Steam Hammer 1

[KANN]

How is steam hammer prevented in submarine piping, or maybe, is it prevented?

I assume the slope of piping on submarines changes as the vessel navigates. Where steam piping is installed, how is the piping designed to prevent steam hammer when the steam and condensate flow can vary from co-current to counter-current two-phase flow?

Is there a specific velocity that is not exceeded for sucessful designs? Is there a maximum condensate depth, below which slug flow (steam hammer) will not be initiated?

 

[davefitz]

Technically, "steam hammer" is caused by the change in momentum of a large mass of high pressure steam that is suddenly stopped by a fast closing stop valve. Steam hammer is a problem on some very large central station power plants during a steam turbine trip event or closure of a bypass valve, but in those cases the steam pressure is over 2400 psig and the pipe length is over 200 ft long. The operating pressure and pipe length in subs is less than these problematic values.

For the case of 2 phase flow, I would assume the subs would use a centrifugal steam-water seperator that ensures the mixture is accelerated over 5 G's so that the sub inclination is nor too much of a factor.

 

[KANN]

I am also thinking of steam hammer as occuring when the steam velocity is great enough over a pool of condensate that a slug of condensate is formed and carried into an elbow. In this cases there is no valve event. It is a matter of condensate removal and limiting the steam velocity where the condensate removal can not be complete due to the piping configuration and the change in orientation of the piping due to the vessel's movement, or both.

It is my thought that perhaps only superheated steam is used in a sub and saturated steam conditions are minimized. Yes, I agree a steam-water separator would reduce the problems, but that would probably mean quite a number would be needed in a vessel where space is at a premium.

 

[itsmoked]

Just guessing, but I would presume the runs are probably VERY short, running high velocities, and can be well insulated so condensate is probably not a big problem.

 

[electricpete]

Almost any steam system saturated or superheated will have means to minimize accululation of condensate within steam lines.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[mauner]

KANN,

Your question from 2-1/2 years ago was finally answered!

I am curious - why did you want to know about submarines?

I have many years of experience with steam plants and I agree with electricpete. All steam piping system must have drain traps to remove the condensate especially during start up.

A dead leg of pipe in a super heated steam system can fill with condensate causing water hammer.

 

[rmw]

Most submarines in my country's navy are nuclear, and nuclear reactors do not as a general rule generate superheated steam, so I suppose that is out, and good piping methodology is in.

However in light of steam systems in general and steam generators in particular, knowing some of the dive and climb angles that a submarine reactor might be suddenly subjected to makes your question an interesting one.

I certainly would hate to see some of the land based power generation reactors I have worked near be subjected suddenly to such inclinations. How about you Pete?

rmw

 

[KANN]

mauner -

Well, I don't agree that my question was answered, but there have been a number of recent responses.

My primary interest is to better understand how steam piping is designed to prevent condenstate slugs from developing. I have been involved with steam system design for process plants and central plants for almost 25 years, not exclusively, but as part of my facilities design work. My work has been with 400 psig or lower pressure steam systems. I know the importance of, and the basic design approaches for removal of condenstate in saturated steam systems. What interests me is the design of the unusual piping arrangements which could result in the development of condensate slugs - vertical risers and countercurrent steam/condenstate flow such as can occur in long up sloping distribution mains. Though this is addressed to certain extent by ASHRAE, technical handbooks, and steam equipment providers (steam trap manufacturers for example) the physics of what is going on in these flow situations in not clearly explained, in my humble opinion. Much of the basis for the guidelines in common use today appears to stem from the early ASHRAE research which was limited to low pressure steam and small pipe diameters. To begin to understand the flow dynamics involves the study of multphase flow. So, over the past several years I've tried to get many of the journal articals (scholarly and trade) relating to steam multphase flow and learn what I could from them.

As a result of the reading I am of the opinion, generally speaking for typical distribution pressures up to 150 psig, that if steam flow approaches 2000 to 4000 fpm in vertical pipes and a little higher in horizontal pipes with a sufficient condenstate film layer, then slugs of condensate can become entrained in the flow.

