Skip to Content

Register Login

Industry Online Support

Technical Forum

Navigation

Pitfalls of PID controllers

Created by: hdhosseini at: 5/31/2011 9:27 PM (4 Replies)

Rating (3)

Thanks 1

5 Entries

5/31/2011 9:27 PM	Rate (1)
hdhosseini Diamond Expert Joined: 1/28/2009 Last visit: 2/21/2023 Posts: 6794 Rating: (1341)	Hello friends, This is the presentation based on the function of modular PID controller from SIEMENS.In this post, I upload the user code for this application.Because of copyright issues I do not upload the actual codes. FB19 (PID) , FB13(LMNG_C) and FB22(ROC) Best regards Hamid Hosseini Attachment prog.zip (273 Downloads)
	Last edited by: hdhosseini at: 6/29/2013 10:54 AM Last edited by: hdhosseini at: 6/3/2011 9:06 AM Last edited by: hdhosseini at: 5/31/2011 9:48 PM Demo Channel on Youtube
Suggestion To thank Quote Answer

5/31/2011 9:45 PM	Rate (0)
hdhosseini Diamond Expert Joined: 1/28/2009 Last visit: 2/21/2023 Posts: 6794 Rating: (1341)	Check the attached trend snapshot: 1- Jump happended due to auto to manual swithover-LMNR_ON deactivated in FB13 2-LMNR_ON activated in FB13, mode change from manual to auto works bumpless, process experiercing a ramp instead of STEP 3-LMNR_ON activated in FB13,sudden jump in setpoint branch is exerted to process, process handles this smoothly. 4-LMNR_ON activated in FB13, auto to manual and manual to auto smooth changeover 5-LMNR_ON deactivated in FB13,P part selected and de-selected.semi-impulse in output of algorithm 6-LMNR_ON deactivated in FB13,sudden change in P value may cause jump in LMN and lead the process to oscillation. Some solution will be discussed later regards Attachment first.jpg (343 Downloads)
	Demo Channel on Youtube
Suggestion To thank Quote Answer

6/3/2011 9:54 PM	Rate (1)
hdhosseini Diamond Expert Joined: 1/28/2009 Last visit: 2/21/2023 Posts: 6794 Rating: (1341)	In this section I want to show the severe effect of power failure on loop which is still in auto mode after power return.Normally, while power recovers, we set the loop in manual. If operation requires the valve be fail to close, zero send to analog output and if the operation requires the valve be fail to open,100 send to analog output. Check the attachment as I simulated the case that the power fails, valve closing and PV gradually became zero. In case of power recover, algorithm set auto (big problem is here),Big error fed to PID algorithm and Big jump in LMN is inevitable. This jump is transferred to the control valve and the process which is not desirable. See the attachment Attachment Bigjmp.zip (183 Downloads)
	Last edited by: hdhosseini at: 6/29/2013 10:56 AM Demo Channel on Youtube
Suggestion To thank Quote Answer

6/3/2011 10:21 PM	Rate (0)
hdhosseini Diamond Expert Joined: 1/28/2009 Last visit: 2/21/2023 Posts: 6794 Rating: (1341)	The solution is provided here.There is option of activating conversion of step change ramp in OB100.This will reduce the error and then the value on the control valve will increase(or decrease) gradually as shown in the attachment. Attachment Nojmp.zip (321 Downloads)
	Last edited by: hdhosseini at: 6/29/2013 10:50 AM Demo Channel on Youtube
Suggestion To thank Quote Answer

6/29/2013 8:32 AM	Rate (1)
hdhosseini Diamond Expert Joined: 1/28/2009 Last visit: 2/21/2023 Posts: 6794 Rating: (1341)	Most of the time problems related to regulation and PID controllers doesn't come the controller itself.There is a wrong common belief even between "automation engineers" that PID controller should predict what is wrong in the process or mechanical defects.My answer to this issue is always disappointing and try to explain about the mechanism behind industrial controllers and their limitations.Normally,PID controllers when there is no defect in transmitters, positioner and actuator and the governing rule handle the process as designer expected. In most of the time, if the instrument side is OK, problem is from mechanical defects or operation error.As example I can leakage from compressors, opening of bypass line when not expected and finally,valve passing and mismatch between action on positioner and expected restriction made by mechanical part of valve.The golden rule in such case, is to check history of loop via trends ,check all parameters to be normal and check it by setpoint changes.If no success after all these tests, more inspection needed in site as PID controllers can not predict defects. Best regards Hamid Hosseini
	Demo Channel on Youtube
Suggestion To thank Quote Answer
This contribution was helpful to 1 thankful Users ali1357

|

Share this page:

Share this page on...

Follow us on