Sizing PSV for Centrifugal Pump 3

[chemter]

Hello,
I am sizing a relief device for a situation in which a centrifugal pump could overpressure a vessel. Generally speaking, I have seen cent. pump curves that denote the y-axis as Total Developed Head (TDH) and others as differential head. Are these really implying two different values? If a pump curve reads a TDH of 30' at zero flow, does this assume that there is no pressure downstream? Also, when determining the required relief rate for a PSV using the cent. pump curve, how is the flow adjusted to account for static suction head? Is the amount of static suction head (in feet) simply added to the pump curve?
Any resources or articles on sizing PSVs for centrifugal pumps would be greatly apprectiated. I have read many manuf. sites, but none seem to go into detail about these types of situations; where you are dealing not only with a shut off pressure, but also with different suction and discharge heads and how they may affect the pump curve. Thank you for your time.

 

[SeanB]

You have to calculate the maximum discharge pressure from your pump. Generally, you calculate it 2 ways:
1. Normal Discharge Pressure + Maximum Suction Pressure OR
2. Normal Suction Pressure + Pump Shutoff Pressure.
You should calculate both and the greater of the two will be the maximum developed pressure from the pump. The required relief rate should be the rated flow of the pump.
I hope this helps.

SeanB

 

[djack77494]

chemter,
Total developed head (TDH) and differential head are the same. If a pump curve indicated 30' TDH at zero flow, then the discharge (downstream) pressure is not atmospheric; it is at its maximum value (shutoff). The pressure would be =
Suction pressure + TDH/2.31*sp_gr
where TDH = head in feet and sp_gr = fluid's specific gravity at flowing conditions.

Frankly, you are asking enough questions about this suject to make me nervous. The asking of questions is good, but if you are sizing safety valves then I think you are not sufficiently knowledgable about your systems to be able to do so without help. Please get a person experienced in sizing PSV's to help you.
Doug

 

[chemter]

Thank you SeanB & djack. I just have one follow-up question, though. As SeanB mentioned, the required relief rate should be the rated flow of the pump (at the relief valve's set press. plus overpressure). My question then is can this flow from the pump curve be increased to account for any static suction head that may be present upon discharge from the relief valve or is the flow from the pump curve the max flow possible (ignoring any friction losses and/or static discharge head present for the time being). For example, say at the absolute allowable relieving pressure of 114.7 psia for the RV, the pump curve gives a value of 100 gpm. There is also 50' of static suction head. Would it be incorrect to account for this and move to the right on the curve and increase the required relief rate? Thanks for the help.

 

[SeanB]

Chemter,
The flow from a centrifugal pump is independent of the static head pressure. Static head only adds to the discharge pressure of the pump. AS for the required relieving rate, use the pump design rated flow, not the flow at relieving conditions.

 

[pleckner]

Let's look at the dynmaics here.

For a properly designed system, relief from a centrifugal pump is caused by the pump circuit being blocked-in. The pump head rises along the curve. The actual pressure at the pump discharge will be the pump head plus the static suction head plus the source pressure. The pressure at the PSV will be this pump discharge pressure minus any discharge static head. There is no frictional loss since there is no fluid movement.

Assuming the suction static pressure and source pressure are constant at this time, there will be a point on the curve where the pressure in the system causes the relief valve to open. It could be anywhere on the curve or at complete shut-off; strictly a function of the pump curve. Once the PSV opens, flow begins with a bang! The pressure will fall along the curve and flow will increase. But so will frictional losses since we are now flowing. The pressure at the PSV may very well reach a point that will fall below the PSV blowdown, the PSV closes and flow stops. The cycle repeats itself.

The maximum flow through the PSV needs to be found using basic hydraulics, balancing the pump curve with system losses. This is the exact same calculation used to size the pump in the first place except now you have the pump curve to work with. There is no way to know, without calculating it, whether the maximum relieving rate is equal to, less than or greater than the design capacity of the pump. Assuming just the pump design capacity could be a big mistake or, it could be conservative. Only the true hydraulics of the system knows for sure.

The use of the pump design rate has been used because the theory goes that I couldn't possibly lift the PSV at the design rate because my system normally operates at this rate. But the system only operates at this rate because the fluid normally has a place to go and now it doesn't. Also, the system pressure still must drop to blowdown before the valve will close.

Do it right, calcualate it. On a spreadsheet or using some canned program, it doesn't take long.

 

[SeanB]

Pleckner,
You make a good point about the relieiving rate. Using the rated flow usually tends to be conservative in my experience, but I am sure there are cases where it could be undersized. It doesn't hurt to do a quick check.

 

[chemter]

pleckner,
Thank you very much for your response. It was very helpful.