Steam Pressure Reduction vs Temperature 4

[BronYrAur]

It has been a little while since I studied my thermodymanics, so I need a little help here. Reducing steam pressure through a PRV is an isenthalpic process (no enthalpy change), correct? So If I have 100 psig (338 deg F) saturated steam and reduce it to 15 psi, what changes? I assume that I will have superheated steam at 15 psi, right? How can I determine what the temperature is?

Where I am going with this is that I currently have a steam to water heat exchanger being served by 15 psig steam. However, this 15 psig steam is coming immediately out of a PRV that reduces it from 100 psig. I want explore the possibility of just running a 15 psig main and not having to reduce the pressure, but a saturated 15 psig main is only about 250 deg F. I assume that I have a higher temperature steam now, but how high? Is that even inportant in the grand scheme of heat transfer, since I know that the latent heat is the driving force - at least I think it is.

Any thoughts?

 

[UtilityLouie]

Typically you can find a steam property calculator that you can find the enthalpy of the steam into the PRV - then you enter your new pressure with the enthalpy of the entering steam.... and voila... you should have your steam temperature on the downstream side of the PRV.


A good steam state calculator is here....

http://www.chemicalogic.com/steamtab/companion/default.htm

and it's free...

Hope that helps....

 

[katmar]

You are correct about the temperatures. At 100 psig saturated the enthalpy of the steam is 1189.6 btu/lb. The enthalpy will remain the same as it expands to 15 psig, giving a theoretical temperature of 300 deg F, compared with saturated 15 psig steam temperature of 250 deg F.

But this increase in temperature may actually reduce your heat transfer, rather than increase it. A portion of your exchanger area will be used to transfer the sensible heat from this superheated steam, bringing its temperature down from 300 F to 250 F without any condensation. The heat transfer coefficient for sensible heat transfer is usually much less than it is for latent heat transfer (condensation). You may well find a 5 to 1 ratio of heat transfer coefficients.

So, depending on your temperature differences, you could end up with a situation where a significant portion of your area is consumed by sensible heat transfer, and you are actually getting only a small amount of heat transferred in that area.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[Latexman]

Katmar,

If BronYrAur's water at the steam inlet end of the exchanger is less than (or ~ equal to) the saturation temperature of 15 psig steam, condensing will start immediately. There will not be a dry desuperheating section of the exchanger. Instead, thr superheat will transfer into the condensate film itself.

Good luck,
Latexman

 

[TBP]

In order to calculate superheat, you'll need the dryness fraction of the steam upstream of the PRV. Very few plants have a means of determining this value. In any event, if you're dealing with "normal" (general industry) conditions - saturated steam under 300#, superheat after PRV stations is much more of a theoretical issue than a practical one. Typically, the steam isn't 100% dry to begin with, and the bit of superheat you get across a PRV provides nice dry steam downstream. Once, we actually installed a thermometer immediately after a PRV station - 125# down to 10#. At high loads, there was some superheat, but at low loads, essentially nothing. A few feet down the line, any superheat was gone entirely.

If you're starting off with any amount of superheat ahead of the PRV station, that CAN be an issue. However, I've never seen an operational problem due to superheat in installations with pressures like you're dealing with. Botched condensate piping on the other hand ...

 

[katmar]

Latexman,

Whether you regard the heat being transferred from the superheated steam as being transferred to the condensate film or to the tube you are still looking at sensible heat transfer, which is slower than latent heat transfer. Steam at 300 deg C will not condense at 15 psig until its temperature has been reduced to 250 deg C, and this is the slow step.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[Latexman]

katmar,

The local heat transfer coefficient for desuperheating steam with a falling film of condensate is quite large. There is essentially no resistance. He can include the duty of desuperheating in the total heat duty, use the temperature difference between sat'd 15 psig steam and his water, and he'll be just fine. I've sized dozens of condensers this way with satisfactory results.

Good luck,
Latexman

 

[katmar]

Latexman,

I disagree that you can condense superheated steam with condensate that is at the saturation temperature, but it is actually irrelevant. What I was trying to get at was that BronYrAur would not improve his situation by using superheated steam.

And as TBP has pointed out, the practicality of the matter is that there would be essentially no difference.


Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[Latexman]

katmar,

My point was that your image of condensing superheated steam is flawed. As long as the water at the steam inlet end of the exchanger is less than the saturation temperature of 15 psig steam, condensing will start immediately and there will not be a dry desuperheating section as you alluded. For somewhat similar reasons, that’s why a glass of iced tea sweats in < 100% relative humidity.

