Delta T for Steam on Heat Exchanger

[BenjaminM]

Good morning,

I'm looking at heating water using a heat exchanger with steam.

I have dozens of examples of sizing heat exchangers with liquids on both sides, but I have not found any references to using steam on one side.

I know how much heat is in a pound of steam.

I have estimates of overall heat transfer rates of steam condensing and water, 250-400 Btu/h.ft2.*F.

I know what temperature I want the water coming in and going out at. I know the temperature of the steam entering.

What I don't understand is what the delta T for this situation is. Would you assume the condensate is the temperature of the steam? Would you calculate the log mean temperature?

 

[TBP]

The last steam to water HX I was involved with was for 20 USPGM water at 35*F water inlet and 135*F outlet, and used 15 PSIG (250*F) steam at the HX inlet. Both the inlet & outlet of the hot side were taken as 250*F, as there is typically no subcooling of the steam condensate in these applications. The LMTD of my HX was calculated at 159.5.

For the steam input required, I used "Hook-Ups - The Design Of Fluid Systems" by Spirax Sarco. (I believe that it's still in print.) USGPM X 500 X Delta-T/1,000 = 20 X 500 X (135-35)/1,000 = 1,000 #/hr of steam required. I've used that book for almost 40 years, and it has never let me down.

 

[TiCl4]

Myrdale,

This is a mixed-phase condensing type application. Exit condensate temperature is under your control based on how liquid-full you operate the steam side (shell?).

By operating partially liquid-full, you break the exchanger into two parts for analysis. The first is a constant T condensing application that is reduced to the square footage available to the steam based on chosen liquid level. The second is treated as a liquid-liquid exchanger for the remainder of the Hx area. My understanding of this operation (with a liquid level) is that it will typically require more Hx area for any given duty, but there is an OpEx reduction in total reduced steam requirements.

I have not personally sized one of these, so there are likely some intricacies in calculating the condensate-side film coefficients to determine the overall U for the liquid-section. However, there are Hx companies that will do this sizing for you and already have specific designs for this, so I suggest checking in with them to get more specific guidance.

If you design the system to have no liquid level in the exchanger, then exit temperature will be approximately the steam temperature - i.e. the most sensible heat of the hot condensate is wasted. It's an easier analysis this way, as you are only dealing with condensing steam, but it will require more OpEx to operate due to the wasted heat from the condensate. The loss in efficiency can be made up, in part, by recycling the condensate back to the boiler feed water tank.

 

[don1980]

Some subcooling of the condensed steam may occur, but for design purposes that should/can be ignored. The design intent is to use the heat that comes from condensing the steam. Thus, for design purposes the steam side and condensate are in thermal equilibrium (same temperature). The temperature of the steam and the condensate is the saturation temperature of those fluids.

The superheat in the inlet steam is also ignored because: (1) it represents a trivially small amount of heat, (2) it's not useful - it diminishes rather than enhances the exchanger performance, and (3) it's quickly eliminated when the steam enters the exchanger, as long as the amount of superheat isn't excessively high. If the superheat is very high, then install a de-superheater on the steam line upstream of the exchanger so that it doesn't impair the exchanger performance.

 

[EdStainless]

Are you returning the condensate to the system?
In many cases people don't want the condensate back because of the risk of contamination.
In these cases using direct steam injection is the most efficient way to heat the water. (companies like Pick)
This is very common with food products.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[pierreick]

Hi,
As per suggestion above ( EdStainless ) consider the document attached to support your work ( direct injection).
Good luck
Pierre

 

[TiCl4]

Ed,

Today you taught me that Pick is an actual company. All this time I’ve called and heard them called “pick heaters” as a general term - I didn’t know the source of it! Kinda like Kleenex.

Also, to other’s responses on ignoring sensible heat - are exchangers that operate with a non-zero liquid level to recapture some of that heat not widely used? I thought large applications used them (power plants?).

 

[EdStainless]

Yes they do. Feedwater heaters (actually re-heaters) have three distinct zones, de-superheat, wetted heat transfer, and condensate flooded.
Each of these contributes to the overall heat transfer and is controlled.
There are many coal fired power plants in Russia designed to run at one power level that used direct stream injection for heating feedwater.
Ver efficient but not very flexible.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[BenjaminM]

Thank you all for the quick detailed responses.

