28 replies to this topic

[Afshin445]

Dear All,

In many design criteria and practice I found two below formula for estimation of Shut-off pressure of Motor Driven Centrifugal pumps in absence of any Perfoemance Curves:

Shut-off pressure = 1.25 x Diff. Pump Pressure + Design Pressure of Upstram Vessel + Static Press. (From HLL).

Shut-off pressure = 1.25 x Diff. Pump Pressure + Static Press. (From HLL).

I want to know which one of above formula is correct. In this regards, I want to know how those formula is developed.

I appriciate if someone advice me in this regards.

Regards,

Edited by Afshin, 13 January 2010 - 07:05 AM.

[shan]

The first one is applied to the suction drum with pressure and the second one is for suction drum without pressure (such as tank).

[ankur2061]

Afshin,

I am reproducing what Shell "Design Engineering Practices (DEP) has to say about Shut-Off pressure (SOP):

The shut-off pressure is the pressure at the discharge of a centrifugal pump with the suction pressure at the maximum possible value and the discharge system closed.

The shut-off pressure SOP is determined by the equation:

SOP = SP + HPS + HPO - HPD, in which

- SP = set pressure of the pressure relief device on the pump suction system (Notes 1 and 2)
- HPS = hydrostatic pressure of the liquid above the pump suction (Note 2)
- HPO = pump differential pressure at no flow and maximum pump speed and highest density as per process design (Note 3)
- HPD = hydrostatic pressure of the liquid above the pump discharge

NOTES:
1. In the exceptional case of the shut-off pressure condition and the relief condition having a common cause, the relief pressure shall be taken instead of the set pressure.

2. Consideration shall be given to the designing of systems which are not liquid-full but could be filled completely within a short period, due to low vapour volume or high pump rates, so as to withstand pump shut-off pressure. Each case shall be studied individually and the Principal's agreement is required. As a guide, operator action can be assumed within 30 minutes after an appropriate reliable warning that overfilling is taking place, see API RP 521.

3. For steam turbine driven pumps and variable speed electric motor driven pumps, the maximum possible speed at start-up and during operation of the pump shall be carefully assessed since this can influence the design pressure considerably. Differences of more than 25% between differential head at normal running speed and maximum running speed are not unusual.

Shut-off Pressure is normally always something that the pump vendor provides as is obvious from the pump curve of Q vesus H. So it is not an absolute must to specify shut-off pressure & if possible it should be left to the pump manufacturer to specify this value. In cases where it becomes unavoidable & some value needs to be provided I have used the above equation for specifying the SOP. The SP in the above equation becomes equal to atmospheric pressure (1.013 barA or 14.7 psia) if the suction system is open to the atmosphere.

The notes accomapanying the equation are a bit tedious to understand but the equation is very straight forward.

Hope this helps.

Regards,
Ankur.

[fallah]

SOP = SP + HPS + HPO - HPD, in which

- SP = set pressure of the pressure relief device on the pump suction system (Notes 1 and 2)
- HPS = hydrostatic pressure of the liquid above the pump suction (Note 2)
- HPO = pump differential pressure at no flow and maximum pump speed and highest density as per process design (Note 3)
- HPD = hydrostatic pressure of the liquid above the pump discharge

How can you evaluate HPO value without having pump curve in hand?

[Art Montemayor]

I agree with fallah’s basic question. The HPO cited is basically the same value (& definition) as that for Shut-off (or “dead head”) pressure of a centrifugal pump.

However, in all honesty, I can also understand Ankur’s response to an over-simplified request. I believe that Ankur has assumed that the OP has an operating centrifugal pump with an installed PSV directly on the discharge of the pump. That may – or may not – be the case. What the Shell engineering practices formula infers is that there is a relationship between the safest highest pressure attainable in a centrifugal pump and the Shut-off (or “dead head”) pressure for that pump. This is true under some circumstances – but also may not be true for the involved installation. It depends.

If the OP is dealing with a proposed, new pump then the standard (and most accurate) answer is the use of the manufacturer’s performance curve for that pump and the installed impeller at the specified speed. However, if you don’t have a performance curve for an installed pump and you simply don’t want to “rev-up” the subject pump and shut off the discharge pressure to find out the empirical Shut-off (or “dead head”) pressure, then you are left with only the estimated method Shell proposes (or something like it).

