9/11/2017 3:35 PM  
Joined: 6/20/2017 Last visit: 9/14/2017 Posts: 5 Rating: (0) 
We have a ROBICON GENIIIE HARMONY VFD equipment in Cartagena Refinery in Colombia, and this equipment has defined “Integral Gain” parameters. However, we don’t know what is the units for this parameter: resets per minute, resets per second, seconds, … You answered me: "We use resets per second. The term is represented by 1/time, so the larger you make time, the slower the integrator will react." However, I still have the following doubt: you say that you use resets per second, but you also say "so the larger you maketime, the slower the integrator will react". Our New Question:So the larger I make "integral gain" parameter (represented by 1/time), ¿the faster the integrator will react? (since time is inverse to the “integral gain” because is represented by 1/time). 
Last edited by: Jen_Moderator at: 9/12/2017 9:20:49 AMNew subject after splitting 

9/11/2017 4:11 PM  
Joined: 9/27/2006 Last visit: 6/18/2019 Posts: 8321 Rating: (1704) 
Hello Haider; The implementation of PID (or PI) controllers varies amongst the different manufacturers, which can lead to some confusing discussions. What you are looking at right now if the dichotomy of Reset Time (in sec. per repeat) versus Reset Rate (repeats per second). In this context, "reset" refers to the time given to the integral component to match the action of the proportional part of the control algorithm. So the smaller the time allowed (or the larger the rate, since it is the inverse) the stronger the integral action. One great ressource I follow for PID controls is controlguru.com. The following extract can appear trivial, but start following the links and you will soon discover a lot on cntrol theory, as applied in PLCs: Reset Time Versus Reset Rate Tr = 1/Ti No matter how the tuning parameters are expressed, the PI algorithms are all equally capable. But it is critical to know your manufacturer before you start tuning your controller because parameter values must be matched to your particular algorithm form.Commercial software for controller design and tuning will automatically address this problem for you. 
Last edited by: dchartier at: 9/11/2017 4:23:34 PM 

9/11/2017 4:39 PM  
Joined: 6/20/2017 Last visit: 9/14/2017 Posts: 5 Rating: (0) 
Thank very much fot your help. With your answer I understand that, because the Perfect Harmony parameter is in resets per second (1/time), So the larger I make "integral gain" parameter, ¿the faster the integrator will react? Note: The Perfect Harmony does not work with Ti (reset time), only with proportional and integral gain parameter. 
9/11/2017 4:52 PM  
Joined: 9/27/2006 Last visit: 6/18/2019 Posts: 8321 Rating: (1704) 
Hello Haider; I do not have the manuals for Robicon drives. But if integral gain uses time units (e.g. seconds), the larger the integral gain (the more seconds you allow for the integral action to act on the elimination of the error in the control algorithm) then the slower your integral action becomes... Think about it. Hope this helps, Daniel Chartier 
9/11/2017 6:14 PM  
Joined: 6/20/2017 Last visit: 9/14/2017 Posts: 5 Rating: (0) 
The integral gain is in repeats per seconds, not in seconds. Then it is the oppossite? Please see this manual: https://www.yumpu.com/en/document/view/54265551/nxgcontrolmanual/115 
9/11/2017 6:51 PM  
Joined: 9/27/2006 Last visit: 6/18/2019 Posts: 8321 Rating: (1704) 
Hello Haider; Sorry, the document you proposed does not indicate whether the integral gain i in sec. or in reset/sec. But if you are sure of your units, then use the value that you need to make the loop as tight as you can. Just remember that integral time (sec) and integral gain (resets per sec) are inverse values, and are simply chosen by the writer of the PI algorithm based on how easy it is for him (her) to adapt the mathematics to the platform he/she is using. The manual also ponts to an autotune function for the Perfect harmony drive, maybe you should study the information and test it for your application. Hope this helps, Daniel Chartier 
9/11/2017 6:57 PM  
Joined: 6/20/2017 Last visit: 9/14/2017 Posts: 5 Rating: (0) 
Thank you Daniel. The document does not specify, however in consultation with SIEMENS (as I show above), they answered that "repeats per seconds" The term is represented by 1/time. 