Water hammer in system with two MOC? 1

[KLee777]

How does one analyze a piping system that involves two or more different types of pipe? The system I'm working on involves FRP connected to HDPE...then back to FRP. Pipe sizes are also different, but I can handle that.

Any suggestions?

Thanks,
KLee777

 

[BigInch]

This is not something I started doing yesterday. Been there, done that and you are wrong... if the area difference is "significant". If the area difference is not significant you are correct. My only point was that you cannot arbitrarily neglect reflected waves. You must analize it (or know from experience) that reflections are insignificant before you can make your statement, and the statement is case specific. Start up of a pump may give a significant reflection from a diameter change while a flow rate change made via an FCV may not, so the reflection is not even the same for different control actions. Do you neglect reflection effects from pump starts? I made a special case, just for you, that I have running now that shows a 20% over pump shutoff head maximum pressure reflected back to a pump station off a diameter change. Assuming that MAOP of your pipe equals shutoff head, can you can show me what code you're using that would ALLOW you to neglect that?


Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[sailoday28]

Big Inch--The fact is that any hammer or fluid transient analysis dealing with a discontinuity such as area change treats the discontinuity as a quasi steady flow. Now my experience is limited to one dimensional transients.

I'd be interested to know how your program or analysis deals with a sudden area change. What conservation equations are used?

Of course one can model reducers, etc when dealing with area change, but that is not the case of a discontinuity. The characteristic equations for that model can also be a type of nightmare for the analyist.

Regards

 

[stanier]

Idelchik gives for an abrupt change at pages 208 & 216 respectively

Kup=(1-Aup/Adown)^2 for an expansion and
Kup= (0.5(1-Aup/Adown)^0.75)/((Aup/Adown)^2)

This is adopted by AFT's Impulse for computing the loss at the change in diameter. A separater node for such changes can be used to define the boundary between materials and diameters.

I am not about to argue with the likes of Idelchik.
Idelchik, I. E., Handbook of Hydraulic Resistance, 3rd edition, CRC Press, Boca Raton, FL, 1994.

Geoffrey D Stone FIMechE C.Eng;FIEust CP Eng

http://www.waterhammer.bigblog.com.au

 

[BigInch]

As described in "Exxon-Mobil Pipeline Hydraulics, 2003"

These equations are related to the velocity of sound in each segment.

Amplification Factor for transmitted wave = ((2*A1)/a1) /(A1/a1+A2/a2).

Amplification Factor for reflected wave = (A1/a1-A2/a2)/(A1/a1+A2/a2)

Where A = respective pipe areas up (1) and dn (2)
a = velocity of sound in respective pipe segments




Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[KLee777]

"I was reaching the same conclusion. I'm originally from Texas, so I still remember what a load smells like...If I was their client, I'd have to bust their something or other."

Hold on there, BigInch. Geez. IN NO WAY was I saying that I am totally ingnoring pressure reflections, much less surge in general (what kind of engineer would that make me)? Check out my other threads and you'll see I'm not one prone to neglect--I ask questions so that I'll know what I *don't* know! (Thanks to stanier for recognizing this.) However, I may not have been as clear as I should have been in my post and I apologize for that. Perhaps it has something to do with these 16 hour days I've been working....

The area change in this system is *very* small (4.0" I.D. pipe going down to 3.94" I.D.). I was originally more concerned with the change in MOC anyway. Since I'm no expert in hydraulic calculations, I discussed this at length with several of my company's fluid transient "experts," that is, once I was finally able to get in touch with them. All agree that any reflections due to area change alone would be virtually negligible, especially since the area change occurs ~5,600 ft from the pump, and we're pumping wastewater at <3 ft/sec. It's not like this stuff is screaming through the pipes.

Turns out we're leaving the calculation to the more experienced guys, even though it comes with a bigger price tag for the specialists. That's what we do when we want to get it right. Although, I'm not relinquishing this altogether--I'm going along for the ride so that I can learn, too. It never hurts to be more informed.

"Then KLee says that his client wants a very detailed analysis, and yet he apparently proposes to reflections. Am I reading that wrong?"

