Cooling Tower and Condenser Water Pump NPSH Problem

[nespakbsd]

Hello

I have a problem

A cooling tower is installed as the attached drawing with reference to the condenser water pump and HVAC plant room Finished Floor Level.

The NPSH(r) of the pump is 22.5 feet. The water flow rate is 968 gpm.

The place where the system is installed has an atmospheric head of 33 ft.

The head loss due to piping and fittings is 41 ft.

The dimensions given in drawing are in inches.

The pump is posing problems as it is not giving the required water flow rate (due to NPSH problem) and whenever it is shut off, water flows out of cooling tower.

Any ideas??

 

[Artisi]

How did you calculate the head loss of 41ft?
What diameter is the inlet pipeline, total length and are the 90 deg. fittings bends or elbows?
Is the pipework drawing a plan or elevation view?

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.)

 

[nespakbsd]

Thank you for your reply

The head loss calculated is with the aid of equivalent pipe length tables for fittings and valves etc and with the assumption of 4/100 ft head loss.

The drawing is a side elevation drawing.

 

[Artisi]

Can't see 41ft head loss.
You need to expand your information into something meaningful if you want answers:
"The head loss calculated is with the aid of equivalent pipe length tables for fittings and valves etc and with the assumption of 4/100 ft head loss" doesn't tell anybody anything as your calculation might be wrong.

Your call - if you want help, supply reasonable information.

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.)

 

[nespakbsd]

Thank you.

 

[RossABQ]

I think that is about the worst arrangement I've ever seen. A submersible or cantilever pump at the CT basin would be a far better arrangement. Are you subtracting the losses on the inlet piping and the lift from the pump's head? While your net lift is about 7', it has to get up and over the hump while clearing all the air in the line on startup.

 

[BronYrAur]

If you have approx 33' from atmosphere and a net lift of 7', you are down to 26' available to overcome friction and have enough left over for NPSHr. Since you need 22.5 at the pump, you only have 3.5' available for friction. I too have no idea where the 41' of loss came from, but you don't have enough reserve for much of any friction loss. I didn't even mention vapor pressure loss which is another foot or so rounded up. So you really only have 2.5' for friction.

Is it possible to raise the tower or lower the pump?

 

[Artisi]

Looks like the whole set-up needs re-engineering including a Pump requiring a lot less than 22.5ft NPSHr at the duty (assumed this is at the nominated flow) and a proper engineering calculation of head losses. My understanding is 1000usgpm thru an 8" sch.40 pipe has a Head loss of somewhere around 1.6ft/100, therefore 41ft h/l equates to over 2000ft of pipe.

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.)

 

[SAK9]

Your system has got a basic flaw which is the cooling tower is located at the lowest point in the system.So the system will self drain every time you turn off the pump.As suggested by Rossabq,you need to move the pump to a location below the tower.In order to prevent the system from draining ,you can modify the tower inlet into an inverted U loop with AAVs at the high point.Quick shut off motorised valve also is an less than ideal option

 

[chicopee]

I second SAK's proposal for relocating the pump. Just looking at the diagram, one can see the basic flaw without calculations.

 

[Artisi]

I have to disagree somewhat with the assumption that the pump is in the wrong location, there is no rule to say the pump needs to be at the source. For instance, a domestic water supply system booster pump could be located miles and miles from the supply source. As for draining back whenever the pump is stopped is cured with a NRV.
With the OP's system there appears to be something radically wrong with the setup, and until such times full engineering data is supplied it is crystal ball gazing


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.)

 

[RossABQ]

A booster pump has (usually) positive suction head, not a lift on an open system. The pump doesn't "have" to be right next to the basin, but having a large lift is very bad practice. Adding a NRV will only add to the suction losses, and few of them are leak-tight after a couple months' operation.

 

[ChasBean1]

Yes, I have an idea. The cooling tower water level usually has to be at least 10-20' above the CW pump suction in order to just get a zero pressure inlet. If your tower water level is almost 7' below the pump inlet like you show and you're looking for any kind of acceptable NPSH value, forget it.

 

[Artisi]

There is absolutely no reason for 10 - 20 positive head on the inlet side of a cooling tower pump unit, yes it's ideal to have a positive head but not imperative.
However, a correctly engineered installation where there is no alternative other than a lift on the inlet side is / can be done to give trouble free operation.
This is not saying I endorse such an installation, but it is achievable.

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.)

 

[SAK9]

My point is use suction lift only when absolutely necessary like in the case of drawing water from a well.In a CT circulation loop,the most reliable location for the pump is below the tower basin. If you were to use suction lift,air entrapment or a malfunctioning foot valve can put your system out of action.That is why in 99% of CT installations pumps have flooded suction.

In regards to system drain off,can you explain how you will stop it with a NRV?

 

[Artisi]

SAK9 -- No argument from me on where the ideal CWP should be located, but we aren't discussing an ideal situation here - the OP has / is presenting a completely different scenario although he has gone very quiet on the subject.


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.)

 

[ChasBean1]

Ok, don't worry about what I said.

 

[GHartmann]

Unfortunately that is a very unusual installation. I always used the Crane table for water flow through pipe to get rough estimates of dP. Suction lines not under pressure should normally be larger than the discharge and have substantially less pressure drop.

Having 41 feet of suction loss is approximately 18 psi! That's very high. If that is correct, your suction line is too small, too long, and has too many turns.

One of the previous posters was correct in suggesting that every time you stop the pump the suction line will drain to the basis. You can add a foot valve (special type of NRV) to remedy this situation.

I have linked / attached a couple of basic guidelines to help your review.

[link undefined]

http://r.search.yahoo.com/_ylt=A0LE...tent.pdf/RK=0/RS=ZXQGkYYMnVTij39YnRAjGRcSGnU-

[/url]

Good luck

G.Hartmann
BSChE NC State Univ

 

[GHartmann]

Sorry I did not make the link work correctly.

[link undefined]www.pdhonline.org/courses/m134/m134content.pdf[/url]

G.Hartmann