29 replies to this topic

[Christiano]

Any inputs to the following is highly appereciated

Project Background:
1400 km stabilised crude (52") and fuel gas (22") line operating @ 70-75 bar with 10 pumping stations and elevations ranging from 1m to 800m ASL.

1. Surge Relief: planning to keep 3 +1 gas turbine driven mainline pumps at pump stations of 10MW each. Sectionalising valves enroute are slow closing and hence relief not considered. What kind of surge relief systems could be considerd in your exp for pump trips?? Is it 2 way bladder tanks or a normal relief tank to receive the surge for pump trips. Also could you let me know if gas turbiine driven pumps require surge relief as the trip time(coast time) is more than motor driven pumps. Any advice would be helpful.

[ankur2061]

Raj,

Since you are using a gas turbine driven pump and since the trip time (coast) is long your surge presure wave in the reverse direction (towrds the pump) will be of smaller magnitude. You may require to discuss your requirement for surge protection with your pump supplier considering your pump trip will not be instantaneous.

Surge pressure calculations can be done using an excel sheet

The maximum surge pressure in any pipeline is the sum of 2 components:

1. The instantaneous pressure increase at the moment of total flow blockage

2. The subsequent gradual pressure rise due to the "line packing" effect.

The magnitude of the instantaneous surge can be calculated using the Joukowsky's equation:

P_s = ρ * a * Δv

where:
Ps = Surge pressure, Pa
ρ = liquid density, kg/m³
Δv = velocity change, m/s

The wave velocity "a" is given by

a = sq root (1 / ( (1/K + d/t_w + 1/E)*ρ)
where:

K = liquid bulk modulus, Pa
d = pipe ID, m
t_w = pipe wall thickness, m
E = Young's modulus for pipe material, Pa (for commercial steel Young's modulus is considered 210*10⁹ Pa)

When protecting pumps from surge due to pressure wave the relief system should be quick acting such as bursting discs or rapid response relief valves.

Hope this helps.

Regards,
Ankur

Edited by ankur2061, 19 December 2011 - 12:18 AM.

[Christiano]

Thank you ankur..

[kkala]

The reader should be aware of some additional info on the subject, mentioned in http://www.cheresour...-for-pipelines/, Maop for pipelines, post No 10. Link to Saudi Aramco crude pipeline design is useful for general knowledge. Relief set pressures resulting from surge analysis differ from case to case. Surge analysis through specialized software may be necessary for a complex 1400 km line, requiring a lot of pipeline data and patience for a rather precise result. A local Process Dept used TLnet for pipelines 20 - 45 km, Aramco uses other software in addition (as you can see in the link).

[ankur2061]

Dear All,

A nice article on liquid pipeline surge is attached.

Regards,
Ankur.

Attached Files

•  Liquid_Pipelines_Surge.pdf   89.69KB   659 downloads

[Christiano]

Hi Kala/Ankur

Thanks for your inputs.

I am using Hammer V8 (from bentley) to model transient analysis for pump/pumps trip. Planning to taking a stoppage time of 2 min for pump+turbine of this magnitude (10 MW each).

Please note for remote block valves I am adopting a philosophy of slow closing isolation valve (after determining the min closing time) which are FO type. Could you suggest if this is sufficient for remote block valve stations and further surge analysis not required!!!??

Regards,

Rajesh

[ankur2061]

Rajesh,

A closure time of 30 to 60 seconds is considered to be more than adequate for the purpose of minimizing a pressure surge wave. In many projects where as an engineering consultant we did a surge analysis this value proved quite adequate. Stoner software by Advantica was used for the surge analysis.

You still need to check out using the "Hammer" software by providing as an input a valve closure time of 30 to 60 seconds.

Regards,
Ankur.

