[TRNSYS-users] PID controller parameters setting - Cooling coils Type508
George Kyriakarakos
gk at aua.gr
Wed Mar 30 00:17:26 PDT 2022
Dear Clement,
Setting the right PID parameters can be tricky if you do it by hand and trial and error runs.
On the other hand it is a very straight forward optimization problem. I would suggest that you couple TRNSYS with GenOPT and use Particle Swarm Optimization.
If you have license for TRNOPT then this is pretty easy to setup just follow TRNOPT manual, if not you will have to follow the following guide:
https://sel.me.wisc.edu/trnsys/demos/genopt-type56.pdf
This guide was prepared since you cannot optimize type56 variables through TRNOPT, but it can be used as a base to understand how to couple TRNSYS and GenOPT if TRNOPT is not available.
As far as PSO parameters are concerned you can use the ones from the publication below – they work really well with engineering problems. You can also get insights on how to formulate cost functions for the optimization process which include technical constraints. For your problem you just need an equation for the outlet air temperature oscillation which you will try to minimize.
https://www.sciencedirect.com/science/article/abs/pii/S0196890415005610
Since you have gone down the TRNSYS route, I would strongly advise you to learn how to do optimizations with heuristics. Even though your current problem is an easy one, optimization through PSO which can handle up to 100 variables both discrete and continuous is an invaluable tool in anything that you might want to do in the future.
Best regards,
George
--
Dr. George Kyriakarakos, Agricultural Engineer
PhD, MSc Energy Systems & Renewable Energy Sources
Agricultural University of Athens,
Dept. of Natural Resources and Agricultural Engineering
Iera odos street, 75, Athens 11855, Greece
Tel. +30.210.5294046 (direct), Fax:+30.210.5294023
Mobile +30.6942.046895
e-mail: <mailto:gk at aua.gr> gk at aua.gr
Skype: georgekyriak
Linkedin: http://www.linkedin.com/in/georgekyr
From: TRNSYS-users <trnsys-users-bounces at lists.onebuilding.org> On Behalf Of Clement Nevers via TRNSYS-users
Sent: Tuesday, March 29, 2022 4:20 PM
To: trnsys-users at lists.onebuilding.org
Cc: Clement Nevers <nevers at insa-toulouse.fr>
Subject: [TRNSYS-users] PID controller parameters setting - Cooling coils Type508
Hi everyone,
I have encountered some problems when the PID controller(type23) was used for controling the outlet air temperature of cooling coil (type508).
The fluid flow rate of cooling coil is conected with a outlet of flow diverter (type11). Thus, the fluid flow rate of cooling coil is determined by the control signal of the flow diverter. I build this mechanism in order to model a cooling cooling with a bypass in fluid side (3-ways valve flow controlling).
Then, the PID controller is used for keeping the outlet air temperature (Controlled variable) of cooling coil at a setpoint, the control signal is basically the fluid proportion of the flow diverter.
But no matter how to adjust the parameters of PID, the outlet air temperature oscillation cannot be eliminated.
So, I have the following questions :
How to determine the parameters of PID controller, such as Gain constant, Integral time and Derivative time, to avoid Controlled variable oscillation ?
Is there a better way to accurately model this system?
Thanks in advance,
Regards
NEVERS Clément
Research Assistant - LMDC, Toulouse
mail : nevers at insa-toulouse.fr <mailto:nevers at etud.insa-toulouse.fr>
Département du Génie Civil
INSA Toulouse
135 avenue de Rangueil
31077 Toulouse Cedex 4
tel : +33 (0)5 61 55 95 13
web : www.insa-toulouse.fr <http://www.insa-toulouse.fr/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.onebuilding.org/pipermail/trnsys-users-onebuilding.org/attachments/20220330/b52e9daa/attachment.htm>
More information about the TRNSYS-users
mailing list