Cooling tower with VFD control

Hi,
I am new at this forum but I have been working with PLC for some time.
I have a 5 cell cooling tower VFD control that I am currently programming using a SLC504.

What I would need some help is how should I control the available fans creating on a "on service" status the lead fan, the second fan, and so forth.

I would like to control or stage each fan based on the PID output (say >60%) and with PID error of >3 Deg. f, at that point a seconf fan will start then if we need any additional fans the 3rd will start using the same setpoints after the second fan was started and runnig. The same would apply for taking the fans off (say <30%) and PID error <3 Deg f.
Thank you

Fulvio

The main problem I have seen with these fans is rain getting in motors and rusting bearings (we do get a little rain in Seattle)

ASSUMING
the total load of cooling towers never goes below 10% say
and
at minimum speed on fans you can still get airflow

it COULD be possible to run all five on one VFD at speeds proportional to load

This would keep all five spinning and water out of them and may simplify controls and possibly save on energy.

If my setup I would have spare VFD in case the sole unit goes out
OR bypass switches (sigh added cost)

Dan Bentler

were also going to need some more information for example what type of vfd are you using? and what type of external frequency reference signal does it use? Is there one VFD for all? or one VFD for each? If its the first case you will have to have all 5 running at the same speeds so cannot switch one then the other etc. May also be handy to know if there is only one temperature senser and what type of analog output card your using? This affects the scaling ranges also the signal the temperature senser returns is also important for the same reason.

the VFD drives are for each fan motor. The drives are ABB ACH-550-uh with a connected 60 HP motor. I am using 1746-NIO4 and 1746-NI4 for input and output. The input temperature is scaled from 0 to 150 deg F and rescaled for PID use to control the Cooling tower temp, for now, the speed Demand goes to all "in service" fans. The output from the PID is rescaled for the NIO4 output. We are replacing the existing two speed fans with VFD, during the past 10 years we did not notice any problem of with bearings do to fan not operating when the tower is in service.

I've done a few cooling towers. I recommend that you run all fans at the same speed and that you have a minimum speed for the reasons Dan enumerated. When the tower temperature drops below the minimum the fans are off. The tower feed pump doesn't need to be on a VFD as it is generally on or off - but I do recommend using a soft starter with a start ramp for centrifugal pumps.

strange I couldn't find any literature on that drive one question I can help out is the " In service detection " On the AB drives there is usually a fault contact and drive ready contact. I would patch those in to your PLC inputs as well as having a maintenace\auto switch in series with the either the drive ready or fault contact. Ron Beaufort once posted an example code for in service on a 12 pump application using indirect addressing a few modifications on this example could probably be used in your case. as far as controlling each seperately on a PID control from the same set points that one is beyond my skills. However I would think only PI rather than PID would be needed and a quarterly amplitude decay rather than critically dampened process model sounds sufficient.

The drive is ABB ACH550. The "in service" is operator selectable by a switch, I have seen other cooling tower setup that will send the signal to all the drives or motors. Maybe I was making this more difficult that is necessary.

The set up includes a max and min temp setpoints that will "start" the drive, the PID will take control of the output with a min signal of 10%

I've done more than a few cooling towers, staging control and PID.

We always turned fans on/off on a deviation from setpoint. So, 3 degrees over setpoint would drive a timer that would add fans. 3 degrees under would drive a timer to reject fans.

All VFDs were driven by one PID that had the same setpoint as the staging. The VFDs had a minimum speed, mostly because the fans had a gearbox between the motor and fan. The gearboxes had strange wear characteristics at low speed. The belt driven fans didn't have the same problem, but the minimum speed was still there for consistencies sake.

Thanks Exabmorgan

That is what I am experiencing problems with. To schedule fans that would take into consideration how many are in service and how to add/subract each fan with 3 deg from setpont. If I start from fan 2 "3 deg from set" how should I set the rest?

The ABB ACH550 has an internal macro called PFC (Pump, Fan Control) which sets up a PID loop for your process and then sequentially controls the remaining motors based on demand. The Macro is designed to use a single drive and the rest of the motors would be across the AC line but you could easily modify it so the relay controlling the lagging motors would be a run signal on a VFD set for 60hz. The effect would be similar and you would get soft starting and soft stopping. You could even set the top speed to hp match the motor to the fan potentially getting a little extra out of each fan.