Yes, you are wrong, the client does not oppose, WE do in this case--but if the client wants the transient analysis, we'll provide it. And please, don't automatically assume I'm a guy...it gets annoying sometimes. I do appreciate the rest of the discussion, but I ask everyone to please watch their tone and/or assumptions when posting so as not to offend. The first quoted statement kind of left a bad taste in my mouth, even with advance apology.

Thanks,
KLee

 

[BigInch]

KLee,

You're right! I'm, not as diplomatic as I should be all the time. I (and a few select others) have "buttons", of which apparent ignoring transient analysis is one and it obviously gets me ballistic, however I am not saying this as an excuse for poor choice of words. What I will say though is that it wasn't ment for you. It was aimed at sailorday's statement, "treats the discontinuity as a quasi steady flow" (to whom I will also offer the same appology) but, only after he gives us his equations. I still believe he is only thinking steady state pressure loss in the short section in relation to steady state pressure loss in the other longer segments, which I think everybody pretty much has that down after fluid mechanics 1001 and is completely irrelevant to transient reflections and transmissions from a reducer/expander. Actually what I understand about the hydraulics of his statement is so obvious, that I can't believe he is making a comment about transient analysis at all.

Sorry, I didn't mean for you to be a victum of friendly fire . I assumed you were just asking for advice when you made no attempt to debate the issue one way or another.


Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[sailoday28]

I'm going on vacation for about a week, but will give a fuller explanation of "my???" equations later.

Regards

 

[KLee777]

Thanks, BigInch. I understand that everyone has their buttons, I have mine too! Just wanted to say that I didn't respond right away because I simply did not have the time. Sit back and give it a few hours next time to see if the poster will respond. Otherwise it feels sort of like an attack (love the machine gun guy, by the way)!

Thanks again to all for their input. This forum has always been a place for good information and debate, and I hope it continues to be such.

Regards,
KLee

 

[BigInch]

I don't mean to take things to a personal level. Its just my way to talk about things directly and impartially as I see it. I have grand delusions of Dennis Weaver in "McCloud". I'll see if I can't keep them in the holster a little longer.

Well, I tried. Didn't do too well. I need professional help. Just shows you it wasn't personel, just my way with everybody. I'm just a in a china shop. You might want to check out my latest response over on,

http://www.eng-tips.com/viewthread.cfm?qid=161768

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[BigInch]

PS I wouldn't even count anything less than a full diameter change as an "area change", but, as Stanier says, watch the different effects from Youngs modulus variation with the FRP's t/D. I'm still greatful for that comment.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[sailoday28]

u=velocity A =area p= pressure

1[---------------------]23[--------------------]4-

Consider stations 1 far to the left of sudden change in are 2 to 3. Stations 1 to 2 have the same diameter and stations 3 and 4 , the same diameter but different from 1 or 2

For simplicity negelect friction and elevation changes and assume sound speed is constant through out.
A right characteristic traveling from 1 to 2 travels at dx/dt =u+c, along that characteristic du+dp/rho/c=0
(u2-u1) + (p2-P1)/(rho*c)=0. equation 1
Along a left traveling characteristic which travels at dx/dt=u-c from 4 to 3. Along that characteristic du-dp/rho/c=0
(u3-u4)-(p3-p4)/(rho*c)=0. equation 2
Quasi steady conditions are taken across the discontinuity.
Between stations 3 and 4 if flow is left to right (p3-p4)/rho=(u4^2-u3^2)/2 equation' 3
and from continuity u3*A3=u4*A3 equation 4

4 equations and the unknowns are p2, p3,u2, u3
It is assumed that conditions at 1 and 4 are known,
For reflections
1 and 4 will be impacted by characteristics traveling back from 2 to 1 at dx/dt =u-c
and 3 to 4 along characteristics traveling at dx/dt=u+c
Equations 1-4 are easily solved with algebra.

The above consider highly incompressible fluids.
With compressible flow the curves of x vs t anre not nearly as straight as for incompressible.

I'll stay glued to the computer for a little while as I leave tomorrow for a week.