Allow me to point out some basic field facts. A centrifugal pump doesn’t necessarily have to have a PSV on its discharge (or its suction). In the greatest majority of all the installations that I have made in the past 50 years, I have rarely employed a PSV in the centrifugal pump’s discharge. I simply design the pumping system to withstand the Shut-off pressure and be done with it. It is mechanically impossible for a centrifugal pump to surpass its Shut-off pressure. For many, many years during the last century there was no published data on a centrifugal pump's MAWP. I know this sounds ridiculous, but so benign were the applications then, that this valuable information was rarely published. You could obtain it, but you had to make a special request from the pump manufacturer. I don't try to minimize the importance and seriousness of knowing your centrifugal pump's MAWP and potential need for PSV protection. I merely mention an illustration of this simple deletion of a PSV as a solution - if the pumping system were designed to withstand the shut-off pressure.

I don’t know where Afshin got the equations he refers to. Therefore, I certainly can't reveal how thery were developed. If anyone else cannot identify where these equations come from, I seriously doubt that they can either. I believe those equations do not apply to anything resembling the shutoff pressure, and therefore, I would not recommend he use those equations.

[breizh]

Afshin ,

Let you consider the Norsok standard attached :
http://www.standard..../P-CR-001r1.pdf

hope it helps
Breizh

[fallah]

SOP = SP + HPS + HPO - HPD, in which

- SP = set pressure of the pressure relief device on the pump suction system (Notes 1 and 2)
- HPS = hydrostatic pressure of the liquid above the pump suction (Note 2)
- HPO = pump differential pressure at no flow and maximum pump speed and highest density as per process design (Note 3)
- HPD = hydrostatic pressure of the liquid above the pump discharge

It should be noted,SOP as per above would be in point of pump discharge line where being blocked.Thus,for having SOP just in pump discharge HDP should be removed from above equation.

[Afshin445]

Dear All,

Thank you for your responces. As I understand from Shell practice, for estimation
of shut-off pressure I must add increase in pressure by pump in no flow condition to
set pressure of pressure relief device in upstream vessel and Static pressure from HLL.

Something I can understand is in which scenario upstream vessel pressure
rich to it's set pressure and vessel liquid level is in higest point and also pump discharge line is
blocked. Is this scenario is really happen for pump or this is only a conservative
approach for pump Shut-off pressure estimation?

Regards
Afshin

Edited by Afshin, 17 January 2010 - 05:10 AM.

[fallah]

Something I can understand is in which scenario upstream vessel pressure
rich to it's set pressure and vessel liquid level is in higest point and also pump discharge line is
blocked. Is this scenario is really happen for pump or this is only a conservative
approach for pump Shut-off pressure estimation?

If adequate protections haven't been provided for stopping the flow into upstream vessel during discharge line blockage (e.g. upstream SDV closing and pump shot-off on HHA,...) and supposing the operator don't perform any action on HHA (usually in 5 minutes duration),mentioned vessel take its upstream pressure and may reach to its design pressure (set point).

[Afshin445]

In scenario you explained, discharge line of pump is blocked and it's mean we don't have liquid stream outgoing from vessel. If the flow upstream of vessel still running, in case of pump outlet line blockage the vessel level reach to it's HLL but the vaopur side of vessel is still open to downstream and how vessel reach to it's design pressure when we don't have vapour accumulation in vessel?

Edited by Afshin, 17 January 2010 - 06:32 AM.

[fallah]

In scenario you explained, discharge line of pump is blocked and it's mean we don't have liquid stream outgoing from vessel. If the flow upstream of vessel still running, in case of pump outlet line blockage the vessel level reach to it's HLL but the vaopur side of vessel is still open to downstream and how vessel reach to it's design pressure when we don't have vapour accumulation in vessel?

As i said in my previous post,when operator don't perform any action on HHA and no SDV being at the inlet to stop the flow,if the outlet other than pump suction being existed on the vessel,mentioned vessel take its upstream pressure and may reach to its design pressure (set point).

Anyway,if you are interested to continue the discussion on the probability of the scenario,it is better that has been done on a specified sketch/PFD/....

What did you mean when you said:
"...in case of pump outlet line blockage the vessel level reach to it's HLL but the vaopur side of vessel is still open to downstream..."

[kkala]

I found two below formula for estimation of Shut-off pressure of Motor Driven Centrifugal pumps in absence of any Performance Curves:
Shut-off pressure = 1.25 x Diff. Pump Pressure + Design Pressure of Upstream Vessel + Static Press. (From HLL).
Shut-off pressure = 1.25 x Diff. Pump Pressure + Static Press. (From HLL).
I want to know which one of above formula is correct. In this regard, I want to know how those formula is developed.