Check the ABB Instruction Manual for PFC Macro. It's probably just what you want and no PLC or extra hardware is required.

Here is the basic logic, I'll just make up some numbers.

Start Fan #1 when Temp + 3 > SP for Timer #1 (120sec)
Start Fan #2 when Temp + 3 > SP for Timer #2 (240sec)
Start Fan #3 when Temp + 3 > SP for Timer #3 (240sec)
Start Fan #4 when Temp + 3 > SP for Timer #4 (240sec)
Start Fan #5 when Temp + 3 > SP for Timer #5 (240sec)
(so the "3" in any of the above lines would be operator accessible for tuning, as well as the timer values)
Stop Fan #5 when Temp - 3 < SP for Timer #6 (120sec)
Stop Fan #4 when Temp - 3 < SP for Timer #7 (120sec)
Stop Fan #3 when Temp - 3 < SP for Timer #8 (120sec)
Stop Fan #2 when Temp - 3 < SP for Timer #9 (120sec)
Stop Fan #1 when Temp - 3 < SP for Timer #10 (240sec)

So Fan #2 can't start until Fan #1 is running, Fan #3 can't start until Fan #2 is running, and so on. Timer #2 starts after Fan #1 is running(and the temperature is too high), Timer #3 starts after Fan #2 is running, etc.

And in reverse, Fan #4 can't stop, until Fan #5 is stopped, etc. Same with the timers, Timer #10 starts when Fan #2 is stopped (and the temperature is low enough).

Keep track of the these numbers as "required" fans.

Then that count of the # of req fans feeds into logic that tracks running and disabled fans. That way you can account for fans that are in maintenance mode, forced on by the operator, etc and alternate the fans so they can get the appropriate runtimes.

I revised the existing program and graphics during expansion of a cooling tower system that used the temperature PID output to stage the towers on and off. The system was designed to minimized energy consumption while maintaining cooling water at setpoint and included adding VFDs on the tower pumps. When we were done the towers staged in 12 steps: 1st pump, 1st fan, 2nd pump, 2nd fan, up to 6th pump, 6th fan. The tower sequence rotates automatically, with the time left at two hours to minimice buildup on stopped fan blades during cold weather. Within the sequencing, I also have my usual logic to skip out-of-service towers - HOA switch not in auto or pump or fan fault.

The last fan to turn on ramps from minimum to maximum speed as the PID output rises from the fan on setpoint to the next pump on setpoint or 100% for the last fan. The previous fans in the sequence run at maximum speed. Similarly, the last pump ramps from minimum speed to a fraction of the speed band determined by the hotwell level PID loop while the previous pumps run at PID speed. Maximum pump speed is below 60Hz to limit water flow to the tower design rate after the spray nozzles were lowered two feet.

During winter weather, only one to three towers are enough to reject the process heat. By mid-summer, the outlet temperature will rise an acceptable 5 degrees C above setpoint as the airflow reaches humidity saturation.

Mike

I'll add that this application is one that is easy to underestimate. I'm sure this seems like a "duh" to the experts that have posted here.

Another bent I didn't see mentioned is to keep track of hours on the motors so you spread the load. The motor with the least hours is the one started next. The motor with the most hours is the one stopped first. I'd probably put this concept to the side until the rest was worked out.

Paul B said:

I'll add that this application is one that is easy to underestimate. I'm sure this seems like a "duh" to the experts that have posted here.

Another bent I didn't see mentioned is to keep track of hours on the motors so you spread the load. The motor with the least hours is the one started next. The motor with the most hours is the one stopped first. I'd probably put this concept to the side until the rest was worked out.

Click to expand...

That insures that all motors fail at the same time. But, most do that way. You need to exercise them but keeping the same amount of hours IMHO is not necessary.

In addition to my information posted above on the ABB PFC Macro, you can select with or without rotation to keep running hours balanced between motor/pumps. By selecting rotation, after a selectable number of hours, the lead pump shifts over one position. It keeps shifting around so eventually all pumps spend the same amount of time as the lead unit.