PID replacement

[(8{)} ( .)]

Hello,
After a lifetime of hating PID, I have decided to try to replace it with something else. My first stop is my ever handy Logix 5000 development program. Within it are the IMC, CC, and MMC instructions.

My first shop test will be some heating plates. These are 6x6 plates, with 450 watts of heating, and a thermocouple installed in the plate.

There are two of them and sandwiched between them are components that are relatively non thermally conductive. (IP prevents more detail.) The distance between these plates is approximately 1/4 to 3/8 inch, so there is heat transfer between the two halves.

Has anybody any experience with heating, or better yet, heating like above, with the IMC CC or MMC instructions? Or something else? Bang Bang is not quite accurate enough. Once again, I am not interested in PID.

Thanks

I've been using the IMC almost exclusively for the past year. I had a PID loop that we literally spent weeks trying to tune. The expert from Germany threw up his hands as well.

On a whim I plugged in an IMC instruction in spite of having zero experience with it. I did an auto-tune and had the system running perfectly within the hour.

(8{)} .)
Yosi

 

[JoeM3]

Ok, now this is what we need, some helpful stuff. Is there anything specifically I should know about? Any "gotchas"? Like I said I am going to be shop testing this before deploying it? Thanks!!

 

[drbitboy]

I've been using the IMC almost exclusively for the past year. I had a PID loop that we literally spent weeks trying to tune. The expert from Germany threw up his hands as well. ...
Yosi

What was the process? Was it non-linear?

 

[(8{)} ( .)]

What was the process? Was it non-linear?

The process was a 8000 liter poacher. We needed to keep the temperature constant but the problem was that it was fed by another poacher which would dump hot potatoes into it. The process was constantly being disturbed by the hot potatoes and an undetermined amount of hot water that went along with it.

As I said, I did an auto-tune that took about half and hour to complete and it's been working perfectly for a year now.

Yosi

 

[(8{)} ( .)]

Ok, now this is what we need, some helpful stuff. Is there anything specifically I should know about? Any "gotchas"? Like I said I am going to be shop testing this before deploying it? Thanks!!

The only real "gotcha" that I have encountered is the situation where loops have long dead times. This generates an error during auto-tune which forces the auto-tune to abort.

Typically I run these loops in a 250 ms periodic task. When I encounter the above error I move it to a slower periodic task. I've never experienced this error in a 5000 ms task.

Other than that the terminology takes a little getting used to. I don't remember off the top of my head but the names "auto/manual/software manual" which we were used to were replaced. I can look up the equivalents for you next week.

Another possible "gotcha" for you is that this instruction doesn't exist in ladder. I program it in FB but you can also use ST.

Hope this helps,

Yosi

 

[JoeM3]

Thanks! Is there any issue with running these in continuous tasks? I have all the Rockwell languages so programming this in FB is no problem. Actually working on it right now. Have you tried a RMPS instruction in front of it? The deadtime in my process is not long.
Thanks again!!

 

[(8{)} ( .)]

Thanks! Is there any issue with running these in continuous tasks? I have all the Rockwell languages so programming this in FB is no problem. Actually working on it right now. Have you tried a RMPS instruction in front of it? The deadtime in my process is not long.
Thanks again!!

Couldn't tell you since I've only used them in periodic tasks. Actually if I'm not mistaken the last time I programmed a PID loop in a continuous program was in a PLC-2 with the horrible PID subroutine.

I haven't used the RMPS on the PV but I have used it on the SP. Had no issues.

Hope this helps,

Yosi

 

[JoeM3]

Yosi,
I ended up rewriting things and placing the FB in a periodic task anyway.
Thank You so much for your help.

 

[DaDaDadeo]

Use "Timed Proportional Control"

I use TPC for digital output temperature and pressure control for steam. It's very effective and the rudimentary programming will work exactly the same across different control systems.

http://www.plctalk.net/qanda/showthread.php?t=123425

Cheers!

 

[Peter Nachtwey]

What we have here is a lack of understanding the system.
A PID will work most of the time. There are three things that hinder a PID.
1 non-linear systems.
2 long dead times.
3 users.

The OP hasn't indicated that the first two are a problem so why is there a problem with a PID?

BTW, IMC is just a way of computing gains for the PID.

If you want a method that is very good and handling non-linear systems then try Sliding Mode Control. Sliding mode control is good for on-off type of control with SCRs. However, SMC does not like dead times. If there is a dead time then a model must be used and chances are if you had a model, a PID would work.

Another method of control is MPC or model predictive control. However, it is complicated and requires a lot of processing power. MPC doesn't use gains but it does require a pretty good model.

One my Peter Ponders PID YouTube channel, the second most popular video is "Sliding Mode Control with a Smith Predictor"

https://www.youtube.com/watch?v=uhLMyOlwCoM&t=3s
A model is necessary for SMC+SP to work but it is very robust.

 

[Tom Jenkins]

Here is an alternate I have used successfully on non-linear processes with long dead times.

 

[JoeM3]

Tom, Good Stuff, thank You

 

[JoeM3]

DDDD,
Thank You good stuff!!

 

[JoeM3]

Hi Peter,
This is not about whether or not PID could be made to work or whether it is a appropriate engineering choice. I have been "making it work" for a great deal of time. This is about raising the bar and moving forward, leaving the past behind, if possible. Put another way, I think that if you looked around you could find a appropriate transportation application for a Model T. Yet in 2021, they are only in shows and museums.

 

[Mispeld]

smh