Programming Advice

We are seeking programming advice on a new cooling water application. Our objective is to maintain a Temp. Setpoint of the water leaving the cooling towers entering the process. There are four new cooling towers, each having eight fans. Water leaving the cooling towers goes to a common line with an Analog temperature transmitter which will be our only input to control the fans. We want to start and stop fans in pairs of two with on/off control to maintain our setpoint. We are trying to accomplish this using a Allen-Bradley SLC 5/05 ethernet processor.


We are seeking a wide variety of programming schemes to accomplish our objective.

One possibility would be to use comparison instructions. This would
involve taking the analog input from the temp. transmitter and
scaling it to your desired engineering units. The input could
then be compared to a pre-determined number by which certain
sets of fans would be turned on. Logic may need to be written
so fans would remain on for a certain amount of time to prevent
rapid on/off control.

In addition to time delays for minimum on and off times you probably want to add a deadband in the compare instructions.

BTDT

You probably want to pick the two fans to run based on their accumualted run-time, balancing out their wear. Of course, once a fan is started, you want it continue to run, even if it is no longer the "youngest".

Here's a link to an old thread where Terry improved on my technique for doing just that.
http://www.plcs.net/dcforum/DCForumID5/1050.html

PID, with some modification

If you made a PID loop out of this, then the
'percent output' would be 'how many fans I
need to operate.

This then, could be fed into a scheme not
unlike the traditional 'pump loading, pump
unloading' concept for sumps and accumulators.

...or not...

Another thing to consider is whether or not to use 2-speed motors. You could chose how many fans you wanted to run at one time, and if the temp is higher than normal switch to high speed before starting additions fans.

Another thing to consider might be VSD's for the fan motors- I'm not sure what KW size you're talking about (I assume huge) but price for VSD's for "reasonable" size motors is not too scary these days- you can then have pretty well infinite control with precise energy requirement matched to the load.

You said you only have one analog input to measure temp. How do you know which fans to run? Is water coming out of each tower all the time? Are the fans that are running tied to any flow control?
Also, are the fans started by motor starters? If so, do you monitor "fan running" with aux inputs? That way you could monitor failure and jump to the next fan if one fails.