Opinions: Tank Heating Control

Bullzi

Lifetime Supporting Member
Join Date
Jun 2012
Location
Colorado
Posts
1,530
Hi Everyone,
I am about to complete a panel that controls 2) 12kW 480VAC Tank Heaters. Thanks to everyone that helped me on my previous thread about designing of the panels. I have designed the panels to use Crydom SSR's for the control.

I have never done a tank heating application so I am looking for some opinions on how to best control the temp. I was planning on using a simple ON/OFF control scheme. I need to maintain the tank at about 85-90 Deg. F The tanks is a 30k Gallon steel tank and is not insulated. It is in a semi cold environment. Winter lows around 20 F.

My Question is: Is it better to have a narrow band of ON/OFF Temps and maybe be able to use one heater at a time to maintain the temp or is it better to have a wide band of ON/OFF Temps and use both heaters to heat the tank?

Just looking for ideas as I start the PLC program from someone that may have done this kind of thing before.

Thanks for the help!!
 
You say you want to maintain 85 to 90 Deg.
What dead band can you live with ?
I would use either PWM On Off control or Phase angle control either one set as a proportional controller with PID control I think this would give you the best control for the tank. you could set it up to only use one heater when the temperature difference is small and use both heaters when you have a large temperature difference There are many possibilities here.
 
Here is a quick question: Do you know if 24KW will be enough to keep the tank of 30,000 gallons of liquid 90F when it is 20F (or less) out? Is there flow into/out of tank? Depending on the dimensions of the tank, there can be a whole lot of cooling happening there.

Where are the heaters in relation to the measurement?

The SSRs that you have are discrete DC controlled in ON/OFF manner correct? ie. they do not do proportional control from 4-20MA or 0-10/0-5VDC?

I would use pwm control on a discrete output. Some PLCs have a "PWM" function built in, if not you can do it simple with a few timers. I would experiment with both heaters at the same time, and also with one heater leading and second lagging (ie, first heater scaled 0-50%command 0-100% heater, second heater 50-100%).
 
It takes the same amount of energy to heat the tank whether you use one or two heaters. Turning them both on at the same time will provide the fastest temperature increase and may be necessary when the ambient temperature is low.

My questions are deeper. What is the fluid? Mostly water? It is good to know the heat capacity. Second, does the fluid level change. It will take less energy to heat a tank that is half full than full. This will change the gain and time constant.

The question about the sensor placement is critical because it affects the dead time.

I could make an "Advanced Control" topic where "we're going to engineer the heck out of this"

It would be gross overkill but educational.
 
2x12kW for 30000 gallons in a non insulated tank? Are you sure about this?
For simple on-off heating applications mechanical contactors are more than enough. Pid control has nothing to offer you here.
Either way I think your heaters will probably always stay on as the pace at which the heat is removed from the liquid in the tank will be higher that the energy transferred by the heaters.
 
It takes the same amount of energy to heat the tank whether you use one or two heaters. Turning them both on at the same time will provide the fastest temperature increase and may be necessary when the ambient temperature is low.

My questions are deeper. What is the fluid? Mostly water? It is good to know the heat capacity. Second, does the fluid level change. It will take less energy to heat a tank that is half full than full. This will change the gain and time constant.

The question about the sensor placement is critical because it affects the dead time.

I could make an "Advanced Control" topic where "we're going to engineer the heck out of this"

It would be gross overkill but educational.

Do it. I'll pay attention.

To OP: The position of the heaters is pretty critical too. Are these band-type heaters or just elements dipped into the tank?
 
just guessing

for 30000 gallons/24 kw heater starting temp 85 deg to achieve 90 deg
it will take ~28 hours.
 
Don't understand why you guys don't go to SI units, it makes calculations easy.

1 Watt is 1 Joule for 1 second.

The specific heat capacity for a material is how many Joules is required to heat 1 kG by 1 Degree Kelvin

1 Litre of pure water has a mass of 1kG.

The specific gravity (SG) of a material tells you the mass of 1 litre.

So at its simplest:
30000 US Gallons = 113562.35 Litres
113562.35Litres * SG of 1 = 113562.35kG
Using waters specific heat capacity of 4.186Joules per Gramme or 4186 Joules per kG
113562.35 * 4186 = 475371997.1 Joules required to raise temperature by 1 Kelvin (Centigrade)

24kW is 24000 watts is 24000 joules per second
475371997.1 joules required / 24000 = 19807 seconds = 330 mins per degree or 5.5hours

So as kallileo said, are you sure about the 24kW, the energy losses from an uninsulated tank are going to increase that 5.5hours considerably. Your heaters are going to be ON a lot more than they will be OFF. And if you ever start with cold liquid it would take 6 days to warm up the tank.


The hard part in the calculation is knowing the energy loss from the system.
Really I would also want to know the mass and heat capacity of the tank material.

