NPSHa Cargo Tank Unloading

[tmartin125]

The marine vessel being designed has multiple cargo tanks and multiple cargo pumps. Initially during unload all pumps unload one tank (same tank) until level of tank is low then operators open second tank. These cargo pumps will cavitate (NPSHr>NPSHa) when pumping from single tank at low cargo level, i.e. less than quarter full. How does calculation of NPSHa change when two tanks are open at once at varying levels of cargo? Is the level of the full tank to be used for the static head or is it an average of the two? Does the higher level tank make up for the lower level and allow for more cargo to be removed from the tank without cavitation? Is say 90% of the cargo pumped from 2nd tank and 10% from first?

 

[racookpe1978]

Assuming you're pulling from the absolute bottom of tank 1 at 25% full. NPSH available is 25% of the full height of tank 1, right?
Now, if the suction is at HS, and tank 1 height = HT1, then maximum NPSH available (is proportional to) = HT1-HS

If tank 1 and tank 2 were connected at the bottom, but were not being pumped from either tank, they'd eventually even out (Ht1 = Ht2) because oil would flow from the higher to the lower (if no check valve of course).

So, as long as you could pump out tank1 faster than tank2 could empty into tank1, the NPSH of tank1 at the pump suction will be that of the current relative liquid level of the highest tank cross-connected, minus flow losses through the cross-connect pipe, minus (or plus!) "ship elevation changes (from bow to stern) of the two different tanks".

 

[LittleInch]

Are these tanks manifolded together at the pump inlet? Can you draw a quick schematic of this and post it with some idea of dimensions, line size, flow etc as it is difficult to work out what is happening.

Yes the calculation will change, but how much by is difficult to judge. The full tank will be at a higher level and will flow more than the first tank (possibly even 90/10). This will increase the frictional losses in the second tank line compared tot he first tank line until the common pressure at the inlet header matches the calculation of higher tank plus higher friction losses in tank 2 to the lower head plus lower friction losses in tank 1. Where that comes in terms of head difference will vary by pipe size, flow rate, length of piping etc and will keep on changing as the tanks empty. As the ratio of friction losses to tank head difference will vary between any two systems, it is not possible to come up with an easy calculation.

The two fixed elements are atmospheric pressure and the presumably common inlet header pressure into the pumps. Everything else is variable depending on your system and pump configuration, but opening the second tank will increase the NPSHA.

My motto: Learn something new every day

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

 

[tmartin125]

There is a sump at the tank bottom where the suction bellmouth is located so yes the static head of the tank is the full 25%.

The cargo tanks are entirely separated structurally. They would only be connected through the cargo piping however I expect with continuous suction from the pumps that cargo would not be allowed to flow between tanks.

So with NPSHA=+/- Z+(P-Pv)*(1/SP GR)*2.31-HF ..... the z value to use for two tanks pumped simultaneously would be of the higher tank? Also unsure how to determine the lowest level the pumps can get any one cargo tank or the rate at which either tank would empty during two tank situation.

I tried to find literature online which discussed this situation but so far have been unable to find.

 

[LittleInch]

This is a dynamic simulation exercise so you really need a dynamic tool. either that or work out for varying levels of tank 1 and 2 what the predicted flow rate is from both tanks. It doesn't matter which tank you use to calculate the change in NPSHA at the moment you open the second tank, but the flow rate from the second tank will be higher than the first. If level in tak 1 is H1, etc then the ratio of friction losses will be H2 / (h2-H1) ASSUMING ALL PIPING LENGTS AND DIMENSIONS ARE THE SAME. As they probably are different, you have a different equation to set up to factor that in. When you have two tanks the total volume flow should then increase as the inlet pressure rises, unless this is has a negligible impact on the overall TDH of your pump.

Too many variables for us to work out

My motto: Learn something new every day

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

 

[tmartin125]

Thank you for your comments and analysis. This is a difficult problem... I don't have a dynamic simulator.. Will have to do this the hard way with excel