PID Loop issues

Hey Guys,

Hardware:
SLC 5/05 processor
20 to 140 temperature probe in cooling tower basin.
Target basin temperature 72 Degrees.
Average temperature coming into cooling tower is 90 degrees

Problem:
I have a PID loop setup for my cooling tower, but I am getting this very annoying going too fast and going too slow effect on the tower. I am trying to get the drive to stay very steady with temperature change and not this dramatic speed up and slow down. I dont know if I have the PID loop setup correctly, but can someone tell me what I should do to start getting this thing tuned correctly?

Can you provide a trend of the loop control vars and output?

let's do this the easy way ...

I dont know if I have the PID loop setup correctly,

Click to expand...

please post your complete RSS program and we'll help you check the setup ...

can someone tell me what I should do to start getting this thing tuned correctly?

Click to expand...

if the setup isn't correct, then you're just going to be going in circles ...

suggestion: let's check the setup first – and THEN work on the tuning ...

I will be on site with the client tomorrow having the issue, I will get a trend of the PID Loop and I will send the code over tomorrow. I know another PLC programmer has changed it since I was last there 2 years ago.

--Ryan

The first problem I see is your target set point. You are going to have a tough time holding 72F in a cooling tower in Los Angeles year round. In fact, it will be darn near impossible. I just looked up LA wet bulb temperatures and your worst case wet bulb is 70F. The most efficient modern towers have a seven degree approach, meaning that you should be able to reliably get 77F water out of the tower during the summer. Winter of course is better, but you simply aren't going to consistently make 72F water year round with a cooling tower in LA and all the tuning in the world isn't going to change that.

Next question: Are you controlling the fan speed of the tower or are you controlling the water feed rate to the tower? I assmue its the fan as that is by far the most common way, but you know the old addage about assumptions so I'm asking.

If you can answer that question and then post the program as Ron suggested then we can help you. Also, if you can provide a plot that will help as well.

The first problem I see is your target set point. You are going to have a tough time holding 72F in a cooling tower in Los Angeles year round. In fact, it will be darn near impossible. I just looked up LA wet bulb temperatures and your worst case wet bulb is 70F. The most efficient modern towers have a seven degree approach, meaning that you should be able to reliably get 77F water out of the tower during the summer. Winter of course is better, but you simply aren't going to consistently make 72F water year round with a cooling tower in LA and all the tuning in the world isn't going to change that.

Click to expand...

This is in the city of Wilmington not too far from the ocean. I can get the tower to be in the range of 70 to 76 with very little issues. We have a fairly consistent ocean breeze and that helps a little bit.

Next question: Are you controlling the fan speed of the tower or are you controlling the water feed rate to the tower? I assmue its the fan as that is by far the most common way, but you know the old addage about assumptions so I'm asking.

Click to expand...

As this is a confidentiality thing, I can only say so much about the project. I have Tank A and Tank B. Cooling tower is between Tank A and Tank B. We pump 100 Gallons per minute through the cooling tower. We pump 105 gallons per minute out of the cooling tower into Tank B. Pumps are controlled by pressure transmitters at the bottom of the tanks. I am not worried about the pumps. I am controlling the cooling tower fan speed to maintain a set temperature of 76 from 6AM to 6PM and a temperature of 72 from 6PM to 6AM. Our Machines run most of the day. The only thing I can say is that we have a retort system. As far as the product is concerned, thats top secret. The retorts enter a cooling phase and send back water to the cooling system at usually 80 to 86 degrees with an occasional 90 to 92 peak. Tank A Equalizes the temperature and sends it to the cooling tower. We want a cooler temperature for the water so we can push more product through the machines as fast a possible. I am having these temperature swings in the cooling process and my return tank starts heating up and the retorts start becoming less efficient and i have to start pumping CITY water which is illegal. I have had return water temperature as high as 90 degrees because the cooling tower is being erratic.

So thats a look at the big picture of the project.

I will have to dump parts of the program out so I don't breach confidentiality. This PLC runs all the retorts, cooling tower/pump vfds, and waste water system, as well as retort boiler. So please bear with me as I get it ready for you guys.

I had an outside contractor modify the system when we added the wastewater, and it has been weird ever since.

The trend will reveal all!

I'd say it's under damped, but who knows.

Mostly we just want to see your PID and its configuration as you indicated you weren't sure if you had it configured correctly, and we've seen plenty of cases where helping someone just get the PID programmed correctly solved a huge part of their problems.

Also a trend will really help in suggesting how to tune the system.


At 100 GPM tower feed and ideal conditions your tower is operating at about 30 tons, but based on your stated temperature extremes your cyclic cooling need is somewhere between 43 and 61 tons. Tank A gives you a bit of a thermal flywheel, provided the tank capacity is large enough. Dual tank or baffled tank systems are very common on cooling towers. Since your cyclic cooling load is so much higher than your tower's capacity then I'm going to advance the suggestion that you consider the overall design of the tower system and not just the control, especially since you are trying to maximize production. Cooling towers are dirt cheap and so there isn't any logical reason to futz around with an undersized system at the expense of production capacity.

Towers also usually don't really require the advanced control schemes that sometimes end up being applied to them. Its not unusual to just run the fan at full speed when temperature > sp, and turn it off when temperature < sp with some small hysteresis band at the sp. If sp <= wetbulb + 7 then the only thing a VFD gains you is soft starting/stopping of the fan and a little energy savings in the winter time, aside from that the VFD doesn't help the actual tower performance. Large tower systems have multiple fans and its not unusual to simply sequence fans on/off. This simplifies the control quite a bit. What I'm saying is that you might get equal performance out of the tower system if you simply had two or three fan speeds and staged between them. I've done quite a few cooling towers and haven't seen one yet where a VFD on a PID performed any better than one where fans were staged up/down in steps. I'm just throwing that out there for your consideration. We'll still be happy to help you with the PID.

If you know the temperature of the water coming back from the process you may want to use that to start cooling more agressively than your tank temperature would indicate. You may have a sensing lag that is causing you an issue. You should be able to come up with a disturbance rejection contribution to the control output based on returning water temperature. You know the hot return is going to heat up the tank. Why wait for that to happen?

But, as Alaric said, you may just not have enough peak capacity to handle the returning heat, in which case no amount of control will help you.

Keith