[kkala]

I am using Hammer V8 (from bentley) to model transient analysis for pump/pumps trip. Planning to taking a stoppage time of 2 min for pump+turbine of this magnitude (10 MW each). Please note for remote block valves I am adopting a philosophy of slow closing isolation valve (after determining the min closing time) which are FO type. Could you suggest if this is sufficient for remote block valve stations and further surge analysis not required!!!?? Regards, Rajesh

A locally applied rule of thump suggests valve closing in 5*d s, d=pipe diameter in inches (e.g. for d=10", valve to close in 50 s). This rough preliminary indication is for pipelines within a plant or refinery, needing more detailed examination in "suspected" cases. But what about long distribution pipelines?
Suppose closing a valve at main line, 60 km far from pump discharge; pressure wave velocity (celerity) is estimated at 1200 m/s. Time for pressure wave to travel to the pump and return is 60*2/1.2=100 s. Any closing time shorter than 100 s has same result as a sudden closure. To "smooth" surge, valve closing time should be "preferably several times 100 s" (*). But most probably the valve should not close too late ("after the fiesta"), so a compromise has to be made between closing time and surge pressure through the aid of suitable software. The latter can also help in specifying required protection against developed overpressure.
Now "remote block valves" may not be on main line, yet feeding line having mentioned valves on it may be long. Check of their closing time and consequential surge is worth while, at least in some representative cases to access general situation.
Mentioned Process Dept (post 16 Dec 11 by kkala) performed a detailed investigation on these 20 km or 45 km pipelines through TLnet. Of couse lines were very simple (one start, one destination) compared to an operating 1400 km network.
Note: Yes, valves have to be specified as failed open. In the local investigation, 10% surge pressure over design pressure was considered acceptable (this depends on piping code).

(*) Schum's "Theory and problems of Fluid Mechanics and Hydraulics", SI edition, McGraw-Hill, 1977, Chapter 11 - Forces developed by moving fluids, problem 53 (page 221).

Edited by kkala, 17 December 2011 - 07:11 AM.

[ankur2061]

Kostas,

Please explain the rationale behind the locally applied rule of thumb of "5*d" seconds. Is it given in TLNET or anywhere in any literature for long distance pipelines?

In fact a lot of literature suggests valve closure times of 5, 10, 20 seconds for pipelines as can be seen from the following link:

http://cla-val.com/pdfs/B-59.pdf

Ankur.

[kkala]

Kostas, Please explain the rationale behind the locally applied rule of thumb of "5*d" seconds. Is it given in TLNET or anywhere in any literature for long distance pipelines? In fact a lot of literature suggests valve closure times of 5, 10, 20 seconds for pipelines as can be seen from the following link: http://cla-val.com/pdfs/B-59.pdf Ankur.

Above complies with rule, for small line diameters (d up to 4"); so does following for higher diameters.

Rajesh, A closure time of 30 to 60 seconds is considered to be more than adequate for the purpose of minimizing a pressure surge wave. In many projects where as an engineering consultant we did a surge analysis this value proved quite adequate. Stoner software by Advantica was used for the surge analysis.
You still need to check out using the "Hammer" software by providing as an input a valve closure time of 30 to 60 seconds. Regards, Ankur.

Mentioned rule of thump is flexible enough to increase closure time with diameter, which is an advantage.
However, for constant fluid velocity, water hammer intensity does not depend only on line diameter but also on line length. Hence rules of thump as above have quite limited validity, as explained in the today's post No 8 by kkala.
TLnet is not expected to contain such a rule. I believe something similar will be found by searching international standards. But, as already said, caution is needed on its application.

Edited by kkala, 17 December 2011 - 10:33 AM.

[ankur2061]

Kostas,

A small request. While mentioning local practices and local rules of thumb also mention that these may only be applicable locally and may not apply globally and the reader needs to exercise caution in considering these for his or her purpose.

Regards,
Ankur.

