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?

 

[rmw]

It will come as a shock to the manufacturers all over the world of high pressure feed water heaters, who design their DSH sections very carefully in order to balance high velocities against tube vibration and dry wall margins so that the steam leaves the DSH section with a modicum of SH so as not to cut the tubes in the high velocity DSH section. The velocity is as high as it can be without causing vibrations so as to maximize the HTC in that zone so it can be DSH'd to get it to the condensing zone where the real HT is. The tubing is below saturation temperature in this zone until the very end. These Hx's typically have a 0ttd or a -1 ttd.

See

http://www.yuba.com/power.htm

or a good cutaway view here

http://www.babcockpower.com/index.php?option=products&task=viewproduct&coid=20&proid=35

Now, I don't know where Perry got his information, nor where he is coming from, but on this one, he has quite a set of detractors.

Harvey, maybe truth is just hard to find. I think the myths are flowing freely.

rmw

 

[katmar]

rmw,

The Perry formula surprised me too. However, what it means is not that there is always direct condensation but that direct condensation is a possibility under the right circumstances. I would not have expected the heat flux to be higher with superheated gas than with condensation unless there were extreme conditions. But I have not tried to calculate any HTC's to test this.

With the conditions described in you last post (high velocity and temperature) the tube surface temperature will approach the vapor temperature and there will NOT be direct condensation. It is therefore sensible (sorry for the pun) to design for dry gas at high velocity and take all the precautions you mention.

It is one thing to get sufficient area into the exchanger by making the conservative assumptions Latexman recommended, but it is entirely a different thing to design the exchanger for long and safe operation. I would never design an exchanger of this nature myself, whereas I have designed plenty of simple saturated vapor condensers. I would leave this to the specialists - and preferably one who would give a guarantee.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[Latexman]

rmw and katmar,

What type of heat exchanger are ya'll speaking of for a desuperheatercondenser?

Good luck,
Latexman

 

[rmw]

Harvey,

The Perry formula given is essentially a statement of the skin temperature for the conditions specified.

First, let's us for the sake of this discussion say that the skin temperature of the HT surface we are taking about is always less than Tsat. Yes, the skin temp in some of the DSH zones of the Hx's I mentioned are above Tsat for parts of the zone, but not for all of it. But that is irrelevant.

Let's carefully look at what Perry is saying. First he addresses sensible heat transfer with the tube surface being above Tsat which we are not dealing with. Then he states that condensation will occur directly from the superheated vapor when the skin temperature is less than Tsat. He is right in saying this. BUT, HE DOES NOT STATE AT WHAT RATE THIS HEAT TRANSFER OCCURS.

I have done the calculations, and in English units, it is at a rate around a U value of 52 at reasonable velocities while the condensing U value for the same Hx after the steam is cooled to Tsat is in the 450 range.

These calculations were done (first manually and then with web tools) for a Roberts type rising film evaporator common to the sugar industry. The steam velocities were nowhere near the range of the type of Hx's I gave links to above, (Latexman, they are called 3 zone heaters, each having a DSH zone, a condensing zone and a condensate sub-cooling zone) which even with the velocities that they get to (to the point of causing hydrodynamic whip-vibration of the tubing) only attain a "U" value for the DSH zone in the range of 160.

The Roberts evaporator studied obviously couldn't have that much velocity because once the steam reached saturation and became wet, it would cut the copper tubed calandria to ribbons. That is the same reason the designers of the high pressure heaters linked to above are so concerned about maintaining a 5F dry wall margin over saturation as the steam exits the DSH zone, because they don't want the (steel and stainless steel) tubing cut to ribbons. I have personally witnessed many heaters where conditions changed and the DSH zone was destroyed by the presence of moisture in the DSH zone. Note for the record that the tcoolant in the DSH zone is below Tsat for the heater pressure and the tube skin temperature in the region of the steam exit from this zone is below tsat as well. But I digress.

Now to Perry/Latexman. They are both absolutely right. The superheated steam will (eventually) condense, but the 64,000 dollar question is; at what rate? Latexman hits the nail on the head when he uses the phraseology "sit there, cool and condense". Saturated steam, on the other hand, as it enters the heater and contacts a cold tube, condenses immediately. No "sitting there" involved.

