Acceleration losses in Diaphragm Pump running in Parallel

[bmw318be]

Hi,

I am studying a case of the metering pump running in parallel with the following service

1. 1056 GPH
2. 3 Pumps running in parallel at the same time
3. Suction line is about 40 Meter:
Section 1: From tank outlet nozzle is 3", 16 feet
Section 2: From 3" to common header is 6", 98 feet (This meant to be sharing suction header)
Section 3: From 6" to pump suction is 2", 7 Feet

4. Atmospheric Tank
5. Stroke Per Minutes is 120 SPM


Using the metering pump manufacturer guides, the Acceleration losses through all Pipe section is 30.73 PSI or 2.11 Bar or 21 meter This based on 1056 GPH.
Link

https://www.dropbox.com/s/w0onpq84u...GLE-PUMP-APPLICATION-Rev-B 1056 GPH.xlsm?dl=0

If i running the pump in parallel, shall we considered 1056 x 3 = 3168 GPH Total Flow, this turned out to be very large acceleration losses of 92.19 PSI or 6.35 Bar or 60 meter.
Link

https://www.dropbox.com/s/atmuf9fsc...GLE-PUMP-APPLICATION-Rev-B 3168 GPH.xlsm?dl=0

For this types of pumps, the NPSHr = 5 meter so the NPSHr < NPSHa

Can I get your opinion of the following

1. if the losses shall be multiple by 3 ? would install the inlet dampener bladder help to reduce the acceleration losses by 3 to 4 psi according to Blacoh inlet stabilizer.

https://www.blacoh.com/downloads/literature/Solving_Pump_Inlet_Problems_P21E11_010.pdf

Would the calculated if dampener was installed, we just need to calculate the lines losses after the dampener

2. Currently maximum pipe size is considered 6 " with the space constraints. Bringing the pump closer is also not feasible as this is the retrofitting jobs.

P& ID layout

https://res.cloudinary.com/engineer...ketch_4000_LPH_3_pumps_in_parallel_lyijhy.png

]

 

[LittleInch]

That spread sheet calculation loks very high for what is a realtively low flow and very low velocity. The only thing I can see out of place is the diameter of the piston. If this is a single piston then its stoke is over 8 inches.

Looking at the Milton pump website this size pump at that SPM is only good for 35-40 GPH [edit] (depends on the exact model)

So everything else becomes out of range very fast.

Please go back and check that the model you entered in the vendors spread sheet can actually do the GPH [edit] you want as I don't think it can by two orders of magnitude.

If you enter the raw data into the equation you attach you get about 5 ft of head so about 2 psi for the 6" section.

But as said above your pump you entered looks wrong

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Latexman]

LI,

OP is talking GPH and you are talking GPM. Is that the problem?

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[bmw318be]

Hi Sir Littleinch,

The formula of acceleration losses does not account the types of pump.

You can take a formula from

http://madhavfluid.com/entries/general/acceleration-head-loss

The variable dependent on the

Q, flowrates ( i am not sure if i shall total the pump in parallel or just considered 1 pump). Both would be high since this is long suction and high flow
.

http://madhavfluid.com/entries/general/acceleration-head-loss

ha=LVnC/Kg

Where

ha = Acceleration head loss in feet

L = Actual suction pipe length in feet

V = Mean flow velocity in suction line in feet/sec

n = Pump speed in cycles per minute

C = Constant (Depends on pump type)

K = A factor representing the reciprocal of the fraction of the theoretical acceleration head which must be provided to avoid a noticeable disturbance in the suction line

g = Gravitational constant (32.174 ft/sec2)

V= 0.0485 * Q/D2

Where:

V = Mean flow velocity in suction line in feet/sec

Q = Flow rate in gallons per minute

D = Pipeline inside diameter

Constant

Pump Type

0.200

Duplex single-acting (diaphragm pump)

0.115

Duplex double acting

0.066

Triplex (single or double acting)

0.040

Quintuplex (single or double acting)

0.028

Septuplex (single or double acting)

0.022

Nonuplex (single or double acting)

It shows the acc losses would be very high die to length and flow total

 

[LittleInch]

Latexman - My error now corrected. I did actually mean GPH. These are metering pumps where you're supposed to be pumping low numbers (<10) of litres per minute. 1056 GPH translates to 70-80 litres per minute. That's not a metering pump.

See the B type pumps which were used in the spread sheet...

http://www.miltonroy.com/wp-content/uploads/PD3641-Milroyal-B.pdf

Wrong input date = wrong answer

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Latexman]

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[LittleInch]

bmw,

Do the calculation according to your equation and let me know what you get.

I got 5 ft head. Now maybe I made a mistake in the numbers - don't know.

1056 GPH is a very low flow in a 6" line.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[bimr]

The rule of thumb for sizing metering pump piping is to size the piping for a flow of 3 times the pump capacity.