Having read the contributions to the subject, useful indeed, I would like to add following views.
1. Centrifugal pump shutoff (shut-in per Norsok) head falls below 1.2-1.25 x operating head at the selected operating point. Of course the latter should not be far from best efficiency point. This rule is statistical, has exceptions, but has been verified for most curves of refinery pumps. We have to use such a rule in a new project, seeing that piping design proceeds in advance of pump orders / data.
2. Max pump shutoff pressure comes directly from this rule and the above formulae compute it, as explained by shan. Actually we can consider the first formula only, since Design Gauge Pressure=0 for an atmospheric tank.
Datum is suction centerline, so static pressure shall consider the perpendicular distance between it and HLL (HHLL can be used instead, being more conservative).
3. Let us try to feel discussed "Design Pressure of Upstream Vessel" by example: Forced circulation boiler, operating during safety valves blowing off at steam drum (I think this is requirement by code, but not certain). Max shutoff pressure of circulation pump includes design pressure of steam drum (this is operating pressure in the example)+ max static pressure of water to pump centerline+pump differential shutoff pressure.
4. We can add the requirement "shutoff head = 1.25 operating head, or lower" in our pump data sheet. For refinery pumps transferring hydrocarbons, the factor 1.25 can be even a bit lower. Some pumps may not be able to realize the requirement, and this will be assessed during bid evaluation.
This is usually preferable to no shutoff requirement, which would increase checking labor, time and design changes.
5. Shell formula (brought by Ankur) is understood as SOP=SP+HPO+HPS, only if HPS represents the static pressure due to height difference between suction relief device and suction centerline; SP actually representing design pressure of Upstream vessel. The negative HPD is not understood, but at any case it is a rather small quantity and neglecting it would be on the conservative side.
One can only imagine that if the closed discharge valve lies at a height creating HPD static pressure in respect to suction centerline, pressure on this valve would be given by Shell formula.
5. I wonder why the flat pump curve is preferable, while less flat curves would give more satisfactory control, e.g. combined with a control valve downstream of the pump. Probably this matter (not known to me) is related to stable pump operation in a wide range of flows beyond its best efficiency point.

[Afshin445]

[quote name='fallah' date='17 January 2010 - 07:20 AM' timestamp='1263731428' post='35578']
[quote name='Afshin' date='17 January 2010 - 06:16 AM' timestamp='1263727564' post='35576']
As i said in my previous post,when operator don't perform any action on HHA and no SDV being at the inlet to stop the flow,if the outlet other than pump suction being existed on the vessel,mentioned vessel take its upstream pressure and may reach to its design pressure (set point).

Anyway,if you are interested to continue the discussion on the probability of the scenario,it is better that has been done on a specified sketch/PFD/....

What did you mean when you said:
"...in case of pump outlet line blockage the vessel level reach to it's HLL but the vaopur side of vessel is still open to downstream..."
[/quote]
Two scenario is probable:
1- Discharge line or valve in discharge line of pump is closed.
In this case vessel is in operating pressure,because of liquid side block , level in vessel reach to HLL and pump DP is in no flow condition (Can estimated as 1.2-1.25 DP in normal condition).

2- Upstream vessel is in it's relief condition (Design Pressure), liquid level in
vessel is in normal level and pump DP is normal.

Norsok standard is compatabile with two a.m. scenario, but I have problem with Shell
DEP, because Shell considered Vessel is in it's Design Pressure and Pump DP is in no flow condition. How this two possible in one scenario?

Edited by Afshin, 18 January 2010 - 03:38 AM.

[fallah]

Shell considered Vessel is in it's Design Pressure and Pump DP is in no flow condition. How this two possible in one scenario?

Pump DP is in no flow condition,no outlet other than pump suction line in relevant vessel,leading to pressurization of vessel by upstream pressure (design pressure).

[Afshin445]

How vessel pressure increase when vapour stream is open to downstream? In fisrt scenario we assume only discharge line of pump is closed and there is no blockage in vapour line.

Edited by Afshin, 18 January 2010 - 04:52 AM.

[fallah]

How vessel pressure increase when vapour stream is open to downstream? In fisrt scenario we assume only discharge line of pump is closed and there is no blockage in vapour line.

If you look at my 3rd post you will see my statement as:

"Anyway,if you are interested to continue the discussion on the probability of the scenario,it is better that has been done on a specified sketch/PFD/...."

Without any sketch/PFD/... i would prefer to consider worst situation,in this case surge drum (without any overhead line),and may be SHELL also done the same.

Now,after many posts to be submitted you consider separator and this is another story respect to what i did consider.

Anyway,possible scenario would be estimated with having adequate document/information about the case in hand.