The "sitting there" Latexman describes is the blanketing effect that gives the "air-bound" symptom that TBP mentioned in his 26May post. Oh yes, it is transferring heat while it is "sitting there" but at a much lower rate of heat transfer, sensible heat transfer rates (single phase as Perry puts it) instead of condensing (he might have said two phase for condensing since a change of state of the steam takes place).

Since the goal of most Hx designers is to minimize the cost of the Hx, minimizing its surface area is one of the most cost effective ways of doing that.

Since the U values for condensing, in the 400-600 range, are significantly higher than the best DSH zones that money can buy, in the 150-160 range in heaters where velocity is the limiting factor, it wouldn't take a smart designer long to figure out that adding more surface to account for the sensible heat transfer necessary to reduce the steam to saturation temperature isn't the thing to do.

The study where I got the numbers I quoted above was done to justify the addition of desuperheaters to the turbine exhaust steam headers supplying an evaporation station at a sugar mill. The calculations showed that easily 1/3 or more of the calandria was devoted strictly to desuperheating the incoming exhaust steam. It was easy to see this in the sight glasses on the body above the top tube sheet. The area of the top tubesheet nearest the steam inlet was virtually dead, very little of the percolation effect common to this type of rising film evaporator. A little, but not much.

The backside, however, once the steam was cooled to saturation, was very lively, with sugar juice jetting up higher than you could see through the sight glass. This was a newer evaporator with several good, clear sight glasses, and easy to make the observation.

After the desuperheaters were added, and the exhaust was brought down to Tsat, two things happened. The overall evaporation rate for the first stage evaporators (Pre's they are called in our part of the world) increased dramatically, and the visual observation in the sight glass showed that the lively boiling was uniform across the entire surface of the top tubesheet. Gone was the dead zone that was there during the previous grind.

Latexman,

http://www.processassociates.com

has (or had) in their 'tools' section the capability to model heat exchanger performance, and/or calculate HTC's. I challenge you to go set up a couple of different scenarios. One being pure condensation where your inlet conditions are right at Tsat, and another where your inlet conditions are above Tsat keeping steam side pressure and Tcoolant constant.

I think if you do it right, you will see something different from what you are stating in this thread.

I have been there, done that, and I think I know the result you will find. BronYrAur, if you are still following this thread, I recommend the same for you. Mr Perry, I recommend the same for you too.

I have enjoyed this discussion. It has cleared out a lot of cobwebs.

rmw

 

[25362]

For those interested in a different (heat flux) approach there is an article in the Dec. 29, 1980 ChE issue, titled Superheated vapor condensation in heat exchanger design by Foxall and Chappell.

 

[katmar]

rmw,

Thank you for a very well reasoned post - you have obviously put a lot effort and time into this and I appreciate it. Your experience with the Roberts unit proves that the heat flux was lower in the dry zone. This was what I expected when I said earlier that the heat flux in a dry zone would only exceed that in the condensing zone under extreme conditions.

I dug out a copy of the Foxall and Chappell article referenced by 25362. Their examples do show higher fluxes in the dry zones, but this is only because of the temperatures they used. The heat transfer coefficients they calculate are much lower for the dry zone than the condensing zone, but because they selected coolant temperatures so close to the condensing temperature and because they had large amounts of superheat the temperature effects outweighed the HTC effect. In their steam condenser example they condensed superheated steam at atmospheric pressure with coolant at 200 deg F. With a condensing temperature of 212 F there is only 12 F of driving force in the wet zone, but in the dry zone the vapor temperature was 500 F giving an overall temp difference of 300 F.

Probably the most important lesson to take out of all of this is to remember how dangerous it is to make assumptions in engineering. A properly formulated and calculated solution is the only way to be sure.

regards
Harvey

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[Latexman]

rmw,

The Babcock url wasn't working when I asked what type those units were. It is working now, and I see they are horizontal, vapor-in-shell units. The Robert’s rising film evaporator sounds like a vertical, downflow, vapor-in-shell unit to me. In these cases, the vapor is usually steam.

