Impeller trimed and Flow increased 4

[MIANCH]

Hi,
We have trimed the impeller of a centrifugal pump and after triming the impeller this pump flow has increased and load has reduced.
I'm confused with this as I know by triming the impeller will reduce flow and reduce load on motor.
Could any one share his experience with me if have this type of problem.
thanks

 

[BigInch]

How much did you trim it?

Sounds like you have significantly affected the curve, perhaps even too much. Load is essentially the product of head x flow.
Head will reduce with a trimmed impeller. If you reduce it too much, even more flow at such a reduced discharge head could easly result in a reduced total load.

 

[1gibson]

Was it at full diameter initially? If the impeller intruded into the volute area slightly, you could lose efficiency from friction there.

Also if a better underfile was performed, you could trim a little and potentially increase head.

If you trimmed because you were close to overloading the motor, it's possible that your speed increased.
If no speed increase, then for some reason your efficiency has gone up. Not sure I would classify that as a problem.

 

[Artisi]

Recheck your flow measurements / calibration of the measuring equipment.

 

[Artisi]

The only explanations that would make any sense to me, assuming what you are telling us is in fact correct and measurements are accurate is,
1. it is a poorly designed impeller and there is too much overlap of blades effecting the discharge off the impeller into the volute, trimming back on diameter may have overcome this problem, ie opened up the discharge area.
or
2. Trimming back the impeller has altered the discharge angle of the blade which has changed the flow characteristics.

A little more info might help- pump size and type, impeller diameter and trim amount.

 

[BigInch]

Might also be interesting to know precisely how much volume of metal was removed.

 

[rmw]

And, how are you measuring the flow? I have less confidence in flow measuring equipment than most other types of instrumentation. If you reduced your head significantly, and your Flow Element isn't pressure compensated, it may need recalibration.

And BigInch, can I get eye burn from looking at your signature?

rmw

 

[ccfowler]

MIANCH,

You have not mentioned the type of drive for your pump, and I suspect that you have no useful data on the actual shaft speed for your pump either before or after trimming the impeller. If you don't know the shaft speed accurately, it is very easy to make erroneous evaluations of pump performance.

I share rmw's doubts regarding the likely accuracy of flow measurement. It would be no surprise to find that instrument and test inaccuracies could indicate an increased flow rate when reality is just the opposite.

I've seen seemingly careful tests of pump performance carried out over a period of years with the intention of evaluating deterioration of pump performance due to wear, but review of the resultant data showed performance deterioration followed by performance improvement with no repairs or modifications to the pump. The only certain conclusion that could be drawn from the tests was that the pump was actually running, and it was performing tolerably well. The apparent variations were all within a range that was obviously due to the reasonable uncertainty inherent with the instrumentation that was used (all excellent quality and carefully calibrated). Simply stated, if your instrumentation is good for +/- 5%, don't try to conclude anything from an apparent difference of 3% between two tests.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[BigInch]

What did Mom always tell you.
Never look into the welding flash!
Don't stare at a solar eclipse!
Don't go outside when it's lightning!
Listen to your mother!

PPE is now required in this and a few other forums.

 

[Artisi]

and make sure you always wear clean underwear.

 

[MIANCH]

Hi,
Pump Data is here,
V=60m^3/hr
DeltaP=8.83 barg
Shut off pr=14.8 barg
BHP=34kw
Drive HP=39kw
Suction temp=355^0c
Service = Gas oil
Before triming pump was not able to pump more than 35m^3/hr and after triming pump is able to pump 50m^3/hr before the trip load that is 65 amps.
Hope you will make some more clear assumption which I can't do it for understanding this problem.
One more thing biginch asked about amount of material (weight) that idon't know because my company did this job with prevate workshop.
Thanks
MIANCH

 

[BigInch]

Inlet and outlet diameters?
Impeller dimension before trim,
and after trim?
Operating speed?
would help.

 

[katmar]

At that temperature could you have been having NPSHa problems before? Has the temperature changed? Any changes made to the suction or delivery piping at the same time as the trimming was done?

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[JJPellin]

I probably have a different perspective on this than most of the others. If this was a pipeline application, it might be that the pump is run with no control valve. This would be a case where the pump would run at the intersection of the pump curve and the system curve. Cutting the impeller down should reduce the pump head which would reduce the flow rate in this type of system.

