Pump discharge line size calculation 4

[Asisraja D]

Hello professionals
Thank you for going to help me.
We have pump with capacity of
180 m3/hr , head = 45 meters.
Pump nozzle size is 100mm x 80mm
But in our plant we already installed piping for suction with 150mm inlet piping and discharge with 250mm outlet piping to the plant
(This is cooling water supply and return line both are the same size as 250mm)
So I calculated the velocity method (Q=A X V) with the known pipe size and flow rate I get 1m/s.

Few more details for your reference:

1.Pipe material is Mild steel "C" class pipe
2.Pipe length from cooling tower sump to plant supply (250mm) EOL is 80 meters.
3.Pipe length from Main header to sub-header supply (200mm) EOL is 30 meters.

My doubt is all about discharge nozzle size with discharge piping size.
Did we take wrong piping size for discharge line?
Is there any rule of thumb or design considerations for pump piping as applicable to liquid flow.

I have attached my hand sketch for more clarity.

 

[Asisraja D]

I have attached the pump specifications sheet with this.

 

[pierreick]

Hi,
1 m/s velocity is too low for water!
Pierre

 

[pierreick]

Hi,
Some guidelines for velocity in pipes.
Pierre

 

[Asisraja D]

pierreick hm this is really too low velocity
Thank you sir.

 

[LittleInch]

"Did we take wrong piping size for discharge line?
Is there any rule of thumb or design considerations for pump piping as applicable to liquid flow."

Piping size is always a compromise between lower cost of the pipe material /welding / fittings vs increased pumping costs over time.

For pipelines, say >10km long, a start point for velocity is about 2m/sec which has proved over time to be the correct balance between low pipe cost and low pump cost / energy usage. Many people prioritise CAPEX over OPEX, but if you add in lifetime energy costs, you often find you go up a pipe size and lower velocity.

Piping on the other hand a higher velocity - prob around 3m/sec, is usually most economic due to the high cost of fittings (bends, elbows, tees, flanges and most importantly valves of which there are a lot more on piping than pipelines.

You normally want to flow at 1m/sec minimum as this velocity will blow your bubbles / air pockets out. Less than 0.5 you could get air building up at high points and dirt at low points.

At 180 m3/hr my initial guess would have been a 6" / 150mm pipe, possibly 8" / 200 NB if the run was quite long. 130m isn't long, so yes the 250mm looks a bit too big, but your energy bills for the pump will be quite low. If this works 24/7/52, then you might get payback for the bigger pipe in 5 -7 years. You normally want to go up at least one if not two pipe sizes compared to the discharge flange size of the pump and at least one pipe size higher for the inlet / suction line, especially if its at low pressure / gravity fed.

however I guess now your pump was sized based on a low frictional loss in the bigger pipe, so you can't change it now without changing the pump for a more powerful one.

but this is why you do design concepts to find the most appropriate size for your particular circumstances. Any numbers are just an initial guide. There are no fixed rules saying it must be this velocity or not more than say 3m/sec. You sometimes see lines running at 5-6m/sec if it makes sense. Surge can be an issue over 2.5 to 3m/sec, but just look at it.


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

 

[FluidPowerUser]

I agree with LittleInch's comments.

In the world of fluid dynamics, the general saying is the pressure is cheap and flow is expensive. That is certainly true where moving large volumes of fluid comes into focus.

With 45m of head available, the primary objective is to minimise the head loss between the pump and the cooling tower and if you have a need for pressure to supply sufficient flow to the equipment, then head loss prevention is critical.

It is assumed that you have done the calculation to check for head loss from friction losses and bend losses etc. With all of the losses due to friction, fittings, bends and drop legs etc, if the total head loss is too high, then you naturally have to consider larger diameters, with the data available to justify the decision. After all, larger diameter means heavier supports and brackets and the CAPEX gets high. If you go with marginal flow performance and have turbulent flow and higher pressure at the pump outlet, the OPPEX is going to be higher and the higher input energy will likely shorten the working life.

 

[Asisraja D]

LittleInch sir I hope it will not affect our process requirement now due to low friction loss. We have less piping distance than you mentioned >10km. As you told "There are no fixed rules saying it must be this velocity or not more than say 3m/sec". We should consider our velocity need (PIPING COST). This is really sense so I can reduce my velocity even more little upto 0.5m/s so I can get the same flow rate but the problem is the piping cost and other fittings will increase. Finally I can get the same flow rate ay EOL with this 1m/s velocity but with high cost in CAPEX.

 

[Asisraja D]

FluidPowerUser sir In the world of fluid dynamics, the general saying is the pressure is cheap and flow is expensive. As you told we have achieved our desired pressure to omit head loss but with high energy and piping cost. Yes it is really lot of materials and energy cost involved in this. I still have to understand some deep concepts in this. I will e-mail this thread. I hope it will definitely give clarification. Thank you sir .

 

[FluidPowerUser]

Have you calculated the head loss in the system so that you can make a justified selection of the bore size?

If there is no requirement for heat exchange, then go as slow as you can to keep the head losses to a minimum. How deep do you want to go in terms of the concepts?

 

[Asisraja D]

FluidPowerUser sir we didn't calculate head loss and this head 45mtr is taken by our managing director (he is also not sure about this design because I did conversation with him so he is not clear about this design and he took it based on his experience)

We selected bore size based on cross sectional area calculation because we would have to take few branches from the main header so the 250mm pipe meets our requirements so we proceed with this size that's it apart from the factors there is no calculation is involved in this selection.

If there is no requirement for heat exchange, then go as slow as you can to keep the head losses to a minimum.