[kkala]

Since the matter of valve closure time has been touched, a relevant link is enclosed. But I believe only surge analysis can give a rather reliable answer, as said in previous posts; unless somebody is highly experienced and conservative. Reading the matter may also lead to this conclusion.
The link is http://www.eng-tips.....cfm?qid=151365

[kkala]

Kostas, A small request. While mentioning local practices and local rules of thumb also mention that these may only be applicable locally and may not apply globally and the reader needs to exercise caution in considering these for his or her purpose. Regards, Ankur

I bet mentioned "rule of thump" can be found as a practice of some reputable engineering company, probably with a different expression. As clarified in http://www.cheresour...-control-valve/, local practices do not contradict international practices, they are influenced by them and can help interpretation on specific points.
Things are actually quite simpler, yet anybody has the right to doubt; so origin of information had better be clarified. But just imagine to write a personal conclusion (e.g. post No 7 by ankur2061, yesterday ) and then add "personal conclusion, reader needs to exercise caution in considering these"! This would not be proper, and neither would to report something similar after any reference to "local practice". Anybody can assess the information, "local practice" just informs about the origin.
Of course authority increases as we proceed from personal conclusion to local practices and then to international practices; this is understood. And has to be remembered, when "according to local practices" is read. Any information has its own value and usefulness, but it can be also assessed. Usually the problem is not the practice itself, but how it is applicable to the specific case.
However, can some contradiction be specifically detected between mentioned local practice and some international practice? This would be constructive and helpful to point out. Otherwise argumentation is for nothing.
Greek legislation cannot contradict EU legislation and this has influence also on practices. Only a time delay can happen between a new EU Engineering Practice and the issue of corresponding Greek Practice.

Edited by kkala, 17 December 2011 - 06:03 PM.

[ankur2061]

Kostas,

The time mentioned by me in post # 7 of 30-60 seconds is not based on personal opinion it is based on surge analysis using Stoner software. By just mentioning TLNET does not make anyone knowledgable on surge analysis and by the way it is not "rule of thump" it is "rule of thumb".

As far as betting on rule of thumb being found as a practice of some imaginary reputable engineering company, the forum does not encourage gambling. Constantly mentioning local practices and local rules of thumb also indicate half-baked knowledge and the propensity to bluff.

And as you mentioned "nobody has the right to doubt" cannot be accepted. Anyone has the right to question and ask for justification and if the person is honest he or she will accept his ignorance or lack of knowledge and say so.

Since valve closure time has a major impact on the surge analysis let us get the fact right in terms of practices followed by internationally recognized companies instead of saying local rules of thumb. Here is what Shell standards have to say regarding valve closure time

3.1.2 Valve closing

The closing time for valves up to and including nominal size DN 600 SHALL [PS] be no more than one second per 25 mm of nominal diameter. For larger sizes the maximum closing time SHALL [PS] not exceed 30 s, unless otherwise specified by the Principal.

NOTE: ESD valves in LPG/LNG marine loading arms SHALL [PS] be able to close within 5 s; see DEP 30.06.10.20‑Gen. This may be a reason to select a hydraulic system instead of a pneumatic or electrical system.

The final valve closing times SHALL [PS] be determined on the basis of surge calculations, and the flow rates may have to be reduced to avoid excessive surge pressures generated by rapid valve closure.

As far as I am concerned this thread is over and you can keep writing a novel on the subject.

Edited by ankur2061, 18 December 2011 - 01:31 AM.

[kkala]