So even if BronYrAur has the superheated steam predicted by the textbook calculation, he will not have any heat transfer problems if he switches to 15 psi saturated steam as he is looking to do. He’ll have the same *real* temperature difference he has now, but with less heat duty, i.e. no superheat to remove.


Good luck,
Latexman

 

[rmw]

Latexman,

If you are right, some major big name heat exchanger manufacturers are wasting lots of time and effort making their heat exchangers with DSH sections, a very expensive part of the Hx. Could it be that they are all wrong and you are right. I think they will be quite shocked.

I firmly disagree with you as well. Katmar is right. The heating steam does not begin to condense and give up its latent heat until it reaches saturation temperature via a sensible HT process regardless if this heat transfer is against a dry wall or a wet wall.

Some of the wet wall you envision, if it exists at all, will be revaporized in the process and HT surface will be required to recondense this vapor as well, robbing the HT of surface area.

And, you are incorrect on another point. The one benefit of SH steam, if there is any, is that the delta T of the superheated portion of the steam is at a higher value than the saturated steam so at least a portion of the SH heat transfer is at a higher LMTD for what that little bit of driving force is worth.

An iced tea glass sweating on a less than 100% humidity day is not related to superheat. It is related to partial pressure relationships of the vapor in the air surrounding the tea glass.

rmw

 

[Latexman]

rmw,

Under conditions similar to those in this post and as I prefaced each time before, provided the condensing surface is cooler than the saturation temperature of the steam, I know I'm right. I learned it from the guys in the world's largest chemical company that *tell*, not ask, the heat exchanger manufacturers what to build.

What's in it for these companies to make smaller heat exchangers for companies that have no heat exchanger expertise? Now do you get it?

I repeat, provided the condensing surface is cooler than the saturation temperature of the steam, condensing begins immediately.

What is the velocity of the superheated steam flowing at the boundary of the heat transfer area? Zero, right? Well, if it's velocity is zero, it's going to sit there, cool, and condense! When the first bit condenses the volume of the steam collapses to the volume of condensate which decreases pressure and induces more steam into the heat transfer surface. And on and on and on!

The superheat is removed by a mechanism of condensing and re-evaporation at the condensate/vapor interface. For superheated steam (within limits at constant heat level), condensate loading is lower, the film thickness is less, and the condensing coefficient is slightly higher. Desuperheating occurs at the expense of the temperature difference between the superheat temperature and the saturation temperature.

Good luck,
Latexman

 

[katmar]

Latexman,

I like to believe that I am open-minded enough to know that learning something to be a truth while at university does not necessarily make it the truth. Some of the most fundamental truths I learned as a boy, I later found out were just mythology. So I have been doing some serious thinking about what you have experienced with superheated steam condensing. But we should also bear in mind that money and prestige also do not turn myths into thruths.

My own hands-on experience with direct contact (rain tray) condensers has proved to me that you are correct when you say that there is essentially no resistance to heat transfer when condensing superheated steam into cold condensate - even in the presence of non-condensibles. But these direct contact condensers are always fed with cold water that only achieves a 3 or 4 degree C approach to the saturation temperature.

It is impossible to condense steam (superheated or not) into condensate at the saturation temperature. This is one truth that I (and hopefully you) do accept. So it means that the condensate film in your model has to be significantly sub-cooled. In your model, if you use a higher HT coefficient for transfering the heat from the superheated steam then you must off-set this with a lower temperature difference between the condensate and the cooling water on the other side of the tube wall. In the classical model of a tubular condenser the condensate is assumed to be at the saturation temperature (unless the condenser is specifically designed with a sub-cooling zone).

This balance between higher HTC and lower temperature difference in your model probably comes back to giving the same net effect as the classical model (which would have low HTC and higher temperature difference). Or it could be that your experience with superheated steam is explained by TBP's very pertinent comment that in real life superheated steam from let down valves is not a problem because of the quality of the steam before being let down.

Whichever model we use, at least we agree that BronYrAur will not be any worse off if he feeds his exchanger with saturated 15 psig steam.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[Latexman]

katmar,

Is it impossible to condense steam (superheated or not) into condensate *at* the saturation temperature? On a macroscopic scale, yes, but at the vapor/liquid interface there is a dynamic equilibrium situation going on. Water molecules are continuously going back and forth into the liquid phase (condensing) and vapor phase (evaporation) depending on what direction they are headed (chaos), what energy state they are at due to their past history of collisions and thermal treatment (thermodynamics), local pressure gradients (fluid mechanics), etc. This is what “vapor pressure” is.

