Skip to Content

Register Login

Industry Online Support

Technical Forum

Navigation

PI parameters setting for temperature control

Created by: Milosh at: 12/16/2014 8:37 AM (18 Replies)

Rating (1)

Thanks 1

Result pages:: |; 1 | 2; |

19 Entries

Entries per page: 10 | all

12/16/2014 8:37 AM	Rate (0)
Milosh Silver Member Joined: 8/16/2010 Last visit: 9/25/2024 Posts: 637 Rating: (12)	Hi all. In one project I use PI regulator for controling the temperature of press tools. I use FB41 (CONT_C) and FB43 (PULSEGEN) to turn on/off solid-state relays which powers the heaters. CONT_C parameters: 1. CYCLE = 100ms 2. SP_INT = 150 °C 3. GAIN = 1 4. TI = 5m 5. other default PULSEGEN parameters: 1.INV = LMN from CONT_C 2. PER_TM = 10s 3. P_B_TM = 1s 4. STEP3_ON = false 5. ST2BI_ON = false 6. SYN_ON = false 7. CYCLE = 100ms 8. other default With this settings regulator make skip (positive) to 170 °C. That is +20°C from setpoint value. Which parameter GAIN, TI to change to make better regulation? First, to try to decrease GAIN?
	Last edited by: Milosh at: 12/16/2014 8:45 AM Last edited by: Milosh at: 12/16/2014 8:45 AM
Suggestion To thank Quote Answer

12/17/2014 11:58 AM	Rate (0)
hdhosseini Diamond Expert Joined: 1/28/2009 Last visit: 10/6/2024 Posts: 6852 Rating: (1365)	Milosh Hi all. In one project I use PI regulator for controling the temperature of press tools. I use FB41 (CONT_C) and FB43 (PULSEGEN) to turn on/off solid-state relays which powers the heaters. CONT_C parameters: 1. CYCLE = 100ms 2. SP_INT = 150 °C 3. GAIN = 1 4. TI = 5m 5. other default PULSEGEN parameters: 1.INV = LMN from CONT_C 2. PER_TM = 10s 3. P_B_TM = 1s 4. STEP3_ON = false 5. ST2BI_ON = false 6. SYN_ON = false 7. CYCLE = 100ms 8. other default With this settings regulator make skip (positive) to 170 °C. That is +20°C from setpoint value. Which parameter GAIN, TI to change to make better regulation? First, to try to decrease GAIN? Hello Milosh, Just some suggestions , check them in practice if they help: PID parameters adjustment, portion of the I-Part is very small and it is practically a P controller.Try PID in FB41 side with the most conservative values you can use for D-Part. As your PER_TM is long, why don't you activate "SYN_ON " to do conversion whenever an update happens in LMN of FB41. I suggest checking a sample here (Pid_test.zip ) and check values of PER_TM and P_B_TM because they are project specific. I hope these comments help you! Hamid Hosseini
	Last edited by: hdhosseini at: 12/17/2014 12:00 PM Demo Channel on Youtube
Suggestion To thank Quote Answer

12/17/2014 1:36 PM	Rate (0)
Milosh Silver Member Joined: 8/16/2010 Last visit: 9/25/2024 Posts: 637 Rating: (12)	I know that you would answer. :) First entry, I don't understand...if you can explain a little bit...to be clear. Please. Is thet some kind of experimental setting of parameters. Second and third entry is clear. I didn't know the specific details about SYN_ON input. Do I need to do reset of controlers (CONT_C, PULSEGEN) in OB100 with parameter COM_RST, or it's not needed???

Suggestion To thank Quote Answer

12/17/2014 2:24 PM	Rate (1)
IBN-Service Posts: 8946 Rating: (999)	Milosh ... Do I need to do reset of controlers (CONT_C, PULSEGEN) in OB100 with parameter COM_RST, or it's not needed??? COM_RST makes sense with PLC-Startup scan... Do you just try to use all Parameters you can get without reading the Manual...? I think we can not do a remote - optimation of a PID here. At least you must give us a curve of a step response of controlled System.
IBN-Service Posts: 8946 Rating: (999)
Suggestion To thank Quote Answer
This contribution was helpful to 1 thankful Users hdhosseini

12/17/2014 2:52 PM	Rate (0)
Milosh Silver Member Joined: 8/16/2010 Last visit: 9/25/2024 Posts: 637 Rating: (12)	Ok... I read the manual, but some parameters is not clear for me. So, I try to get help from experienced colegues. Also, I do my project and I didn't have possibility to set parameters and test it in my company. All that things I must do on side... I can describe how the process come. With mentioned parameters press tool is heating about 25min, and after that time temperature is 150°C. Then output of CONT_C (LMN), start to decrease, but this time of decreasing is to slow. because of that temperature goes high to 170°C (solid state relay turn on and off). When it reach that (170 °C) temperature, it's start to decrease. I also see, that some experienced colegues make diagram (on HMI) with measured value and setpoint value in time. So, on Y axis is measured value and setpoint value, on X-Axis is time. After looking on diagram they make some setting to the PID which is almost optimal. How? :) Thank's a lot for help...

Suggestion To thank Quote Answer

12/17/2014 3:44 PM	Rate (0)
Marko Bursic Platinum Member Joined: 9/8/2009 Last visit: 7/17/2023 Posts: 1410 Rating: (153)	Milosh Hi all. In one project I use PI regulator for controling the temperature of press tools. I use FB41 (CONT_C) and FB43 (PULSEGEN) to turn on/off solid-state relays which powers the heaters. CONT_C parameters: 1. CYCLE = 100ms 2. SP_INT = 150 °C 3. GAIN = 1 4. TI = 5m 5. other default PULSEGEN parameters: 2. PER_TM = 10s 3. P_B_TM = 1s 7. CYCLE = 100ms PULSGEN: (P_B_TM ) meaning at least 10% in range 0-10% on, and/or 100% in range 90-100% If you are using SSR you should change this parameter to something like 100 or 200ms. CONT_C: CYCLE = 100ms vs PER_TM = 10s, during one PWM on/off cycle of SSR your PID is executed 100 times, what for? Like, set output to 50% and before is set to this value you change your mind 100 times each time?! My recomandation: 1. set P_B_TM to 100ms, 2. execute C_CONT every 10s or more 3. disable integration, disable diferentiation, set setpoint to 120, begin to optimize with Ziegler-Nichols method
	Last edited by: Marko Bursic at: 12/17/2014 3:45 PM
Suggestion To thank Quote Answer

12/17/2014 8:44 PM	Rate (0)
Milosh Silver Member Joined: 8/16/2010 Last visit: 9/25/2024 Posts: 637 Rating: (12)	Marko Bursic Milosh Hi all. In one project I use PI regulator for controling the temperature of press tools. I use FB41 (CONT_C) and FB43 (PULSEGEN) to turn on/off solid-state relays which powers the heaters. CONT_C parameters: 1. CYCLE = 100ms 2. SP_INT = 150 °C 3. GAIN = 1 4. TI = 5m 5. other default PULSEGEN parameters: 2. PER_TM = 10s 3. P_B_TM = 1s 7. CYCLE = 100ms PULSGEN: (P_B_TM ) meaning at least 10% in range 0-10% on, and/or 100% in range 90-100% If you are using SSR you should change this parameter to something like 100 or 200ms. CONT_C: CYCLE = 100ms vs PER_TM = 10s, during one PWM on/off cycle of SSR your PID is executed 100 times, what for? Like, set output to 50% and before is set to this value you change your mind 100 times each time?! My recomandation: 1. set P_B_TM to 100ms, 2. execute C_CONT every 10s or more 3. disable integration, disable diferentiation, set setpoint to 120, begin to optimize with Ziegler-Nichols method Ok, thank you all. My next test will be: CONT_C parameters: 1. CYCLE = 10s 2. SP_INT = 150 °C 3. GAIN = 1 4. TI = 5m 5. other default PULSEGEN parameters: 1. PER_TM = 10s 2. P_B_TM = 200ms 3. CYCLE = 100ms 4. SYN_ON = TRUE 5. other default
	Last edited by: Milosh at: 12/17/2014 8:45 PM
Suggestion To thank Quote Answer

12/17/2014 8:50 PM	Rate (0)
Milosh Silver Member Joined: 8/16/2010 Last visit: 9/25/2024 Posts: 637 Rating: (12)	Marko Bursic 3. disable integration, disable diferentiation, set setpoint to 120, begin to optimize with Ziegler-Nichols method http://en.wikipedia.org/wiki/Ziegler%E2%80%93Nichols_method This is the method I need to use?
	Last edited by: Milosh at: 12/17/2014 8:51 PM Last edited by: Milosh at: 12/17/2014 8:51 PM
Suggestion To thank Quote Answer

12/17/2014 10:55 PM	Rate (0)
Marko Bursic Platinum Member Joined: 9/8/2009 Last visit: 7/17/2023 Posts: 1410 Rating: (153)	It's the most common method, you will find it in every student book, very easy and any automatitian should know it. Tostrech into fewsteps and where to be prudent: 1. Integration shall be disabled (I part), Diferentiation shall be disabled (D-Part), onlyproportional part is active. 2. Begin to increease Kp gain, slowly and look at the response. IMPORTANT: when the system comes to a staeady state, make some disturbance, like cool down the furnace /or/ make a change of setpoint value. 3. If the systemreacts slowly and comes to a steady state overlong time thenthe gain is too small, increase gain 4 If system isoscillatingfrom 0% to 100% violently, then the gain is too big. 5.When the system oscillates anddoes not stop oscllating neither escapes to unstable oscillation, then you have foud the critical point of the system. IMPORTANT:If system oscillates not symetricaly arround setpoint is normal(or evennot close to setpoint) So, this is called the critical gain, also ultimate gain. Now you should measure the oscillation period (critical period/freqeuncy, ultimate period). With all these data you can calculate optimal Ti, Td, Kp parameters. 6. Enable integration, enable differentioation, set parameters and test.

Suggestion To thank Quote Answer

12/18/2014 12:05 AM	Rate (0)
Marko Bursic Platinum Member Joined: 9/8/2009 Last visit: 7/17/2023 Posts: 1410 Rating: (153)	Also a good thing is to add DISV value. If you know what ouptut is needed at steady state of your system, it is good (not only good, its preferable) to set this value. Regulation techique is a study of time varying process, therefore all equations are differential equations without constant parts, all parts have differentiation from 1st order to n-th order, no part with zero order differential. All the study is based arround a working point (where the system equations are linearized - therefore arround setpoint) at smallest deviation possible.Real systems are not linear, but they behave linearly if looked at very close interval. Therefore study of regulation does not include static characteristics, bias point - work point, that's another study. A classic regulator has the setup of twodifferent mathematical worlds - static by means of DISV(or bias, offset, work point) and dynamic (P,I,D) If you set DISV, then regulator will output this value even if Kp is 0 and will brought the system to steady state, if there are no distrubances the temperature shoul be near setpoint. With apropiate parameters your regulator will only add or subtract this value to react at distrubance (dynamic process).

Suggestion To thank Quote Answer

|

Share this page:

Share this page on...

Follow us on