 

[bmw318be]

Hi Little inch Sir,

I got much higher using the theoretical formula above: We split the pipe into 3 different portion

3", 6" and 2"

Link below is the excel calculation

https://www.dropbox.com/s/1e9jbhydgpmzd0h/Calculation 3 section P.xlsx?dl=0

Hi Littleinch,

The metering pump does meter such a very high flow, please refer to attached in highlight red. Maximum capacity is 4203 LPH or 1110 GPH

https://www.dropbox.com/s/jugw051jg5j1mxj/primeroyal N PLASTIC MAX 4200 LPH.pdf?dl=0

I like to confirm the acceleration losses would be so much as per calculated.

 

[georgeverghese]

What type of pump is this - duplex single acting ?

The tricky part is what value for the constant C have you used when all 3 pumps are running (for the acceleration head loss calc) at 3168gph? It would be lowest when all 3pumping heads are out of sync and highest when they all are in sync.

Cannot see these calcs with this dropbox software..

In my experience, taking credit for a pulsation dampener purely to reduce acceleration head loss is a bad idea - it is much better to go for multihead pumps (minimum of triplex double acting) if you really have a NPSH problem.

 

[LittleInch]

bmw

You're using different factors and numbers to me

I) The equation you quote state n is no of CYCLES. You've use strokes. There are two strokes per cycle hence the figures you have in your spread sheet need to be halved
2) You use a c factor of 0.4. The largest number in the list you supplied was 0.2, but Ok 0.4 may be applicable
3) The pump you highlight is not the pump that was listed in the calculation supplied. Those 1" pumps are much smaller than that monster of a metering pump you've chosen

In your initial post you say a pulsation damper reduces the pressure "by" 3-4 psi.

The text in the attachment states it reduces it "to" 3-4 psi.

Not that it needs to be pretty big ( 15 x volume of pumps liquid chamber) The pump you quote is 500 cc so you need a 75 litre damper.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Latexman]

bmw318be,

There are still unit conversion issues. 4203 LPH = 1110 GPH, not 110 GPH.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[bmw318be]

georgeverghese (Chemical)31 Jan 18 06:46
What type of pump is this - duplex single acting ?

Single acting
The tricky part is what value for the constant C have you used when all 3 pumps are running (for the acceleration head loss calc) at 3168gph? It would be lowest when all 3pumping heads are out of sync and highest when they all are in sync.

I am using 0.4 simplex as this 3 pumps in parallel, for simplicity, I like to calculate the furthest pump running which is along
. Suction line is about 40 Meter:
Section 1: From tank outlet nozzle is 3", 16 feet
Section 2: From 3" to common header is 6", 98 feet (This meant to be sharing suction header)
Section 3: From 6" to pump suction is 2", 7 Feet

I attached the tabulation using the formula in excel above

https://www.dropbox.com/s/1e9jbhydgpmzd0h/Calculation 3 section P.xlsx?dl=0

LittleInch (Petroleum)31 Jan 18 10:05
bmw

You're using different factors and numbers to me

I) The equation you quote state n is no of CYCLES. You've use strokes. There are two strokes per cycle hence the figures you have in your spread sheet need to be halved

Can elaborate why halved, the metering pump would run in Stroke per minutes which is 140 SPM
2) You use a c factor of 0.4. The largest number in the list you supplied was 0.2, but Ok 0.4 may be applicable
Noted
3) The pump you highlight is not the pump that was listed in the calculation supplied. Those 1" pumps are much smaller than that monster of a metering pump you've chosen

Yes, the calculated acceleration losses would still be the same as this factor of the Stroke per minute, flow and Length, pipe size

In your initial post you say a pulsation damper reduces the pressure "by" 3-4 psi.

The text in the attachment states it reduces it "to" 3-4 psi.

Yes, my type, is it necessary to be so much, because the nitrogen is compressible, the recommended is the bladder types which normally sized according to the size and as near as possible.

Not that it needs to be pretty big ( 15 x volume of pumps liquid chamber) The pump you quote is 500 cc so you need a 75 litre damper.

Can you confirm if the Losses along the 40 meter is correct to be

LittleInch (Petroleum)31 Jan 18 10:05
bmw

You're using different factors and numbers to me

I) The equation you quote state n is no of CYCLES. You've use strokes. There are two strokes per cycle hence the figures you have in your spread sheet need to be halved
2) You use a c factor of 0.4. The largest number in the list you supplied was 0.2, but Ok 0.4 may be applicable
3) The pump you highlight is not the pump that was listed in the calculation supplied. Those 1" pumps are much smaller than that monster of a metering pump you've chosen

In your initial post you say a pulsation damper reduces the pressure "by" 3-4 psi.

The text in the attachment states it reduces it "to" 3-4 psi.

Not that it needs to be pretty big ( 15 x volume of pumps liquid chamber) The pump you quote is 500 cc so you need a 75 litre damper.

Latexman (Chemical)31 Jan 18 10:28
bmw318be,

There are still unit conversion issues. 4203 LPH = 1110 GPH, not 110 GPH.

Thanks, typo

 

[LittleInch]

The metering world may work in strokes per minute as many pumps are multi cyclinder, but the equation requires cycles per minute. Two stokes for a simplex pump equals one cycle or RPM.

