Compressible Flow of Real Gases

[JoeChem]

Hi All,

Interested in calculating flow and pressure drop of compressible steam in complex distribution piplines - lots of pipe, elbows, globe/ball valves, etc. Basically, the system involves a steam boiler to single large main header (10 inch) to many small (2 inch) branch lines. I want to stay away from ideal gas assumptions if at all possible. Does anyone know of a good reference for dealing with real life industrial compressible steam/gas flows? I am not real interested in purchasing software as I would like to avoid the "black box approach" to the problem.

Thanks much in advance

JoeChem

 

[rbcoulter]

What range of pressures and temperatures will the steam be in? This will determine the suitability of various
PVT models.

 

[JoeChem]

The steam is delivered to the plant from an electrical generation plant about 1/4 mile away. Steam enters through an 8 inch line at 430 psig. The line increases to 10 inches at a point where water is injected to lower the pressure down to 120-125 psig and about 360 F. The main runs for about 30 ft and then a total of (28) 2 inch branch lines of varying lengths and fittings are taken off the top of the 10 inch main. Flow to each 2 inch branch line is regulated by a 2 inch globe valve located about 2 ft above the main header. The globe valve manually controls the branch line feed pressure at 30-40 psig. At the end of the branch lines are perforated pipe distributors. Just prior to the pipe distributors, the steam pressure usually is around 5-10 psig (depending on the particular geometry of the line). Not sure of the steam temperatures in the 2 inch branch lines - surely is cooler due to expansion. This is the system I am interested in modelling. If more information is required let me know.

Thanks for the interest.

JoeChem

 

[rbcoulter]

Basically, the only problem I see here is that at the higher pressures the steam deviates from the ideal gas law by about 20% (per my steam tables). I would calculate the pressure drops using a constant specific volume of steam over sections of pipe where the pressure term (P1-P2)/P1 is not greater than about 10%. Adjust the specific volume term for each section.

The globe valves are tricky. They obviously take a big pressure drop. I would contact the manufacturer's for any data they may have for the flow of steam through their valves.

At the lower pressures I believe the ideal gas law could be used..

 

[ccfowler]

Will the flow rates in the supply line and all of the various branches always be great enough to assure that the steam remains above saturation at all times? If not, condensation can make your modeling much more "interesting."

I've found that modeling such systems using the equivalent length method for the various fittings is remarkably accurate and convenient. I prefer setting such a problem up using a spreadsheet (usually Excel)--I don't like the magic black box method either.

How steady is the source temperature and pressure? Do these vary much with power plant load? (I'm assuming the steam comes an extraction point on the turbine(s).)

How uniform is the steam condition leaving the desuperheater at the 10 inch pipe? Will some of the branches receive saturated steam while others get superheated steam?

This seems like an interesting problem to set up and study. Enjoy!

 

[JoeChem]

Thanks for the feedback!!! I find it very helpful to discuss this problem with other engineers.

Dividing the systems into sections where the pressure drop is low enough to assume constant density is an interesting idea. Should work well and is easily entered into an excel spreadsheet. I will see if I can find some information about the globe valves.

The steam supply pressure and temperature vary a bit but not too bad. I have not spent a lot of time in this plant so my view point is somewhat limited on this. I do not know how uniform the steam condition is following the desuperheater. You make a good point that some of the smaller lines may be receiving steam of varying quality. The pressure loss across the 2 inch globe valves leads me to believe condensation will not be a problem in the smaller lines.

Some more questions...

All of the adiabatic compressible flow models that I have found are based on a stagnation condition starting point. Is it reasonable to assume that the steam condition in the 10 inch main header is a stagnation condition? If so, then the flow will start with an insentropic expansion across the 2 inch globe valve, correct?

Can the flow choke if a perforated (orifice) pipe distributor is at the end of the system? I am not sure but I do not see how a choked condition can occur at an exit orifice.

Thanks for the discussion,

JoeChem

 

[rbcoulter]

I guess the models assume that the velocity of approach is small compare to the pressure head. Maybe it is just a matter of calculating the velocity head at the inlet of the globe valve and compare it to the pressure head.

I don't think you have isoentropic flow because you are not recovering the work of expansion going through the
globe valve. This appears to be adiabatic flow (the pipes and valves are insulated correct?).

I think you can have choked flow at an orifice discharge but I don't think your pressure is high enough to cause it. There are examples in the Crane handbook on determining if you have choked flow at outlets.

The concern here is that the globe valves take a big pressure drop and I don't know if even the standard adiabatic flow equations in textbooks would work because they are derived assuming the ideal gas law.
Maybe should calculate the z value before and after the globe valve to ensure ideal gas law can be used here.

