Sizing for Smaller Pump Impeller 1

[RJB32482]

We have a centrifugal pump that is pumping too fast for our process. The medium is water and currently it is pumping at 635 GPM. We need it to pump 425 GPM of water to our reactor. Our maintenance staff wants to trim the impeller instead of just throttling back a valve in the discharge line (or a flow orifice). I discussed with our pump vendor a procedure to get the correct pump head vs. flow rate so they can give us a new impeller size. My thoughts were:

- Read the pressure gauge right after the pump at the current flow rate of 635 GPM
- Throttle back the ball valve after the pump AND after the pressure gauge until the flow rate to our reactor is 425 GPM (by local flow meter).
- Whatever that pressure gauge reads at the 425 GPM rate is the new head that we need for impeller sizing (at full open ball valve after pump)

The vendor is saying actually we need a gauge after the throttled valve to get the head we need to size a new impeller.

From your experiences, which one of us is correct? Can you please explain why if my procedure is incorrect?

Thanks in advance.

 

[StoneCold]

The vender is right about the pressure measurement. You need to know how much head the pump needs to put up at the new flowrate, and that is downstream of the valve, not upstream of the valve. Personally I would install a VFD and slow the pump down to get the exact flow. Because if you trim the impeller and miss you have problems. Also things change around here faster than the weather.

Regards

StoneCold

 

[SNORGY]

My recommendation is to obtain the Vendor's pump curves that show the family of characteristics for various impeller sizes and to plot the piping system head curve - as you currently understand it to be - onto that family of curves.

If maintaining the same pump differential and delivery pressures is not important (and I assume it isn't, based on what I read between the lines regarding the application), then an impeller trim by itself can get you where you need to be. Otherwise other adjustments in the overall system might be required anyway.

Regards,

SNORGY.

 

[RJB32482]

Thanks for the replies. Can one of you give me a technical reason why measuring the pressure after the valve is what I need and measuring it before the valve is incorrect? Need to understand this concept more about centrifugal pumps.

Thanks!

 

[Chance17]

Downstream of the throttle is a measure of "system requirements". It is the head pressure to push 400+ gpm flow through the system. As a minimum you must always meet system requirement.

Upstream of the throttle is "pump rating".
The pump rating has to meet or exceeds system.
If not you don't get 400+ gpm flow.

 

[JJPellin]

The pump will operate at the intersection of the pump curve and the system curve. If you take the pressure reading upstream of the pinched valve, you will be measuring the pump curve with the existing impeller trim, running in a system with a pinched valve. Neither of these is going to exist in the future system with a trimmed down impeller. You want to know the system curve with 425 gpm flowing and no pinched valve. That can only be measured downstream of the pinched valve. Once you know the system requirement at that flow rate, you will design the pump curve to cross at that operating point.

If your system is very stable with no major variations in pressure drop, trimming the impeller is the easiest, quickest and cheapest way to get what you want. But, if the system may be changing over time, I agree that a VFD would give you more flexibility. Based on what you have provided, I would agree that trimming the impeller sounds like a good option. If you have a very tight flow requirement (ie 435 gpm is OK, 450 is too much, 410 is not enough), then you may want to leave a little extra meat on that impeller at the first pass. You can always go back and trim more if you still get too much flow. In other words, if the calculation, based on the affinity laws, shows that a final impeller diameter of 8.450” is correct, I might trim it to 8.500” and test again. It is hard to put that metal back onto the impeller if you take off too much.


Johnny Pellin

 

[MJCronin]

I agree that a VFD seems to be the best way to go....

Capital expenditure justification can be made on the energy savings alone.

Additionally, with a VFD there is futute flexibility built-in if process conditions change in the future.

Go back to the pump vendor and see if he will give you a set of "VFD pump curves" with an operating range for your particular pump.

Not all pumps can (safely) operate in the range that a VFD can put out.

My thoughts only

-MJC

 

[RJB32482]

Thanks everyone for your input. Learned a lot after reading these posts.

I did look to the affinity laws and saw that flow and impeller diameter is a 1:1 ratio. So if I want to cut flow from 635 GPM to 450 GPM, I would need to cut the diameter of the impeller 70.9% (450/635*100). We currently have a 7.0" impeller in this pump, so the affinity laws are saying we need a (.709*7") = 5" impeller.

Am I doing the affinity law incorrect or does that look right to you guys?

Thanks again.

 

[JJPellin]

That does not look correct. The way you used the affinity laws would correspond to a smaller impeller running a lower flow at the same head. You will not be running at the same head. Once you take the measurement for the head required by the system at the lower flow, contact the manufacturer of the pump and ask them for the impeller trim needed to achieve that head at the lower flow.

You are still looking for the intersection of the system curve with the pump curve. The affinity laws can approximate the new pump curve. But, you still have to account for the system curve. I use a simple spreadsheet program to generate a new curve for a change in speed or impeller diameter. I have attached it. You can generate new curves for different impeller diameters until you find one that crosses at your new operating point.


Johnny Pellin

 

[Artisi]

Is the 7" impeller full size or is it already trimed?

There will be a minimum diameter to which you can trim - so be careful.

 

[SNORGY]

JJPellin:

Nice spreadsheet. Is it public domain? Can I request using a copy?

Regards,

SNORGY.

 

[JJPellin]

Anyone is welcome to use that spreadsheet in any way they see fit.

Johnny Pellin