Another PID thread...

Alen bradley

Member
A
Join Date
Apr 2013
Location
qatar
Posts
23
Hi all,

I have pretty much read all the PID threads on this forum which helped me a lot. I am working with the following system and trying to get a stable PV.

The PID is used for water flow control to a mixer. I am not been able to get a stable PV response as you can see in the Figure. The yellow is the calculated setpoint and green is the process variable.
ran95e.jpg

The first thought that would come in to mind is that the PV is responding to changing SP, but even if I keep SP constant, the PV would be oscillating. I am not too sure which control action to change. Will setting up a dead band help?
Following is the PID control Block.
29zoepw.png


Thanks
 
Last edited:
I'm not sure what your definition of stable is , but that looks pretty good to me.

Especially for flow control.

I would setup the PID to use engineering units for the set-point and process variable.

You would have to set values for the setpoint min/max in the PID configuration. (Smax and Smin)
 
Mickey said:
I'm not sure what your definition of stable is , but that looks pretty good to me.

Especially for flow control.

I would setup the PID to use engineering units for the set-point and process variable.

You would have to set values for the setpoint min/max in the PID configuration. (Smax and Smin)
Click to expand...

Thanks for feedback! Have a look at attached picture, I have a more accurate data over long period of time now.

2rp9tso.png


Yellow is setpoint and green is PV. Data is over a period of 2 hours.
 
Last edited:
As others have stated your control looks great and your PV seems to be within +-1 of you set point in your units. This is actually good. What is the accuracy +- of your instrument? I would suspect your instrument may just need some dampening. Either in the instrument itself if it is capable, or you could use a smoothing routine on your input to the PLC. If your CV is stable and not overreacting, and your QA on your end product is good, I'd say leave well enough alone, you will drive yourself nuts to find a "perfect" tune. But again, your only as good as your instrument reading the process in loop tuning anyway.
 
I think it is remarkably stable given that the setpoint is constantly changing.

Why is the setpoint constantly readjusted by being 'calculated'?
 
danw said:
I think it is remarkably stable given that the setpoint is constantly changing.

Why is the setpoint constantly readjusted by being 'calculated'?
Click to expand...

It is a setpoint calculated by the program based on how much material is in the mixer, basically the more the material, the more the setpoint for water flow in to mixer. Will having a constant setpoint over a certain time period help:smas:
 
Yes, a more constant set point would be very useful for getting things smoothed out. That said, I would be happy that the PV is tracking as well as it is given the fluctuation in the set point.

Bubba.
 
I don't know if there is a way to do it with that PID block, but for flow control I like to limit the rate of change of the PID output. In some cases, it looks like your setpoint jumps, which triggers an instant jump in the PID output and causes some instability. Also, when your setpoint jumps like that, it starts the 'I' component winding up and can lead to overshoot if you don't have some kind of windup limit.
 
If you do need to have a fluctuating setpoint, I would suggest using a ramp or a low pass filter to gradually change the value, thereby avoiding the jumping around. This will avoid the issues CapinWinky mentions above.
 
Finally we find out why the set point jumps around.

Alen bradley said:
It is a setpoint calculated by the program based on how much material is in the mixer, basically the more the material, the more the setpoint for water flow in to mixer. Will having a constant setpoint over a certain time period help:smas:
Click to expand...
I would use a low pass filter to smooth the set point.
Yes, having a smooth set point will make a big difference if you are using the derivative gain.
 
Why is you PV so notchy? Does you feedback device have very low resolution? Is your PV in floating pt or integer? If you are trying to control to better than 1, you need floating point don't you? What is the time we are looking at on the chart?
 
Peter Nachtwey said:
I would use a low pass filter to smooth the set point.
Yes, having a smooth set point will make a big difference if you are using the derivative gain.
Click to expand...


Peter,
I'm almost afraid to ask since I usually lack the technical knowledge to understand 90% of what you post, but I can't afford to be ignorant forever. So here goes nothing. :unsure:

You seem to be saying that an unsteady set point will have a greater effect if the derivative term is in play. My question is, wouldn't the end result be similar with PID or PI control? To my eyes the set point changes shown in the second graph are not drastic enough to introduce a great deal of instability with either PI or PID control. Am I missing something specific, or just not grasping the concept in general?

Thanks,
Bubba.
 

Similar Topics

J
Another PID thread
Hi Guys, I have been trying to tune a PID for a tank heating system in RSLogix5000. The system is a reasonably large volume of water at about...
Replies
8
Views
2,158
JonAW
J
A
Limit PID output based on another PID output
Hi all, I posted recently about a HVAC-style process control application I'm working on, and have a related question that probably warrants its...
Replies
12
Views
3,905
ASF
A
D
yet another pid question
This is my first post here...so HI I own a small injection molding operation (thermosets) and I've had good success replacing the relay logic in...
Replies
9
Views
2,378
dougpdf
D
wildswing
Curious: What if I copy a PID tag to another PID tag in RSL5000
Just curious, trying to make set up of 20 PID tags identical. What if on first scan I COPY MyPIDtag01 to MyPIDtag02 Length 1 COPY MyPIDtag01 to...
Replies
4
Views
2,917
wildswing
wildswing
M
Sorry, another PID question!
Hi, I've been searching the Internet for days now and there's so much confusing information about PID terms. I'm not new to the general concept...
Replies
11
Views
2,606
JRW
JRW
Back
Top Bottom