cooling a di process water storage tank

[cmh2k]

The company I am working for would like to do the following. There is a ~1500 gallon storage tank. In the tank (maintained at 80 to 85% capacity) is about 1200 gallons of di water, continuously cycled by 2 US Electric Motors 10 hp pumps at a rate of 20 and 22 gpm (there are two separate pumping circuits). Make up water is added at a rate of 4 gpm as needed to maintain the capacity. The tank which is made of plastic is coated on the outside with a heavy hard insulation. Coiled around the outside of the tank,under the inslutlation layer(hopefully in contact with the outside surface of the tank, is around 600 feet of clear plastic 1/2 inch tubing wrapped around the middle third (or so) of the tank. I'll come back to the tubing in a moment. The processes the company uses the water for calles for water at around 20 C. What they found was, although the makeup water is 20 C when added, the continuous cycling of the water raises the temperature of the water to around 34 C in the cycle piping and 36 C in the tank. I hope I still have someones attention. To shed this excess thermal energy they added the hose coiled around the outside of the tank and then covered the tank with the insulation and ran cool city water through the hose. Here is what I was told about thier results. The pressure difference between the ends of the hose was 40 psig at one end and 0 psig at the other. The flow rate with the city water was 0.75 gpm. They were able to produce a 4 C temperature difference between the inlet and exit of the hose. A suggestion has be made to use chilled propylene glycol (-5 C?) instead of cool city water. I'll add, the hose has a volume of ~24.4 gallons, which is ~ 212 lbs of propylene glycol. I'm sure the heat transfer between the tank wall and the hose is probably extremely low. I'm hoping someone might be willing to suggest a mass flow rate or a pump size to move enough chilled glycol through the tube to have the desired cooling effect on the stored tank water. The tubing is there, they want to use it, there is certainly several easier ways to accomplish this but there it is. I welcome all suggestions and criticisms. If there is other information you would like please ask.

Matt

 

[tickle]

Cool the water at the user off take point(s), this is probably the simplest and cheapest method. The operating costs are likely to be huge using the system that you have in place. Plastic tubing and plastic tank do not sound like they are conducive to good heat transfer.

What is the ambient temperature around the storage shed? What temperature is the mains water? What is the water used for - is it a pharmaceutical use? Do you need to ensure that there is no possibility of cross contamination between coolant and DI water?

It sounds as though the coil has been added and not designed. And will probably never do the intended duty.

Do you have access to the coil at any point? Can you parallel up the coil (decreasing pressure drop and increasing flowrate and cooling)?

 

[ClydeMule]

In my opinion you should not worry about trying to cool that huge insulated tank with 1/2" plastic tubing. The heat transfer is pretty lousy, and over time you are going to dump a lot of water down the drain. You are also going to have to expend a lot of energy to pump glycol thorugh that tubing, and to get it down to -5 C.

Use one of the existing pumps, or add another one that will pump the DI water to a chiller. You just will have to make sure that the chiller is DI compatible. You probably will have to go to a process chiller mfg. The evaporator will have to be stainless steel or titanium.

Another option is to get a standard packaged chiller that will pump water through a heat exchanger, and have your pumps run the DI through the other side of the heat exchanger. This will give you the ability to shop around a bit more and not have a special chiller built for you.

For the previous chiller setup you would cycle the chiller based on the water coming into the chiller, assuming there isn't a large temperature rise between the tank and the chiller inlet.

Let me know if you would like to get names of some mfg's. I actually build a lot of DI compatible systems myself, but I am willing to help with no strings attached.

Regards,

Clyde

Hey look for the "coil in tank" thread in this group. Mr. Montemayor explains this situation extremely well.

 

[IRstuff]

In addition to all the aforementioned issues, being on the outside of the tank, the tubing probably winds up sucking in a big chunk of solar load, so the effective heat transfer coefficient is NOT effective at all.

But, it would still seem to be a losing proposition, since the water will have to cooled many times over before it actually gets used.

The suggestion of point-of-use chillers seems to be more efficient to me. Something along the lines of a scaled up bottled water chiller? They're relatively efficient and you only cool the water once.

TTFN

 

[saxon]

cmh2k, Thermo 101, "Any energy entering a system that is not used as work will be converted to heat". On your pumped recycle loops the excess energy from the pumping is being converted to heat. On the return lines to the tank from these pumps install heat exchangers that will remove the excess heat. Automatic Control can be accomplished by installing a closed loop temperature control using a thermostat and CV in the exchanger system. As for the design details of the control loop you'll need to take a close look at your system requirements.

Hope this helps.
saxon

 

[cmh2k]

Thanks to everyone for such great suggestions. As some background, you can think of me as an engineering intern who has been handed this problem. I apologize for my mishandling of jargon or any misunderstanding of fundamentals.
I would like to discourage the use of the 1/2 tubing which I'm sure was added not designed. I like the point of use idea the best. Why chill all 1200 gallons if you use only 500 gallons in the daily processes? I don't actually know what volume of water is used per day, someone at the plant should know. Let me answer the questions in order read.

Tickle,
1) The temperature of the area where the tank sits may get above 100 C on hot summer days.
2) Forgive my lack of understanding, I'm not sure what the "mains water" is. The temperature of the tank is around 36 C, that of the di makeup water is around 20 C.
3) The use of the di water is for pharmacuetical processes. I'm not sure specifically what processes. I do believe however that cross contamination would be a problem.
4) The coil is only accessible at either end.

ClydeMule,
Any helpful hints you can throw my way would be greatly appreciated. I have people who insist I figure out the flow rate of the glycol to cool the storage tank. I'm not sure how to go about this. Here is what I think I know.