[Afshin445]

How vessel pressure increase when vapour stream is open to downstream? In fisrt scenario we assume only discharge line of pump is closed and there is no blockage in vapour line.

If you look at my 3rd post you will see my statement as:

"Anyway,if you are interested to continue the discussion on the probability of the scenario,it is better that has been done on a specified sketch/PFD/...."

Without any sketch/PFD/... I would prefer to consider worst situation, in this case surge drum (without any overhead line), and maybe SHELL also done the same.

Now, after many posts have beeen submitted you consider a separator and this is another story with respect to what I did consider.

Anyway, possible scenario would be estimated with having adequate document/information about the case in hand.

I refer you to my prior post as bleow:

"...in case of pump outlet line blockage the vessel level reach to it's HLL but the vaopur side of vessel is still open to downstream and how vessel reach to it's design pressure when we don't have vapour accumulation in vessel?"

And I think when we talking about vapor and HLL that it means we are talking about a separator - not a Full Liquid Surge Drum.
In this regards, I have a problem attacheing a file here but I am talking about a typical Pump downstrem of the liquid side of a separator and I think the formula we are talking about is valid for pump independent of type of upstream vessel.


Thanks

[fallah]

Anyway generally ,i think the case in which pump DP blockage and reaching pressure of its suction drum to design pressure may simultaneously would be happened is given as worst case scenario.

Edited by Art Montemayor, 23 January 2010 - 10:51 AM.

[akslzf]

The basic question is "How to estimate the Shut-off pressure of a Centrifugal pumps in absence of any Performance Curves".

"1.25 x Diff. Pump Pressure ++++" is not a wise estimation, as the shape of the Q-H curve depends very much on the geometry of the impeller.

Getting back to the basics, a centrifugal pump converts kinetic energy into pressure energy and the kinetic energy is obtained by the rotating impeller.

For illustration purpose, Let us consider a 0.265m dia impeller revolving at 60 rps.
The tip speed of the impeller (v) is 50 m/s. To convert kinetic energy into pressure head, the fundamental formula is p = v^2/(2*g)=> 50*50/2*9.81 ~= 125m
If we assume the fluid to be water, ~10m of water = 1 kscG. Therfore, 125m of water = 12.5 kscG.
If the pressure at the suction flange of the pump is, say, 1 kscG, SO pressure of the pump will be 13.5 kscG, theoritically.

[kkala]

"1.25 x Diff. Pump Pressure ++++" is not a wise estimation, as the shape of the Q-H curve depends very much on the geometry of the impeller...
For illustration purpose, Let us consider a 0.265m dia impeller revolving at 60 rps.
The tip speed of the impeller (v) is 50 m/s. To convert kinetic energy into pressure head, the fundamental formula is p = v^2/(2*g)=> 50*50/2*9.81 ~= 125m
If we assume the fluid to be water, ~10m of water = 1 kscG. Therfore, 125m of water = 12.5 kscG.
If the pressure at the suction flange of the pump is, say, 1 kscG, SO pressure of the pump will be 13.5 kscG, theoritically.

Interesting example of estimating shutoff, in the ideal case where exit fluid speed = impeller periphery speed (π*f*d, f=60 rps, d=0.265 m). Even for closed impeller, pump curve Q-H must give a more precise figure. The example accidentally verifies the rule, but if f=30 rps (1800 RPM) shut off pressure would be calculated at 4.2 kscG (v=25 m/s).
However the empirical rule of shutoff is statistically valid for refinery process pumps, although there are exceptions. As said previously in the thread, it is very useful in new Projects. Piping design goes ahead of pump data from suppliers so something has to be assumed.
I have read somewhere that turbine pumps have a coefficient grater than 1.2 . So I assume that the rule is statistically valid only for process & feed pumps (Ns<2000), not for turbine (Ns=2000-5000), mixed flow (Ns=5000-10000), axial flow (Ns=9000-15000) pumps. Last data is taken from Perry. It would be useful if somebody could advise on this assumption.

[Afshin445]

From fisrt my main question was that. How this two case simultaneously can happened in one scenario?

Is really happend or only is conservative approach for estimation of pump shut-off pressure?

Edited by Art Montemayor, 23 January 2010 - 10:52 AM.

[kkala]

From fisrt my main question was that. How this two cases simultaneously can happened in one scenario?
Is really happend or only is conservative approach for estimation of pump shut-off pressure?

