Determine NPSHa of centrifugal pump 2

[kkav]

Hi Everybody,
There is a pump which transfer liquid from bottom of a distillation tower. In order to specify NPSHa,it is needed to determine liquid head (level) on the bottom of tower. For this, there are some options as below:
1)Normal liquid level (50% on level transmitter)
2)Liquid level on the lower tapping of level transmitter
3)liquid level on the lower tapping of level gauge
4)liquid lavel on low level alarm
5)liquid level on bottom tangant line
Which option to be considered? Are there a standard that specified it?
Thanks

 

[ione]

If I were you I’d make reference to the most binding scenario, that is the one which foresees the lowest point (bottom tangent line).

Take also a glance at the link below

http://www.pumps.org/content_detail.aspx?id=2128

 

[cheute79]

The minimum NPSH required (by the pump) might define the LLA. Else why the LLA?

 

[MJCronin]

Cheute79 is correct..... The most probable reason for a LLA on a distilation tower is pump protection

 

[chemebabak]

KKAV's question is "Which level should be used as a basis for NPSHA calculations?"

There is no definitive standard, however good engineering practice dictates that the Low Low Level Alarm (plant/pump shutdown alarm) is the highest level to be reasonably used for NPSHA. Many engineers prefer the bottom tangent line. Some are overly conservative and select the bottom flange.

For best practice I recommend using the BTL. The liquid level could be maintained slightly above the LLLL before a trip. Best is to add ~3ft safety margin on top of any NPSH required or deduct 3 ft from the calculated NPSH available.

 

[PetersPhotography]

I agree with chemebabak - the bottom tangent line is what I'd recommend using. However, the 3' safety margin for NPSH seems excessive.

For example, a 6' high bottom tangent line would give you, at most, 6' of NPSHa for a bubble point fluid. Say the centerline of your pump is at 12" (fairly low, considering the steel mount), and now you're left with at most, 5' of NPSHa. There is a potential for appreciable pressure drop in the upstream strainer - a ft. of head is common for a Y-Strainer - so now you're at 4' of NPSHa, without accounting for line and fitting losses (which should be almost negligible for a bubble point fluid).

4' of NPSHa is do-able with a horizontal or vertical in-line pump, depending of course on differential head and flow. However, a 3' safety margin would push us to 1' of NPSHa, and in this range, a can pump is necessary. Can pumps are a pain to maintain and cost significantly more than vertical and horizontal in-line pumps.

Am I making any incorrect assumptions in this analysis? I'm new to the field (a very recent grad) and I'm just trying to learn.

 

[chemebabak]

PetersPhotography asked if his NPSHA analysis was correct. Please review the following.

NPSHA is calculated this way:
NPSHA =
Absolute pressure in vessel +
static head -
vapor pressure of liquid -
friction loss through pipe.
All pressure units must be converted to ft or m.

A vessel which has a BTL 6 ft above the pump centerline (3 ft above grade is typical) does not have 6 ft of NPSHA. It will have 6 ft + (Absolute pressure in the vessel - vapor pressure of the liquid at pumping temperature - friction loss through pipes and fittings) * 144 in2/ft2 / density, lb/ft3, at pumping temperature. Let us say that the pressure in the tank is 60 psig and the fluid is water at 300°F. Also let us say that there is 2 psi loss through the pipes and fittings. At 300°F, the vapor pressure is 67 psia and the density is 57.3 lb/ft3.
NPSHA =
6 ft +
(60 psig +14.7 psia atmospheric
- 67 psia
- 2 psi) *144 / 57.3 =
6 ft + 14.3 ft = 20 ft NPSHA
Deducting a 3 ft safety marging NPSHA = 17 ft.

A low NPSHA can be increased by elevating the vessel in detailed design phase. This is a common practice. However, NPSHA must be compared to NPSHR which is provided by the pump manufacturer prior to detailed design.

Another question asked in this thread is what is the purpose of the LLLA - low low level alarm. This alarm would cause a plant shutdown, not for the purpose of protecting the pump, but to prevent gas blow by to the liquid storage tanks. The LLA alerts the operators to a low liquid level, such that they may observe any potential plant shutdown or pump cavitation.

 

[PetersPhotography]

The problem with chemebabak's response to my question is that water is not a bubble point fluid at 60 psig and 300 deg F. My analysis was for bubble point fluids. As I'm typically working with NGLs and even LNGs, I'm often working with bubble point fluids.

