Valve open/ close pulse signal from analog PID output

Hi,
I am facing a problem. I want your help to sort out the problem.
I am programming a PID temperature control loop in 9030 PLC. PID loop has only analog output. But the temperature control valve is motorised & it will take only open or close pulse signal. One way to do that is to compare PID output with position transmitter of the control valve & generate open or close pulse signal. But here, Temperature control valve do not have position transmitter. Can any body suggest how to generate valve open/ close pulse signal from analog PID output in this case.

Sunith

Wow, so let me get this right.

You're trying to control the position of a valve but "valve do not have position transmitter". O-k-a-a-ay.

And the valve "will take only open or close pulse signal" and you've only got an analogue output. O-k-a-a-ay.

So you've no way of finding out what position the valve is in at present, and you've no way of telling it to go to any other position.

There seems to have been quite a major breakdown at the design stage, never mind the PLC code. There doesn't seem to be any alternative to getting an appropriate digital output to provide pulsed signals, counting them in the PLC, and since you've no feedback, hoping the device doesn't stick or jump and you don't lose count. I don't envy you.

Regards

Ken

Page 51 of this manual should help you.

http://web1.automationdirect.com/static/manuals/d2user/ch8.pdf

What you need is called "boundless valve control". I think this come up here before. ABB have a nice implementation in their Commander 300 process controllers. The manual is available on the ABB web site (sorry, don't have a link, search on Commander 300).

Nope, I don't see how we can offer answers yet.

Can anyone say what the PV is (temperature sensor somewhere, as yet unmentioned? non-existent valve position feedback?). Can anyone say what the controlled variable is? (Valve position only seems controllable by a digital pulse stream and so far there's no evidence that this is available.)

Loop set-up and tuning are way in the future as far as I can see. We need to measure something and we need to control something. What are they?

Ken

PV is Temperature. Variable to be controlled is also Temperature.

Sunith

Not so Simple

Time proportioning control is not the answer, as suggested by Tark, as the valve will remain opened.

I think the concept is partway there, although you need it on both open and close pulse.

If you take the rate of chage of the PID output, and turn it into a pulse width, it should work.

ie, if PID is increasing, pulse the open, if decreasing, pulse the close. If you work out the time the valve takes to operate 100% range (T100), then
every 10 seconds, check
T = (((PIDOUT@-10sec) - PIDOUT)% /100) x T100

That will give you a pulse width to acheive roughly the same change as the PID.

If T is negative, then close pulse. If T is positive then open pulse.

If the PID output reaches the limit, then open or close the valve all the way as required.

Well you get the idea. Tell me if you think Im an idiot.

By any chance is this a Modutrol type valve,ie, power open, spring close, or spring open, power close?

If it is, then a shorter or longer pulse will work,right?

David

I learned something here today.

I learned what the technique is called.

What you need is called "boundless valve control".

Click to expand...

Here are other threads where someone is doing a similar valve control.
http://www.plctalk.net/qanda/showthread.php?t=19141&page=1&pp=15
http://www.plctalk.net/qanda/showthread.php?t=21277&highlight=valve
http://www.plctalk.net/qanda/showthread.php?t=21112&highlight=valve

The trick is to think of the valve as the integrator for a PD controller. That is a big hint.

Well, I think something is lost in the translation. Before expanding more efforts I would research the control valve some more. We need to the know the control valve works. From your description it almost sounds like a servo motor instead of a typical control valve. Any documentation or link to manufacture?

Sounds to me that the valve is meant to be opened or closed, but is slowly driven to those extremes. If there are open/closed limits, he can measure the time required for full travel, then start at one end or the other and maintain a total value of time for pulsed movement--add to when pulsed open, subtract when pulsed closed. That will provide a ROUGH position indication. Scale this position to 0-100 and scale the analog output to 0-100 and pulse the valve to cause the values to match.

I think this is how those fancy pants car power seat and mirror memories work. I don't think GM puts position indicators on seats and mirrors so how else could they do it.

This is sloppy, at best and the position value should be reset and restarted when at either limit, maybe even forced to a limit occasionally, when possible, for the purpose of position resetting.

There is no real substitute for KNOWING where controlled equipment is positioned at all times via feedback devices. If we assume, no matter how eloquently (is that a word?), we're asking for touble.

Why is valve position of any consequence ? all I care about is that when I open the valve , fuel gas or other heating medium is introduced into the process , and the temperature rises - if I use a basic honeywell controller , then there wouldn't be any feedback for valve position , only perhaps a calculated position for display .

If we take a modutrol type valve , I believe that normal travel time is 90 seconds and that pulse width is 1 second for optimal operation , the rest is easy .
I think if you search for it , burnerman posted the complete solution not so long back .

This is sloppy, at best and the position value should be reset and restarted when at either limit, maybe even forced to a limit occasionally, when possible, for the purpose of position resetting.

Click to expand...

This is just not necessary. The loop is closed by knowing the PV, you do not need to know the valve position. It is usually easier if you do, especially if you've got a nice valve, but not necessary. Even if you do know the position, it is only useful if you also know the installed valve characteristic, which may be far from certain.

One of my two points was that in the original post Sunith Bakundy didn't tell us what the PV was. Oh yes, he said it was a "PID temperature control loop" but there was no indication that he was directly measuring the temperature. This could have been an indirect consequence of other actions.

Does anyone know yet what's flowing through this valve? Have we enough faith in Sunith's description to assume it's the heating fluid and not the heated fluid?

Finally, could his description be a valve with a stepper motor for positioning? How does he use an analogue output signal to 'pulse' this?

I still don't know enough about the problem.

Ken

Greetings Sunith,



this might be a little too simplistic for your particular application ... if so, just disregard this post ... and as others have said, we’re probably going to need a lot more detailed information to nail this project down ... in the meantime, this is just another idea to think about ... as I said, if it doesn’t seem to fit your application, just ignore it and move on ...



I’ve been wondering if the same “non-PID” approach that I’ve been using to control the Hotrod’s air flow would work ... the basic idea is to either open the valve - or close the valve in steps ... and the size of the steps can be varied depending on whether the deviation from the setpoint is either large or small ...







you’ll notice that the Hotrod has two dampers which are used to control the amount of air flowing through the tube ... the damper at the left is positioned by a pneumatic actuator controlled by a PID loop ... the damper on the right is positioned by a gearmotor controlled by a simple time-proportional “pulsed” signal ...



in actual operation, I either use one damper or the other - but never both at the same time ... (the unused damper is always kept fully open) ...



this first trend shows how the gearmotor damper responds to a stepped setpoint change - and to a “load” change produced by turning off one of the fans ...







now let’s compare the response of the first trend to the next one which uses the pneumatic damper and full PID control ...








as you can see, there’s not really a lot of difference - at least in this limited “bench-top” student application ...



if you’re interested in following up on this approach, I’ve attached a .PDF printout of the “gearmotor” subroutine file below ... the “meat-and-potatoes” part is contained in rungs 0004 through 0008 ... hopefully the documentation will be enough to explain what’s going on ... if not, post any questions that you might have and I’ll try to answer them ...



good luck with your project ...