HOW TO CALLCULATE FLOW RATE OF PUMP WITHOUT USING ANY APPARATUS? 3

[tangochem]

I have a centrifugal pump working on suction lift for cold water at 20 C. its specs are as follows.

Suction dia = 5 inches
Discharge dia = 4 inches
Motor RPM = 1460
Motor Power = 20 kw
Impeller dia = 14 inches


from these specs I calculated head and pressure of the pump as follows:
1- Peripheral velocity of vane =
=( RPM * Dia of impeller in inches) / 229
= (1460 * 14) / 229 = 89.26 ft/sec

2- Now the head at shutoff = (Peripheral velocity)2 / 2g
= (89.26)2 / 2*32.2
= 123.7 foot of head.
[Ignored all the friction losses and suction lift
that is not too high]

3- The discharge Pressure is then calculated as = Head * 0.433
=123.7 * 0.433
=53.57 Psig

So from such values, how I can calculate Flow Rate of pump just with calculations without using any apparatus

 

[SNORGY]

That's an interesting problem. I am not a "pump guy" and I haven't checked your calculations or methodology in detail, but I am pretty sure that without making several more unsubstantiated assumptions (for efficiency, rise to shutoff, head at run-out), you will not be able to predict the flowrate from equations alone.

At least, not easily.

My first instincts would be as follows:

1. Assume your head-capacity curve does not cross over and above your theoretical hydraulic power curve.
2. Assume a % rise in head to shut-off from BEP. Some folks like to use 25% but I think it's optimistic. My gut feel would suggest 10% to 15%.
3. Assume a maximum pump efficiency. For centrifugals, in the absence of data to the contrary, I would personally assume 75%.
4. Subtract 15% from the shut-off head you appear to have already calculated and apply 75% efficiency and full power (bhp) at that point. Apply the equation P = Q*H*SG/(3960*n), algebraically rearranged to solve for Q. You now have two points on your pump curve defined: shut-off and BEP.
5. Assume a concave down parabola with vertex at BEP and crossing the "Q" axis at (0,0) for the pump efficiency curve left of BEP. Derive it's corresponding equation.
6. Assume a concave-down parabola with vertex at BEP and crossing the "Q" axis (n=0) at a point ~5*(Q@BEP) for the pump efficiency curve right of BEP. Derive its corresponding equation.
7. Take the efficiency corresponding to ~3*(Q@BEP) and choose this as your run-out condition. Apply the equation P = Q*H*SG/(3960*n), algebraically rearranged to solve for H.
8. Find an equation in the form of Q=k*H^2 for a simple concave down parabola with vertex at the pump shut-off point that passes through the three points defined above.

Now, this is in no way founded in theory or textbook. It's just what I personally would do to "rough-in" something that might come approximately close to modelling a typical centrifugal pump if I was given nothing more than what you have to work with. I might be close; I might be out to lunch. In any case, you would have to then go back and check several other points along the locus of your hypothetical curve to validate the rather bold assumptions made, particularly at Steps 1 and 2. With a spreadsheet, you could tweak and tune things to get a fit.

The only true way to figure things out is to get the as-tested pump curve.

Regards,

SNORGY.

 

[SNORGY]

Hmmm...

So double checking my last post with some representative pump curves (maps) by typical vendors, showing the "eyes" (BEP) of the curve...

6. Maybe ~4*(Q@BEP) would be a better assumption for n=0 right of BEP.
7. Maybe ~2*(Q@BEP) would be a better "run-out" point to assume.

Like I said...it's something that would need some fiddling and experienced judgement to arrive at something make sense.

Regards,

SNORGY.

 

[Pumpsonly]

There is no way to estimate the pump performance to any decent accuracy with those data. Contact the manufacturer and get the performance curve.
If you failed to contact the maker, the best you can do is get more operating data.

You can at least eliminate one uncertainty by putting suction vacuum gauge and discharge pressure gauge on the pump nozzles to find out the actual differential head.

You will also need to measure the motor current and make an estimate of the power out put by comparing with the motor full load current.
If it more than 80% of the of the full load, the accuracy is much higher.

Then use the basic formula Kw = Q(m3/h)x H (M)/367/ pump effcy.
Here come the difficult part in guessing the pump efficiency.
For the size of the impeller and nozzle size, it would be some where between 60 to 70%.

Good luck..

 

[Pumpsonly]

One more point You will have better luck posting it to the Pump Engineering forum

 

[tangochem]

Well! "SONERGY" Following your suggestion, I have almost done but here another question arises. If i check the current drawn by Motor with the help of an ammeter, do you think that it will help me to predict the efficiency while the motor is designed for 20KW?
Further, how this process will help me to estimate actual bhp?