Regards

Regards

 

[sailoday28]

Please correct my equation 4 to read
and from continuity u3*A3=u4*A4 equation 4

 

[BigInch]

I thought there was something else. I'm still not sure why you neglect reflections. Apparently you are neglecting them simply because the sign is negative and of relatively small amplitude. Correct? That assumes that the only effect you are interested in examining is the maximum pressure of the pipeline.

I will now ask you to consider what the impact of that reflection might be on equipment at the beginning of the pipeline, if, as Stanier mentions, no vacuum, or that might be trying to maintain a constant discharge pressure or flow, or for a typical scenairo I see a lot of, the effect on a high speed, high pressure pump with a 1500 m static head that has just started a runup? I don't think you can say, neglect reflections, then justify it by saying, "Well. The pipeline must be flat and it must be incompressible flow.

Hope you had a good vacation.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[stanier]

In respect of the "button pushing" I would like to enquire of Klee777 why the analysis is only done if the client wants it. Or did I get that wrong.

Standards and codes require surge to be included in the design pressure. So the designer has to do the analysis to determine the design presure and hence wall thickness of the pipe.

Surge analysis is not optional. It is not something to be guessed. Yes it is complex. There are tools available for the engineer to use to determine the surge in a system. It beggars belief that companies will spend a small fortune of software for fancy presentations, risk analysis and stock control but wont give engineers the tools to address design parameters that are mandated in standards and codes. I am more bemused by the ignorance in the piping industry of these requirements.

Geoffrey D Stone FIMechE C.Eng;FIEust CP Eng

http://www.waterhammer.bigblog.com.au

 

[BigInch]

A client I have did a transient analysis for several large refined product storage sites as part of an initial design contract, but the transient analysis was done only for the design flowrates. The analysis was limited to examining the effects of closing critical valves and stopping pumps and for the design of a surge control system, but nothing else. We have found that doing an additional comprehensive transient analysis covering the entire range of flowrates and all possible operating scenarios from start_up to maximum possible flowrate can be quite valuable. Design deficiencies and bottlenecks can be identified. Operating temperature ranges of pumps can be determined. Intermediate operational scenarios can be identified and control points developed. Control sequences can be verified, corrected and even the timings adjusted. Pump, compressor and instrument parameters can be evaluated. Pressure setpoints for smooth start-ups can be determined. Optimal pump configurations and horsepower settings can be predetermined. And when the comprehensive transient analysis is finished, the model can be attached to a training simulator for pipeline operators, so they can be trained in what to expect when they turn the key or a particular pump doesn't start on time and the pipeline pressure upstream goes high. Common possible operational mistakes can be identified and recovery methods examined. Leaks can be simulated and spill volumes at various points can be determined to assist preparation of emergency procedures. Proposed upgrade modifications to the pipeline can be checked in advance. The original transient analysis can be the starting point to a highly leveraged operational tool useful for the life of the pipeline.





Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[stanier]

BigInch,

You certainly know how to push buttons. Fancy wanting to do more engineering to optimise the design, construction and operation of a facility. In today's modern world dominated by "project managers" this is surely heresy and you will burnt at the stake.

$1 spent at concept design stage saves
$10 spent in design saves
$100 spent at detail design saves
$1000 spent in procurement saves
$10000 spent in construction saves
$100,000 spent in operation saves
$1,000,000 when the lawyers are called in


Geoffrey D Stone FIMechE C.Eng;FIEust CP Eng

http://www.waterhammer.bigblog.com.au

 

[BigInch]

All those numbers are correct, except for the last, where you must multiply by (1+ RAND(20)/100MM project_cost). For offshore work you can probably 2X to 4X those.

Today its easier to justify a project by using the "avoided legal cost" method.