I keep my eyes peeled for better design guidelines for steam piping design. It occurred to me that submarine designers would have to face the issues I mentioned and deal with them effectively, as well as many other maritime vessels. So, that't the background as to why I asked the question I did with regard to submarines.

 

[electricpete]

One comment - in terms of the US Navy, I believe submarine propulsion plant details are classified.

You certainly have a right to ask. But the response may be somewhat limited for that reason.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[kirsner]

Kahn--I don't think the answers submitted to your question have been very responsive. First of all, the correct term for what you're asking about is "water hammer" in a steam system. "Steam hammer", as Davefitz points out in the first responce, is something else. Second--you've got the wrong mental picture for what causes water hammer--it is not due to a condensate slug being blown down a steam pipe by the velocity of steam and then being abruptly stopped. Almost all destructive water hammer is caused by "condensation induced waterhammer". Slope still matters in condensation induced waterhammer, but in a different way than you might imagine. I recommend you consult the articles page at

http://www.kirsner.org

if you want to learn more about water hammer in steam systems.

wayne kirsner

 

[KANN]

Mr. Kirsner:

Yes, I agree, “water hammer” in steam piping describes the transient condition I have in mind. Since using the term “steam hammer” in my original question 2-1/2 years ago I did learn that my use of the term was incorrect, thanks in no small part to your fine website and helpful articles which I read, shortly after making that original post. It is regrettable that I carried the use of the term steam hammer in my 28 March post above. Thank you for the correction.

However, I’m not sure how to understand your second point. While I do have a mental picture of steam velocity (alone) producing water hammer, “steam flow driven water hammer”, I do also understand that there are at least two other causes: the introduction of “cold” water into a heated vessel resulting water hammer by the expansion of steam, and the introduction of steam into a vessel with “cold” water resulting in condensation-induced water hammer. Although expanded steam generated, or condensation-induced water hammer is more damaging, and the design and operating practice to prevent those types of water hammer is a necessity, I am specifically interested in steam flow driven water hammer.

What I understand, in part, from your response, is that water hammer is not caused by steam flow velocity (alone). Or, if not that strong of a statement, that at least it would be rare for steam flow driven water hammer to be destructive.

It would be instructive to me, and probably others in this forum, if you would elaborate more, or clarify, what you think is the significance of steam flow driven water hammer.

Certainly much of the “first level” technical literature portrays the elimination of steam flow driven water hammer to be an objective of design. As “first level” I am referring to piping handbooks, ASHRAE handbooks, and steam trap manuals. The familiar picture is that of a pool of condensate drifting lazily down a sloped pipe suddenly being lifted into a menacing slug of water mass capable of destroying any object in its path. In seems to me that this often presented scene of condensate pooling on the bottom of the pipe only occurs under very low velocity conditions. In “working” steam lines with higher velocities there are several multiphase flow regimes that could be present. My first “enlightenment” of this probably came via Hal Finkelstein’s booklet “Steam Distribution and Flow”. As you well know, there is much research over the past 5 or 6 decades examining multiphase flow and much specifically funded by the NRC to steam and condensate interaction.

What is of the most interest to me at this time, and the reason for my original post, is to gain a better understanding of the design considerations for steam/condensate counter-current flow. Both nearly horizontal and vertical flow orientations are of interest to me. Spirax Sarco, TLV, Yarway and other steam trap manufacturers, for nearly horizontal countercurrent steam/condensate flow, recommend drip legs every 50 ft, 1” in 10 or 20 ft slope, and “lower” steam velocity. How much is “lower” is not discussed, but I have found some suggestions that 2000 to 4000 fpm may be a “lower” value. But on what is that suggested velocity based? The basis of the slope and spacing recommendations are not discussed. Could the spacing be much greater? Why or why not? My understanding is that steam flooding lines in oil production is the delivery of wet steam with few (any?) steam traps.

