High Pressure Centrifugal Pump Trip.

[chemicalengr]

A recent centrifugal pump re-build left my client with pump curves (x3) that are "flat" at their lower flow range, to the point that the minimum stable flow is on the flat part of the curve. The pump may be required to periodically operate at the minimum flow, thereby making a high pressure trip for pump protection nearly impossible to implement. API RP 14C recommends that a high pressure trip be present on the pump discharge.

A low flow trip could be implemented as there are flow meters on each individual pump discharge, but there is some concern about the accuracy of the meters.

Each pump is equipped with its own minimum flow control valve and the service is seawater at high pressure (~ 360 bar). The discharge piping is rated for the pump discharge pressure. Therefore, the concern is pump protection as these are large expensive pieces of equipment.

Does anyone have any experience with a similar situation or thoughts on a potential solution?

 

[jistre]

It would end up using more power than using pumps with a more suitable curve, but could you use a recirculation loop to allow the pump to move more water and slide down the curve to a point where you have a large enough pressure differential between the operating point and the flat part of the curve to allow your high head shutoff?

I'm curious, why did the rebuild change the curve to create this situation? Was this a rebuild to allow the pumps to be retasked and reused? It just seems like a pump rebuild shouldn't take pumps that operate at acceptable points on their curves and flatten the curves so they ride the curve in undesirable places.

Is starting over with more suitable pumps possible? I'm pretty sure I know the answer, but I had to ask!

 

[JJPellin]

For a "large", "high pressure" pump, the risk of pump damage from running below minimum flow would tend to be greater than the cost of lost energy by spilling back some flow to achieve a better operating point. I work in a large oil refinery. We almost never set up a pump such as this to trip on low flow or high discharge pressure because the upset to the process could create a dangerous situation. But, we would often program the minimum flow spill-back to open automatically to protect the pump. We would also set an alarm on low flow to warn the operators if the pump strayed into a low flow condition.

If you do want to have a shut-down system to protect the pump, I would use the flow rather than the pressure. Even if the flow meter has questionable accuracy, for a pump with a very flat curve, the flow would almost certainly be more accurate than the pressure in determining the onset of low flow.

In any case, consider the economics. If spilling back some flow costs you 100 HP you would waste less than US$40,000 per year (even if you ran that way 24/7). But, running the pump to destruction at low flow could cost US$250,000 to repair. These are numbers that might be typical in our plant. You need to consider your own energy and repair costs.

I was not aware of any API RP recommending high pressure trips. This seems very strange and I would take exception to it in most of my applications.


Johnny Pellin

 

[chemicalengr]

Thanks for the advice.

jistre, the proposed pump curves prior to rebuild showed head rising to shut off, but the FAT curves did not demonstrate the same phenomenon. The pumps have been purchased, therefore, starting over with new pumps is not possible. The pump rebuild was to provide higher flow and head at the BEP.

The issue with running the minimum flow lines continuously is that the metallurgy was not designed for continuous flow and failures have occurred in the past.

 

[davefitz]

If money was not an issue ( !), a min flow recirc control valve designed specifically for this case ( valve , piping + metallurgy) would allow continuous operation at the min flow case . However, the measurement of min flow must be assured as accurate- for the flow element on the main line , a small non condensable vapor bubble in the DP sensing lines can indicate adequate flow when in fact zero flow is occurring.

So, a back-up high pump outlet pressure pressure switch would also be used to protect the pump, perhaps with a time delay equal to the stroke period of the min flow recirc control valve.

 

[MJCronin]

"The discharge piping is rated for the pump discharge pressure."

Sounds like a huge piping design mistake IMHO....What code covers the piping design ?

Any pump discharge piping should always be designed for pump dead-head conditions.

Both B31.1. and B31.3 allow a higher allowable stress values for this infrequent condition.

-MJC

 

[chemicalengr]

MJCronin,

"The discharge piping is rated for the pump discharge pressure."

Sorry, that is just a mistake in my wording - the piping is indeed rated for the pump dead head pressure.