The really hard part is understanding why anyone would have an un-insulated tank, it is throwing away money.
 
Ditto what BryanG said. Knowing the range of possible time constants would be helpful to determine how fast a PWM should be.

Controlling the 2 heaters is easy. If the output of the PID is less than 50% then only one heater is required. There really isn't a need for rapid turning on or off the heater so I would do this I would not use the traditional PWM routine you normally see.

If the PID output is less than 50% I would add the current output to the accumulated output value. If the accumulated output value is 50% or greater I would turn on one heater and subtract 50% from the accumulated output put.

If the PID output is greater than 50% I would add the current output to the accumulated output
Since the accumulated output is greater than 50% the first heater is turned on or kept on and 50% is subtracted from the accumulated output. If the accumulated output is still over 50% then I would turn on the second heater and subtract 50% again.
Try it this on paper. You will see it works well.

BTW, if you have 4 heaters you can do the same thing only subtracting 25% for each heater.

Simple

If you are tricky you can add some sort of method to even the on time.

More complex but doable.
 
So other than the questions about the sizing of the heaters themselves, you have a dead simple heater control application.

Turn both heaters on if PV < SP, turn both off if PV >= SP. If you're using transistor outputs for your signal to the SSRs, you don't even need a deadband. Although you may consider one if you're displaying the state of the outputs so as not to confuse operators.

Anything with a PID or PWM is a waste of time for a slow heating application like this IMO. I've used variations of PID driving PWM, and just straight proportional PWM. In each case, it was simpler and resulted in faster heating to just go ON/OFF.

Now that's not to say it wouldn't be fun to go PID & PWM!
 
Last edited:
Oh, I don't know why.

You need more juice to do what you want to do, actually, a bunch more. In a tank that size, with enough power, On/Off would work fine.
This is 24Kw in a 300 gallon tank holding around 160 degrees, and works fine. There is fresh water introduced. This is in a factory in Indiana that is not heated. I know, it was cold there!

heater.jpg
 
Last edited:
Thanks for all the replies everyone,
Let me see if I can answer some questions:
  • Fluid: Bio-Diesel
  • The heaters were sized by others I was just told to make them work with my system. So I hope the heater Manufacture error-ed on the side of caution.
  • The PLC is a 1769-L33ER.
  • The tank may be insulated I need to check on that. When I was there this summer the tank was not insulated but I know they were insulating the pipes so hopefully the did the tank too.
  • The Heaters are placed about 8" above the bottom of the tank in 2.5" pipe.
  • The Temp Transmitter is about 12" above the heaters.
  • I also just double checked the temp range they want me to hold and it is about 70 +/- 10 so I have a fairly large dead band there.
  • The SSR's are straight switching SSR's. No 4-20 or any other proportional Control.

Hope I answered all the questions.
Thanks for the help!!
 
Thanks for the information it helps
You said in the first post it was a 30,000 gal tank now in the last post it only 3,000 gal that huge difference.
Is this a horizontal or vertical tank it does make a difference
The fact that it’s Bio-Diesel means that it’s most likely using it for engine fuel or heating oil
the thing I see that needs to be addressed is you cannot let the thermostat get uncovered. If it is not submerged in the oil it will cause the heaters to come on full power and over heat the oil this could cause a huge fire problem. Without having the tank dimensions it looks like the minimum volume in the tank would be about 1,000 gal to insure that the thermocouple remains submerged.
If it drops below that you may have a problem heating it. (no Temp feedback)
So you will need to monitor the level in the tank as well as the temperature.
A simple ON OFF control will do the job nicely. Set it up to turn on the heaters below 65 Deg. And off above 75 Deg.
It looks like they just want to keep the oil temperature up enough so they can pump it and use it
I believe that Bio-Diesel gels at a higher temperature then regular diesel
 
Turn both heaters on if PV < SP, turn both off if PV >= SP. If you're using transistor outputs for your signal to the SSRs, you don't even need a deadband. Although you may consider one if you're displaying the state of the outputs so as not to confuse operators.

This may work in some situations, but it's a bad idea on principle... with chatter and what not. An hysteresis is always good to have.
 

Similar Topics

Working on an industrial wastewater treatment plant... Do any of you have experiences with these? I'm changing our input signal from a...
Replies
18
Views
6,961
Hey guys. We've been learning Ignition and going through all the courses on Inductive University. I really like the software and the things you...
Replies
20
Views
1,026
Anyone use WEG vfds on here? Thinking of trying CFW300A01P6T2NB20.
Replies
0
Views
762
I would like to use this thread to ask for opinions on methods/practices. I am starting on a fairly large complete rewrite of code for several...
Replies
11
Views
2,373
Looking into using a Stride Managed network switch in an upcoming build versus a Stratix 5700. Just looking for others experiences, comments or...
Replies
10
Views
4,952
Back
Top Bottom