PID Control

J

Thread Starter

jmorgan

I have an A/B SLC 5/03 trying to tune a PID loop on a MAU. Trying to maintain 68degF DAT I have a gas burner but controling the air mixture damper instead of the gas valve. My system is laggy (slow acting) I over shoot by 8degF with the settings that I have in now. Gain 70, Ti 60, Td 6.9. What happens, since this system is slow,is when I tell the damper to close and go to low fire there is approx. betwen 10-25 sec before this occurs. In the mean time the heat exchanger heats up so I have additonal heat blowing into the space that I have no control over when in low fire. How do I take this into consideration when trying to tune this loop? Any suggestions would greatly be appreciated.
 
M
I assume you are controlling the Mixed Air Temp ( OA + RA) via a Mixing box, If so use a MA temp sensor via a PID loop maintain that @ around 55
Deg F. The use the DAT with a Space sensor (if possible, as reset) via a PID loop to maintain a DAT. The space will reset the loop and helps from
over shooting. If the space is large, for example a large production area, then look to use the Return Air sensor, this is generally
representative of an average of the space temp.
Feel free to e-mail if you run into any other problems

Mark Massa
BCS
 
M
ELIMINATE any Derivative. Lengthen your integral term.

You have a dominant transport lag and Derivative will induce an oscillation. I agree Derivative should always be used for temp control but you have a transport medium and this changes things significantly.

Trying to control an adiabatic process is next to impossible. I suggest you play with things like output rate limit algorithms but make sure it ties in with the Integrator or you will have integral windup which also causes overshoot.

Mitch
 
B

Blunier, Mark

> I have an A/B SLC 5/03 trying to tune a PID loop on a MAU. Trying to
> maintain 68degF DAT I have a gas burner but controling the air mixture
> damper instead of the gas valve.

I'm not sure what you mean by controlling the air mixture instead of the gas valve. Something has to be controlling the gas valve. If the gas
valve is controlling based off the air flow, the air flow will need to change slow, or the air fuel ratio will go to unsafe ratios. This is
both a dangerous setup, and one the makes tight temperature control unlikely. A safer, and better control is to send to send your signal
to the air damper, and use the same signal (appropriately scaled) to the gas valve control as feedforword signal, so the pid on the gas valve
trims the flow.

> My system is laggy (slow acting) I over shoot by 8degF with the
> settings that I have in now. Gain 70, Ti 60, Td 6.9. What happens,
> since this system is slow,is when I tell the damper to close and go to
> low fire there is approx. betwen 10-25 sec before this occurs.

Without knowing your temperate spans your gains don't mean a lot, but your numbers look wrong. Instead of gain do you mean band? A gain of 70 is
enormous. A band of 70 is small. An integral gain of 60 (min?) is huge, and would be very slow acting.

> In the mean time the heat exchanger heats up so I have additional
> heat blowing into the space that I have no control over when in low
> fire. How do I take this into consideration when trying to tune
> this loop? Any suggestions would greatly be appreciated.

You could also use a cascade control, where the outlet temp controls the temperature of the inlet to the heat exchanger, or the temperature
difference across the heat exchanger, and then use that signal to to control the damper.

Mark Blunier
Any opinions expressed in this message are not necessarily those of the company.
 
M

Michel Lemieux

Did you solve your problem ?

If not I might have a couple off comments & questions.

What is the controlled end device ? Is it the gas valve ???

What type of setup do you have ( Type of sensor & their location, type of controller )

A little schematic also helps.

Are your value expressed in the following
Gain = PB ( proportional band )
IT = Integral time in minutes or seconds ?

One of the most common case of hard to control supply temperature with a gas burner is that the equipment selected is oversized for the application. This makes stable operation in low demand situation almost impossible since the
minimum threshold of operation of the unit is already above the actual demand. The minimum threshold is defined by parameters as minimum fire requirements & minimum stack flue velocity.

Also, the real range of operation of a gas valve should not be calculated in control signal input or position, but rather in pressure variation. With a pressure gauge after the valve you will find out that minimum pressure to maximum pressure happens within the first third of the stroke or less. Re-scaling the PID output within that rotation range will help the time reaction of the system a lot.

I have seen some case when low demand control case was so problematic, that they added a small electric heater with SCR's until demand range crossed in the true modulating range of the gas unit. As an example some gas humidifier will
modulate their capacity from 50 to 100% with full control span input. There is no modulation below 50 %

Hope this helps.

 
A
The best way to get very fast response is to use the feed forward or bias at the output of the PID controller. If you know where you want
the controller to go on a given transient then position it there with the output bias and let the PID algorithm trim the fast open loop
signal.
 
Top