Easy Way To Tune Pid

aju

Member
Join Date
Jan 2005
Posts
22
We Always Found Out That Tuning Pid Always Create Some Problem Can Any Body Suggest Easy Way To Tune Pid With Less Efforts An Dexact Way By Consuming Less Time
 
PID Tuning

I find it fairly simple as follows:
First set I and D gain values to zero or minimum value.
Adjust P gain to get an acceptable response.
Gradually increase I only enough to elliminate any offset.
Re-adjust P gain to get desired response.
Re-adjust I gain if you have any offset.
Add some D if you find it provides any improvement (usually it won't)
Thats it.
Regards..............John Gaunt
 
John is correct, I was taught by an expert in PID, although he was used to mechanical PID, as john said start with no I & D, make the prop band large & reduce until it goes through stable & then becomes unstable, then introduce I & if required D, however I have never found that the derivitive is required or even made much difference.
 
aju said:
We Always Found Out That Tuning Pid Always Create Some Problem
Yes, beware of the so called experts. Many have experience with only a particular system and think the way they tuned that system applies to all. How can one give advice unless they know what the system is?

aju said:
Can Any Body Suggest Easy Way To Tune Pid With Less Efforts An Dexact Way By Consuming Less Time

Perhaps. So what kind of system are you tuning? For the system I get involved with I write auto tuning software so my customers are spared most of the pain and agony.

John Gaunt said:
Add some D if you find it provides any improvement.
What does the derivative gain do? What improvement are you looking for?

John Gaunt said:
(usually it won't)
Why not?
 
Peter Nachtwey said:

From www.simplesolvers.com

--- snip ---
[font=Verdana, Arial, Helvetica, sans-serif]Many control systems will, in fact, work quite well with only Proportional Control. Integral Control examines the offset of SETPOINT and the PROCESS VARIABLE over time and corrects it when and if necessary. Derivative Control monitors the rate of change of the PROCESS VARIABLE and consequently makes changes to the OUTPUT VARIABLE to accomodate unusual changes.
--- snip ---

I guess it depends on what your working on. I work on an offshore production platform and we use Fisher pneumatic controllers on almost everything. Many only have a proportional band control but a few have a time delay adjustment for repeatability. ( would this be I, D, or neither ? )

I looked this up because of this discussion. I look forward to gleaning some good information from you guys.

Thanks,

Mike

[/font]
 
Don't believe everything you read on the internet.

castlerock said:
From www.simplesolvers.com

--- snip ---
[font=Verdana, Arial, Helvetica, sans-serif]Many control systems will, in fact, work quite well with only Proportional Control.[/font]
[font=Verdana, Arial, Helvetica, sans-serif]
[/font]
Which systems? I know temperature systems, and velocity systems don't work properly with proportional only control. I do mainly position control applications. I know that proportional gain alone is unsatisfactory so I wonder what these simpleminds are sovling. It bothers me there is so much bad information out there. It is obvious to me these guys don't have experience in these types of systems and are making general statements that are misleading to rookies.

castlerock said:
[font=Verdana, Arial, Helvetica, sans-serif]I guess it depends on what your working on.[/font]
[font=Verdana, Arial, Helvetica, sans-serif]
[/font]
YES!
 
Peter Nachtwey said:
[/size]
Which systems? I know temperature systems, and velocity systems don't work properly with proportional only control. I do mainly position control applications. I know that proportional gain alone is unsatisfactory so I wonder what these simpleminds are sovling. It bothers me there is so much bad information out there. It is obvious to me these guys don't have experience in these types of systems and are making general statements that are misleading to rookies.

All of our Fisher controllers control levels and pressures. We do however have a Fisher temperature controller on a glycol reboiler but we rarely, if ever, touch it because it just works.

But I'm glad you mentioned temperature systems. We have emersion heaters to heat water for heat exchangers ( to do away with fired vessels for separation ) that are controled via PLC-5. I'll get our programmer to explain them to me the next time he's in the field.


To Mickey:
Thanks for the link. I'll be checking it out.


