Re: [TRNSYS-users] Controlling heating/cooling equipment

[Michaël Kummert <michael.kummert@gmail.com>]

Knut,

I often use Type 23 (PID) for systems such as the one you describe. If it is
acting like an on-off controller you probably need to tune the parameters,
and possibly reduce the time step. In my experience, for cases such as the
one you mentioned, you can sometimes do with a time step of 15 min or so if
you accept some excursions above/below the setpoint but rarely with one
hour. The required time step depends on the system but for most HVAC systems
I tend to use something between 1 and 15 minutes in "temperature control"
mode. 
The PID parameter tuning can be time consuming but by starting with a low
gain, no derivative action and an integral time equal to the time step you
can usually get a valid starting point (if somewhat sluggish). Then you just
get a cup of coffee (or sometimes a large pot), find a typical situation in
which you want to achieve a good control (e.g. one day), and start playing
with the P, I and D constants. One useful thing is to look at outputs 3, 4,
5 and 6 of the PID. They give the unsaturated control signal and the
breakdown between the proportional, integral and derivative parts. This can
help you understand the impact of changes in the parameters or why it is not
doing what you think it should do.  

Type 22 was an attempt to provide a controller that would not require fine
tuning but I have to say that I have only been able to use it successfully
in a few cases where there is no/little thermal mass in the controlled loop
- it also has a strong tendency to blow-up simulations. In other words I
tend not to use it. 

Hope this helps,

Michaël Kummert
ESRU - University of Strathclyde


> -----Original Message-----
> From: knut.erik.enerstvedt@niva.no
> [mailto:knut.erik.enerstvedt@niva.no]
> Sent: 19 September 2008 08:37
> To: trnsys-users@engr.wisc.edu
> Subject: [TRNSYS-users] Controlling heating/cooling equipment
> 
> Dear all,
> 
> I wish to simulate a particular cooling system, but find it difficult
> to
> control it the way I want.
> 
> I use Type 56 in "temperature level control", as the main objective is
> to
> simulate operation of the cooling equipment. The equipment delivers
> cooling
> to the building through chilled water and a fan coil, and the amount of
> cooling delivered is controlled by adjusting the water flow rate
> (variable
> speed pump, control signal between 0 (min) and 1 (max)). Therefore I
> need a
> controller with a continuous control signal, and not just an on/off
> signal.
> I wish to have, as far as possible, a continuous and smooth operation
> of
> the system during the period of day when cooling is needed.
> 
> Type 23 PID controller is not ideal to use for this task, as its
> control
> signal is based on the difference between the zone temperature and a
> set
> point temperature and goes to zero immediately when the zone is cooled
> to
> below the set point. Then at the next time step the zone temperature is
> above the set point and cooling is turned on again and so on. This
> results
> in a lot of oscillating on/off operation of the system, and I wish to
> avoid
> this.
> 
> I have tried Type 22 iterative controller, hoping to achieve just the
> right
> amount of cooling at each time step, but have encountered some
> difficulties. The pump flow rate can not be too low, so I need to use a
> "threshold for non-zero value" of about 0.3. This causes the pump to
> never
> start at all, even when its control signal obviously should be above
> 0.3. I
> also find it difficult to get TRNSYS to converge properly when using
> Type
> 22.
> 
> I will be grateful if anyone has some advice on this.
> 
> 
> Kind regards,
> Knut Erik Enerstvedt
> 
> 
> 
> 
> 
> -----------------------------------------------------------------------
> -----------------------------------------------
> NIVAs hovedkontor har flyttet til nye lokaler i CIENS -
> Forskningssenter
> for miljø og samfunn; Gaustadalléen 21, 0349 Oslo. Meld deg på vårt
> nyhetsbrev på www.niva.no
> 
> 
> _______________________________________________
> TRNSYS-users mailing list
> TRNSYS-users@cae.wisc.edu
> https://www-old.cae.wisc.edu/mailman/listinfo/trnsys-users