The time mentioned by me in post # 7 of 30-60 seconds is not based on personal opinion it is based on surge analysis using Stoner software.
It is a personal conclusion based on some worked examples. There are cases when the conclusion is not valid, see posts No 8 and 12 by kkala (yesterday). Yet it has value, as a try to systematize knowledge.
Practices, including local ones, have more value. They are based on broader experience.
By just mentioning TLNET does not make anyone knowledgable on surge analysis and by the way it is not "rule of thump" it is "rule of thumb".
As already noted, TLnet (mentioned for information) is irrelevant to this rule of thumb (yes, thumb is the right spelling).
As far as betting on rule of thumb being found as a practice of some imaginary reputable engineering company, the forum does not encourage gambling.
The rule is believed to be imported,"Local does not mean limited to Greece, I have seen them applied here. In fact they have much wider influence (and they are influenced)" (http://www.cheresour...control-valve/ . Search by an expertised willing person could promote the topic, even though this rule of thumb is an improvement over stated "30-60 s" (as already shown).
Constantly mentioning local practices and local rules of thumb also indicate half-baked knowledge and the propensity to bluff.
False accusations, deserving no answer. Not documented at all, in my opinion. Criticism on specific "local practice" could be constructive, not generics. Let the assessment of local practices to readers, when referred.
And as you mentioned "nobody has the right to do doubt" cannot be accepted. Anyone has the right to question and ask for justification and if the person is honest he or she will accept his ignorance or lack of knowledge and say so.
"yet anybody has the right to doubt; so origin of information had better be clarified" has exactly opposite meaning.
I agree to the text following the misunderstood sentence, noting that this should be valid for all members without exception and that good faith from both parties is necessary for the conversation through posts.

Edited by kkala, 18 December 2011 - 06:06 AM.

[kkala]

....Since valve closure time has a major impact on the surge analysis let us get the fact right in terms of practices followed by internationally recognized companies instead of saying local rules of thumb. Here is what Shell standards have to say regarding valve closure time

3.1.2 Valve closing The closing time for valves up to and including nominal size DN 600 SHALL [PS] be no more than one second per 25 mm of nominal diameter. For larger sizes the maximum closing time SHALL [PS] not exceed 30 s, unless otherwise specified by the Principal. NOTE: ESD valves in LPG/LNG marine loading arms SHALL [PS] be able to close within 5 s; see DEP 30.06.10.20‑Gen. This may be a reason to select a hydraulic system instead of a pneumatic or electrical system. The final valve closing times SHALL [PS] be determined on the basis of surge calculations, and the flow rates may have to be reduced to avoid excessive surge pressures generated by rapid valve closure.

As far as I am concerned this thread is over and you can keep writing a novel on the subject.

I would have been glad to know that the subject of local practices (not the main subject) has been over. Opinions from both parties have been developed, so that readers can be aware. Following are a few additional points on closing the subject.
1. I had sent post No 15 in reply to Ankur's post No 14, when I saw that post No 14 has been edited, adding some more test. I had not noticed it, so the present post replies to this additional test, framed above.
2. kkala has not initiated the subject (novel?) of local practices ; however kkala ought to respond to queries by Ankur, who has started the subject.
3. Concerning the framed text on the top, following seems to make an adequate response.

...A closure time of 30 to 60 seconds is considered to be more than adequate for the purpose of minimizing a pressure surge wave. In many projects where as an engineering consultant we did a surge analysis this value proved quite adequate. Stoner software by Advantica was used for the surge analysis.You still need to check out using the "Hammer" software by providing as an input a valve closure time of 30 to 60 seconds...

Mentioned rule of thump is flexible enough to increase closure time with diameter, which is an advantage. However, for constant fluid velocity, water hammer intensity does not depend only on line diameter but also on line length. Hence rules of thump as above have quite limited validity, as explained in the today's post No 8 by kkala.
TLnet is not expected to contain such a rule. I believe something similar will be found by searching international standards. But, as already said, caution is needed on its application.

Shell is much less conservative, contrary to both posts mentioned; but they also say that final closing time shall be determined by surge calculation.
Closing time t (before surge calculation) can be a vague criterion, differing from organization to organization. Is it applied t=5*d sec (d in inches) too conservative? Post No 8 by kkala indicates a case where t=multiple of 100 s, to avoid surge (though it does not concern refinery area).
Consequently all lines subject to water hammer risk should undergo surge analysis, when this is possible. This shall define developed surge pressure and then way of protection.

Edited by kkala, 18 December 2011 - 11:07 AM.

