Pump curve doesn't cross system curve 1

[EngMecBra]

Hi everybody,

Process team have informed me a rated head of 41m, then I have selected a pump.
Now I have been informed that the new head is only 15m.
The problem is that the pump has already been bought! Neither with minimum diameter impeller the pump curve crosses new system curve!
What kind of problem will occur in operation? Will these pumps operate?

See attached drawing.

Tks!

Rodrigo Alves

 

[77JQX]

Short answer - you need another pump. Longer answer - you'll probably burn out the motor due to overloading (depending on the pump particulars and horsepower curve for the pump (which isn't shown)). Even longer answer - even if you don't burn out the motor, you'll probably destroy your impeller in short order with cavitation.

 

[Compositepro]

A valve or other restriction can easily change your new system curve to be similar to the old curve. This will waste a lot of power, and buying a new, more appropriate pump, will be less costly in the end.

 

[LittleInch]

If you do nothing then your new pump will continue pumping until the pump curve meets the system curve. As this is beyond the end of your pump curve then you will probably exceed the motor rating or experience high flow and vibrations.

Your option is to instal some sort of flow control valve and accept the loss of energy or buy a different pump. See the "system curve" post below for a similar discussion on alternatives. You could replace the motor with a lower speed unit.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[chicopee]

About lowering the RPM or increase the resistance to the flow? Talk to the pump supplier about an exchange without reimbursement from the supplier.

 

[Artisi]

It's all been said, wrong pump for the duty, well wrong pump at the current speed.
Assume it is 4 pole 60Hz, check with the pump supplier or manufacturer and discuss 6 pole motor or post a pump performance curve of the pump showing H.Q,P,NSPHr - and it can be reviewed here by our intelligent, helpful and always ready to unravel the confusion members.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[mechengineer80]

Provide a recirculation line across the pump. But again there will be a lot of waste of energy.

 

[LittleInch]

Why use a recirc line?? Low flow isn't his problem, too much head is. All it needs is a flow restriction, orifice plate or control valve.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Artisi]

Come on guys, why encourage the OP to waste something like 12 or so kW using the wrong size pump. The cost's over a year paying for the excess power burnt off thru a control valve would probably pay a couple of times over for the cost of replacing the pump rated for the duty.

Of course if the OP has it fixed in his mind that the existing pump must be used then the effort is wasted in trying to be logical about the problem - so stick on a valve of some configuration and throttle it until you get the required flow.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[petrochemical87]

This pump needs to be changed with a pump which is more compatible with your system curve. Installing a flow control valve/orifice to achieve such a huge change in the system curve will cause a lot of energy wastage.

 

[BigInch]

The pump curve goes all the way to zero head. Extend the curve. Consequently the pump curve does intersect the system curve. Just not at the BEP. There is no reason why it won't run there. Perhaps not the most efficient operation, but maybe not all that inefficient, (some pumps get even a little more efficient to the right) and it will be maintenance intensive, but as long as the NPSH and power are available and the electrics are rated for it, it will run

http://www.slideshare.net/bin95/143-don-truncentrifugalpumpsonendofcurve

... for awhile. Let the OP evaluate whether he can afford the maintenance and early retirement, the pump's and maybe his own too.

What he didn't say was what flowrate he wants to run. Since he does say his new head requirement is 15 m, that would imply at 130 m3/h. SO, if he can slow the pump's speed down, he can certainly run at 130 m3/h with head reduced to 15 m.


Independent events are seldomly independent.

 

[DRWeig]

Process team have informed me a rated head of 41m... Now I have been informed that the new head is only 15m.

Did the process team give you that 41m requirement in writing or an e-mail? At my company, they'd be getting their paychecks dinged to pay for a new pump.

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[BigInch]

Let me change, "some pumps get even a little more efficient to the right", to
"Some pumps don't lose so much efficiency when moving to the right"

Independent events are seldomly independent.

 

[QualityTime]

This is really not that big a pump. Why don't you check with the pump manufacturer for the correct pump and see if you can reuse the motor.

The other alternative is to see if a slower motor speed can do the job. Standard motor speeds are 3600 rpm, 1800 rpm, 1200 rpm, 900 rpm 600 rpm. The motors get bigger in physcial size the slower the rpm.

You might be able to install a vfd to decrease the speed. But you have to change the motor to be inverter duty rated.

 

[CaracasEC]

This will be a troublesome pump if operated on the new process requirements...

 

[TenPenny]

The fix depends on your timeline and priorities.

First off, looking at that curve, I doubt the pump will do what you want by slowing it down, by the time you get it slow enough (even with a trimmed impeller), your flow rate will be beyond runout.

So, the quick and easy solution is, as suggested above, throttle that baby to add in enough head to bring it onto the pump curve.

The 'proper' solution is to get the right pump. Depending on your pump supplier and what model and size pump, you may be able to convert it to a smaller pump relatively easily, or if your pump supplier is a good partner with your plant, you might be able to work a good deal to trade for the correct size pump. How much piping, foundation work, etc is done?

 

[BigInch]

slowing down from say 1800 rpm to 77% rated speed, around 1400 rpm, will get about as close as you can get. Not too bad. Not too bad at all. See new curve at 77% rated speed attached.

Independent events are seldomly independent.

 

[QualityTime]

BigInch could you post it in pdf format. I can't open the file. You put a big effort into this.

EngMecBra should now go to the pump supplier to get pump curves developed for different speeds. He will now see with accuracy where the system head curve will intersect with the pump curve. His alternatives to slowing the speed down is to put on a sheave and pulley or install a vfd. VFD's do have the advantage of unlimited number of starts and stops on the motor

 

[1gibson]

VFD is not at all a bad idea. They are not magical energy savers, but in this case it should work out much better than throttling.

I don't think the problem is that the pump has already been bought, that can usually be addressed when the difference is reasonably small (trim the impeller a little bit.) The problem is some major change or oversight that was discovered very late. I'm curious what that could have been.

This looks way different than the usual "safety factor on top of safety factor" pump oversizing. Based on that system curve sketch, it looks like elevation is the main component that they changed on you (or, I am putting too much faith in the sketch, which I think is very possible.)

What are some of the system details? Did somebody put parts of the system at a different elevation than originally planned? Did someone forget to consider suction pressure at the pump, and size it for discharge pressure instead of differential pressure? I know I'm not the only one who is curious.

 

[BigInch]

Not a whole lot of effort.
New flow coordinate is Old_Flowrate * new_rpm / old_rpm
New head coordinate is Old_Flowrate * new_rpm[sup]2[/sup]/ old_rpm[sup]2[/sup]

Independent events are seldomly independent.