Using a PID on a cooling tower? Anyone applied one here?

I’m just curious, has anyone ever implemented a PID control for an evaporative cooling tower system?



I’ve never bothered to apply a PID to one but use a 4 state control based on water temperature deviation from the setpoint and duration of the deviation.



State 0: Tower fans off, tower feed pump off.
State 1: Tower feed pump is on
State 2: Tower feed pump is on, one tower fan is on
State 3: Tower feed pump is on, both tower fans are on.

From State = 0
At setpoint + 1 a timer starts. If it ever times out, the state increments.
At setpoint + 2 a shorter timer preset is used.
As setpoint +3 the state unconditionally increments.

From State = 1
While PV remains at SP+1 a long timer counts. If it times out, state increments.
While PV is at SP + 2 a shorter timer counts. It if times out – state increments.

and so on….. you get the picture of how it works (states are decremented in reverse)


As you can tell, there are three discretes, a feed pump and two fans. That’s it. No VFD on the feed pump or on the fans, only soft starters.

Typically on a cooling tower precise water temperature control is not necessary, but rather just staying within a band. The 4 state control works quite well.


However, I’m right in the middle of building yet another cooling tower control system, so I thought I would pick everyone’s brains.

I'd say stick with what you have. PID can be a pain to tune, and I'm not sure what you will gain by using it.

If it ain't broke, don't fix it.

Well said Tom,
If you would like a fancy term for what you have done, technicaly you have a Fuzzy logic control system.

Ton,
The last thing an incessant tinkerer like myself wants to hear is to "stick with what you have."



You know that nagging urge that there is a better way to do it, or at least a way with a bigger coefficient of coolness?
Thats what I have on this project.

Oh well, I've done lots of these, I know that the state control works great. You are right, a PID would be an unnecessary pain in the rump on this - besides, in the end, I would have to divide its output into bands and assign a state to the bands anyways. Mostly I was curious if anyone had approached it from that angle.

Allen, I wouldn't go straight to calling this one fuzzy logic - its a very basic state machine (both are technically correct but I usually think of fuzzy logic as a little more sophisticaed) - I could do it with three temperature bulbs wired straight to three motor starters - except that I prefer to give the customer a little more functionality and some alarming as well as alternate fans.

I did a cooling tower a few years back much the same way you do. I did add a t/c input for the outside air temp. When the temp gets close to 32 we turn the fan on first and the spray pump on last. This would keep the tower from freezing. By the time the spray pump started the tower was warm enough not to freeze.

We also did some without a t/c and just put a summer/winter switch in to reverse the logic.

We have a system which uses PID to maintain a constant temperature on the discharge from the tower but the basic control works pretty much like yours. The PID loop controls a 3 way valve which allows water to bypass the tower.

Alaric,

We have a system of 4 towers that work together to control the head pressure on an ammonia refrigeration system. Each of the 4 towers has its own pump and seperate sump. Each has 3 fans. All 12 fans are controlled by single VFD. This VFD is on a PID to control the fan speeds. Of course there are manual controls as well. It's kind of neat to hear the fans spin up when the load increases. The pumps are controlled by across the line contactors that are programmed to start and stop at different pressures. Our system is different than most because our refrigeration load can vary wildly. From 0 to 600 tons in a matter of 10 seconds. Our head pressure is maintained accurately and efficiently with this system. The VFD saves a lot of wear on the fan drive belts. Your system is probably much different. I think Charles has a great idea of adding an external temperature sensor. I am going to look into that.

BD

I liked Charles's idea also, its something that I had thought about adding anyways as I had a spare RTD channel. This one is a 270 Ton tower system, and I wasn't planning on putting a VFD on the fans but rather just staging one fan on or both fans on. The fans are direct drive and on soft starts.

Worked at a place where they
had 4 chillers and two cooling towers -
all pumps and the tower fans were VFD and controlled by PLC.

IF your are buildig a new one
(ASSUMING it is big enough for real savings)
I would definitely use PLC control and VFD
you could
control circ pumps - eliminate need for throttling and save power
control fans (leave em running to keep water out of bearings - we have some water problems in Seattle) Save on power
Instead of using cooling towers to cool the recirc can you extract heat and heat building, preheat hot water etc etc?

Dan Bentler