RE: Controler - old strategy - on off to early

[Michaël Kummert <kummert@xxxxxxxxxxxxxxx>]

Arko,

A short answer is: You should not use Type 2 without a deadband (what you
did by choosing Input5=Input6=0).
In your case, choosing a reasonably large deadband should do the job, e.g:
dTHigh (Input 5) = +2
dTLow  (Input 6) = -2
You probably did not try negative values for Input 6. Setting dTLow equal to
-2 means that the controller should switch OFF when (Th-Tlow), i.e.
(Tset-Tzone) is lower than -2, i.e. when the zone temperature is 2°C higher
than the setpoint.

A more technical explanation:

Type 2 uses TRNSYS iterations to calculate the control signal. If the
building walls are cold, you can end up with the controller oscillating
around a "wrong" value with your settings. You can easily understand what's
happening using the "TRACE" feature (right-click on a component, select "Add
TRACE" and look to the listing file after running the simulation).

Here is an example of what you could end-up with (It assumes Type 56 is
before Type 2 in the deck file and Nstick=3, but any odd value would give
the same results):

Time | Iter | Tzone | On/Off
   1 |    0 |  20.3 |   1
   1 |    1 |  24.9 |   1
 --- Convergence ----------------
   2 |    0 |  26.7 |   0
   2 |    1 |  23.3 |   1
   2 |    2 |  26.7 |   0
   2 |    3 |  23.3 |   0 (Nstick)
--- Convergence -----------------
   3 |    0 |  20.0 |   1
   3 |    1 |  24.7 |   1
--- Convergence -----------------
   4 |    0 |  26.9 |   0
   4 |    1 |  23.7 |   1
   4 |    2 |  26.7 |   0
   4 |    3 |  23.7 |   0 (Nstick)
--- Convergence -----------------
   5 |    0 |  20.2 |   1
   5 |    1 |  24.8 |   1
--- Convergence -----------------
   6 |    0 |  27.0 |   0
   6 |    1 |  24.7 |   1
   6 |    2 |  27.0 |   0
   6 |    3 |  24.0 |   0 (Nstick)
... Etc...

The end-of-timestep values for the zone temperature will be:
24.9 - 23.3 - 24.7 - 23.7 - 24.8 - 24.0, etc.
So the zone temperature does not reach 26°C
Note however that if you maintain steady-state inputs for the Type 56, you
will eventually reach the setpoint and even overshoot it.

Some other thoughts:

- Even without deadband, you could force the
  controller to oscillate around the right value by
  using the value of Tzone at the last time step and
  not the value at current time step. You can do that
  by inserting a Type 93 between Type 56 and Type 2.

- You could use a better (smoother) controller,
  unless your objective is precisely to study
  the effect of the controller you used.

- You might want to reduce the simulation time step
  to reduce the oscillations.


Regards,

Michaël


_________________________________________________________

Michaël Kummert

Solar Energy Laboratory - University of Wisconsin-Madison
1303 Engr Res Bldg, 1500 Engineering Drive
Madison, WI 53706

Tel: +1 (608) 263-1589
Fax: +1 (608) 262-8464
E-mail: kummert@sel.me.wisc.edu

SEL Web Site: http://sel.me.wisc.edu
TRNSYS Web Site: http://sel.me.wisc.edu/trnsys




> -----Original Message-----
> From: owner-trnsys@relay.doit.wisc.edu
> [mailto:owner-trnsys@relay.doit.wisc.edu] On Behalf Of hgjh kgkgf
> Sent: Monday, October 14, 2002 23:38
> To: trnsys@relay.doit.wisc.edu
> Subject: Controler - old strategy - on off to early
>
>
> Dear TRNSYS users,
> I would like to discuss the behavior of the
> On-Off-Controler.
> I  want to have a roomtemperature in a shower of 26
> Celsius in wintertime. The airflow is 12 /h. The
> supply air temperature is 32 Celsius.
> The Controller is set to the following parameters and
> inputs:
> 1 	2
> 5	100
> INPUTS:
> 1	2 	3	4	5	6
> 26	10	20	0	0	0
>
> I did the loop with the Input control function, no
> monitoring temperature is external supplied and the
> Lower Input temp is connected to type 56.
>
> Regardless, if I change the Oscillation from 5 to 19,
> or if I change the upper dead band to 2, the
> Controller starts to go on and off before the room
> reaches 21 Celsius.
>
> I would like to know the reason for that. Is it
> possible, that the Controller want to reach the
> desired roomtemperature through a curve like an
> e-function? Or are there Osscillations not showern,
> because the airflow is so hight, that I could reach
> the desired temperature at once?
>
> Thank you for any ideas!
>
> Arko