So, your response, if I understand it correctly, is very interesting. I hope you can write a bit more from your experience with steam flow driven water hammer, and entice some more responses from others.

 

[kirsner]

Kahn. Steam driven slug flow--which is commonly thought of as waterhammer in a steam system--is technically not "water hammer". Water hammer, whether it occurs in a plumbing system or a steam sys-tem creates a pressure pulse proportional to Joukowski's Eq : P= rho c v where rho is the density of the fluid (~60#.c.f), c is the speed of sound in rhe fluid (~4000 ft/s), and v is the velocity at which the fluid is moving before it is halted. To avoid confusion, I sometimes use the expression " rho- c-v event" when I'm specifically talking about water hammer to an academic .
A slug of water picked up by high steam velocity, blown downstream, and striking an elbow creates an overpressure also, but it's proportional to P = rho v^2. The physical difference (from water hammer) is that the water is not stopped abruptly so that it is not forced to compress on itself, instead it sloshes around a corner taking a much longer time to change its momentum: t = Length of slug/v. The stopping time in a water hammer is proportional to L/c—100 to a 1000 times quicker.
Now, given that most of the sources giving guidance on trapping steam lines that are pitched uphill don't understand what causes water hammer, I wouldn't give them much credence. Even the research quoted in ASHRAE is not useful except in small single pipe systems since it assumes that the steam mass flow equals the condensate mass flow.
So, what do you do if you’ve got to run a steam line up hill against condensate flow?
First--I don't think it's that big a deal (I haven't investigated any accidents where i found that uphill pipe slope was the problem) but i don't have a good rule of thumb either for trap interval versus slope. It's an open channel flow problem where one doesn't, I suppose, want the uphill steam-water surface shear force to overcome the gravity force causing the water to flow downhill so that water builds up and plugs the pipe. Even if it did, however, it still would probably not result in a water hammer
Now i think this problem could be easily solved and guidance given to design engineers if it was funded. There's a lot of research on countercurrent flow in the nuclear field where they're worried about pump-ing cooling water into a steam filled line, but if that research can be translated to apply to this common HVAC design issue, to my knowledge, it hasn't been done. Apparently, it's not an interesting enough problem for researches to get funding.

As for the steam flooding idea in oil production--that is saturated steam conduits without traps--i've never heard of this--is it real?

wayne kirsner

 

[davefitz]

steam flow driven waterhammer is a large problem on modern combined cycle power plnats, specifically in the hot reheat dump/bypass line to the condenser. This is typically a 14" dia pipe , slighlty sloped to the downstream direction. The spray attemporator typically oversprays at low steam flows , due to incorrect control logic. The steam veloctiy is on the order of 200-250 fps in this line.

At very low steam flows, and with excessive spray water flow, the excess water accumulates on the bottom of the 150 ft long pipe. The water level can build up to several inches high due to the slight slope. If the steam control valve then opens wide in response to a high reheater pressure event, then the water is accelerated to 250 fps and forms large slugs . When these slugs hit the elows at 250 fps they cause severe piping damage.

The oil and gas industry has a lot of experience with 2 phase flow in pipelines, as so they have "Baker plots" and other correlations that will predict what 2 phase flow regime will occur in the pipe. For example, in offshore Gulf of Mexico drill rigs, the output of oil and gas is comingled and transported together in the same underwater pipeline to the onshore seperation plant. Europeans have other correlations.

 

[kirsner]

Davefitz--Just to avoid confusion, let's call this a rho v^2 event since it's not, technically, a waterhammer. Can you elaborate on the damage done. i calculate an order-of-magnitude maximum collison pressure of 811 psi for a water slug hitting an elbow at 250 fpm--that assumes the slug gets going at the steam velocity which I doubt is a good assumption other than to bound the problem. Are there any pictures or descriptions of damage?
wayne kirsner (

http://www.kirsner.org)

wayne kirsner

 

[davefitz]

At one plant, the 24" pipe was permanently deflected over 8", damaging the bumpers, supports, and nozzle connection to the condenser. The entire pipe run needed to be replaced, about 150 ft long. After the pipe was repaired, they failed the pipe a second time after the spray control valve was placed in manual and the operator forgot about it.