"Way back when" in the early '80s, my boss was picking up some complaints from the project management group about my offshore pipeline projects taking too much engineering and design time. During those years we were making numerous short pipeline tie-ins between existing lines and new platforms, so each engineer did 4 to 6 of those a year, so I knew quite a lot of data was available. I went to the Accounting dept. and got the financial records for all projects done in the previous 5 years. My projects were taking 33% more time than other engineers, but fortunately were still being finished within overall schedule. Thinking I was in deep... water, I had to come up with a way to justify my engineering costs. It seemed to me that the construction managers never gave me as much trouble as the other engineers, (one they cruicified on a daily basis) so I started checking my project's construction costs. I found my pipelines had little if any extra field work charges that were not weather related and their costs, when including construction, were 20% less per diameter-mile with the average cost being around $1xE6 less. I had to assume that the savings were due to the extra engineering hours spent doing detailed dimensional checks on the platforms, double checking all drawings, making accurate material take-offs, requisitioning proper quantities and tripple checking all mark numbers against drawings and requisitions. 2000 hours extra was returning 1 million dollars, when the average engineering hour was around 25.00 Spending $1.00 on engineering netted a payback of around $20. I presented that case and got a salary increase. Don't let anybody tell you its not true.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[rconner]

After reading through this lengthy thread, I wonder if the original inquirer yet knows exactly how to determine the design pressure when different pipe materials of construction are all thrown together? I guess while it might be judged over-conservative by some, in the absence of more definitive field research I kind of like the first thoughts of BronYrAur, that if I understood the post correctly may have been to determine a conservative surge as if the pipeline were at least a little more simplistic system of one stiffer material and go with the higher number (at least on paper this might be harder to at some point in someone else's forensic work judge unsafe?) More detailed analysis might well require the services of a "sure 'nuff" specialist or expert in pipeline transients, who additionally is willing to stand by their detailed surge analysis of composite pipelines involving multiple materials. I noticed similar or related subjects were brought up in a couple other threads a few years ago, including

http://www.eng-tips.com/viewthread.cfm?qid=125930&page=10

and

http://www.haestad.com/hmicom/lists...ction/thread/messid/10829/v/n/searchid/76066.

I suspect different materials of construction, and I will note also variations in stiffness of encasements and penetrations of buildings/structures etc. are probably more a rule than an exception in many contemporary systems.

 

[stanier]

rconner,

The scenario can be readily modelled in Impulse (

http://www.aft.com)

with all changes of diameter, materials, abrupt diameters etc.

The original poster has taken the advice and is going to get it analysed using this software by people that know what they are doing. The original post has been answered and the rest is just cats whaling on the fence post at night.

Geoffrey D Stone FIMechE C.Eng;FIEust CP Eng

http://www.waterhammer.bigblog.com.au

 

[sailoday28]

u=velocity A =area p= pressure

1[---------------------]23[--------------------]4-

Reflections must be included. I only gave a simple portion of an analysis.

Refering to my previous post, the pressures velocity etc are solved at stations 2 and 3 and say time t1.

The reflections from 2 to 1 along characteristic dx/dt=u-c
and u1-u2-P(p1-p2)/rho/c=0
The reflection from 2 to 3 along dx/dt =u+c and hammer formulation u4-u3+(p4-p3)/rho/c=0.
The above equations are solved for pressure and temperatures and the new time(s) that they occur.

The equations stated above and on the previous post were simply given to show how the discontinuity of area is accounted for.

Typical MOC will approximate the characteristics dx/dt as +or - c. This is because the velocities, u are generally much less than the sound speed, c.

Big inch "I will now ask you to consider what the impact of that reflection might be on equipment at the beginning of the pipeline,"


u=velocity A =area p= pressure

1[---------------------]23[--------------------]4-


Consider known condtions of velocity and pressure and station2 and 3 at time t. This would have been obtained from the previously solved equations

Let us assume that at station 1, similar to a H-Q curve we know the Pressure1, velocity1 relation.
Along the reflected characteristic dx/dt=-c
U1-U2-(P1-P2)/rho/c=0 P2 and U2 are known and with this equation AND a relation between P1 and U1, the pressure and velocity at station 1 are determined along with the new time.
Let's say the pressure at4 is know. The the equations along the dx/dt=+c characteristic are solved.


I hope this is answering your question(s).
I have tried to give a simplifed explanation to the use of the discontinuity. I have assumed that use of characteristics dx/dt and du+/=dp/rho/c were already known.

Believe me, using constant sound speed and low velocities and relatively incompressible fluids makes analysis simple compared to compressible flow.
Regards