I believe I have been clear in saying my experience and comments are on vertical, downflow, vapor-in-tube condensers. In most of my experience, the vapor is usually an organic, but has been steam from time to time.

Condensing outside the tubes is very different than condensing inside the tubes. In fact, condensing outside horizontal tubes is different from condensing outside vertical tubes.

Let’s list a few differences between the Robert’s vertical, downflow, steam-in-shell unit to a vertical, downflow, steam-in-tube condenser I am familiar with:
[ul]
[li]Inside the tubes, the condensate flows in layer form down the length of the tube. The condensate film is “thinned out” by the flowing vapor which increases heat transfer.[/li]

[li]Outside the tubes, the condensate is impeded by tube supports. Most of the condensate falls off the edge of each baffle providing poor contact with the cooling surface where the larger h exists.[/li]

[li]Outside the tubes, especially if crossflow baffles are used and in the high velocity section where the vapors enter, condensate is stripped from the tubes by the vapors flowing through the unit. Again, this provides poor contact between the condensate and the cooling surface where the larger h exists.[/li]

[li]The mechanical design issues are different too. With outside the tube condensing, the shell could be several hundreds of degrees higher than the tubes where condensate exists and flow induced vibration is a mighty issue. Not so for inside the tube condensing.[/li]
[/ul]
It appears to me we are not discussing an “apples-to-apples” comparison. I know I’m right with my discussion on vertical, downflow, vapor-in-tube condensers, so I would recommend BronYrAur understand these differences when reading them and apply the correct one to his application

By the way, when I said "sitting there" my intent was not in describing a blanketing effect that gives the "air-bound" symptom that TBP mentioned in his 26May post. I was referring to the “boundary layer” at the tube wall that exists during turbulent flow. In this “laminar sublayer”, the velocity at the wall is zero. This is the classical textbook model that is taught in every fluid flow course. So, since the velocity at the wall is zero and if the tube wall is < Tsat, the steam will condense, even if the steam in the “turbulent core” is superheated. This includes at the steam inlet. In a vertical, downflow, steam-in-tube condenser once the film is formed, it stays formed and is not impeded by tube supports or baffles and is not stripped off the tube wall by vapor flow. We may be describing the same thing, just using different words.

In the vertical, downflow, vapor-in-tube condensers I’ve designed IF condensing started at the beginning of the tubes and the tubeside fluid was superheated steam, the h would be in the 400-600 range you mentioned for condensing, or maybe even higher, not the 52 you said you calculated.

I also studied the “Process Tools” at

http://www.processassociates.com

and they do not currently have a tool that handles a condensing heat transfer coefficient. They say it is “coming soon”. When this tool arrives if it is as simplistic as their existing tools, I doubt it will have a flash calculation to determine when and where condensation occurs anyway, but I’m speculating. Anyway, since their current tools do not handle condensing heat transfer, I must decline your challenge.

I do have a question I’d love the answer to though. If you have “been there”,

http://www.processassociates.com/process/tools.htm#heat

, and “done that”, what tool did you use exactly?


Good luck,
Latexman

 

[rmw]

Latexman,

I'll give a brief answer because I am packing for a trip to South America that I have to leave early tomorrow AM on.

I kept the print outs of the results of the processassociates runs in a file, but due to a recent office move, they are still packed, if not in storage, so I can't put my hands right on them. I do remember the condensing result, because it verified the HTC for condensing that I had seen used in vendor literature and on various websites but had not done for myself.

You have me curious, so I will check it when I can, but it won't be right away.

Both Yuba and TEI (division of BPI) have vertical heaters, but you are correct in surmising that they are steam outside the tube, and have the presence of the tube supports. They don't use flow through except in the DSH zones.

I read your thorough and complete post through once, but want to re-read it before making any comment. It was chock full of deep material that is going to take me a couple trips through at some later time when I can get my mind into it.

And, Harvey, yea, the "per degree F" component of the HTC formula can generate higher heat flux if the delta T is high enough, but in most cases that I deal with, the degree of SH is low enough that it is a penalty, not a benefit. The goal is to get the HT by getting the latent heat, so the SH present has to be dealt with separately.

Good night all, more later.

rmw