All I'm saying is that putting your vessel up 9' in the air isn't always feasible. Even more, we are often working with existing facilities, where the vessel is already installed. In those cases, a 3' safety margin is impractical, I think.

 

[chemebabak]

PetersPhotography, yes you are correct that water is not a bubble point fluid at 60 psig. Per the analysis and steam tables water is a bubble point fluid at 67 psia.

Nevertheless, the analysis can be modified such that the pressure in the vessel is at 67 psia. Then the calculation becomes NPSHA =
6 ft +
(52.3 psig +14.7 psia atmospheric [67 psia vessel pressure]
- 67 psia [vapor pressure of water at 300°F]
- 2 psi [assumed line losses]) *144 in2/ft2 ÷ 57.3 lb/ft3 [density of water at 300°F] =
6 ft - 5 ft = 1 ft NPSHA
Deducting a 3 ft safety margin NPSHA = -2 ft.

Please keep in mind that all fluids, even bubble point fluids, are subject to line and fitting losses. It is an error to state that line and fitting losses "should be almost negligible for a bubble point fluid".

I understand that in certain cases it might be impractical to raise a vessel to increase NPSHA, please keep in mind that a safety margin is just that - a safety margin. This is not law. This is best engineering practice. Hydraulics Institute recommends multiples of NPSHR for minimum NPSH. Compared to multiples of NPSHR 3 ft is a small safety margin.

Since you mentioned that you are a recent graduate I offer this free and friendly advice: any engineer can design any system with or without any design margin and that system could potentially still work without any problems. Safety margins and other design considerations ensure that the potential for problems is lessened. Engineers should ask ourselves "what risk are we willing to take?" Then design based on that risk.


------------------------------------------
Babak Firoozi, PE

 

[PetersPhotography]

chemebabak, thanks for your reply. Now that your example is working with a bubble-point fluid, the resulting NPSHA (-2 ft) is in-line with what I was expecting. Unfortunately, "higher safety margins" usually result in "more $$$$," which is why a 3 ft NPSH margin hasn't been feasible in the pump projects I've worked on thus far.

But I absolutely agree with you that, if the economics allow for it, a safety margin is certainly a good idea (considering that, over time, the pump suction piping/strainer can begin to plug and the effective ID will decrease). Also, a safety margin gives operations more flexibility in where they operate on the pump curve (higher flow -> greater NPSHr).

What I meant by line and fitting losses "should be almost negligible for a bubble point fluid" is that, when I design the suction pump piping for a bubble point fluid, I:
- Select a pipe ID that minimizes pressure drop due to frictional losses
- Minimize the use of fittings (i.e. tees, elbows, reducers) as much as possible
- Minimize the length of pipe between the vessel outlet nozzle and my pump suction nozzle
- Use only full-port ball isolation valves

In doing those things, I'm able to make line and fitting losses negligible (I'll usually see a greater pressure drop through my strainer than in all of my suction piping/fittings combined).

I apologize if I caused any confusion - you're right, it did sound like I was implying that line/fitting losses were low only because I was working with a bubble point fluid, which is obviously incorrect.

Thanks for your advice!

 

[rmw]

The steps you mentioned are good ones. Don't forget designing the vessel outlet to minimize entrance (to the piping) losses.

Your use of the word "strainer" in a bubble point fluid gives me real heartburn. I suppose it is there because it is necessary. If it is not ABSOLUTELY necessary, get rid of it.

One other one you didn't mention was designing the piping to minimize places for vapor pockets - using eccentric reducers at the pump suction, for example. If you are not using reducers, your piping is too small.

rmw

 

[djack77494]

It's nice to have a design that will work under any/all anticipated conditions. Still, it may be overly conservative to assume that you will operate a bubble point liquid pump at a level below the bottom level instrument tap while the pump is at end of curve flowrate. Yet we often think that is a reasonable design. If practical considerations push me hard enough, I'd be inclined to specify a substantially higher liquid level and/or limit the maximum acceptable flowrate. You can often take a design that seems unacceptable and end up with plenty of available NPSH by doing this. Is it wrong? I don't think so as long as you have the instrumentation and controls to ensure that you are able to limit operations to within the desired envelope. Also, few bubble point liquids would require an upstream strainer. If you're dealing with vacuum tower bottoms or some other heavy, dirty fluid, then yes you do need one. But if you're talking LPG or NGL, I wouldn't even think about a permanent pump suction strainer.