1) There is some confusion as to whether we are cooling a temperature difference of 36 C - 20 C = 16 C or a difference of 36 C - 34 C = 2 C. The latter is the temperature difference between the two pumping legs and the storage tank.
2) To cool 1200 gal of water 16 C will require the removal of around 610000 Btu.
3)To cool 1200 gal of water 2 C will require the removal of around 36000 Btu.
4) Assuming perfect heat transfer (ha) and using
q = (C)(m)(T2-T1)
C ~ 0.56 Btu/lb F
T2-T1 = 16 C (61 F)
q = 610000 Btu
we find,
m ~ 17900 lb propylene glycol.

OK, is this idea just stupid or am I. I'm already frustrated. I've been looking at heat exchangers at flateplate.com. If you can suggest a couple of mfg's I really would appreciate it.

I really like the point of use idea and I'm going to try to push the people I'm working with in that direction. Thank you all for your patient help. Please add any thing else you think might be helpful.

Matt

 

[ClydeMule]

IN my opinion, this is one of those cases where you just have to look at the system to decide what to do, before you whip out the slide rule. Look at the sizes of things, do they match up?

From what you have described it doesn't sound like it. A 1/2" PLASTIC hose wrapped around a PLASTIC tank that is 3-5' in diamter. Doesn't match. Even if it was copper tubing and a copper tank, there just is not enough in contact with the hiuge surface area of the tank.

So how to get more heat transer with you current setup?

1. Increase the flow:

Already it takes 40 psi to get 0.75 GPM through the tubing. It will take you 80 psi to get 1.25 GPM through the same coil. Not much help, even though you spent more money on pumps and electricity.

2. Decrease the temperature of the glycol/water.

23 F glycol is not cheap to make. And if you can't flow enough, and have huge thermal resistanace you spend a lot of money on refrigeration and on a big pump we already mentioned.

With that being said, I hope we have shown that using the existing setup is not a good idea.

Sources of heat:

When trying to size a chiller for cooling a tank, here are the loads you must analyze:

1. Heat Gain from space: Ambient air heats up the water. Usually smaller than the other heat loads, and is hard to evalutate so it is somewhat ignored.

2. Heat gain from pumps: In this case all of the HP of your pumps is converted to BTU.

3. Make up water- The makeup water, if warmer than than the desired temp is a heat load. 500 X GPM X DT. DT equals incoming water temp minus desired temp.

4. Pull down heat load- This is when you fill up a batch, and have to pull it down from xyz temp to 123 temp in xyz minutes. This does not seem to be your case.

5. Chemcial Reaction- N/A in your case.

Having said that, we need to figure out where to cool this thing:

1. Cool the makeup water-. Easy, soultion, but tough to do if the flow is intermittent. Chillers like constant flows. Also it is such a small flow relative to a large volume you might have a hard time getting to the temp you want if you are trying to be accurate with temp control.

2. Cool the water in a side arm- Use the recirculating pumps that you already have paid for to pump water through the cooling system. Saves from having to buy another pump. Make sure though that the additional pressure drop imposed by the cooling system doesn't adversely affect flow.

3. Cool the water as it leaves the tank- Like in case #1, this is tough witout constant flows. However, you have tight temp control.

Hey they are kicking me out of the building. To be continued...

Give me call 760-489-9920 ask for Jeff.

Regards,

Clydemule

 

[tickle]

Pharmaceutical water! This adds new complexity to the requirements. I could write a lot about it, but it is probably all out there.

For the heat exchangers you need to consider double tube sheet HE's or double pipe. PHX are generally NOT acceptable because you cannot drain them properly.

Is there a requirement for the water to be heated up to sanitise? Do you have UV sanitisers fitted in the loop.

I designed a system for DI water that was held hot and used hot by most users. When cold DI was required then the water was cooled at the user point.

For end user HE's these need to be after the isolation valve and be self draining so that there is no hold up.

I would search for the requirements for pharma water. The chances are that the standard acceptable requirements for pharma water have changed since the installation. Any thing that you do to the system should be in accordance with these requirements.

How is the DI water made?

For my information - How do you control tank breathing?

For your information - By Mains water I mean city water from the distribution main.

 

[rmw]

Without commenting on whether or not I believe that the set up you have will actually transfer the heat necessary to do what you want, in order to be able to transfer all the heat you can, you should consider the use of a heat transfer mastic.

Suppliers of metal jacket heaters for metal tanks use it to fill in the air voids between the jacket surface and the tank surface, and enhance the heat transfer across what would have been an air gap.

Since you have cylindrical tubing only in contact with the vessel wall at the tangent point of the tubing, you need something to fill in the air gaps, and transfer the heat to the tank wall.

Try googling Tranter, and/or Mueller, or plate coil or jacket heat exchangers. One of those should get you a source for a good heat transfer mastic.

Good luck.

rmw

 

[greg87]

Since most of the temperature rise is due to the addition of the pump energy, you may want to look at the loop hydraulics to see if you can reduce the flow rate, or the pressure drop. If you can, then you can reduce the pumping power by using a smaller impeller or by using a VFD on the pump motor to reduce the rpm. The flow rate may be higher than it needs to be (there’s a thread under pharmaceutical that addresses mimum flow). Even a modest decrease can reduce the power considerably, since pressure will decrease as well. The biggest rise in temperature will be when there is minimal or no use of DI water, and therefore no "cool" makeup water is coming into the system, for instance over wk ends if your facility doesn't operate 24-7. If there are periods of no use, a VFD could reduce pump rate during those periods (based on pressure or flow rate monitoring. So reduction of heat input would be the first option.
The amount of heat to be removed is relatively low. A small exchanger, using your chilled glycol, would take care of the heat load. Using a double tube sheet shell and tube would ensure that there is no cross contamination. Exergy now makes a little double tubesheet unit. You could put it on the return line of one loop, or on a slip stream off of a loop