Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations waross on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Pipe network pressure drop analysis 2

Status
Not open for further replies.
dmeet

dmeet

Mechanical
Mar 6, 2007
5
0
0
US
Hello everyone.

I have a question for pipe network analysis. This can be done by computer programs but I would like to know if there is a hand calculation (spreadsheet) method available.

Even if someone can point me to a reference book, I would be greatful.

I tried searching for this on internet but was unsuccessful.

My question is.

1. If we have a pipe network, how do we find the total head developed so that the pump can be selected.

2. If we have heat exchangers in parallel, how do we find out the pressure drop in each line (is it similar to electrical resistance in parallel) and afterwards the total head developed for the pump calculation.

3. If a pump is pumping to a single pipe which then splits in to two lines going to two tanks (open to atmosphere) how do we know what will be the flow rate in each tank if these lines dont have a flow control valve. How can we find out the total head developed by the pump. To make this question interesting what if the pipe size going to these tanks are different and that the fittings are also different for both the pipes and that the tanks are at different elevations.

4. For a cooling tower there are several discharge points for a single line, how do we find out the pressure drop and the flow rate in each outlet if these outlets dont have a flow control valve or even a butterfly valve.

Till date I was using Crane technical paper 410 to do the calculations for pressure drop, but that was for a single input and single output pipe with just one pump connected. there were cases of 2 x 100% pumps, but I simplified the calculation by neglecting one pump. I took the additional tee in the calculation for pressure drop.

The reason for asking all these questions is that they were in my head for a long time and now that client has asked us to perform similar calculation, I thought it is the best time to ask other questions too.

Thanks

Dharmit
 
Replies continue below

Recommended for you

Sort by date Sort by votes
I've seen several people try really hard to do a pipeline model in a spreadsheet and it is always a failure. The reason for this is that there is not a closed-form solution to the mix of equations. You (or the program) guesses at a dP and a resistance for each line segment and then solves the equations for each segment in the system, looks at the results, adjusts the inputs, and tries again. I have one gas network that my model (MNET) has to iterate over 200,000 times to satisfy the constraints, that is a lot of hitting F9 in Excel.

David Simpson, PE
MuleShoe Engineering
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem
 
Upvote 0 Downvote
Pipe networks work just like electric circuits: Flow is analogous to current and head loss is analogous to voltage drop. The flow into a node (tee) must equal the flow out. The head loss around any loop of pipe must equal 0. The kicker is, unlike electrical circuits, the relationship between flow and head loss is non-linear; the resistance to flow depends on the flow rate.

Take your example number 3. You know the flow rate into the tee. The head loss from tank 1 to the tee must equal the head loss from tank 2 to the tee. (Don’t forget to include the surface elevations in the tanks.) So you have to change the flow in each line until two conditions are met: 1) the flow of the two lines out of the tee equals the flow into the tee, and 2) the head loss across both pipes are equal. Finally add this head loss to the head loss from the pump to the tee and you have the total discharge head loss. This example is relatively easy to do with excel. But only slightly more complicated networks get very messy to solve this way.

If the client wants you to perform these calculations then I suggest you get software to do it. I use AFT’s Fahom software which costs around 1,500. Still, even with the software you can get into a lot of trouble if you don’t have a good understanding of what’s going on “behind the scenes.” So go to the library and study any book on fluid dynamics. Learn how to do simple models by hand before using software to solve more complicated ones.
 
Upvote 0 Downvote
Quite a lot of questions with answers that would be difficult to fit inside the forum. I would suggest you get a copy of the following text. It is written specifically for gas networks, however the equations just need to be changed a little bit and they will work for liquids. It is really the best reference text I have seen on network analysis.

"Simulation and Analysis of Gas Networks", A.J. Osiadacz, Gulf Publishing Co.

In the meantime you might occupy yourself with this page. I think it has the answers to most of what you're interested in.


 
Upvote 0 Downvote
Thats really good advise I received.

As for starters, for small projects, I can take consideration of advise from khardy and do the calculation. Though it will be a bit tedious doing a manual calculation, but it will clear my doubts and I will gain confidence in the analysis.

I downloaded the Epanet software, thanks to Stanier, and am trying to work with it. It is a good utility for starters and will help me to check my answers from manual calculation.

I went on the webpage mentioned by BigInch and that will help me too.

Thanks for the post.

If you have any other comments please let me know. It will help me to improve the knowledge.

Dharmit
 
Upvote 0 Downvote
This is the gathering system from hell. 50% of the lines are looped, and in only about 1/3 of them do the loops terminate at the same place as the original pipe (so the loops really become cross-connects). There are two off-system delivery points that have very similar hp and they fight pretty hard.

The system is really two systems with about half the gas on the top of the mesa and half in the wash with a 14-inch line between them. It only takes 1/2 psi change in suction pressure at either of the sites to change direction of flow in the 14-inch. Less than 1/4 of the inch-miles of pipe are piggable and all the others have standing water in them (they are fiberglass so corrosion is not an issue).

