ESDV Closure Times to Prevent Surge

[ProcessRookie]

Hello all,

I'm currently completing a study whereby I'm looking to rule out the possibility of surge occurring in certain systems. The valves in question are ESDVs on large bore lines (24") and I have the data sheets which specify achieved closure times of 25s. The pressure originates from a set of high pressure pumps upstream which are designed to stop on high pressure.

To me this looks like a no brainer as I can't imagine a pumped system being at risk of surge with valve closure times of 25s, however, I am looking for a rule of thumb to use in my explanation. Are there any rules of thumb for screening for surge issues based maybe on closure time, flowrate and line size/valve trim size for instance? It is not necessary for this work to do any Joukowski or other such surge calculations all I require is a quick rule of thumb.

Thanks in Advance

 

[dicksewerrat]

Run the calc for surge. then look at the downstream side of the valves for a vacuum issue.

Richard A. Cornelius, P.E.

http://WWW.amlinereast.com

 

[LittleInch]

The key is actual velocity. Whilst I would hesitate to use any of this to discount the possibility of a surge event taking place, anything below 2-3 m/sec rarely exhibits significant surge, but it all depends on yur system.

The other key factor is length of pipeline between the valves and the pumps - "upstream" could be 10m or 100km. Again anything much past about 20km is often far enough to attenuate any surge, but it is difficult to judge with no real information.

Pump run down time, how much pressure you have between pump shut off head and design pressure, the fluid being pumped, it's bulk modulus all have an impact on surge pressures.

given that most ESd valves are ball valves and that most of the closure effect is in the last 20% (5 seconds) I can easily imagine a very significant surge from this, but with no data provided it's impossible to comment further.

My motto: Learn something new every day

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

 

[BigInch]

What LittleInch said. Over 3m/s can often mean the difference between significant surges or not, but is highly system dependent. A water system can be far more critical than a gasoline pipeline, because of the water's 30% more mass, etc., and velocity head increases as the square of the velocity, so tread easy over 3m/s. 6m/s+ and you've almost always have something requiring extended analysis.

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

 

[e43u8]

ProcessRookie,

A rule of thumb is normally used and in most cases will lead to an acceptable rersult is: "one second per one inch valve size"; then with lack of adequate info and as a quick estimation if the valve size is same as the line, 25 seconds closing time for a 24" valve size could be reasonable...

 

[BigInch]

I've always thought that 1in/s is a normal close rate and always used <= 15s for ESDs.

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

 

[ProcessRookie]

Thank you all very much for your replies. BigInch, do you mean to say that the rule of thumb is closure times which are faster than the values you have quoted are at risk of surge? if so, something like this is all I need, if you have a reference or somewhere I can do some reading. My lead engineer has already agreed on my approach I just need a reference to back up what I say. FYI guys pumps operate at 90 barg and are situated about 30m downstream of ball valve ESD. Fluid is LPG, specified closure time is 20s and the pipework is 900 lb/sqinch. There is no way I'm doing any detailed calculations it is outwith the scope, and vacuum upstream is not an issue it a screening exercise for vibrational fatigue due to surge using the EI guideline.

Best Regards,

PR

 

[ProcessRookie]

Additionally, flowrate is not so high. Velocity is less than 3m/s.

 

[LittleInch]

Can you get your story straight. Are the pumps upstream or downstream of these valves? Why is vacuum not an issue, bearing in mind that "vacuum" for LPG is actually between 7 and 15 bar depending on what type of LPG you are transporting.

Unfortunately your idea that you can find "I just need a reference to back up what I say" is not going to be easy as each system needs to be looked at on its own and the consequences of getting it wrong are too high for anyone to say in general this isn't a problem. How your pumps start stop and how your system is configured and designed all has an impact. If the pumps keep running or take a long time to slow down then you could easily get surge issues, whether form high pressure or having the LPG vaporise on you then collapse again - Now that can generate some big spiking pressures...

BTW do you mean ANSI class 900 or 900 psi rated? This appears to have no relationship to the discharge pressure of the pumps.

Can you describe a bit more what this study is for "screening exercise for vibrational fatigue due to surge using the EI guideline" I have difficulty believing that surge events contribute any significant part of a vibration study.

My motto: Learn something new every day

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

 

[BigInch]

Don't remember exactly where. Probably SaudiAramco. But try looking in RP-14C, or API 521.

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

 

[SNORGY]

Either way, certainly part of the ESD philosophy would (should?) be to de-energize or ramp down the pumps to full stop prior to completely closing the ESDV, particularly if the ESDV is on the suction (upstream) side.

As for the rule of thumb, I would think that if you keep the closure time somewhat longer than the time it takes a pressure wave to travel through the fluid twice the length of the pipe involved, the closure is no longer "rapid" and the surge or hammer is quantifiable but often tolerable.

 

[ProcessRookie]

LittleInch, they are upstream of the ESDV, sorry about the typo however it was nice of you to dig me up about that rather than be sensible and give constructive feedback. You have obviously never carried out any high level consultancy type engineering analysis or any vibrational analysis therefore I don't think you can contribute much to this post, thanks anyway.

BigInch, thank you, I'll have a read of those sources, much appreciated as always.

 

[LittleInch]

Well it does make a bit of a difference...(U/S or D/S) How exactly do you expect people to respond if you're
a) drip feeding information,
b) making errors,
c) accusing people of not knowing what they're talking about without any information
d) clearly looking for free information to complete a paid for task
e) Not reading posts properly or answering perfectly reasonable questions

My point, not that you seem to recognise it, was that if you're worried about vibrational analysis from what should be a very rare event (ESD trips are not normal) then it seemed unlikely to contribute to what is normally a long term issue, but instead of trying to address the point you simply dismiss the notion as not worth replying to or providing any further information.

If your ESD trip is guaranteed to stop the pumps and I mean stop rotating, before the valves have got even half way closed you might be correct, but without looking at the whole of the system and not simply a 20 m section of pipe from pumps to the valves I don't believe it is possible to simply rule out the possibility of surge without doing at least some basic transient runs. That's my opinion - take it or leave it, but please don't insult my knowledge or experience or anyone else who bothers to reply - it won't get you many friends on this site.

My motto: Learn something new every day

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

 

[BigInch]

ESDs can be with pumps on or off. Usually (especially according to Engineer Murphy) pumps are on when the ESD trip occurs, then the pump shut down switch is triggered. If pumps are off, then theoretically the pump, or pump station is already isolated. ESDs should "isolate the pump station from the pipeline" (pipeline codes), so there will (must) always be one on suction and one on discharge.

Closure time on a pipeline of any appreciable length is almost always shorter than the 2x sonic travel time down a pipeline to the initiating valve, although sometimes you do find remote controlled, relatively fast acting valves at intermediate locations. Lesser reflections are also bounced back from reducers, bends, control valves, orifices and other incomplete flow obstructions (which are proportional to the reduction of flow area). As well as the worst problem typically comes on surges created between pump and the ESD of a discharge isolation valve. Those distances are extremely short (at sonic velocities) so there are usually 4 or 5 wave reflections generated in quick sequence before most of the energy is dissipated. When a valve closes there is a positive pressure generated on the upstream side and a simultaneous negative pressure wave initiated on the downstream side. Each reflection is (very roughly) 1/2 the energy of the previous wave in most engineering systems, but it does depend on energy dissipation due to pipe friction effects at the surge velocities.

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

 

[MortenA]

@ProcessRookie Quote: There are no such "rules of thumb" or quick fix. Just tell you management thats "its a no-brainer"