PID tuning - forced hot air

[PlasticsDude 0]

What is the general idea regarding PID tuning for a forced hot air system (air blowing through an electric resistance heater, thermocouple at the outlet of this heater). Do I want to use P, I, and D, or maybe not the I, or not the D...and what about gain? Thanks for any help.

[geniusintraining 2]

Try here http://www.controlguru.com/ and here http://www.learncontrol.com/tutorial/ the funny thing is...I have been reading the past week or so on this site (listed above) and I think on my next PID setup, I will try and develop a reasoning for my madness and try the logical approach, not the WAG theory, you guy's are starting to rub off on me.... Edited 17 Mar 2007 by geniusintraining

[Peter Nachtwey 15]

Good. There is no need to guess. I would pay more attention to the controlguru site than the other site. The controlguru's is rooted in basic and sound control theory. The other (most) sites that push ZN are usually just copying some info from some other source and don't really understand the underlying theory. I can mathematically derive the equations that the controlguru uses. The equations are not empirically derived like ZN. That is why I point people to the control station site.

[Gerry 13]

A couple more questions... is the air flow constant? is the inlet air temperature constant? Answers are probably "no". A calculated feed forward taking into account the inlet temp and air flow should do most of the work.

[PlasticsDude 0]

Airflow is constant. Inlet temperature is ambient factory temp (close enough to constant).

[mugs 0]

I'm almost complete with a forced hot air PID system, in a very low flow application. I used P and I. An older Omron, C200HS PID instruction, with P at 600 and I at 100, no D. The system works perfectly. My problem is that I am trying to control a downstream temperature at around 1000 degrees, but the heater cannot exceed 1300, and I have about a 300 degree loss between the two points. I'm trying to find a solution from a controls standpoint, while my partner is looking at reducing the massive heat loss. Its only over about 12", but its not easy, at least for us...

[paulengr 44]

Based on your description, you are running very near the full output of your system. A PID controller is a linear device inherently. But you are making it operate probably some place close to a nonlinear situation, or at least running in saturation (output=100%) most of the time. At best you are going to have problems.