Maximum Pump Capacity 1

[TreeEng]

I realize this is a simple question, but I would appreciate any help you could give.

How is the maximum capacity of a pump determined from the pump curve? Do the suction and discharge pressures affect this analysis?

 

[TD2K]

There's a couple of ways to answer your question, depending how I interpret it.

The maximum capacity of the pump is about where the pump curve ends. You can extrapolate the curve somewhat to higher flow rates but the head may quickly drop off below what you believe the pump should be putting out. Be aware that driver hp requirements (unless you have an axial type pump) and NPSHR will continue to increase and pump reliability will suffer.

The actual flow through the pump for a set of conditions is set by the dP or head across the pump. Discharge pressure minus suction pressure is the pressure rise across the pump. You then convert this to head by the equation 2.31*dP/SG where SG is the specific gravity of the fluid being pumped. You then use the curve to read off the resulting flow rate for that head. For viscous fluids (greater than about 50 cP), you should correct for viscosity. Compared to pumping water, head will be decreased and driver load will increase when pumping viscous fluids.

 

[TreeEng]

I am pumping a non-viscous fluid between two vessels. I would like to find the maximum possible pump-out rate so that I can determine the appropriate vacuum protection for the source vessel. Would using the point where the pump curve ends be appropriate for this? I have seen other methods which try to minimize the pressure drop across the pump to maximize the flow based on the process conditions/tank ratings. Thoughts on this?

 

[quark]

As TD2K suggested, I would agree for checking the dp across the pump (when the fluid in source tank is at it's highest level and discharge valve fully opened). This is the minimum dp prevailing in the system and there will be maximum discharge at this condition.

Note that with less dp discharge will be more and viceversa.
In my view, considering the discharge at the end of the curve wiil be highly redundant. This condition may occur when discharge is full open without any losses (just check what the dp is at the end of the curve, if it is marginal, you need not go upto this point)

I have a gut feeling that the higher volume in the source tank will offset damage due to pump out of liquid initialy. (i.e the external pressure of water vs crushing due to vacuum)

Good Luck,

 

[abcmex]

I will try to define your problem: You want to determine the maximum flow during the transfer between two vessels. With this information you will design an adequate venting on the source tank to avoid excessive vacuum.
First probably the worst case will be at the begining; the tank is full, you have plenty of NPSH.
Your pipe line has a pipe curve (flow vs pressure drop)and the pump has also a curve; make a grafic of them and see where they cross. The pump will probably operate at this point and you can design a proper venting. In very short lines you probably will need a orifice plate as restriction to avoid pump damages.
The pump curves end mostly where the NPSHr tends to be too large as TD2K mentioned already

 

[Montemayor]

TreeEng (Chemical):

I have done your application at plant engineering level, perhaps, hundreds of times. I have always documented my calculations and project MOCs and never found any challenges or questions raised. The method, assuming that we are dealing with a source vessel that is nominally atmospheric design (uses a conventional conservation vent/vacuum breaker), is as follows:

1. Read the "end-of-the-curve" gpm on the centrifugal pump's performance curve; agreeably, the pump will never get to this condition but this is a conservative estimate and the errors are not worthwhile debating.

2. Manufacturers have advised me that up to 10% of the performance curve (at the end) is theoretical, extrapolated, or ficticious; nevertheless, the amount of liquid volume converted to essentially the atmospheric air required to displace it is very, very small (relatively) and well within the capacities of nominally small vacuum breakers.

3. For example, a 1,000 gpm displacement of liquid causes a vacuum capacity need for 134 cfm or 8,000 cfh at essentially standard conditions. A vacuum breaker for that capacity is insignificant when compared to what it is protecting. It isn't practical nor cost-effective to do any further calculations or "engineering design" on its capacity. The answer will be conservative, will work, and will be the most cost-effective.

Unless you have a pressurized case or your configuration is other than what I've assumed, I would do no more than the above. The effects of NPSH, the varying pressure difference, the temperatures, the densities, etc. will not have any significant effect on the ultimate size of the vacuum breaker. You will ultimately install the nominal size that the manufacturer (such as Protectoseal) offers for the capacity range desired anyway. The manufacturer is not going to design a "custom" or special vacuum breaker for your application - at least I hope that's not the case. I've often come up with 2" to 3" sizes in most liquid transfer cases that have been more than ample in protection required. If we're talking about the difference between a 2" and 3" vacuum breaker, I don't think this is worth more conversation. However, if you're looking at a 14" to 20" vacuum breaker, we need more details and basic data.

Hope this experience helps.






Art Montemayor
Spring, TX

 

[TreeEng]

Thanks to all for their help, it is much appreciated.