[Christiano]

Hi

Thanks for your contribution. While doing a preliminary check on the design pressure of the line, attempting to limit the MAIP (Max Impulsive pr) to 110% of design pressure I found the closure time is very high for the farthest block valve. I have block valves every 25 kms as a prelim guideline and the farthest is 150 kms from pump station. Now when I calculate the min closure time to limit the surge to 110% DP, I require about 1500 seconds to safely limit the MAIP to the above value.Could you please Comment on this?

IN addition to the above, could you comment if any specic considerations are there for ESD closure time which is at pumps discharge?

I am using the below equation:
P = (0.070) (V) (L) / t + P1

V-wave velocity in f/s, L-length in ft, t-time in sec, P1-operating pressure


Bulk Mod Kb 1.25E+09 N/m2 Density 840 Kg/M3 Flow Rate 2250000 BLD 14916 m3/h Pipe DIA 54 IN 1.4 m Velocity of sound-Vc 1218 m/s 3997.3 f/s Line Velocity 2.8 m/s 9.2 f/s Valve colsure time 1500.0 s 1500.0 s Change of velocity 0.002 m/s2 0.006 ft/s2 Length of line 150,000 m 492,126 ft Pressure inlet 70.07 bar 1030.0 psi MAIP 84.44 bar 1241.29 psi

[ankur2061]

Raj,

I was not inclined to come back to this post because information was provided by kkala without any backup but since you have come back to this post I will respond.

The first post by me after your original post, gives the calculation equation for surge pressure. Go back and look at the equation and let me know if you require any clarifications.

You need not have a valve which is closing very slowly for the entire travel of the valve. Automatic valves can be provided a 2-speed actuator which will allow closure of the valve faster over the first 70-80% of the valve travel and then slow closure of the valve over the last 20-30% of the valve travel which is critical for the build-up of a surge pressure wave. This is as per guidelines of "Petronas".

Regards,
Ankur.

[Christiano]

Ankur,
Thanks for responding back, I understand your views as you have made it very clear in the posts.

WRT my post the calculation formula provided by you doesnt account for valve closure time, it bacsically predicts the max surge due to an immediate closure.

I guess the issue is because of the length of the lines, the farthest block valve is 150 KM long from pump station.

Regards,

Rajesh

[ankur2061]

Raj,

If you have access to Perry's Handbook, 8th Edition, the water hammer and time for valve closure is provided under the heading "Hydraulic Transients", page 6-44, Section 6- Fluid & Particle Dynamics. Refer equation 6-205 which provides the time required for the pressure surge wave to travel from the valve to the source and back. If L is this distance then the total length traversed by the wave will be 2L.

Regards,
Ankur.

[Rnair]

Ankur,

How do you find placing a check valve between block valve stations as my wave travel time and thereby min valve closure time becomes smaller?

[ankur2061]

Rnair,

I have no idea about placing check valves between block stations and have never encountered this before. Nevertheless, I can immediately foresee two problems with check valves:

1. Difficulty in deinventorying the pipline during a shutdown, since you will have to separately deinventory the upstream and downstream of the check valve.

2. Pigging of the line may become difficult with a check valve unless a special pigging device is used which can pass through the check valve which I am not aware of.

Regards,
Ankur.

[kkala]

A. While doing a preliminary check on the design pressure of the line, attempting to limit the MAIP (Max Impulsive pr) to 110% of design pressure I found the closure time is very high for the farthest block valve. I have block valves every 25 kms as a prelim guideline and the farthest is 150 kms from pump station. Now when I calculate the min closure time to limit the surge to 110% DP, I require about 1500 seconds to safely limit the MAIP to the above value.Could you please Comment on this?
IN addition to the above, could you comment if any specic considerations are there for ESD closure time which is at pumps discharge?
B. I am using the below equation: P = (0.070) (V) (L) / t + P1
V-wave velocity in f/s, L-length in ft, t-time in sec, P1-operating pressure
Γ. Bulk Mod =1.25E+09 N/m2, Density=840 Kg/M3, Flow Rate=2250000 BLD =14916 m3/h, Pipe DIA= 54 IN = 1.4 m, Velocity of sound-Vc=1218 m/s=3997.3 f/s, Line Velocity=2.8 m/s=9.2 f/s, Valve clοsure time=1500 s, Change of velocity 0.002 m/s2 = 0.006 ft/s2 (??), Length of line 150000 m = 492126 ft, Pressure inlet 70.07 bar=1030.0 psi, MAIP 84.44 bar=1241.29 psι.
Note: kkala has made insignificant modifications to facilitate response.

