Valve open/ close pulse signal from analog PID output

Temperature is measured by S type thermocouple. Milli volt signal fron thermocouple is going to thermocouple input card of PLC. Application is for Hot Blast temperature control in Blast Furnace which is done by injecting cold blast (150 deg. C) to hot blast (1080 deg.C)to maintain the hot blast temperature around 1050 deg.C. In the cold blast line connected to hot blast header, this temperature control valve is located. Hot blast Temperature is controlled by this valve. This valve is motorised. PID output will drive this valve. In Siemens S7-300/400 PLC, a particular PID block is available which can give open/ close pulse signals to the motorised valve without any position feedbach of the valve. In a few earlier projects, we have used that. But in 9030 PLC, problem is that no such PID block is available.

Sunith

The problem with basic step control as in the previous post by Ron, is it is really just PI control, not PID.
Temperature generally required the derivative part.

If you check my earlier post, it does consider most of the points everyone has made in these posts, and is a simple conversion to a PID PLC block output, just adjust the time base depending on the speed of the loop.

Tried method for controlling a heat valve from PID.

Try this thread,
http://www.plctalk.net/qanda/showthread.php?t=19141

I use the analog output (0-100%) to control a heat valve actuator.
(Steam valve with electric actuator, One input to Open, One input to close, No positioner feed back, 120second Cycle time)
The method gives an assumed position but is normally close enough for operator indication.
Works on theory that if temperature is too hot then
close the valve a bit, If too cold then open it a bit.
By a bit I mean a pulse output open or close of 1% of total valve cycle time. (100% = total valve cycle time from Closed to Open)

Nothing wrong at all with using just PI control - I reckon on tuning 200 ish loops a year , and only once have I used D with a temperature control loop , this was with an electrically heated lab dryer where the was considerable delay and very little heat loss . Generally my processes have large losses , and these losses are fairly constant , so in this case , D is not necessary .

Looking at this project , I would have though that the software that burnerman posted would have been ideal.

Strain your brain

10baseT said:

Nothing wrong at all with using just PI control

Click to expand...

It depends on the system. All your systems may tune well with just a PI but it is wrong to assume everything tunes well with a PI.

10baseT said:

- I reckon on tuning 200 ish loops a year , and only once have I used D with a temperature control loop , this was with an electrically heated lab dryer where the was considerable delay and very little heat loss .

Click to expand...

The number of gains required does not depend on heat loss there was something else. What? Knowing the what is important in being able to tell if the system requires a derivative gain.

You guys should monitor the control guru site.
http://www.controlguru.com/postpics/heatexchanger.pdf

No one has made a comment on my suggestion that PD control be used. Actaully, I think FatController was the first to have the right idea by differentiating the output of a PID. At least I hope that is what he meant. Now what is the difference between differentiating the output of a PID or just using a PD?

Peter , I don't know whether you have a problem with comprehension , or if you are just trying to show how clever you are , but please take the trouble to read .

I am not completely stupid , why do you think I mentioned only with regards to my process , and stated a little information regarding that process.
Can you tell me where I mention Gain , unless of course "200 ish" is how you think I refer to Gain ? I only stated that I have used D once , and the circumstances in which I used it . You have gone on to state what I have already said - of course it is important to look at the process , and how it reacts before making any determination .
When we "know" our equipment , we are able to very quickly tune a loop with recourse to pen and paper and sheets of calculations , we have a good place to start already in mind .
I have no trouble tuning these loops very quickly an obtain excellent response and control - as I said because I know the process and have done it a few times before . It is also very easy to diagnose equipment faults such a valve "stiction" or defective condensate trapping in steam installations .

Again , please take the trouble to read before trying to shoot someone down .

Does this system need an integrator gain?

10baseT said:

Nothing wrong at all with using just PI control

Click to expand...

This implies a proportional and integrator gain. In this case there is something wrong with just using PI or PID control. A PI or PID controller without the differentiator on the output would soon saturated the output just as FATController suggested in his first post.

FatControl suggested using a PID, but then differentiating the output. Differentiating the output is the key. This is necessary because the valve motor effectively integrates the control signal. If you differentiate function like a PID and then integrate it you get the original function with the exception of an offset but the offsets can be assumed to be zero. This allows one to use the PID gains as intended in this application.

I have Motorized valve Gas Burner (open and close) i want to control with PLC PID Control with RTD input and close and open relay, motor full open to close time is 30sec please suggest the logic to control.

Percy_nk, I suggest you start a new thread instead of reviving a pair of fifteen year old ones