedt at niva.no]

Robin,

I can only give you some general advice about convergence stability.

1) Add some pipes in your fluid loop (Type 31, for example). This will 
increase the thermal mass and make the system more stable.

2) If your controller makes decisions based on variables that can change 
during the iteration process, I would suggest using for example the Type 2 
controller instead. If you need to use your own equation, you can use the 
(converged) values from the previous time step (Type 93 - Utility/Input 
value recall) as inputs to the controller.

3) Try decreasing the time step even further. 0.3 hr is not a very short 
time step.  A controller can make only one decision per time step, and a 
lot can happen in 18 minutes. 


Hope that some of this will help to solve your problem. 

Kind regards,
Knut Erik









 
Dear TRNSYS-users

The following mail is divided into two sections: I.) Problems and II.) 
simulation description:


I.) THE PROBLEMS I HAVE WITH MY SIMULATION:
--------------------------------------------------------------

FIRST PROBLEM:
- The Warning "Message 441: The inputs to the listed units have not 
converged at this timestep. The values at the last iteration will be used 
and the simulation will proceed." appears after every single timestep. 
Every unit except the plotters are listed.

APPROACH TO THE FIRST PROBLEM:
-  I already tried variing the timestep - I decreased it up to 0.3hr and 
increased it to 2hr. That doesn't help
- I increased the "tolerance convergence" up to 5! That helped, the 
warnings are gone, but that is like cheating. Furthermore the results I 
receive differ very much from these I get when the "tolerance convergence" 
is left at 0.001.
- an employee of trnsys suggested to work with Type2 (Differntial 
controller with hysteresis) but I can't use a hysteresis in my simulation


SECOND PROBLEM:
- The outlet temperature of "Type11h-4" and the outlet tempereature of 
"V3" are not the same. (FOR MORE INFORMATION JUST CONTINUE READING)

APPROACH TO THE FIRST PROBLEM:
- I wonder if it's because of the convergence problems I have. 

--------------------------------------------------------------

II.) SIMULATION DESCRIPTION:

The simulation is about a chilled water circle. I added two links to show 
i.) the circle without the controller and ii.) to show the circle WITH the 
controller.

i.) 
http://mitglied.lycos.de/ITF/TRNSYS/NEWNEW_chilled_water_cycle_without_controller_shown.jpg


ii.) http://mitglied.lycos.de/ITF/TRNSYS/NEWNEW_chilled_water_cycle.jpg


The main components of the simulation are
a.) a load (TYPE 682) increases inlet temperature: Toutload = Tinload + 
Q_load / (M*cp)
b.) an absorption chiller (TYPE229) decreases inlet temperature: 
Toutchiller = Tinchiller - Q_chiller / (M*cp)
c.) a storage tank for the chilled water (TYPE 60, one inlet, one outlet)
d.) 4 Valves (Type11f)
e.) 4 Tee-Pieces (Type11h)
f.) my controller (TYPE258) 

The outputs of my controller are:
- One Signal for each Valve
- A set massflow which is the input of "V3"
- Cooling demand which is an input for the absorption chiller.


The Component order follows the massflow:

- External cooling demand = Q_load (input for my controller and the load)
- Controller 258
- V3 
- absorption chiller
- Type 11h3
- V1
- storage
- V4
- Type 11h
- V2
- Load
- Type 11h2
- Type 11h4

The outputs of every component are: outlet temperature, outlet massflow

--------------------------------------------------------------

I hope despite of the long text you can help me with my two problems above 
or give me some helpful advice. Maybe you have some general advice for me, 
what I should take care of.
If you need more information, like source code etc. don't hesitate to ask.

best regards,

Robin Sluzalek




------------------------------------------------------------------------------------------------------------
Meld deg på vårt nyhetsbrev på www.niva.no
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.onebuilding.org/pipermail/trnsys-users-onebuilding.org/attachments/20091030/5f4a38fd/attachment-0001.htm>