Surge Pressure due to sudden valve closure 1

[GPB005]

Problem:A pipe is connected to pump creating 13 bar pressure (shut off pressure 17 bar) has ESD valve at 3.5km distance.Flow velocity is 6 ft/sec. Valve closure Time is 4.5 seconds. The Pipe size is 28 inch. As per Joukowsky equation the surge force comes out to be above 200Tonnes. The above system is modelled in fluid flow software (do not want to mention name of software) with boundary condition at ESD as constant flow.
As per theory calculations if valve closure time is more thatn 7 to 8 seconds the surge should not occur. however, in software even if we go with valve closure time 60 sec or 120 sec we get almost same results i.e no change in surge pressure.
I am a piping engineer... cannot take such high load on pipe and believe something is going wrong in software. Software considers pump to be tripped only after total valve closure i.e end of 5 seconds and till then gives constant pressure. At the end of 5sec the pressure suddenly shoots to 50 bar or more. It means as per software the surge will take place irrespective of pump closure time which contradicts the theory.

Are we going wrong somewhere in simulating the situation in software?

 

[BigInch]

The Joukowsky equation should predict surge pressures in psi, kPa, or Bar. Tonnes is not a unit of pressure.

I'd expect a smaller pressure rise.

Some valves (balls and gate valves) have a faster flow cutoff rate at certain positions, essentially closing off most of the total flow within the last 10% of movement. If one of those valves has a linear proportional operator and takes 10 seconds to close, it essentially is going from full flow to no flow within 1 second.

you must get smarter than the software you're using.

 

[desertfox]

Hi

Look at example 3 on this link it gives sizing a surge absorber by hand, you can use it and compare with your software results.

http://www.accumulator.co.jp/eg/catalog/catalog_20090903.pdf

 

[bimr]

Yes, there does seem to be an error or misapplication of the software. You should not have such a rise over 7-8 seconds and with 6 ft/sec velocity.

 

[Xalii]

The value of the head surge, assuming a steel pipe, is about 25 bar, then I guess the software assume that the pump is a dead end and thus double the surge at the reflection giving the ~50 bar (or 200 tonnes for 28 inch). How correct is that is another question, if there is a quick NRV is might be correct, depending also the type of ESD valve you have as said by BigInch.
You could try to :
1) switch off the pump before the surge arrived to it (half of the return time ie 3 second after the valve is closed)
2) Change the type of ESD to have a better closing law and have a bigger closing time
3) Put a anti surge system as close as possible to the ESD valve.

 

[GPB005]

Thanks All for your inputs.
I have mentioned the Force of 200 Tonnes and Not the Surge Pressure (obviously it should be in bar, kPa etc.). The Pimp has NRV.
However my questions are:
1. Software gives same answer of surge pressure irrespective of Valve closure time i.e even if valve closure is 5 minutes it gives force of 200T. How it is possible?
2. The boundary conditions are ... one end Pump curve is input and at the end of the valve (other end)constant flow is maintained. Are these correct boundary conditions?
3. Let's assume software doesn't give force and it provides only differential pressure. Then how to calculate force :
F = delta P * Pipe cross section area
1. delta P = Max (P1)at time t - Max(P2) at same instance t..... where P1 and P2 are pressure at inlet and outlet of a pipe segment
2. delta P = {Max (P1) - Max(P2)} occuring at any time during surge event.. means acting P1 and P2 are occuring at different time
4. software shows constant pressure during the valve closure time? How it is possilbe as valve closes, as per Bernouli's equation Pressure should get reduced to increase the velocity.
5. When the pump will trip? Cut off pressure is 17 bar and surge pressure reaches above 50 bar (about 74 bar). So does it mean the pump should trip during the event of valve closure or at the end of valve closure? (Software considers it shuts off at the end of valve closure and suddenly gives rise in pressure)
6. How to simulate gradual closure of valve in software?

Most important thing is here we need to give worst possible forces to civil and we are not suppose to recommend how to avoid surge.....
regards

 

[LittleInch]

As ever it's kind of difficult to tell what is going on, but my shot at this is:
1) Wrong inputs
2) Doesn't look right to me. Outlet at constant flow may well be overriding all the other considerations. Normally flow programs work this out based on either a steady outlet pressure from the section and then the flow can vary as things change. You will need to use a transient program to calculate surge properly though.
3) this is end cap force and may act on the pipe if it was a "free end", but normally it is contained so that the force is partly resisted by the pipework the valv eis attached to
4) something is wrong in your valve set up, your program doesn't recognise valve closure versus CV change, or you have a very simple program
5) You've answered you own question
6) You give a varying valve Cv versus closure graph

Worst possible forces normally come from piping expansion loads and any anchors, not surge or transient events, but we can't see you layout or system so it is very difficult to advise.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[MortenA]

Are you pumping downhill?

If its not a true dynamic simulation are you sure that the software takes the closing time into consideration (maybe its just using Joukowsky and reports that independently?) Some valve characteristisk at very "sharp" towards the fully closed position. Have you checked this?

 

[LSThill]

ASME PVP2007-26676
ASME PVP2008-61620

 

[GPB005]

Thanks LSThill ... good papers you have shared . thanks others too for inputs. Actually Surge analysis is carried out by process engineer and i as a piping engineer challenging the software results. hence needed 3rd party opinion.

 

[IFRs]

I think "with boundary condition at ESD as constant flow." is your issue.