The cause was a combination of inadequate pipe slope ( which leads to a high water level during overspray events - per open channel flow equations)and the incorrect use of outlet temperature control of spray water flow. The preferred control method is enthalpy control ( based on a realtime heat and mass balance) plus a mandatory limit that the ratio of steam to water flow may never be less than 3.5 : 1.

A survey of other similar plants revealed that other plants havve had exactly the same type of failure.

 

[ccw]

I liked the way KANN originally framed the question. We all know from "Hunt for Red October" and "Blind Man's Bluff" that one of the primary design critera for subs was "run silent". With all of the banging, whistling, fizzing, hammering, knocking, etc. attributed to the power plant steam/water loops -- How did they do that?.

Another noise source comes to mind that was related to feed water injection into hot water vessels or boilers. I believe this caused a considerable stir amongst Nuclear Plant designers several years ago when it was speculated that these bangs could fatique the feed water supply lines to failure. What were the physics and cure for this phenomenon?

 

[KANN]

Mr. Kirsner:

Your response is very interesting and helpful. In thinking back about the technical information I’ve read regarding water hammer, it seems to me, now, that much of what is often described as water hammer is slug flow, as you differentiate the two events. I hadn’t sorted out the distinctions that you elucidate above. The description of a slug that “sloshes around a corner”, versus a mass of water having to “compress on itself”, helps to make the huge difference in energy dissipation, easy to visualize.

What I understand you to say is that in your experience and understanding, steam driven slug flow, probably does not present that much of a challenge to the piping in terms of pressure, and with regard to a single event (effects of fatigue excluded). In Davefitz’s example there is the collision pressure within the pipe as you estimated. But the damage described is due to pipe support locations and/or strength that could not resist the the sudden mass transfer. Perhaps with different support locations the piping and nozzle damage described could have been avoided. And perhaps this is one reason for the recommended trap spacing, to attempt to limit the amount of mass that may be driven by steam and thereby allow a more conventional piping support design to serve successfully.

If the effect of steam driven slug flow is not to produce a damaging pressure surge, perhaps the design of marine piping is not as challenging, in terms of dealing with the effects of counter-current steam/condensate flow, as I first thought. It still eludes me as to what the basis of the slope and drip spacing is in much of the design literature. Much of the early work, I understand, appears to be from just a few sources: early British steam works (Spirax Sarco seems to have some rich early knowledge resources), district energy industry, and military facilities. And from what I could find from those sources, the derivation of the trap spacing and slope is not described.

I have not designed piping for steam flooding, so I do not know how or whether it is trapped. The best information that I have found on the steam flow in piping for “steam flooding” applications is by Sze-Foo Chien, some of his papers are: “Predicting Wet-Steam Flow Regime in Horizontal Pipes”, JPT, March, 1990; Steam Flow Chart, USMS 023417, Texaco, Inc., 1991, SPE. His work has included predicting the flow regime for a wet steam flowing in a steam distribution network for enhanced oil recovery projects.

Davefitz:

Your experience with and description of the hot reheat dump/bypass line is interesting. If I understand you correctly the line is sloped down in the downstream direction. Is the design slope of the line standard, and if so, would you know the amount? Were the lines supported and designed for a slug flow event, but underdesigned for the actual event due to unanticipated control issues? I don’t follow exactly what you are describing, but it sounds as though the water would be at steam temperature and that this is not a “rho-c-v” event.

 

[davefitz]

Kann:
The 24" reheat dump/bypass was sloped at a typical slope of 0.25 " per foot on the horizontal runs.

