Centrifugal Trapped Liquid Hazard 1

[ARCH01]

A multi-stage centrifugal pump is proposed for an application with 60 psig (ANSI 150#) suction pressure and 2500 psig discharge pressure (Ansi 1500#). The pump is equipped with discharge shutdown and check valves. A question was raised related with the possibility of high pressure water migrating to the suction side when the pump is stopped and discharge piping is isolated. Pump seal is designed for 1900 psig static pressure. Suction piping header has a 150# suction check valve. Being water incompressible the concern exists that suction piping can be overpressure when pump is down and isolated with water trapped at high pressure. Has any of you experienced a piping failure in similar conditions? We're thinking about installing a 1500# shutdown valve to stop any possible high pressure water migration to suction pump. Cost of valve is almost prohibitive. What would be your recommendation?

 

[rzrbk]

Some options I have seen used where some components of the pump (suction flange, seals) are not designed for discharge pressure are:
-Install a PSV on the suction line of the pump, which would then relieve any leakage past the discharge block valve back to the suction vessel.
-Car seal or chain lock the suction block valve open so that there is always a path back to the suction vessel for any leakage past the discharge block valve. The suction valve would then only be used for maintenance when the pump is blocked in and drained down.

 

[ARCH01]

RZRBK

Thanks for your reply. the scenario presented here is that the discharge valve is closed tight (i.e. no leak or reverse flow through it). The suction of the pump is equipped with a check valve. If the pump stops, water will likely flow back from the high pressure 1500# to the low pressure side of the pump which is designed for 150#. Since produced water is basically incompressible, the concern exists that piping might fail. I would like to find out other people's experiences in a similar situation to detemrine if this scenario is credible.

 

[Montemayor]

Arch:

There are some things you describe that are difficult to interpret. Let me explain:

You say the centrifugal has an automatic shutdown valve on the discharge PLUS “check valves” (I suspect you mean two checks). If you say you have an automatic shutdown, it is universally understood you are actuating an automatic block valve on the discharge side, slightly downstream of the pump. This is over and above the 2 check valves. If you don’t mean this when you say you have “discharge shutdown”, then you must tell us; otherwise, we interpret what I stated.

You also say you have a check valve on the suction side (why you have this, you haven’t explained). This is not the conventional way to install a centrifugal pump. A suction check only adds NPSH woes and unnecessary suction turbulence. I don’t see what a suction check contributes on a centrifugal pump unless you’re using a FOOT VALVE in a suction sump. But I don’t think this is your case. Yes, under these conditions you’ve essentially trapped liquid inside the pump and should this liquid (produced water) be able to increase in temperature during its residence while blocked in, you will veritably have a credible and very possible thermal expansion hazard created due to the hydraulic forces increasing when the produced water expands under temperature increase. You must supply a safety device of some sort if it is possible to increase the produced water temperature.

I never install suction check valves. When I do an emergency trip and shutdown of the pump, I block both the suction and the discharge – if need be. The way I protect the pump in case of blocked liquid undergoing thermal expansion is that I also install a thermal relief valve (in the case of benign liquids) or an expansion chamber or a safety relief valve on the pump or close to it to protect it and the seal.

When I don't automatically block the suction side, I allow the suction pressure to float with the source pressure - like in the case of storage tanks. Subsequent action(s) may cause the suction valve to be shut either manually or remotely. Again, this depends on the operations and the safety involved.

The manner in which you describe your operation, you have a perfectly credible scenario – if you can credibly have the blocked produced water increase in temperature while blocked in. A buckling pin relief to the atmosphere may be all you need - depending on whether your produced water is clean enough or not.

I hope this helps.

 

[ARCH01]

Montemayor,
Thanks for your reply.

Let me provide necessary clarifications:

1) There is a shutdown valve and a check valve on the discharge side of the pump.

2) The suction line of the pump connects to a manifold with three laterals for various water supplies. There are check valves on each lateral to prevent mixing incompatible waters.

My concern is not so much related to thermal expansion, but migration of high pressure water from the discharge side of the pump to the suction side in a situation where the pump is not in operation but high pressure liquid is trapped in the discharge piping.

I could probably afford sending the produced water to an open drain system. Therefore, the idea of a small relief valve (perhaps sized only for thermal scenario) or an expansion chamber such as a pulsation dampener might be sufficient.

 

[Montemayor]

Archie:

Thanks a lot for your prompt reply. I have a serious concern any time I see a 2,500 psig discharge on a centrifugal pump with <100 psig on the suction side. You certainly require sound and proven shutdown devices and instrumentation - not only to protect the personnel and the pump, but to protect everything upstream of the pump.

I would expect to see two (2) Check Valves (in series) on the disharge of your centrifugal pump - not just one. I never expect one single check valve to seat every expected time and not to leak. You will find this out if you have a formal, serious PHA or normal Hazop on your installation before you start it up. Two check valves don't necessarily give you redundancy, but they reduce the odds of total failure to arrest the discharge pressure before the main, automatic block valve is actuated shut. I would not do it with less than 2 Check Valves and I recommend you do likewise. Also, please take special care in studying and confirming that you have sufficient time with the selected instruments and automatic block valve to effectively block the 2,500 psig pressure from getting into the low-pressure portion of the pump and anything upstream.

Do not fail to protect the suction and seal portion of your centrifugal pump from the 2,500 psig potential over-pressure by applying a buckling pin or rupture disc. Vent the 2,500 discharge SAFELY away from any possible personnel traffic or exposure.

This is a potentially very hazardous operation if not handled meticulously and properly engineered. I highly recommend you employ an experienced and capable instrumentation or EE to assist - if not take over - in the critical shutdown instrument application. I now understand the application of the suction check valve and its use. I would not proceed on this type of project without at least TWO -not one - Hazops meetings: one preliminary meeting with operators and instrument engineer & techs involved and another meeting prior to having the project released for construction. Both hazops should be thorough and detailed in every aspect related to credible and expected events, scenarios, and operation. It should be detailed in all its documentation and calculations. You are on the right path; make sure everyone is as concerned and educated about the application as you are. Good Luck!

Art