My experience is mainly with vertical, vapor-in-tube, downflow condensers where condensate subcooling is very effective due to falling film heat transfer. The superheat removal mechanism of condensing and re-evaporation at the condensate/vapor interface maintains the condensing-surface temperature at essentially the same level as that obtained with saturated steam. Of course, the temperatures decrease through the falling film to the tube wall. My idea of a little subcooling, 5 to 20^o C, may be significant subcooling to you. We usually like some subcooling to minimize vent losses, especially when the condensate is an organic chemical.

We have operations where the quality of steam is not important, and TBP's description is alive and well. We also have operations where the quality of steam is critical and we maintain significant superheat.

Yes, BronYrAur should be fine if he feeds his exchanger with saturated 15 psig steam.


Good luck,
Latexman

 

[BronYrAur]

Thanks everyone for your help. I see my post led to some lively discussion. I'll verify it with the heat exchanger manufacturer, but your answers have convinced me that switching to saturated steam will most likely not reduce my overall heat transfer capability. I appreciate the time everyone took to reply. Thanks again!

 

[TBP]

The upside of using lower pressure steam in a HX, is that it has (marinally) more latent heat. The downside is that with LP steam is that everything - control valves, traps, piping, the HX itself - needs to be larger, in order to handle the required steam flow.

 

[rmw]

I'm sorry, I just want to be on record as saying "I don't buy it". But I'm not going to expend much more effort on this.

Latexman, you defeat your own arguement and make ours with your statement "It is going to sit there, cool, and condense...." That makes our case. Saturated steam doesn't 'sit there' it condenses when it gets there. The sensible heat transfer mechanism that Katmar discussed is what is going on while the steam is "sitting there." Meanwhile a much lower heat transfer rate is occuring while it is 'sitting there and cooling.' Meanwhile, the overall duty of the Hx is being robbed by the amount of area where steam is just "sitting there." Our case is made. Thanks

I do have one parting comment. We do seem to agree on one thing. And that is your comment that the condensing surface has to be cooler than the saturation temperature for condensing to occur. Boy, we are making real progress.

rmw

 

[TBP]

Is the argument in motion based on the differences between applications like surface condensers under steam turbines vs process HXs?

At the end of the day, in the real world, if you feed superheated steam to a process heat exchanger, it'll behave as if it's airbound until you get past the sensible heat that is superheat.

 

[Latexman]

TBP,

I'm referring to my experience with vertical, vapor-in-tube, downflow condensers.

"If you feed superheated steam to a process heat exchanger, it'll behave as if it's airbound until you get past the sensible heat that is superheat" is exactly the argument. RMW agrees with you. My view is "provided the condensing surface is cooler than the saturation temperature of the steam, condensing begins immediately."

Gentlemen, I quote from Perry's 7th Edition on page 11-11, "If the vapor is superheated at the inlet, the vapor may first be desuperheated by sensible heat transfer from the vapor. This occurs if the surface temperature is above the saturation temperature, and a single-phase heat-transfer correlation is used. If the surface is below the saturation temperature, condensation will occur directly from the superheated vapor, and the effective coefficient is determined from the appropriate condensation correlation, using the saturation temperature in the LMTD."

It goes on to describe how to determine whether or not condensation will occur directly from the superheated vapor by calculating the surface temperature by assuming single-phase heat transfer. Equation 11-26 is given as:

Tsurface = Tvapor ? U/h x (Tvapor ? Tcoolant)

h is the sensible heat-transfer coefficient for the vapor
U is calculated by using h
both are on the same area basis

If Tsurface > Tsaturation, no condensation occurs at that point and the heat flux is actually higher than if Tsurface ? Tsaturation and condensation did occur. It is generally conservative to design a pure-component desuperheatercondenser as if the entire heat load were transferred by condensation, using the saturation temperature in the LMTD.

This is exactly what I was trying to say. I thought I was being clear, but . . . .


Good luck,
Latexman

 

[Latexman]

Tsurface = Tvapor - U/h x (Tvapor - Tcoolant)

Good luck,
Latexman

 

[katmar]

It is generally conservative to design a pure-component desuperheater/condenser as if the entire heat load were transferred by condensation, using the saturation temperature in the LMTD.

So the HX manufacturers who design for a dry desuperheating zone aren't such bad guys after all. They are actually trying to save us some money! ;-)

This has been (for me at least) an instructive discussion, plus the reference to Perry 7th Ed has allowed me to fix the typo in my old 5th edition. Thanks to Latexman for perservering with what he believed to be right in the face of all our arguments. One more "truth" becomes a myth.

regards
Harvey


Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com