I myself is not very good at pumping system.But my General Manager has assigned me to calculate the flow rate on the behalf of some basic data which i have already mentioned in the question

Profound Regards!

 

[SNORGY]

Knowing the current draw will help you but, unfortunately, since both efficiency and flow rate are not known, you won't know the actual work being done by the pump.

Can you rent a Doppler meter to measure pipe flow? I used to do that years ago (Polysonics Doppler Meter; it came in a foam case with the pickups, clamp, grease and chargeable battery for about $100.00 per day). If you are measuring current draw anyway with an ammeter...

Regards,

SNORGY.

 

[SNORGY]

For example:

http://www.techrentals.co.nz/Produc...SX30-Dual-Frequency-Doppler-Liquid-Flowmeter/

Regards,

SNORGY.

 

[DubMac]

Unfortunately, unless this is an academic exercise designed to enhance your theorizing skills, there is no way to come reasonably close to estimating flow with given data.

You have proposed a good start based on limited data, but the one most important piece is missing: specific speed of impeller.

A five gallon bucket and a stopwatch will yield much, much more accurate results.

 

[SNORGY]

DubMac:

I did that in the field once. It is indeed effective.

The other engineers involved got mad at me. But...it was what it was.

Star to you.

Regards,

SNORGY.

 

[tangochem]

well! dubmac! u r quite right. I had performed this method on the the condensate outlet of a tire curing press for measuring the steam flow rate.It gave me better results. But it's not easy for me to use stop watch according to the current situations in the factory

Yes! unless I have data from manufacturer, i cant determine the specific speed of the Impeller. you know! in my parallel pumping system, there are two pumps with different impeller dia and rpm of motor but same dia of suction and discharge ports.

So I liked to make the head same by using the following calculations

Pump: 1

Dia: 14 inch
RPM: 1460

Solution:

1- Peripheral velocity of vane =
=( RPM * Dia of impeller in inches) / 229
= (1460 * 14) / 229 = 89.26 ft/sec

2- Now the Diff. head at shutoff=(Peripheral velocity)2 / 2g
= (89.26)2 / 2*32.2
= 123.7 foot of head.
[Ignored all the friction losses and suction lift
that is not too high]



Pump: 2

Impeller Dia: 14 Inches
RPM: 1400

Solution:

1- Peripheral velocity of vane =
=( RPM * Dia of impeller in inches) / 229
= (1460 * 10) / 229 = 61.135 ft/sec

2- Now the diff head at shutoff =(Peripheral velocity)2 / 2g
= (61.135)2 / 2*32.2
= 58.03 foot of head.
[Ignored all the friction losses and suction lift
that is not too high]

In order to make the heads of Pump: and Pump : 2 similar, i will have to increase the RPM for Pump: 2
Therefore

Pump : 1 = Pump: 2

RPM(1) * Imp. dia(1) /229 = RPM(2) * Imp.dia(2)/229

1460 * 14 = RPM(2) * 10

RPM(2) = (1460 * 14)/ 10 = 2044



Now: if we run the pump : 2 with 2044 RPM the calculations will be as:

1- Peripheral velocity of vane =
=( RPM * Dia of impeller in inches) / 229
= (2044 * 10) / 229 = 89.26 ft/sec

2- Now the Diff. head at shutoff=(Peripheral velocity)2 / 2g
= (89.26)2 / 2*32.2
= 123.7 foot of head. (equal to pump: 1)
[Ignored all the friction losses and suction lift
that is not too high]


But here is the problem of SUCTION SPECIFIC SPEED. Now i need manufacturer's data or i will have to take risk to purchase such motor with 2044 RPM or VFD

regards!

 

[tangochem]

sorry: pump (2) impeller dia is 10 inches not 14 inches

regards

 

[Pumpsonly]

tangochem,
It seems you are more interested to do an academic exercise as suggested by DubMac rather than trying to find out the flow rate.

 

[tangochem]

Thanks to all of you!
Regards!

 

[tangochem]

Hello!

I have 22 bar (discharge pressue) Multistage HOT WATER pump ( Temp. of water is 165 C) in the factory. But now i want to upgrade pressure up to 30 Bars. Please tell me the efficient and the most economical way to enhance the pressure?

Regards!

 

[maddocks]

This really should be a new thread in the pump forum. However, the best way to increase discharge head is to:

a.) approach the pump vendor for a re-rating
b.) Put a VFD on the pump and speed up the pump and drive the TDH higher - this needs vendor verification as it will alter the pump curve, increase flow, power and impeller speed. Other bad things may come out of this.