However, I suspect you have a system that is more like the type I see in a refinery that use some sort of flow control. In these systems, they don’t run to the intersection of the pump curve and the system curve. They run to some constraint. The constraint might be a flow target set by the process engineers. It might be holding a constant level upstream or downstream. Or, it might be something like motor amps. If your operators have a meter to check motor amps, they might be controlling the flow to keep from tripping the motor off. If the flow is controlled by pinching on a valve, then this makes perfect sense. When the impeller was trimmed, the performance of the pump would change based on the affinity laws. As impeller diameter is reduced, the brake horsepower drops as the cube of the diameter. The pump head drops as the square of the diameter. So, if the operators were pinching down on a valve to keep from tripping the motor off, they would be able to get more flow after the cut because of the reduction in BHP.

I could describe this another way. With the larger impeller, you were building more head than you needed and dropping it down across a pinched valve. This is wasteful and inefficient. After the impeller was trimmed, the valve was opened further because the motor horsepower was lower, even at the same flow. Less energy was being wasted by the pinched valve. This energy would then be available to push more flow through the system.

Please confirm if the flow was being pinched back before and after the impeller was cut down.


Johnny Pellin

 

[1gibson]

"Before triming pump was not able to pump more than 35m^3/hr and after triming pump is able to pump 50m^3/hr before the trip load that is 65 amps."

From this, I gather that the limit on flow is due to overloading the motor. Pretty important piece of information if I read this right.

Trimming has reduced your head and power required at a given flow. You are now running further out on the curve before hitting your 65 amps. No mystery here.

 

[ccfowler]

MIANCH,

From the quantitative information in your second posting, it certainly appears that JJPellin's and 1gibson's responses provide the most coherent explanation of the situation. Such a dramatic variation in output could only be be explained by something external to the pump.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[BigInch]

JJ There's no difference in the system curve, pipeline or refinery, if the system curve includes the effects of the control valve, as it should, or whatever other flow element you have there, be it pipe, valve, pump, orifice plate, etc. Each piece is simply a flow element with a specific pressure drop (or increase, in the case of a pump). What the pump sees is only the sum total of all pressure drops in the discharge at the current flowrate, no matter what any given flow element happens to be called, be it pipe, reducer, fitting, half-closed ball valve, or a discharge control valve. Each flow element is simply a pressure drop as a function of its instantaneous flow. The system curve is just the sumation of the head loss of all downstream elements calculated for each and every flowrate the system could possibly see.

 

[btrueblood]

Ok. The OP only posted a single delta-P across the pump, and a single shutoff head figure. Those numbers should've changed when the impeller was trimmed, yet we have no data for that. Thus JJ assumes the guy has no idea what the installed head was before and after...not a bad assumption in my mind, and at least as likely as a reputable pump manufacturer selling an impeller so grossly oversized as to impede output flow and head performance. But, I'm making the assumption of the word "reputable"...

And we all know what happens when we ASSume.

 

[ione]

If the system curve had been changed after impeller trimming, by modifying any of the flow elements fitted in the discharge line, then the pump would operate at the intersection of the new performance curve with the new system curve (and affinity laws become meaningless). In this case it is possible to notice a flow increase as a combination of a trimmed impeller and a reduced pressure drop (system curve rotates clockwise).
IMO It seems strange the OP has not taken into account any discharge line modification

 

[JJPellin]

I should have learned by now to choose my terminology more carefully. Of course a complete system curve includes the control valve. But, the use of a control valve makes the system curve variable and not fixed. It is unlikely that the OP would neglect to notice that his downstream piping was changed at the same time as the impeller trim. But, it would be very easy to imagine that the operators would change a control valve position after the trim because the constraint they were running up against had changed. The real comparison I should have made would be between a system with a constant system curve and one with a variable system curve.

The other consideration I took into account is human nature. If the operators are charged with moving as much product as possible, they will tend to operate the pump up to some hard constraint. If the constraint is changed, they will tend to run up to the new constraint. If the motor trips, this is probably an inconvenience to the operators. They will tend to try to find a way to avoid this inconvenience. If they can do this by pinching on a valve, it would not surprise me that they would do so.



Johnny Pellin