B. Rough check of equation used. This is for water, http://www.plastomat...er-hammer.html .
B1. Let us consider example 10 of Perry, 7th ed - 1997, p. 6-44, Hydraulic transients, additionally assuming a pipe length of 1020 m (not specified in the example). Perry's data: water, 4" pipe sch40, 2.0 m/s, pipe elastic modulus given, water bulk modulus given.
B2.1 Developed surge pressure per Perry: SP=27.3 bar, for closure time t between 0 and 1.5 s. Rough approximation for t = 6 s is SP=6.8 bar (if we accept Perry's formula 6-206).
B2.2 Developed surge pressure per equation used: SP=infinite for t=0 s, SP=70.6 bar for t=1.5 s, SP=17.7 s for t=6 s.
B.3 Conclusion: A more sophisticated formula / software has to be tried for the case. Known software (Hammer V8?) can be the most convenient way out. In case of no available software, application of rigid water column theory could give satisfactory results for quite large t (as in the case). I had met this theory in "Waterhammer Analysis" by John Parmakian, Prentice -Hall, 1955, Chapter I, but it looks difficult in understanding. A modern book of hydraulics may express it clearer.
A. Using Perry's formulas as a rough guess based on your data for the 150 km line: SP=840*1218*2.8 Pa = 28.6 bar for t=0 to 246s; SP ~7 bar for t=1000 s (but better to estimate SP for t=1000 s otherwise, not per Perry 7th edition). At any case closure time t for SP=7 bar seems too long, so a compromise has to be made by installing protection against surge. See post No 8 by kkala (17 Dec 11), example of Saudi Aramco crude pipeline, http://www.cheresour...surge-pressure/. A reliable software is needed to estimate surge pressures as precisely as possible.
The size and complexity of 1400 km pipeline can justify surge study by a specialized firm. This is supposed to be done here for connecting pipes between refinery and pier, distance about 2 km.

Note 1: Check valves at the beginning of discharge lines (apparently closer to pump compared to brach for pigging) protect pump in power failure, if flow tends to revert.
Note 2: Attachement is some water hammer software (with animation), judged good for "education" to those interested. It was in the web, but not found now.

Attached Files

•  hammer.zip   374.46KB   189 downloads

[Rnair]

Thanks KKala for your generous efforts to work the problem and sugesting alternatives. I shall further study the system while doing the transient analysis in Hammer V8 once the block valve stations locations are finalised and shall suggest surge relief systems at farthest block valve stations where closing time exceeds 300 sec.

[kkala]

I wish success to this task of transient analysis and consequential protection. In this case software use is the proper way out.
Being in same room with the team running TLnet for the 20 km crude line (but not participating in it), I understood that software requires even local elevations of pipeline, as well as pipe roughness. On the other hand results were detailed, indicating not only overpressures but also low pressures to occur in surge (below liquid vapor pressure, causing cavitation). If I remember well, type of closing valve (e.g. butterfly, globe, etc) was also taken into consideration (which happens even in the attached "hammer.zip" of previous post by kkala).

Note: I had also heard of "surge tank" to lower overpressure, but my knowledge on it is vague; probably flow during occurring overpressure got free to enter an atmospheric tank (through PCV?, PSV? else?). Local process Dept did not undertake the responsibility to specify protective systems, only to specify surge pressures.