Yes this is for heat transferring purpose for 33kl fermenters. But please clarify me here in this scenario I have 1m/s velocity but same flow rate

1.)so flow rate doesn't depend on velocity?
(In my understanding)
2.) If I want to reduce head loss should I need to select bigger pipe size with higher head ?
(Due to large area of pipes the fluid can flow easily without so much accumulation in small pipe area )
3.) So there is no rule for selecting one size bigger or two size bigger than pump discharge nozzle for piping ? (Like rule of thumb like that it is purely based one our needs like pipe cost and material cost , valves , fitting , space , pipe racks etc)

Can you please explain me sir

 

[FluidPowerUser]

Have you selected the pump performance - (180m^3/hr at 45m head) based on the heat transfer requirements?

Can you change the pump if your calculations show insufficient flow or head?

What is the water return temperature at the pump inlet expected to be? This is for the NPSH requirements and the vapour pressure of the water. Your spec sheet for the pump is showing you need to have 4.2m at the inlet. If the water is cool, then temperature won't come into the equation so much, but if you are expecting hot water at the inlet, then the inlet pressure requirements may change. I know there is a cooling tower, but it's all relative and it's important to be specific about the temperatures in the system.

Are you planning to use the by-pass valve to create a pressure drop in the ring and effectively force flow into the 8" line and through the equipment to provide cooling?

To answer your questions

1) Flow rate and Velocity

Your pump is rated to 180m^3/h at 45m head. You need to design your system around this specification so that your total system restriction, as plotted on a chart, is below the curve of the pump. DO you have the performance curve for the pump?

In the pump, it is the diameter of the impeller that dictates the max pressure and it's the height of the impeller vanes that dictate the flow rate.

The volute of the pump is then designed around the impeller to give the ideal output geometry. In your case, the inlet is 100mm DIA and the outlet is 80mm.

The average velocity at the discharge of the pump will be Q/A

Assuming that pump is operating on its performance curve, the velocity of the fluid will be 0.05m^3/sec (Q)

The area of the 80mm outlet is 5.027x10^-3 m (A)

Therefore the average velocity of the fluid at the put of the tube is 0.05/5.027x10^-3 = 9.95 m/s

You can keep this velocity if you want and stick with the same diameter of tube, or you can increase the diameter to reduce the velocity.

2) Head loss

Research 3 things - Head loss equation, bend loss coefficient and Moody Chart.

The head loss equation will help you to identify the losses and the Moody chart will help you to understand your flow regime (Laminar - mixed flow - fully turbulent). The bend loss coefficient table will help you understand the losses in fittings, elbows and valves etc.

Everything in the line will cause a pressure loss.

For every 10m of height your pipework goes up, the load on the pump will go up by 10m or 1 BAR.

3) There are no "rules" as such, just the laws of physics. As stated, it's a balance between installation costs and running costs. As long as your total system restriction is not more than the 45m head that the pump can produce, you will be OK. If your system losses exceed the 45m, the pump will deliver less flow.

Pump drive power is the flow x head with a bit added for the 80% efficiency of the pump.

 

[Asisraja D]

FluidPowerUser
Have you selected the pump performance - (180m^3/hr at 45m head) based on the heat transfer requirements?

My managing director calculated heat transfer requirements because they have equivalent design sheet and when they calculate they were not involving me at that conversation.( Still I don't know how to do them )

Can you change the pump if your calculations show insufficient flow or head?

Sir we have already installed one standby pump with the same capacity (180m3/hr with 45 MTR head) parallel configuration.
And one more we have exact space for the pump base frame so no more space even 2" gap also impossible for new pump.

What is the water return temperature at the pump inlet expected to be?

The pump is below the sump configuration even though I'll calculate the NPSHA soon.

Are you planning to use the by-pass valve to create a pressure drop in the ring and effectively force flow into the 8" line and through the equipment to provide cooling?

Yes sir it is for future consideration but not for routine operation. ( Will affect cooling performance?)

DO you have the performance curve for the pump?

No sir we don't have pump performance curve and this is really new to me because I have never seen this kind of information in our design but I can get them with our pump supplier. (Does it available on internet like common chart for all centrifugal pumps?)

Research 3 things - Head loss equation, bend loss coefficient and Moody Chart.

Of course sir here lot of new concepts I'm have to go through because this gives me more and more deep knowledge about pumps and their design.

As long as your total system restriction is not more than the 45m head that the pump can produce, you will be OK. If your system losses exceed the 45m, the pump will deliver less flow.

Finally I get the concept of less flow from your above sentence.

Pump drive power is the flow x power with a bit added for the 80% efficiency of the pump.

This one I couldn't understand please explain me little bit.

 

[FluidPowerUser]

The pump performance details can be found in spxflow.com.

The data you provided for the pump shows the efficiency is 80%. That means for every 1000 watts of power you put into the motor, you’ll get 800 watts of output. Efficiency is measure of the output vs the input in terms of energy or power.

 

[Asisraja D]

Thank you so much sir. I have learned so much of knowledge form this thread especially from your conversation. Again thank you so much sir .

 

[FluidPowerUser]

You are welcome and good luck!!!

 

[pierreick]

Hi,
To add to this discussion a document about cooling water, I don't know whether you have it in place.
Last remark is about velocity in pipe, if too low it can promote the deposit of salt on the wall and favor corrosion.
Higher velocity will help for self-cleaning of the pipes.
Good luck.
Pierre

 

[FluidPowerUser]

Yes, Pierreick is correct. Laminar flow is ideal for pressure reduction, turbulent flow is good for cleaning the walls of the tubes and valves. The flow pattern in tubes and pipes is to have almost zero velocity near the walls, so turbulent flow helps to prevent deposits and therefore higher flow rates are required.

 

[georgeverghese]

Cement line carbon steel will enable a longer life span in comparison to plain pipe. Pipe ID will be some what less.

 

[Asisraja D]

pierreick sir thank you for sharing the document. It is really very helpful for me