3 replies to this topic

[chejosh]

Hello. I recently started a new job at a small plant, and I’m the sole engineer here. A PHA was conducted several months before I came and there were multiple PSVs identified as not having calculations. So one of my duties is to complete the calculations.

I have not done much PSV calculations since about 6 years ago. I think I have a good handle on most of it, but scenarios invoking pumps are throwing me off. I honestly haven’t looked at pump curves much since college.

If I have a situation where I’m pumping into a tank, outlets are closed, and I have an overflow scenario. For a centrifugal pump, I take the set point of the PSV, let’s say 50 psig, add in the accumulation and then convert to head, so ~127 ft of head, right?

If the pump curve hits the y-axis at 100 ft of head, that means it isn’t a valid scenario because the pump can’t provide enough pressure to lift the PSV, right?

Let’s say the set point is low, 10 psig, 25 ft of head. What does it mean when the pump curve stops at 60 ft of head on the right side of the curve? And your set point is so low, it doesn’t correspond to anything on the curve.

For positive displacement pump, they pump a set flowrate for a set rpm and the pressure, line size, anything else doesn’t affect it, right? So for the relief rate chose the flowrate at maximum rpm?

I have a question about API 2000 and pumps as well, but this is probably enough for now.

Thank you for any advice.

[Pilesar]

Instead of 'what scenario will cause the PSV to lift' I suggest you evaluate each scenario for its potential to overpressure the vessel. Calculate the required relieving flow rate for every scenario. Sometimes that requirement may be zero relieving flow rate, but you document that anyway. All the scenario calculations together form the documentation. After all the relieving rates are calculated, choose the largest required relief area (not necessarily the largest mass flowrate) as the controlling case. Then you can compare it to the existing PSV to see if the sizing is adequate. 

Relief sizing is such an important role. Get the training you need and don't just 'wing it'.

[chejosh]

Instead of 'what scenario will cause the PSV to lift'

That’s not what I am doing. If the set pressure of the valve is higher than the head the pump can achieve (according to the pump curve hitting 0 gpm on the left side of the curve) then pump deadhead isn’t a valid scenario. I am asking if I am correct in my thinking.

Calculate the required relieving flow rate for every scenario.

That’s what I am trying to ask help with. Pump curves tend to trail off on the right side of the curve and not follow the head down to 0. What happens between the point where the curve stops and the head is 0? Does the pump not work, do I assume maximum flowrate?

[Pilesar]

A centrifugal pump has a maximum pressure increase at the left side of the curve. The discharge stream pressure generated depends upon the suction pressure.

The deadhead pressure read on the centrifugal pump curve can be used to determine the relevance of that scenario to vessel overpressure.

Sometimes it means that the pump just cannot overpressure the vessel.

For the right side of the centrifugal pump curve, see the definition of pump runout here: https://www.introtop...p-terms/runout/

Positive displacement pumps have different flow and pressure characteristics than centrifugal pumps.