The piping operates at a wide range of pressures ,depending on the flow of steam to the condenser . The discharge sparger at the pipe outlet is dischargin to vacuum , so teh pressure vs steam flow is a linear relationship, based on choked flow thru the outlet sparger hole flow area.

At low flows, the pressure in the pipe could be less than atmospheric, but in any case, the saturation temperature will vary with pipe pressure and steam flow. So, if one is sparying water into the steam and controlling the water flow based on holding an outlet steam temperature of 20 F superheat, the saturation temperature must be programmed as a function of current pipe pressure.

At low flows, the spray nozzle does not atomize the water properly, causing most of the water to layout as a stream of liquid running down the pipe invert, and this unatomized and unevaporated water does not generally hit the downstream thermocouple and is not automatically detected. Therefore, at low steam flows there is a tendency to have a level of water running doen the pipe invert.

If the upsteam steam pressure congtrol valve senses a high reheater pressure, the reheat bypass /dump PCV valve will fast open, and cause a large increase in steam to flow down the pipe. This large steam massflow will accelerate the water that had layed out on the invert, and cause it to crash into, and straighten out, downstream elbows.

One override that helps to avoid this is to measure the water flow using a standard flow element in the water supply line, and estimate the total steam flow to the condenser based on pipe pressure ( remember, choked mass flow will be directly proportional to pipe pressure and inversely proportional to the square root of absolute temp). The ratio of water to steam flow should never be permitted to exceed 0.33 : 1.

Another control method that helps avoid the problems with downstream temp control is to use enthalpy feedforward control. The water mass flow and temp is monitored, as is the upstream steam P, T, and flowrate. An online heat and mass balance will predict the correct spray water flowrate needed to result in a 1180 btu/lb downstream enthalpy. The spary water valve ism odulated to flow that calculated amount, with other overrides or trims based on downstream steam temp.

 

[kirsner]

Kann--You got my point precisely. I was just rereading papers on clearing slugs from pipe (i.e. "rho v^2" events) to see what type of accident the researchers were interested in since these incidents are not generally, to my way of thinking, very severe. It appears to be a sudden steam releases from safety valves that have water resting above the valve in the pipe that then procedes to strike an elbow on its way to being ejected. This, to me, is like the accident davefitz describes--a case where there's a sudden very rapid steam flow that pushes a water slug out a pipe. Importantly, however, according to the paper I cite below, the slug can only travel several times it's length in the pipe (less than five) before it breaks up due to entrained steam or air.

davefitz--
.25" per foot is 10 x typical steam pipe slope so the reheat/dump has a pretty good slope

Research I've read (Neumann and Griffith--PVP-Vol 231) states a slug cannot be picked up by fast moving steam at a water depth of less than .25 pipe Diameter. Could the water possibably be running that high in the 24" dump/bypass? Or do you think the .25 factor is incorrect?

wayne kirsner

http://www.kirsner.org

 

[davefitz]

kirsner:
The 0.25 " /ft may be incorrect. The slope may be less than that.

In our case, the type of bypass valve and spray attemporator originally provided was extraordinarily crude, and it is surprising that such a valve and spray system is still marketed. In any case, the atomization of the water was totally reliant on shear forces of the water and steam as it passes thru a perforated plate at the valve outlet nozzle. At low steam flows, there is no significant shear, so there was no significant atomization of the water. The water would then run like a river down the pipe invert, and since it did not impact the downstream thermocouple, the temperature control logic would incorrectly demand more spray water since only hot , non-desuperheated steam would impact the thermocouple.

There is no doubt that the 2 catastrophic failures of the reheat bypass piping was due to water hammer- other tests with conductivity probes and action of a addes steam gtraps confimred the obvious. The later upgrades following the second replacement of the piping system was to replace the bypass pressure control valve and spray attemporator with a CCI Self-drag type system, control logic to inlcude enthalpy feedforward logic plus override limit on water/steam ratio , and thermocouple relocation to the condenser nozzle ( greater that 0.3 sec transit time from spray nozzle)