Reference pressure in closed loop pumping system without tank

[lucaspenalva]

How to determine reference pressure and its location in a closed loop pumping system without tank?

I've drawn a really simple layout so we can work on it.

Imagine a system basically composed by a pump and 2 heat exchangers located 10 meters above pump. No openings to atmosphere and no pressurized tank.

Pump ΔP = 20 mca

How do I know pressure profile in this type of system?

Where is my reference pressure?

How do I know i have enough NPSH for pump?

 

[1503-44]

A good reference pressure to use is the inlet pressure of the pump before you start it up.

To find Pump's running inlet pressure, Calculate the pressure losses around the loop at the design flowrate and subtract the total from the sum of reference pressure plus pump's discharge pressure. The resulting inlet pressure must be greater than NPSHR.

Then to find operating pressures at any point in the loop, start with the same resulting inlet pressure add pump discharge pressure and subtract all lossess up until the point of interest.

 

[Compositepro]

Such closed loops are never used in practice. The loop pressure could go from zero (vacuum) to thousands of psi with just a 10C change in temperature, and is thus indeterminate. All the problems go away with a expansion tank placed at the highest point in the loop where any gas bubbles in the loop will collect. In reality, if a tank is not provided, a gas pocket will form at some point in the loop.

 

[LittleInch]

What on earth is "20 mca"??

I thought I'd seen every pressure or head term but clearly not.

Any way as noted this is not a good example but the system will float on a pressure in the middle betwen inlet and outlet pressure of the pump.

This will also be the non flow settleout pressure.

Pressure profile will consist of three sections of pipe win a gradual fall of pressure then two bigger jumps for the two hxs

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[lucaspenalva]

Sorry guys, "mca" stands for (in my native language) meter of water column.

So in the example I gave, there is no way to predict what will be the pressure at pump inlet?

By the way, what are alternatives to expansion tanks?

 

[hydtools]

Alternate to expansion tank, an open atmospheric tank at the top of the loop. Need some accommodation for fluid thermal volume change.

Ted

 

[GT-EGR]

Your pressure at the pump inlet is going to depend on what pressure you fill your system up to - so if you have a makeup water line, you set that pressure reducing valve to some value based on your system goals.

You could fill it from the bottom, at the pump inlet, at a pressure of +30 mca. Then you would be +30 at your pump inlet, +50 at your pump discharge, and then that pressure would slowly reduce as you went through the losses of the system, then the pressure would drop 10 mca when you rise up to the upper level, then the pressure would drop by the heat exchanger loss, drop more from line losses, drop more from the second heat exchanger loss, then it would increase 10 mca when you drop back down to the lower level where the pump inlet is.

As its been mentioned in a few previous posts, you are missing an expansion tank (and probably a few more items you left out for simplicity).

 

[LittleInch]

No because you haven't said what is the initial pressurisation.

The alternative is a large vertical pipe say 10m high filled with water or a small pipe and a large tank located 10m higher than your pump.

In a closed system temperature has a huge effect on pressure


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[LittleInch]

Or have a constant pressure feed and a pressure relief set higher than your feed supply.

So say 10m supply pressure with 25m pressure relief.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[1503-44]

The reference pressure can said to be "mca" at the pump inlet before startup.

You might be able to simply not fill the system completely full. Check maximum volume of the fluid, if water that will be around 1 degree C. Make your pipe volume large enough to accomodate that volume, plus a "tad" more. You might create that extra space with a short vertical stub of pipe at the highest point of the system, or simply put in a short length of pipe of next highest diameter, maybe in or very near the pump inlet would be a good place, as that would tend to help npsh where you need it most. The pressure in the vapour that will collect at the loop's highest point and with the pump on or off, will always be the fluid's vapour pressure, so add 10 mca to get "reference" pressure at the pump inlet. You may have two phase flow in the highest pipe if you dont use a vertical stub, so design accordingly and keep velocity very low so vapour bubbles are not swept down into the pump suction. Using a vertical stub is better to avoid that. In effect you will have an expansion tank, just made of pipe.

 

[MJCronin]

As stated above, I agree that you need an expansion tank set up high and attached to the suction side of the pump.

The tank serves to compensate for liquid bulk expansion and contraction. It can be pressurized with nitrogen

It also serves as a preferred system "fill point"

The tank also provides a suitable place to add corrosion inhibitor to the system ... ( you were going to do this, right ? )

Most importantly, the expansion tank system will ensure plenty of NPSHa and steady operation ... a good recipe for a long pump life.

MJCronin
Sr. Process Engineer

 

[lucaspenalva]

Thank you all for your feedback. It really helped me better understand thermal expansion tanks and reference pressures in closed systems!

Hope you all have a great week!