 

[pmover]

joechem,

you are in luck today!

i've copies of basic programs to conduct pressure drop in steam systems. actually, the program is from the Power Engineering magazine, in February and March of 1986. perhaps you have access to these magazines? the programs are based on information in the well-known crane technical report 410.

i would post the programs and literature, but there is too much text.

you will need gwbasic or some type of basic software - should be able to somewhere download the software free off the internet.

i'm receptive in sending the documentation, but need email address.
-pmover

 

[JoeChem]

The pressure head of the steam at 125 psig and 350 F is typically at least 2 orders of magnitude greater than the velocity head at the entrance to the 2 inch globe valves (flow rates typically around 2500 - 3500 lb/hr). Looking at it this way makes the assumption of stagnated conditions upstream of the valves seem perfectly reasonable.

Yes, the lines are insulated so adiababtic flow should be considered. I would consider the expansion across the 2 inch valves to be isenthalpic and the temperature of the steam can be determined directly from the steam tables once the stagnation enthalpy is determined...seem reasonable? I will check the z values across the valves to check the validity of ideal gas behavior.

I noticed the the expansion factors (Y) for compressible flow in Crane's manual for a square edge orifice do not have an upper limit like those for nozzles. I checked into this a bit more and found some information that there is no critical pressure for gasses discharging through an orifice. The mass flow apparently continues to increase as the discharge pressure is decreased until 0 absolute pressure. I believe there is also some information in Perry's Chemical Engineers Handbook on this as well.

Looks like it is my lucky day pmover. A copy of the steam flow software will be great. Unfortunately, I do not have direct access to Power Engineering magazine here. My e-mail is joe_vera_2000@yahoo.com.

Thanks for the help and discussion,

JoeChem

 

[KernOily]

As somoeone who has done a ton of this type of analysis, on these very systems, I think I can make some comments.

The method you choose to do the analysis depends on how sharp you need your pencil to be, and how much $$$ depends on the quality of the answer.

If you need a 'reasonable' answer (insert your definition of reasonableness here) then you can use the incompressible assumption as a defensible starting point. You assume no density change and no heat trasnfer over the length of the line, and use the Darcy equation. Run the calc and if your calculated dP is about 10% or less than the inlet pressure, your answer should be OK, within 25% or so.

If you need a better answer, you can use a physical property generator, or the steam tables, to estimate the density change along the length of the line. You split the line into segments and calc the dP in that segment using the density in that segment, again using the Darcy equation. Obviously this means you have to know something about the state of the steam in the line. This is a lot of work; proceed with caution.

Most industrial steam systems do not approach the velocities where a rigorous compressible analysis would be used. The vast majority of systems I analyze are best handled by a correlation like BBM, in my experience.

If you want the best answer, again in my experience, use a simulator like PipePhase because it will do the rigorous heat transfer calcs and phase change calcs, and it will use a decent correlation for the dP, like BBM. This method will predict your condensate rates, vapor and liquid velocities, pressure drops, steam exit quality, blah blah blah. Since you have a piping network, this is the way I would go since it will VASTLY speed up the calculations and leaves you with a model you can run under varying conditions of p, T, and x. For the system you describe I can build the model and run it in a few hours, including properly accounting for all valves and fittings. If your system goes saturated, all bets are off unless you do use the simulator. Don't even waste your time using a correlation like Lockhart-Martinelli and the Baker two-phase plot for this (that particular one happens to be easy to program into Excel). I would hesitate to use the 'black box' terminology here, for these simulators. Pipephase has all its correlations fully documented, so you can run a hand calc to back-check it, if desired. There's just no other way to do this kind of work that gives this level of accuracy for this level of time/effort investment.

In my experience, a steam system that is assumed to be adiabatic can be anything but. I have field measurements of large, fuly-insulated steam distribution systems, currently online, that show a 20% or more heat loss across the system. If you have uninsulated valves and flanges, pipe shoes, pipe anchors, support hardware, damaged insulation, etc., you will in fact have a non-negligible heat loss/phase change to contend with, which, if not accounted for, will muck up your calculations to varying degrees. Been there, done that, got the bite marks on my arse to prove it. ;-)

If you decide to go this way, you should be able to find a consultant to do this for you for $4-6k, depending on the complexity of the system and to what level you need the analysis done. If you email me (pjchandl@prou.com) I can prolly hook you up with somebody in your area, if you want.

Just what I think I know. - Pete Thanks!
Pete

 

[JoeChem]

74Elsinore

Thanks for the "real world view point" on this.

Actually, I believe this flow is very compressible. The piping has been around a long time and is undersized for the current use. Approximately 3500 - 4000 lb/hr of steam at a starting pressure of 35-45 psig is sent through 2 inch lines of varying lengths (50-173 ft) and fittings. The steam is screeeeeaming through these lines. Pressure near the final perforated pipe distributors is typically around 5 psig.

Yes there are some valve/fittings and other items that are not insulated well or at all. The heat loss is definitely not insignificant. The system is certainly not isothermal and I had to start somewhere.

By the way, how much does a simulator like PipePhase cost? This will be mostly for curiosoty reasons as I will not be able to appropriate funds for a simulator. I wish I could get my hands on one though...would certainly make life a lot easier. For now I will have to stick with an adiabatic model and a excel spreadsheet. Pmover has just sent me a copy of a computer program for sizing steam lines...this should be handy...thanks again pmover.

JoeChem