Suction pipe design and impeller material

[stinems]

I need to install an pump to intermittently transfer boiler makeup water from a 6000gal tank (~8ft tall x 11ft dia). The bottom of the tank sits ~18ft below the fill line on the heating boiler, which is pressurized at ~18psig. I was envisioning the transfer pump (centrifugal) on the ground at the foot of the tank with a minimal distance between the tank and pump intake. Most pump designs I've seen have some kind of (eccentric) reducer in the suction line, maybe a few diameters before the actual pump. My understanding is that a reducer is typically needed because you normally want a suction pipe with minimal losses to conserve NPSHa, and so the intake pipe is oversized versus the actual pump intake. In this application I have plenty of NPSHa given very short intake and the water level in the tank will be 1 to 7feet above the pump intake centerline. Is there any reason for the reducer other than the one I gave? My tank has a 2" flanged outlet at the bottom; should I be looking for a pump with a 1.5" intake or will a 2" intake with no reducer do fine too?

Also, this boiler makeup water has some oxygen scavenger and other treatment chemicals that raise the pH a bit to maybe 9 or 10. Are there any pump material considerations (for impeller or casing) with this higher pH? Will a brass impeller/cast iron body work fine, or do I need to go with stainless?

Thanks,
Sam

 

[JJPellin]

For a very short run from the tank to the pump, a line-size suction nozzle (2”) on the pumps should be fine. You don't give enough details for us to fully evaluate this project. But I would offer some precautions. Even if you think you have plenty of NPSHa, still to the math and calculate the value. We recommend a minimum NPSH margin of 5 feet for water. Calculate the velocity through the suction line. It should not exceed 5 feet per second. If your 2 inch pipe is schedule 40, you would only be able to pump about 50 gpm. Consider pipe strain in the design. A short, stiff run of pipe from a tank to a pump may be good for NPSH, but it can result in excessive pipe strain. If the temperature of the water in the tank will change very much, the tank may move and create pipe strain even if none exists originally.

For boiler feed water with no oxygen, we would use cast carbon steel for both the casing and the impeller. But, your PH range looks like a potential problem. We would use 12-14 Chrome (CA6NM) for both the case and impeller if we felt that corrosion was a potential concern. We rarely use stainless steel, but I recognize that it is common in other industries.


Johnny Pellin

 

[GPRnD]

The erosion rate of carbon steel in deoxygenated water (less than 40 ppb), varies a lot depending on the PH. Higher PH is generally better.

For example at 60 ft/s velocity the erosion rate at 135 deg C (275 deg F) can vary from 0.6 mm/year (0.023") with a PH of 9.4 through to >10mm/year (>0.39") with a PH of 7. Adding any amount of chrome to the alloy significantly reduces this. CA6MN is an excellent (if more expensive) choice.

So to some extent it depends how tightly you expect to be able to control your water chemistry and how much life you are looking for out of the impeller (and high velocity casing areas).

 

[stinems]

Thank you for both of your replies. This pump will be cheap and simple, but I'm using this as an opportunity to learn. So I do appreciate your time and thoughts.

*The NALCO NexGuard chemical has a pH of ~13 according to its safety data sheet. It's of course very much diluted, though I do realize pH is a logarithmic scale. However, my impression is that the concentration level isn't tightly controlled by the ops group. They check it every so often and add a drip as needed. The water chemistry control is NOT what I would call "tight".

*This is an intermittent duty pump, it only operates to fill ~700gal (I'm told) between fill marks on the package boiler (1950's vintage Cleaver Brooks gas fired at ~5mmBtu/hr). Also, it's likely the system may end up drained over the summer. All of this should figure into impeller life. If we get perhaps 7-10 yrs out of the pump, that's probably OK.

*The tank will bring in fresh, softened water as makeup but it will also be mixed with condensate return (maybe 180degF??) in that same tank . Given the size of the tank, my gut says the water will have cooled down a good measure by the time it gets mixed with fresh makeup and the rest of the reservoir. I've been guessing 120degF at the most by the time it hits the pump. I use that as the point for vapor pressure, which is 1.7psia or 3.9ft H2O. Atmospheric pressure is of course approx 14.2 psia/33ft H2O. I've been assuming a minimum tank level of 1 ft above the intake, and negligible velocity head. At 40gpm, some rough calculations show <5ft H2O for intake pipe losses (absolute worst case). So 33ft atmos + 1ft min tank head - 4ft vap pr - 5ft intake losses = 25ft NPSHa. Most small centrifugal pumps I've been looking at show NPSHr between 7 and 12 ft depending on flow rate.

*Johnny-your point about thermal expansion of the pipe is well taken. We're running Sched 40 steel between the pump and boiler. I have some hesitation, but I'm wondering if Sched80 PVC would give us the bit of flex we need between the tank and pump. High side of the pump to the boiler will be Sched 40 steel.

*What is the significance of the 5ft/s limitation for the intake velocity? I've been designing for 50gpm which means ~5ft/s in 2" like Johnny said, but really 30-40gpm would work too.

Best,
Sam

 

[JJPellin]

NPSH margin may not be the only limitation to determine minimum tank level. I assume that the pump suction is a simple side nozzle with no vortex breaker or other internal features to break up a vortex. Hydraulic Institute standards would require a minimum submergence of 1 foot above the top of the pipe for every foot per second of velocity at the tank exit. So, no matter what the NPSH margin may be, you would need to keep five feet of water level in the tank above the top of the suction nozzle if you expect a maximum velocity of 5 ft/sec. Running any lower than that would risk a vortex forming and gas being drawn into the pump suction. I could not find a reference for the 5 foot/second suction line velocity limit. Hydraulic Institute requirements limit suction velocity to 6 feet/second at the exit nozzle from the tank. I am used to using a 5 foot/second limit, but do not have a reference for this number.

Johnny Pellin