Yep it takes between 125,000 and 200,000 iterations to solve the model.

David
 
Upvote 0 Downvote
Gottcha'. I see the problem is a matter of setting the acceptable error, which in this case must be 0.5 psi and that is very very small for a large number of loops and two identical compressors. I'm surprized you got a solution at all. Gathering sysems usually arn't planned in advance to give good hydraulic characteristics, but result from a time series of very optimistic geologists filling any given area with far too many holes as fast as they can.

 
Upvote 0 Downvote
Actually, this was the first gathering system in the San Juan Basin CBM "Fairway". Based on the Black Warrior Basin (the only commercial CBM play at the time), the geo-guys predicted 19 MMCF/d peak rate. A very gutsy facilities guy designed the system for 38 MMCF/d sustained rate. When I took it over it was making just over 300 MMCF/d so I looped everything, built additional compression, etc. Now there is pipe in the ground for 300 MMCF/d, but the field has declined to under 30 MMCF/d (12 years later).

My model had 0.05 psi tollerance built in. Any coarser and the reasults got progressively less meaningful. I was using the model to evaluate the pigging schedule, so fairly subtle changes were really important.

David
 
Upvote 0 Downvote
Vacuum system?

The gathering field I worked in had the ultimate optimist geologists in a very tight lower Wilcox zone. They would be thinging 50 MM for 5 years, get 20MM for 30 days, then only 50M for the next 3 years. By the time they figured out it was only getting the small 50, they'd already dug 2 more and I'd already put in 10 miles of 12". But they could drill fast enough to keep up 800 MMSCFD from the system ... back then. Wonder what its doin' now.

 
Upvote 0 Downvote
For item 4, the flow to each cooling tower feeder is setup in the field using an ultrasonic flowmeter or a pitot tube inserted into supplied ball valves on each riser. Internal to the tower, each set of spray nozzles has a additional flow equalizing setup .

Item 2 is a special case of item 1, and can be quite sophisticated if you follow it thru to the bitter end. For the case of a HX, you assume several different flowrates thru each HX, and for each flowrate, calculate the heat transfer and pressure drop. This calculation must include gravity head changes, which become very important for phase change cases ( ie boiling or condensing water). Each HX with its piping will then have a curve of DP vs W (flow) plotted. Since each set of the HX's with their piping are operating in parallel with the other HX's, they must have the same total DP, and the sum of all the flows must equal the total known W. Now plot a curve of DP vs (sum of all W's) for all HX's in parallel. Since you know the total W, read the DP associated with that W and then work back to the W for each HX at the known common DP. This works easily if the flow is stable and montonic in all HX's.

The problem becomes "sopisticated" if there is flow instability- ie, if the DP vs W curve for each HX is not a monotonically changing DP with W. Key phrase- Ledineg static flow instability, or parallel channel instability. This is a problem if it occurs at the range of W's that the process is expected to operate at.

For the case of a simpler piping system without HX's, a simailar process is used. All pipes working in series will have a common W and the DP's are additive, whiel all pipes in parallel have a common DP and the W's are additive.
 
Upvote 0 Downvote
In the practical case where you are setting up a closed loop circulation system or a cooling tower, you provide butterfly vlaves on each line to each HX or feeder which can be initially set at a fixed % open. During initial field operations, the butterflys are throttled to provide the mfr's design DP across the HX , or ultrasonic flow meters are clamped onto the pipe and the butterflys are throttled to provide the mfr recommended flowrate.

There is no way you can accurately predict the DP across a large series of randomly oriented bends in a piping system- the actual DP across each bend is really a function of what the relative configuration of each bend is with its next upstream bend, due to the effects of pre-swirl. So you have to add flexibilitly to the system by adding throttling butterfly valves and 10% mor pump flow than theoretically required.
 
Upvote 0 Downvote
Thanks guys,

Thats good history lesson along with the online calculator to support it.

Is there a book that deals with specifically flow analysis, network analysis only and the finite element analysis.

Thanks

Dharmit
 
Upvote 0 Downvote
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

nesamani1983
  • Locked
  • Question
Hydrotest pressure drop analysis 1
Replies
9
Views
43
nesamani1983
nesamani1983
Eric_eng
  • Locked
  • Question
Pressure Drop in Hot Oil Circulation System 1
2
Replies
37
Views
78
katmar
katmar
Sruthish
  • Locked
  • Question
Pressure Drop across the Mist Eliminator Vessel with Internal Cyclone Bundles.
2
Replies
31
Views
93
georgeverghese
georgeverghese
Asisraja D
  • Question
Need review for pipe wall thickness calculation sheet for pipes under internal pressure 7
2
Replies
23
Views
102
GD2
GD2
sentrifice
  • Question
Welding new pipe to corroded pipe 1
Replies
9
Views
117
GD2
GD2
Back
Top