PSV in the Centrifugal pump discharge

[Homayun]

Hi all,

I have a question. I am trying to assess a PSV in the discharge of a centrifugal pump. But it seems difficult to determine the relief load.
Here is hte system:

A centrifugal pump, some reactors in serial line up, a control valve, a safety valve and a block valve.
THe case I am considering is inadvertent closure of the last block valve causing a bloked outlet case.
How do I calculte the relief load since it is depending on the pump curve?

My approach would be to start with the PSV relief pressure, calculate back the pressure at the pump considering all pressure drops in system and hence determining the corresponding pump flow using the pump curve. But the difficult thing is that the "pressure drop" is a function of the flow, so it varies and it is not fixed.

Anyone got any idea?

 

[quark]

First of all you have to decide why you require a PSV in a centrifugal pump outlet. Is your pumped fluid volatile and cause abrupt pressure changes when it is heated during shutoff head condition? Is your pipe system designed for shutoff pressure?

Generally, the safe minimum flow for a pump varies from 20 to 30% of rated condition. If you, anyhow, want to provide a PSV, the releiving rate should be equal to minimum flowrate.

 

[Homayun]

quark,
I am looking at an existing PSV. So it's not a new deisng.
But the pump is being upgraded and besides, we are going to operate both pumps in prallel,so the flow and the pressure would be higher.

I know that a good designed system, does not require a PSV in the discharge of a centrifugal pump. However this PSV is there because some segments of the piping does not have adequate design pressure.

The question is then how to calculated the relief rate in a blocked outlet event!

 

[SeanB]

Typically from what I have seen, the PSV is sized for the rated flow from the pump.

 

[StoneCold]

Homayun
Ignore the hydraulic pressure drop of the equipment and consider the situation as if the relief valve were right down stream of the pump. You have to relieve enough material so that the pump does not back up on it's curve enough to overpressure your piping. That gives you a relief pressure and a minimum volume. Now that you know the rate of flow through the system you can calculate the hydraulic pressure drop of the system and reduce the pressure relief valve pressure setting to ensure you are below the weakest design pressure. If the pump is real small I would just to what SeanB suggested and not sweat it out. You also need to consider the minimum flowrate of the pump, if it is higher than the calculated relief rate, as Quark suggested. Maybe you should really look at a more through and safe re-design. Just because it is there does not necessarily mean you should use it. The guy who put it in might be a moron.

StoneCold

 

[TD2K]

I'm currently working on a centrifugal pump that has the downstream equipment protected by a relief valve.

Basically, you need to establish the greatest suction pressure to the pump which will depend on your system. Then, you need to identify the weak link in the system and using the pump curve, identify how much you have to flow to run the pump sufficiently down on its curve so that the piece of equipment (or equipments) is not overpressurized. Note the suction pressure can change significantly depending on your system and the scenario(s) you consider to be credible (for example, if the pump is taking suction from a pressure vessel, it is feasible for that vessel to go to relief pressure during operation?)

You mention the PSV is downstream of a control valve and several pieces of equipment. If the weak link you are trying to protect against full pump discharge is upstream of that control valve you have a problem. If the control valve closes for any reason (or for that matter a block valve in the system is closed), it sounds like the equipment and piping upstream of that point will see the shut-off pressure of the pump.

 

[quark]

My idea is similar to that of Stonecold's. First check out what is the design pressure of the system. Secondly, check out what is the maximum pressure developed by the pump. As your control valve is below the PSV tap off, there won't be much problem but it is better to check the maximum pressure after the control valve and before the block valve(when it is opened).

Your PRV set pressure should be in between the design pressure of the system and the maximum pressure after the control valve(this will reduce the problem of nuisance PSV opening in normal case). This will ensure that the PSV opens up whenever the pressure exceeds due to block valve closure and PSV closes during normal operation.

Now come to the pump side. Calculate the system resistance downstream the PSV and add PSV set pressure to it. This indicates your pump operating pressure when the PSV opens up. Check the corresponding flowrate and make sure that it is atleast 20% of the rated condition. If you are using volatile liquids then you should actually check the temperature rise and should plan your safe minimum flow.

For example, the pump shut off head is 7kg/sq.cm and the maximum pressure after the control valve is 4kg/sq.cm. Then your PSV should ideally be set between 4-7kg/sq.cm. Say, the downstream piping of PSV offers a resistance of 1kg/sq.cm. If I set the PSV at 5kg/sq.cm, the pump will run at 5+1 = 6kg/sq.cm. If my design flowrate is 100cu.mtr/hr and flowrate corresponding to 6kg/sq.cm is 20cu.mtr/hr then we can go ahead without any problem.

But these straight values are not possible in actual practice as your system is already built. So, I would rather go with an auto on valve which can be activated by a pressure switch set at the shutoff pressure of the pump. The closing of the auto valve will be manual. If I take care that the pressure switch gives me an audio signal when it activates the valve, my operator would easily come to know it.

Regards,