Defining high pressure alarm and trip pressure values for centrifugal pump's discharge 1

[Henrique Orlandini]

Hi,

I'm working in the instrumentation department, and we need to set the alarm and trip point of a centrifugal pump. I'm having trouble with the high pressure trip value.

According to our pump's curve, the highest discharge pressure would be at shut-off head (348 m) with +5% of tolerance, totalizing 365.4 m. That in bar would be about 35.84 bar, and adding the maximum suction pressure (1.5 bar) would give about 37.34 bar, so we set the high pressure trip at 37 bar, while setting the high pressure alarm at 34 bar using the mininum stable flow of the pump.

Our customer commented that this was the worst case scenario, with maximum suction pressure and maximum shutoff head value. He asked what should happen with the alarm and trip if the pump worked outside those conditions, let's say suction pressure of 1 bar and a shutoff head of 348 +3% (which is inside the tolerance). With my calculations, that would leave the pump with a discharge pressure of 36.16 bar, enough to set the high pressure alarm, but not enough to trip it, leaving it running at shutoff.

I thought of using the lowest possible values of suction pressure and shutoff head tolerance, but that would reduce the working range of the pump, causing it to trip at unnecessary moments.

What could be a good solution? Measure the flow at discharge and trip it if goes at a low level (before 0 m³/s) or measure the delta of pressure between discharge and suction and trip it when this delte becomes too high (although I can't see this working).

Could you help me with this? If you need more info, I would be glad to post if I'm allowed to.

 

[LittleInch]

A high pressure trip should be set to protect the pipework, not the pump.

Centrifugal pumps are essentially a constant pressure machine (posting the pump curve would help_ and with the various tolerances for both the actual performance of the pump, changes in temperature and density of your fluid, variance on set point of your transmitter / pressure switch you will either go to no flow and not trip or the pump will continually alarm or trip during normal operation.

Using a pressure trip for what I perceive to be a no flow situation is, IMHO, a poor decision and one which doesn't normally work well. The only time is does is if you've got a mixed flow pump, but even then sometimes the advise is to start against a closed valve and then you need to inhibit the trip....

The only way to activate a trip on low / no flow is to measure the flow. Flow switches are notorious for not working well souse a flow transmitter - there are lots of quite cheap ones which measure to +/- 3-5%.

You can also set permissives and valve position indication such that if certain valves are closed then the pump trips or won't start until there is a path for the flow

Basically do it properly and measure the flow.
You can also measure pump temperature and trip on that - that would be a second layer of protection

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

 

[Henrique Orlandini]

So flow would solve it, I see.

And sorry, I cannot post the pump's curve due to our company's policy, but this thread could be about a general doubt, since we do this process quite a lot, it's always good to know better ways to set our sensors.

So to close this matter, to measure pressure at suction and discharge, not always in a no flow/low flow situation, for example, to prevent the pump to reach the end of the curve, it's always better to use flow transmitters since it's less dependable of external variables or pressure transmitters could still be used?

 

[LittleInch]

You could use the differential pressure, but it's still a very crude and inaccurate way to trip a pump on no flow. Differences in the actual pump curve versus predicted, ( 3-4%) differences in density of the product due to range in temperature (depends on your fluid and temperature range, but can be 5-10%), variance on the transmitters over time ( 2-5%), variance over electrical supply ( 2-4%),

So you either keep tripping the pump during normal operation or you don't protect it when it needs it.

If you're tripping on no / low flow then measure the flow, not some other variable.

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

 

[Henrique Orlandini]

For other points, such as low pressure, in which the flow would be high, is it still advised to use trip using the flow? Or in that case pressure tripping is acceptable?

Anyway, thank you very much, I got the answer for my main problem, the rest is just me trying to get better at my job haha

 

[LittleInch]

If you're tripping on a variable, it is always better IMHO, to measure that variable rather than inferring it from another reading.

However low pressure is one way to protect against high flow and is commonly used as a trip for excess flow on pipelines as this could come about as a line break causing low pressure, but then higher flow.

Usually on high flow you also get over current protection or vibration tripping the pump before the low pressure alarm / trip kicks in. You then also need to find some way to override the low pressure on startup or stop the alarm going off when the pressure drops as part of the stop signal (link it to the pump start stop command?)

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