Vapor-lock of LPG pumps 2

[rocketscientist]

I have been asked to troubleshoot an LPG system. There are 4 racks of 7' dia. bullets, 5 bullets to a rack. The five bullets in a rack feed a booster pump at a rate of 600 gpm with a 4" line from each bullet. On hot days, above 70 F with the sun out, the pumps vapor-lock. When the bullets drop below 3' they vapor-lock. A hysys model of the system shows that at a starting pressure of 135 psig the highest temperature that can be tolerated is 80 F. Once the bullets drop to 3' at a pressure of 104 psig, the highest temperature that can be tolerated is 63 F.

Operations installed a line off the suction line of the pump that goes to flare. This helps de-gas the pipe. Is there a better way to avoid this problem.

I've suggested re-circulating some of the flow back to the bullets from the discharge of the pump. That is, moving the line from the suction to the discharge. Instead, others are pursuing changes to the control valve.

 

[pmover]

place a shelter above the bullets, insulate the piping, insulate the bullets, etc. it is apparent the need to keep LPG temp low to avoid vapor.

at or near the pump inlets, is the piping arrangement/installation allow for vapor to be created? sudden changes in direction, usage of concentric reducers instead of eccentric reducers.

hope this helps and i'm sure there are others with ideas.

good luck!
-pmover

 

[TD2K]

I'd suspect excessive inlet line losses. 600 gpm in a 4" line is about 15 ft/sec, that's way too high for a pump suction line on a bubble point fluid. Up-size the piping is really the only solution though that's not an easy fix.

 

[Montemayor]

There is a scarcity and misinterpretation of the basic data submitted in this thread.
Regardless of what Hysys may spit out, there is no relationship between the LPG pressure and the highest temperature that can be tolerated - at least not 80 oF. The process facts are as follows:

1) The LPG fluid (a mixture of Butane and Propane, usually 50:50) exists in the saturated liquid state within the bullets;
2) The LPG will remain as saturated liquid within the bullets regardless of the ambient temperature – up to the Critical Temperature. This is clearly seen on a T-S Diagram. As the saturated temperature increases, the pressure will also increase – but the LPG remains a saturated liquid and perfectly pumpable – as long as it can be transported into the casing of a pump.
3) The only obstacle that prevents the LPG from being pumped is the difference between the NPSHr of the pump and NPSHa of the piping system that links the bullets and the pump suction. But this hasn't even been discussed.

I know the above to be proven facts because I have pumped a lot of LPG through the years using many different type and size of pumps and under temperatures from 30 to 120 oF – without any problems. And what’s more, this is being done everyday all over the world. There is no need for a shelter or for insulation on the system.

TD2K is correct. The obvious culprit here are inlet line losses – please refer to the attached workbook to see the explanation given. However, the pump suction velocity is not 15 ft/sec, but rather a conventional 3.5 ft/sec – which should be OK when using conventional, conservative piping systems. However, there is cause to suspect that this is NOT A CONVENTIONAL, CONSERVATIVE piping system. Since the bullets are relatively small (7’ is not much for a pressure vessel) and there are 5 of them stacked together, it seems obvious that the piping is very congested and closely packed with a lot of fittings and valves used prior to even approaching the central collection manifold that surely must exist in order to tie-in the 4 racks of bullets. And then, after going through the central manifold, the LPG has to enter the pump suction line. That is a lot of fluid twisting and turning – all leading up to suction line losses that are to be totally avoided in feeding a saturated liquid pump (especially a centrifugal type).

What is needed is a complete description of the piping configuration and sizing within each rack (isometric drawings) and a calculation of the NPSHa. Once that is achieved, it should be compared to the pump’s NPSHr and the obvious result will be apparent at that time – the friction generated in the suction line (mostly due to turbulence and fittings losses) will such that it caused vaporization of the LPG as it approaches the pump’s impeller and the pump loses suction. In other words, the NPSH available is not enough to conform to the NPSH required. That is why there is a certain LPG liquid level that initiates this action. The answer to pumps saturated liquids is to supply as high a liquid level above the pump’s eye as possible and to apply as low a suction fluid velocity as you can justify –with a minimum of fittings, valves, and flow directional changes.