Control loop from hell (and a tight budget)

[JHarbin]

Normally your Control Variable Output would control the fan speed, and a PID would watch the Process Variable (Temperature) and adjust fan speed to make the temperature = setpoint. In this case though the speed control is crippled - no way to set speed at a certain level because you don't know where it is now - unless there is some way to get a speed feedback signal from the VFD to the Compact Logix PLC, so the PLC has a speed reference point.

I agree that this control is crippled, so I don't have anything constructive to add to the specific discussion.

However, I've controlled applications like this a number of times. I do it by controlling a stable supply water temp using the fan VFD and control the process temperature via pump speed. That is assuming we are talking about a cooling tower when the OP says tower.

 

[DairyBoy]

NetNathan, the deadband is 0.5°C! These guys want it all. That said, when it's behaving, the ripple is currently only about 3-4°C.

TConnolly, "like driving blidfolded": Hey, that's my normal day

JHarbin, I agree; I've previously used variable pump speed from a constant temperature reservoir. No problem. Seems like I'm always beta testing (or just covering-up in the software) yet another twisted interpretation of how things should work, travelling miles before declaring defeat and the appropriate remedial action being applied months later.

Lancie1, to run a net cable to the vfd is a big job; scaffold towers etc and as I've said, there's no local analogue source for retransmission of PV. But, there's one analogue i/p left. Looks like we now have a use for it.

 

[FrancisL]

Fan or Pump?
Trying to understand this I am not clear - is there a Pump as per the earlier diagram, or a Fan as mentioned later?
If it is a Fan then how is it installed - is it blowing air through the heat exchanger?

 

[DairyBoy]

My reference to a pump in the last post was on another job. Wish this was the same. This one is fixed speed pump and variable speed fan. I made an error in agreeing that the diagram was correct; I'm sure that hasn't helped here. Thanks for persevering.

 

[TurpoUrpo]

You have:

-Plate heat exchanger with coolant water flowing on other side with fixed speed. Product on other side with fixed speed?

-Cooling tower, with fan, speed control trough "more, less" discrete inputs. No feedback from speed (you could record your more's and less's and keep track of that in plc).

-I don't know about cooling towers, how much does surrounding temperature affect cooling efficiency? (is it dry, wet or hybrid ct?) You have no input on outside temperature?

-You want to control product temperature trough controlling coolant water temperature. Requirement would be 0.5C deviation?

-Is the setpoint always same, or does needed product temperature vary?

 

[DairyBoy]

-Plate heat exchanger with coolant water flowing on other side with fixed speed. Product on other side with fixed speed?

Yes.

-Cooling tower, with fan, speed control trough "more, less" discrete inputs. No feedback from speed (you could record your more's and less's and keep track of that in plc).

Yes.

-I don't know about cooling towers, how much does surrounding temperature affect cooling efficiency? (is it dry, wet or hybrid ct?) You have no input on outside temperature?

The lowest acheivable temperature will depend on ambient conditions. No ambient sensor. Wet.

-You want to control product temperature trough controlling coolant water temperature. Requirement would be 0.5C deviation?

It's not how I would choose to do it but that's what's required.

-Is the setpoint always same, or does needed product temperature vary?

Setpoint varies by 4-5°C.

 

[TConnolly]

-I don't know about cooling towers, how much does surrounding temperature affect cooling efficiency? (is it dry, wet or hybrid ct?) You have no input on outside temperature?

The lowest acheivable temperature will depend on ambient conditions. No ambient sensor. Wet.

Go to the post I linked earlier (reply #7) and then read the whole thread. There is a detailed discussion of tower performance and how current conditions affect it.

 

[FrancisL]

Is this right?

If so then there IS a long delay. The Function should do the split to raise or lower the fan speed depending on the PID output

 

[Peter Nachtwey]

I wouldn't use a PID

I would simply check every minute and increment or decrement the fan speed depending on whether the temperature is below the SP window or above the SP window. If you want to increase the 'gain' then check the temperature at a faster rate than every minute. Decrease the 'gain' by making the interval between checks longer.

Get that to work first. Be patient because the response will be slow. If the error is big you might try doing two increments or decrements or change the interval to half intervals.

There is no way this is going to be a fast system nor will it be that accurate or able to respond to disturbances quickly. This is dirt simple and honestly I don't think you can do better without spending some money and making your system a whole lot more complicated.

I bet even Tom Jenkins would approve of this method.

 

[chelton]

A simple frequency to current converter could give you the analogue output you would like at a reasonable price. Providing you have transistor outputs on the plc?

 

[DairyBoy]

FrancisL, yes, that's the plant layout, minus pid. Peter, what you have described, without the finesse of the variable gain is what I have had from day one. I am going back there tomorrow morning and will increase the period to at least a minute and implement the error-to-period mod. It's a difficult one to 'play' with because when it's on-load it's in-process and when it's free there's no load. I will also implement some patience. Sorry- image is not the best resolution.

 

[Peter Nachtwey]

Peter, what you have described, without the finesse of the variable gain is what I have had from day one.

I am not much help then.

I have thought about how you can use a PID but it is messy. Basically the fan speed would have to be 'homed' by activating the slow output many times until the fan stopped. You would need to know how many times the speed up output can be activated to get to full speed. If it takes 20 speed up pulses to go from stopped to full speed then you need to keep track of the speed as going from 0 to 20 and map that into 0 to 100% output. So if after homing the PID output is 50% you would need to generate 10 speed up pulses and keep track of the fact that the speed is at 10. If the PID output drops to 44% you would need to generate one slow down pulse and decrement the speed to 19. This would probably work pretty well but there is a little bit of book keeping involved and deciding when a pulse should be generated. I would avoid hitting the increment and decrement speed outputs if the output oscillated between 49% and 50%. Then on top of all of this the PID and system must be tuned. The algorithm for incrementing and decrementing the speed should be worked out before going into the field. This seems like a lot of work for something that will probably never work that well.

I still like the simple method with maybe checking twice as fast if the error is big.

 

[DairyBoy]

I am not much help then.

On the contrary, Peter; your endorsement of the "simple method" is enough for me

 

[pmac10]

from what I've quickly read here, a danfoss fc302 drive should handle this without a plc, connect your pt100, select pid function, enter set point, and the drive will ramp up and down accordingly, obviously there is a bit more to the drive programming but you get the gist.

 

[TurpoUrpo]

from what I've quickly read here, a danfoss fc302 drive should handle this without a plc, connect your pt100, select pid function, enter set point, and the drive will ramp up and down accordingly, obviously there is a bit more to the drive programming but you get the gist.

And when you read some more. You will find that new cabling to drive is no-no.