RS Logix 5000 PIDE problem

[kardel.sharpeye]

Hi Guys,

I had a problem with PIDE function block in RS Logix 5000.
Please see the PIDE block attached.
I tried to control temperature control valve to maintain oil temperature @ 170F.
The problem was when the PV is greater than 170, the Cv output is moving slowly to certain point (around 5%) and then suddenly goes to 0, even though the PV is still above the set point.
I've tried to tune the P,I and D gain. but it didn't help.
When i set the P gain at 10, I gain at 5 and D gain at 5. The PIDE seems work allright but then suddenly it stop working because the Cv output stuck at certain value and didn't change even when the PV has changed.

Could anyone tell me how to solve this problem? Is there any mistake that i made in the PIDE?

Thanks in advance

 

[widelto]

1)what is the purpose of selector ( 2 inputs) at output ?
2) This application is for heating or cooling ?

 

[kardel.sharpeye]

1)what is the purpose of selector ( 2 inputs) at output ?
2) This application is for heating or cooling ?

1. The selector is used to choose between auto (using PID) or manual mode
2. It is for cooling application

 

[widelto]

See attached file.

Kardel:

E= SP - PV FOR HEATING
E= PV- SP FOR COOLING

Have you check this ??

 

[Ron Beaufort]

the PIDE has "built-in" controls for Auto/Manual control - so you really don't need that selector idea ...

going further - the PIDE has about 165 parameters "under the hood" ... we can't give you very detailed answers without seeing how you've got those set up ... the best thing is to post your program file (ACD) and we'll try to help you nail down an answer ...

 

[kardel.sharpeye]

Kardel:

E= SP - PV FOR HEATING
E= PV- SP FOR COOLING

Have you check this ??

Yes, it is for cooling
E = PV - SP

 

[Ron Beaufort]

set the Deadband for 0 and see if that helps ...

 

[kardel.sharpeye]

the PIDE has "built-in" controls for Auto/Manual control - so you really don't need that selector idea ...

going further - the PIDE has about 165 parameters "under the hood" ... we can't give you very detailed answers without seeing how you've got those set up ... the best thing is to post your program file (ACD) and we'll try to help you nail down an answer ...

Ron,

i had a problem uploading the file (it's around 1330 KB before zipped, 1200 KB after zipped)
Do you have any solution how to attach it?

Thanks

 

[Ron Beaufort]

try doing a FILE SAVE AS operation - and set it for an L5K type ... those are usually much smaller ...

(note: I'll be offline for awhile - but I'm sure that someone else will be glad to take a look at it) ...

good luck with your project ...

 

[kardel.sharpeye]

I've saved it in L5K and upload it as zip file (because it couldn't be uploaded as L5K file).
please check it out whenever you have a time


Thanks Ron n Widelto

 

[Ron Beaufort]

well, for the first thing, you don't have the "timing" set up correctly ... basically you've just dropped the PIDE into a Function Block routine which executes many hundreds of times per second – specifically, once each time that the Continuous (Main) Task gets executed ...

suggestion:

create a new periodic task (example: PIDE_Task) and give it a period setting of about 500 milliseconds ... DISCLAIMER: since we don't know much about the operational characteristics of your process, the values that I'm giving are GUESSES for "starting points" ... you'll probably need to adjust these values as your project progresses ...

put the PIDE's Function Block routine into a new Program - within the new task – and make sure that the routine gets scanned/executed ... (ask if you're not sure how to do that) ...

the next thing that jumps out at me is that Deadband setting of 5 that you have entered ... why are you doing that? ...

basically that setting is telling the PIDE: "once you get the inlet temperature (the PV or Process Variable) within five degrees of the target temperature (the SP or Setpoint) then FREEZE the output value (the CV or Control Variable) at its current value ...

the CV will then STAY at that same FROZEN value until the PV comes back out of the Deadband again ...

a very popular misconception is that the Deadband is something like an "acceptable range" for the PIDE to shoot for ... that's wrong ... in many (most?) cases the Deadband is used to "rest" some type of output device (for example: something like an electric gearmotor) to keep it from constantly chattering back-and-forth (and overheating) as the PV fluctuates very close to the Setpoint ...

going back to your original post:

The PIDE seems work allright but then suddenly it stop working because the Cv output stuck at certain value and didn't change even when the PV has changed.

of course I'm not there to see it for myself, but based on that description, the Deadband setting could be causing your problem ... back in Post #7 I suggested that you try turning the Deadband off – with a setting of ZERO ... did you try that – and did it keep the output from getting "stuck at a certain value" the way you described? ...

suggestion:

try taking care of the "timing" issue – and the Deadband issue ... see if that helps ... let up know what you find – and post an updated copy of your program ...

and good luck with your project ...

.

 

[kardel.sharpeye]

a very popular misconception is that the Deadband is something like an "acceptable range" for the PIDE to shoot for ... that's wrong ... in many (most?) cases the Deadband is used to "rest" some type of output device (for example: something like an electric gearmotor) to keep it from constantly chattering back-and-forth (and overheating) as the PV fluctuates very close to the Setpoint ...

Ron,
You are correct.
The problem is the deadband. And yes, i had the misconception on the deadband.
I've changed the deadband to 0 and the PIDE now works well.

Thank you so much, Ron