Upper velocity limit for two-pase steam

[KernOily]

Hi guys,

Need some opinions/counsel/suggestions/insults on the upper velocity limit for saturated (two-phase) steam.

System in question is 50 psig, 65% quality, dirty steam. Rates are 8000-45000 lbm/hr. I am sizing the header and laterals feeding the shellsides of three heat exchangers.

My pipe sizing criterion is to select NPS's that keep the flow regime in the annular, this done to minimize dynamic effects (keep the liquid phase in suspension to prevent the liquid from slugging in the horizontal-to-vertical transitions, i.e. upstream of risers). This also helps to smooth out the behavior of the exchangers and their temperature controls.

Unfortunately I can't trap out the liquid before the steam hits the risers and send it off someplace because I have nowhere to send the condensate (there is no condensate handling system) without a WHOLE lot more work, equipment, piping, and controls.

I am not pressure-drop-constrained; I got pressure to burn (what a luxury!).

The annular regime requirement means the required velocities are 180-220-ish ft/sec.

Now, API RP14E makes some suggestions as to upper velocity limit in two-phase fluids. Using this method with the conservative value of c, i.e. c=100, the upper velocity limit for my system is 240-ish ft/sec. So using this method, I am safe. But...first question is, how noisy is this going to be at 220 ft/sec?

What say ye about this? The steam is "dirty" in that there is the occasional solid particle going by. But it is clean enough to where I think I can justify use of the c=100 factor.

Thanks ! ! ! ! ! !

 

[SeanB]

The other constraint you want to look at is erosional velocity. You should try to stay within 40-70% of erosional velocity.

 

[TBP]

As SeanB has stated, erosion will likely be a real issue. Water droplets moving at steam velocity will do things like chew out the back sides of 90* els pretty quickly. I've seen installations with thermodynamic steam traps (blast discharge) with a pipe nipple then a 90* 150 malleable elbow fail within a year. Looking into one end of the el, held up to the light was like looking through a lace curtain. There were pinholes everywhere on the back side. I don't think your temp control valves on the HXs - or the HXs themselves - will like this situation either.

Here's what I see as a MUCH bigger problem - water hammer. Even in steam systems with excellent quality saturated steam, the piping MUST be sloped to drip legs, and the collected condensate trapped off. If there's no condensate system, then it'll just have to be dumped to the ground, or sewer. Wayne Kirsner (

http://www.kirsner.org)

has some excellent information on water hammer posted on his website. He has investigated a number of water hammer incidents, many of which involved fatalities.

 

[KernOily]

Guys thanks for the replies.

SeanB - Do you have a reference or a cite for the 40 to 70% of erosional velocity?

API 14E is an erosional velocity guideline for two-phase flow. That's why I referenced it in my first post. My design velocities listed above are at about 85% of the 14E recommendations.

TBP - Good to hear from you. Pete C. from Bakersfield here. Are you still up in Washington?

Do you have any idea as to the rates or velocities involved in those situations you described?

Usually we design for velocities around 70 ft/s for these steam systems and that has worked well; no erosion. But those velocities are at much higher pressures (800 psig) so they are at higher densities, so annular or mist flow regimes are easy to get at lower velocities with higher pressures.

Your comments re: waterhammer are well taken. We have several other installations with this exact setup, only at higher pressures, and water hammmer has not been a problem because the velocities are high enough to keep the flow regime moving the liquid out of there. The problem with this new system is it operates at 50 psig, and the density associated with the 50 psig is causing high velocities for the annular flow regime.

If I keep the liquid moving along, so it doesn't collect in the pipe, then I don't have dynamic effects or waterhammer problems. This has been our experience thus far. Also, mixing of liquid and vapor phases at the same temperature is not going to cause waterhammer. This of course is during steady-state operation. Startups are a whole 'nother matter, when of course I have to remove all liquid from the lines prior to introducing steam.

 

[SeanB]

The 40-70% of erosional velocity comes from our in-house design practices. I haven't seen it referenced anywhere other than that. But I have never looked for it elsewhere either.

 

[CJKruger]

KernOily, except for special cases, I have given up on targeting a specific velocity. Design the pipe for roughly 1 psi/100 ft and then check to ensure you are below erosion velocity and not in slug flow. If you are in slug flow, decrease line size (if dP allows) or increase support.

Are you going to or coming from these exchanger ?

CKruger

 

[GregLamberson]

KernOily

Might want to take a look at this thread on velocities:

http://www.eng-tips.com/viewthread.cfm?qid=191447&page=1

Need to use a little caution when applying C-factors from RPI 14E, there are other as important, if not more important factors to consider when looking at erosion.

Greg Lamberson, BS, MBA
Consultant - Upstream Energy
Website:

http://www.oil-gas-consulting.com

 

[KernOily]

Got to resurrect this. The design for this system is entering the final phase and final pipe sizes must be selected.

I have reviewed the above info (thanks a ton for that, guys) and I still don't have any conclusions.

1) The ARAMCO steam data does not say if the steam is saturated or superheated.

2) A max value for wall shear stress is given above as 1000 N/m2. What is the source of this?

3) For the annular flow regime, there are no liquid particles or droplets.

4) Wall shear stress is very low - less than 0.5 psi, this at 200 ft/sec, 160 psig, and steam at 70% quality.

I have no solids, or very few anyway, and no corrosion issues at all. Based on this, I think I can justify the use of high C-values of 200 or more from the RP14E guideline.

I think I'm in uncharted waters here. Opinions/counsel/harsh comments/out-and-out insults welcome.

Thanks guys. Pete

 

[TBP]

I lost track of this thread myself - you're thinking of somebody else. I'm in Alberta. But thanks anyway

With 65% steam quality, there won't be any superheat. You don't want superheated steam showing up at your HX anyway - it'll cause lots of problems. And you WILL have water droplets sailing through this system at 200 feet/sec. A lot of your components are NOT going to like this.
There's no condensate system, so the condensate will just dump to sewer out of the HXs, right? Consider installing some steam separators ahead of the HXs, and dump the condensate that gets trapped off. Anything that doesn't get trapped ahead of the HXs will just pass through them in any event. The total amount of condensate to be dealt with will remain unchanged, and your system components will live longer, happier lives.

 

[rzrbk]

I agree with TBP. You will be much better off if you trap and remove the condensate from the steam, and send it wherever the HX's are draining to.

It sounds like you are trying to design for a narrow flow regime over a wide range of flows.

 

[davefitz]

for erosion issues, use P11 piping for the section with higher velocities.

For minimizing phase maldistibution , 2 techniques are recommended:

a) use a "spider bottle" distributor, as used on large boilers that are "once thru units". The main common feder would be a vertical pipe, usually feeding down. The end of tehis vertical pipe terminates in a enalrge 3 to 1 distributor- each of 3 feeders are oriented horizontal, and inside the 3:1 distributor are 3 baffle plates, which eually segment the inside of teh distributor into 3 x 120 degree sections.

b)The heat exchanger would be vertically offset from the 3:1 distribution device so as to promote a "positve" thermal-hydraulic characteristic; ie , the HX that transfers the most heat will automatically draw the most flow so as to equalize the outlet enthalpy of the 3 HX's.If the HX adds heat to the 2-phase mixture, then the 3: distributor would be below the HX and the 3: collecting header would be far above the HX, so as to generate the natural circulation system characteristic. If the HXis condensing then reverse this geometry.