PID control with time as the process variable

[Terry Woods]

Something really big is missing from this picture (at least, from my point of view)...

Russ, what the heck is your "...discrete level sensor..." detecting? How does that particular condition, the one you are detecting, change?

Is this sensor an "Eye"?

What does it mean if the sensor is ON? OFF?

Is the sensor used to detect the height of the material in the container?

Is this "fiber" as in a string-like material? (Fiber-Optic?)

Or is it "fiber" as in Bran-Fiber (the cereal grain)?

Is it the case that the material does NOT pass in front of the sensor as it fills the container?

How is material removed?

Is there a pile of material and the material at the lower level of the container is removed as new material is added on top?

What exactly are you moving?

Regardless of the answers for the above... your SineWave Scheme does NOT sound like it will do what you want.

As weird as this sounds... I understand what you seem to want... but, apparently, I don't have much of a clue as to what you are really doing!

INPUT! MORE INPUT! (Number-5).

 

[russrmartin]

Russ,
Is this sensor an "Eye"?
Yes.

What does it mean if the sensor is ON? OFF?what the heck is your "...discrete level sensor..." detecting? How does that particular condition, the one you are detecting, change?
On = Fiber level present in reservior between infeed and outfeed is at least as high as the photoeye. Off = not present.
Condition changes as fiber piles up or dwindles away via infeed/outfeed rates.

Is the sensor used to detect the height of the material in the container?
Yes

Is this "fiber" as in a string-like material? (Fiber-Optic?)
It is fiber as in loose strands of cotton type material, mostly about 1" long. Think like the stuffing in a pillow perhaps.

Or is it "fiber" as in Bran-Fiber (the cereal grain)?
No.

Is it the case that the material does NOT pass in front of the sensor as it fills the container?
No. The materal falls past the photoeye. It has been set to detect only compacted fiber, not the light as air fiber that is fed into the infeed. As it piles up in the lower trunk, or reservoir, it is compacted by air pressure.

How is material removed?
Via outfeed rollers, which feed the mat out onto a device which weighs the mat. The speed of the outfeed rollers is determined by a line speed in pounds per hour, which is trimmed by the weight scale.

Is there a pile of material and the material at the lower level of the container is removed as new material is added on top?
Exactly.

What exactly are you moving?
I don't understand this question.

Regardless of the answers for the above... your SineWave Scheme does NOT sound like it will do what you want.

As weird as this sounds... I understand what you seem to want... but, apparently, I don't have much of a clue as to what you are really doing!

INPUT! MORE INPUT!

I have developed my code, and in theory, my code will work. I have tried it with a theoretical input and trended my response. It does give me the response I desire. The question now really is whether this will work in an imperfect system, and whether the response to setpoint changes is linear or not. At the most used line speed, I believe this will work. I will keep this post updated.

 

[Ken M]

I've only just caught up with this thread and I'm really, really disappointed!

When I saw the original title "PID control with time as the process variable" my first reaction was "Yes, yes! This is something for the guys at the Smoke & Mirrors Inc R&D labs to get their teeth in to!" Imagine controlling time with a PID block? Enter your set-point for the date you want and off you go ...!

And what did I find?
This.
Boy, did I feel let down.

Ken

 

[russrmartin]

I've only just caught up with this thread and I'm really, really disappointed!

When I saw the original title "PID control with time as the process variable" my first reaction was "Yes, yes! This is something for the guys at the Smoke & Mirrors Inc R&D labs to get their teeth in to!" Imagine controlling time with a PID block? Enter your set-point for the date you want and off you go ...!

And what did I find?
This.
Boy, did I feel let down.

Ken

Sorry Ken. I think you may have rather wanted to visit the post entitled PID control with time as the CONTROL variable. This may be exactly the type of post you're looking for.

 

[Ken M]

Hmmm, no, I think it was 'process variable' I wanted.
I've no idea what quantity S&M Inc will have to manipulate to effect a change in time, and if they can give me "Back To The Future" in a PID block, I really don't care!

(Analogy: if I want my room warmer the control variable is not the temperature. I adjust the set-point and my gas burner stays lit for longer so more hot water is pumped through my radiators. There's the control. The set-point and process variable will always be the same - the controlled variable is usually something entirely different.)

Ken