A check also appears to be to look a the steady state pressure losses at three times the nominal average flow rate as bimr states. If you've got a big discrepancy then you need to look at your inputs.

This site gives an alternative equation in metres head, RPM and volume per second. I get about 3.4m for the 6" section so maybe 10 ft head.

http://www.ruhrpumpen.com/downloads/products/reciprocating-pumps/135-rdp-technical-paper-npsh/file

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[bmw318be]

Hi Littleinch,

"A check also appears to be to look a the steady state pressure losses at three times the nominal average flow rate as bimr states. If you've got a big discrepancy then you need to look at your inputs"

Does it means the losses divided by 3. Since what i quoted the theory of metering pump has 3 times pulsating flow at the discharge, does it means that the rated flow losses would be compensated by pulsating flow.

Or the losses would be more knowing the flow which is Q in formula is 3 times ?

Sorey to ask , What is bimr states ?

May i know if your calculated has considered the 3 section of the pipe which 1st section is 3" pipe lines. And section 3 is 2 inch to the pump.

If this is considered the losses would be greater isnt it ?

Thanks for valuable inputs and discussion

 

[LittleInch]

What bimr ( one of the posters above) and written in a few guides on this subject say is that if you take a flow value of three times the nominal flow rate, the pressure / friction loss you get is similar to the acceleration head.

So in your case, find the frictional loss in the system which is only for one pump for 3168 GPH (3 x 1058) at a steady state flow and this will be the acceleration head for an average flow of 1058.

To find the acceleration head for three pumps in parallel multiply the total flow by three ( just a coincidence that three turns up twice), i.e. find the steady state losses in the common pipe section for ~9500 GPH ( 3 x 3168).

I just looked at the 6 inch section, but it still shows a reduction.

I think you need to start looking at inlet stabilizer option.....


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[bmw318be]

Hi littleinch,

It makes technical sense the pulsating flow. But in the design, we would be having pulsation dampener after the pump. I would expect the pulsating flow would not be a major.

Assuming discharge dampener effectiveky reduced the pulsating flow, can i day the theoretical calculation of suction acceleration losses for 1 pump is only 1058 GPH ?.

It still high knowing that we have not factored in the NPSHr of the pump which is around 5 meter.

If we have open tank of atmosphere which is 10 meter, we left 5 meter for acceleration losses.

I am wondering if upsizing to 10 inch suctuon would help?


I read some article the inlet stabilizer would be at least 2 times discharge and lower than atmospheric. Would this be useful in almost eliminating the acceleration losses ?
.
Shall we install 1 for 3 pumps or 1 pump each ?

 

[georgeverghese]

Think using 0.4 for C isnt appropriate here - it should be somewhere between 0.2 and 0.066 when all 3 duplex single acting pumps are in operation - I'd use 0.2 for the worst case when all 3 units are rototating in sync. As mentioned earlier, use of a suction pulsation dampener is not workable here since the gas charge pressure has to be 0.7x that of the source vessel pressure, from my memory on vendor design guidelines. 0.7x atmospheric pressure on the N2 charge is not possible.

 

[LittleInch]

A 10 inch suction line will help a lot and may even avoid the need for a stabilser. Your issue is more with the 3" common section than anywhere else.

The article you linked to from blacoh does state that inlet stabilisers are now available designed for inlet below atmospheric

If you need one I would sixe one for the 3 pumps running case.

Maybe you could fit a booster pump much closer to the tank?

You are a long way from the tank so need to be careful.

Why can't you use a multi-cylinder / duplex or triplex pump?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[bmw318be]

georgeverghese (Chemical)1 Feb 18 03:23
Think using 0.4 for C isnt appropriate here - it should be somewhere between 0.2 and 0.066 when all 3 duplex single acting pumps are in operation - I'd use 0.2 for the worst case when all 3 units are rototating in sync. As mentioned earlier, use of a suction pulsation dampener is not workable here since the gas charge pressure has to be 0.7x that of the source vessel pressure, from my memory on vendor design guidelines. 0.7x atmospheric pressure on the N2 charge is not possible.

Hi, what guidelines of 0.7 atm and therefore such long suction is not possible. Can you elaborate. Correct me.if i am wrong javing the dampener on the suction is just like a vessel of tank sit before the pump. The accelerations are eliminated.

Hi littleinch,

The common suction line would be probably 6" the maximum. From the common suction to the pump would be the same as pump 2".

(Quoted)

Suction line is about 40 Meter:
Section 1: From tank outlet nozzle is 3", 16 feet
Section 2: From 3" to common header is 6", 98 feet (This meant to be sharing suction header)
Section 3: From 6" to pump suction is 2", 7 Feet

Having a bigger suction would not work and theoretically if 3 times in parallel synchronizing without dampener, the acceleration losses still high with NPSHr of 5 meter.

So acceleration losses would need to be kept aeound 2 meter.

Having booster pump would not allow in the design as there are total 8 tanks.

What we need the suction dampener which i like to study the feasibility of this.