Mike
 
Last edited:
rsdoran said:
This may be of interest, especially if you have/use Matlab.
http://www.engin.umich.edu/group/ctm/PID/PID.html
Examples are available at the bottom of the page.
Links have been posted to this site many times. Does anyone know what it means?

For instance, in the proportional control section, what values of Kp cause the system to be:
1. Over damped
2. Critically damped
3. Under damped
 
I can not profess to be PID qualified in any fashion...plus I ma kind of slow so would have to think for a bit before answering the questions.

What gets me is nooone ever just states...WHAT IS P, I, and D! They usually go off into details about a specific process/application with formulas...BUT what are the values used?

Here is my VERY simple concept of PID.
P is the present, the output is a percentage proportional to the difference (error)
I handles the past, the error is summed over time and multiplied by a constant I which is summed with the proportional output.
D is for the future, it uses the rate of change over time multiplied by a constant to be summed with P and/or I.

As for using P only the critical aspect is the band, as it can not have a zero error.

I play around with this at least once a week using different things. I have heated water, anti-freeze, oil and chambers employing different options. I have done some things using pneumatics. I have not done much recently with speed, velocity, or positioning due to lack of equipment.

I read everything...AS Peter mentioned some of the Internet info can be WRONG, just plain wrong. I understand more than I know and/or can explain but in due time.....
 
a plain english version

I'm no PID expert by a long shot, many of the post are way out of my league.

But I use this explaination for the mechanics at work:

P= how far out am I
I= how long have I been here
D= how fast am I going

The processor looks at these parameters, applies tuning , and then calculates an output.
 
Peter Nachtwey said:
Links have been posted to this site many times. Does anyone know what it means?

For instance, in the proportional control section, what values of Kp cause the system to be:
1. Over damped
2. Critically damped
3. Under damped

1. Too much P, overshoot, ie overtemp
2. Just right amount of P. On a step change, the PV typically has a small overshoot, but comes right back to the setpoint.
3. Not enough P. PV doesn't meet the SP, ie, doesn't come up to temp.
 
I will take a quick stab at this...It has bee a long time since playing with this. My vibrations book is at work so I can't check this.
1. Dampening Ratio >1
2. " " =1
3. " " <1
 
One Step at a time. I am only asking about proportional gain.

CroCop said:
1. Too much P, overshoot, ie overtemp
Over shoot, Right. Overtemp? Wrong.

CroCop said:
2. Just right amount of P. On a step change, the PV typically has a small overshoot, but comes right back to the setpoint.
Wrong. Critically damped means no over shoot.

CroCop said:
3. Not enough P. PV doesn't meet the SP, ie, doesn't come up to temp.
It depends upon the system doesn't it? On temperature and velocity systems you are right. On a position control system you would be wrong if it weren't for static friction.

Bob O said:
I will take a quick stab at this...It has bee a long time since playing with this. My vibrations book is at work so I can't check this.
1. Dampening Ratio >1
2. " " =1
3. " " <1
Bob, you are right about the damping factor. However, what values of Kp are required to achieve those damping factors?

What values of Kp, for proportional control system in RSDoran's link ,make the system critically, under and over damped?
 
Last edited:

Similar Topics

See picture. I want to add a rung (magenta) into the existing code. Can't figure out how to do this. I select a -||- , right? When I drag/drop...
Replies
21
Views
1,690
Hi All, On my site, the standard template for storing recipes in the PLC is to create a tag called PartRecipe which is an 2D array - for example...
Replies
4
Views
1,430
I would like to do change a kinetix drive, so i believe i just have to put the IP address in the new kinetix and just connect right? Do I need to...
Replies
1
Views
1,175
Hey all, I am trying to check in logix designer in emulator and I need to check the hmi alarms by toggling bits to activate an alarm . But...
Replies
4
Views
2,054
Good afternoon guys, I have a basic question on the bubble level in the picture attached. This is probably really easy, but I don't get it, haha...
Replies
11
Views
3,265
Back
Top Bottom