If you mean the two cases represented through the two formulae mentioned in your first post, answer has been given by shan. The first formula includes the case of second formula (see also post by kkala, para 2 on 13 Jan 2010).
If you consider the two cases of suction design pressure and pump shutoff, answer by fallah covers the subject (see also post by kkala, para 3 on 13 Jan 2010, example of a forced recirculation boiler).
It seems that you find it difficult to assume that suction design pressure, max suction level and pump shutoff can occur simultaneously, as indicated from your two scenaria of your post on 18 Jan 2010. These have reminded me of similar engineering practices of 1985-1990, recommending design pressure of discharge pipeline more or less according to these two scenaria. Formal "single contingency" rule might justify these scenaria, however these are not judged safe enough in my opinion.
For instance, consider scenario No 2, where upstream vessel is at relieving state and pump is operating. In the panic created, somebody might close a discharge valve inadvertently. Then upstream vessel can soon reach HHLL. So upstream relieving, max suction level, pump shutoff should be faced as single contingency. Or perhaps not? I have seen such simultaneous contingencies in the 6 (lucky) years of my work in operations, and they were not so rare as to be non credible.
According to API spirit, exceptions to rules are permitted by using sound engineering judgement; however it would be very hard to convince an inspector today that the mentioned senaria have resulted in a reliable design pressure; simply because the rule defining the discharge design pressure as "relieving pressure at suction + max static pressure (from PRV) + pump shutoff differential pressure" has been widely accepted.
We also apply the first rule written by Afsin at work, giving more conservative results that the senaria. Cost increase, if any, is not judged worth wile before the safety gain.

Edited by kkala, 23 January 2010 - 04:07 PM.

[Afshin445]

Dear kkala,

Thank you for your complete reply.

As I understand from your post is in 1985-1990 yeras, in calculation of Pump Shut-off pressure you considered "single contingency rule" and only one of two probable scenario considered and whichever gave you more pressure has been selected by designer, similar Norsok standard,presented in link in post by brizh dated january,14,10. And in recent years we use more conservative approach, similar engineering practice presented by Shell DEP posted by ankur, dated January 13,10.

Is my imagination in your intention is correct?

Thanks.
Afshin

Edited by Art Montemayor, 24 January 2010 - 09:03 AM.

[kkala]

Dear Afsin

Yes, you are correct, our understanding (in our previous posts of Jan 23rd & Jan 24th) is generally the same on the subject and I am glad to be of some help.

Following clarifying notes in addition may be useful.

1. Our target is to estimate discharge line design pressure (scenario No 2 of Jan 18 does not represent shutoff). Today more conservative practices (compared to max shutoff pressure) can be adopted, e.g. 5% margin over max shutoff pressure (discussed in Dec 09 in the forums). Water hammer plays also an important role in long pipelines, e.g. design pressure was 2.5 x max shutoff pressure in a ~30 km line. I have read in WWW (but not verified) that ANSI incorporates any surge pressure into design pressure, while other codes permit excess of pressure due to surge up to 10% over the design pressure.
2. I saw in 1990 a written rule for discharge (design) pressure more or less complying with the two scenaria of 18 Jan 2010. The rule did not explain the supporting scenaria, only said "do so and so" (details not remembered). Instead of it, I applied the concept of Shell brought by Ankur2061 (without HPD), since the latter was simpler and safer.
I had a brief look at Norsok standard posted by breitz (14 Jan 10), but could not trace a relation to mentioned scenaria (shut-in pressure found in design pressure chapter).
3. Contingency for "modern" shutoff conditions (per Shell concept) may be between single and double (post by kkala, 23 Jan 10). E.g. for scenario No 2, relieving condition at suction may cause inadvertent closure of discharge valve; the two contingencies do not seem independent. There are many similar cases in Operations, one had better be conservative.
4. Today same design pressure is specified for suction and discharge pump line, see ChE Plus forum, Pump suction piping rating (25 Dec 09) by M Orojlu. This could be a bit confusing. It might be proper to consider upstream vessel PRV set pressure + max static pressure at suction, then pump shutoff head, then discharge line design pressure. After that, set suction design pressure = discharge design pressure (I have never worked on such a case though).

Edited by kkala, 24 January 2010 - 11:55 AM.

[kkala]

To continue subject of yesterday, I found a practice of 1970s to estimate piping design pressure downstream of centrifugal pump as the highest of :
α. Suction pressure at normal conditions plus 120% of pump differential pressure.
β. Max suction pressure plus pump differential pressure.
Case (α) is close to Scenario 1 and case (β) close to Scenario 2, as presented by Afshin's post, 18 Jan 2010. It is understood that there was no PRV on the pipe.

Of course safer practices are generally applied today, as indicated in previous posts.

Edited by kkala, 25 January 2010 - 01:00 PM.