Sealing NGL in a Reciprocating Pump

[GaTechTheron]

We are evaluating reciprocating pumps in NGL service. However, there is an inherent issue with sealing these pumps especially when we have a suction pressure in the 300psi range. Basically, we cannot perfectly seal them (as far as I know).

Interestingly, the pump vendor is recommending fabrication of a "distance piece" that will cover and seal the cradle area in order to capture natural gas that gets past the pumps' packing.

For me, this threw up red flags. Why would someone recommend capturing an explosive gas in an enclosed area? What if something came loose from the pumps' push rods, and caused a small spark in the cradle area? Further, if we contain more than a few pounds of pressure in the cradle area, wouldn't it push in the pump's crank case as well? Are there any regulations against this?

Unfortunately, I do not have nitrogen available to flush the area to keep it clear of natural gas. Does anyone have an experience with a better solution?

I appreciate your time.

 

[micalbrch]

The red flags are entitled. The plunger rod seals to the gear side usually seal only against atmospheric pressure. Their purpose is to avoid (unpressurized) oil leackage from the gear or that (unpressurized) liquid can enter the gear. And it is correct that capsuling explosive gas in the cradle area is a bad idea, too.

What can be done? Liquid gas leaking over the plunger seals must be avoided under any circumstance unless you can live with that - and I guess you can't. The manufacturer can provide a plunger seal flushing system which is pressurized. But that means that you must find a liquid which does not "contaminate" you NGL. The amount of flushing fluid that contaminates the NGL would be very, very small but it would be there.

The other idea is a diaphragm pump which is hermetically sealed but it will be difficult to find one for that application. LEWA might be worth a try.

 

[DLite30]

That's exactly the way reciprocating compressors have been built and in natural gas service for a looong time. The important thing is to 1) vent the distance piece to a safe area outside of the compressor building and seperate the lube oil,

and 2) install an automated method of detecting a packing rupture (to shutdown and vent the unit) such as a pressure switch set at a low pressure (say 1 to 2 psi), or a flow switch in the vent line.

 

[Pumpsonly]

A diaphragm head reciprocating pump would be your best choice.

 

[NGLENGR]

We had some reciprocating pumps in isobutane service and they were plagued with VOC emission issues.

Eventually they were fixed by replacing them with API BB3 centrifugal pumps.

The unfortunate part about recip pumps is that they are so good at a wide range of flows in a fixed head situation. Process engineers forget that packing is a terrible sealing solution and spec them out for projects all the time, to my frustration.

We have used vapor recovery systems, but you have to design them to be statically sealed or vented to an appropriate place. Other ideas like Plan 53 style barrier systems have never had much success in our machines.

 

[curtis26]

I would use a double packed stuffing box with a leak detection system for NGL service. This configuration has a primary packing configuration followed by a pressure switchgage and then a secondary packing system. Once the primary packing begins to fail, the pressure switchgage will trip indicating it is time to replacing the primary packing. The secondary packing will continue to seal the pump so that the packing can be changed at your convenience.