TBP, I had worked out the LMTD similar to what you've listed assuming no cooling on the condensate. I'll see if I can track down a copy of the book you referenced. I've seen a similar method in a Warren Control Valve catalog.

TiCL4, it will be low pressure steam on the shell side, around 20 psig. My assumption was there wouldn't be a large amount of liquid standing in the heat exchanger.

Don1980, that was my assumption that there was no significant cooling of condensate.

EdStainless, in this case, the condensate is a closed loop. The water being heated is part of a process, I don't think we'd want to risk contamination with boiler chemicals, etc with the product. But yes for heating tempered water for other processes, we've done direct injection with excellent results (other than condensate buildup over time).

I have not used a Pick brand heater before, but I do have some experience with Hydro-Thermal heaters. They look similar in principle to one of the options I see with Pick.

 

[EdStainless]

Since there will be condensate standing in the bottom you will get some sub-cooling.
And you need a way to control the water level (drains).

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[georgeverghese]

Also take note of the effect of subcooling on the process control scheme you would be using to enable temperature control of the hot water exit stream as described on page 162 and page 778 of "Process Heat Transfer" by DQ Kern, which is applicable to LP steam.

 

[MJCronin]

A quick comment ..... PICK and other injection steam heaters have been around for decades and work well ....

They are noisy as hell ....

Twenty years ago I specified two for heating Grease Washdown water in a steel mill .... the mill steelworkers complained about the noise ...... steelworkers !!!

Just my two cents

MJCronin
Sr. Process Engineer

 

[BenjaminM]

georgeverghese, thank you for the references in Kern. We are probably looking at about 25 psig for the steam supply. I don't have exact details on what style steam trap is going to be used, but I was not expecting to flood the heat exchanger with condensate. We will definately keep this in mind.

I did think it was interesting that Kern state on the next page "all heating services employing relatively air-free steam a value of 1,500 Btu/h.ft2.*F will be used for the condensation of steam without regard to its location." That is five times higher than the 300 Btu/hr.ft2.*F that I am using.

MJCronin the Hydro-Thermal heater I references is akin to a banshee. We rebuilt the heater a couple of years ago and found significant damage to the plug. After replacing the damaged parts, and adjusting positioning of the mixing tube (if that is the correct term), it was significantly quieter.

 

[EdStainless]

In a feedwater heater they use different heat transfer values for de-superheat, condensing, and flooded zones.
Heat transfer is very low in de-superheat, but the delta T is very high.
Just as heat transfer in flooded is high but the delta T is low.
These are two pass HX, so in effect the flooded section is the pre-heat.
They use an internal partition to control the amount that is flooded.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[georgeverghese]

In the chapter on condensation of pure single component vapors, Kern shows a plot of heat transfer rate over horizontal tubes with an average of about 400btu/hr/ft2/degF for h, which is in stark contrast to the value of 1500btu/hr/ft2/degF in the earlier chapter you've mentioned. Think this is to account for the the fact that dropwise condensation is about 6-8x higher than filmwise condensation.
Agreed, temp control by flooding the tubes with condensate is most likely not the way to go here with 20-25psig steam, as you could have subatmospheric pressure in the shell due to subcooling. You then wont be able to evacuate the condensate out of the shell without a pump as Kern points out.
More thermal rating calcs may need to be done to ensure that P_shell does not go below a tolerable low limit even if you employ pressure control as the basis for your temp control of exit water. Work out what T_film and T_surface is for new and clean (no fouling) on tube od.
Perry Chem Eng Handbook 6th edn shows some other alternate schemes for steam heating- fig. 22-163 (b)seems to look good for this case to me at the moment. Sadly this section on temp control schemes has much of this material scrubbed out in the 7th edn.
You might be able to make this HX work with 20-25psig steam with plain TC on exit water resetting shellside PC, else consider these alternate bypass schemes. Or use slightly higher pressure steam. What seems to be a simple engineering problem has a sting in the tail for the unwary.

 

[EdStainless]

Wholeheartedly agree with George's cautions.
I have seen U in condensing service run from 300-700, it depends. 500 is fairly typical.
Tube diameter and cold-water flow velocity are the two big factors.
Tube material almost doesn't matter as long as the walls are thin.
If you are trying to heat the water then you will likely end up with a fairly low U.
Higher heat transfer infers higher water flow and lower delta T.
There are also HEI design guides you could refer to.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed