Pump Head Calculation - Open Cooling Tower 1

[BronYrAur]

I'm a little sketchy when it comes to adding pumps to existing systems. can anyone offer some ideas on this approach?

I have an existing cooling tower on the roof of a 26 story building in Chicago. The tower serves 2 chillers on the 13th floor. The tower sump line drops down into a header on the 13th floor where 2 pumps in parallel pull water and send it to the chillers. The chiller discharge water then goes back up to the tower. Pretty straight forward. I don't know pipe sizes or GPMs, but each chiller is probably 500 tons.

I have a small (20 ton) cooling load on the 25th floor - one story below the tower. Actually it's more like 2 stories since the tower is on top of the 26th floor. I want to tap the tower lines and run condenser piping over to a water-cooled unit on 25. This is where I need a little help.

I can determine the head loss in my new piping and equipment, but how do I calculate what's happening in the tower lines I will be tapping? My thinking is that the suction side will have a static head of the difference between the water level in the sump and the location where I will tap. But do I need to condiser that I have a rather large pump down on the 13th floor that will be sucking water out of the same line that I want to pull from? Do I need to concern myself with this pump?

Likewise on the discharge side, I know I will have a vertical lift and pressure loss through the tower nozzles to overcome, but do I have to consider the discharge pressures of the pumps on 13?

Any help would be appreciated.

 

[BigInch]

Is there any connection on the new discharge with the dischrge from the 13th floor pumps?
Do you want to make a diagram of that?

"I'm all in favor of keeping dangerous weapons out of the hands of fools. Let's start with typewriters."
- Frank Lloyd Wright (1868-1959)

 

[BronYrAur]

I've haven't had much luck trying to attach something to enginering.com, but I gave it another try

I'm a little confused with your question?? Right now there is nothing except supply and return risers between the tower on 26 and the pumps and 13. I want to tap those risers on 25 and run branch piping and a pump over to the new laod.

If nothing existed below the 25th floor, I would have no problem. I could calculate my head to/from the load and to/from the tower.

I lose confidence in what to do when other pumps exist. Do those existing pumps affect the sizing of my new pump?

 

[katmar]

You need to consider the 13th floor pumps only insofar as that the flow they induce will cause a friction loss between your tapping points and the tower.

The head at the point where you tap in the new suction piping will be the height difference between the tower and the tapping point LESS the friction head caused by the flow of the combined old and new pumps in the pipe from the tower to the tapping point.

Similarly, the head where you tap in the new discharge piping will be the static head PLUS the friction caused by the combined flow in the section of pipe from the tapping point up to the tower.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[stanier]

http://www.freecalc.com/pumpfram.htm

 

[SAK9]

You have a system in which three pumps pump into a common pipe( ie pipe returning water to the towers).So the new pump on 25th storey would need to match the head characteristics of the pumps on the 13th floor.For eg pumps on the 13 th storey are discharging at 3 bar.After deducting all pressure drops thru the heat exchanger(condenser),valves,and pipe run between 13th floor and 25th floor, the pressure at the point where the 25th storey tap in point is 3-1=2 bar.The pump on 25th storey will need to deliver water at the tap in point at a pressure higher than or equal to this value( 2 bar).

 

[Artisi]

The point of connection into the return line from your new installation - will have to be exactly equal to the pressure in the return line at this point coming from the existing pumps - if higher or lower it will affect both pumps. The pressure available at the connection point coming from the new line will need to be slightly higher than the pressure in the main return line - you will then need a pressure control valve to regulate the pressure from the new line to that of the old line. Alternately return the flow from the new installation direct to the top of the cooling tower. This way you do not need any control system as both units are independant.

 

[katmar]

Artisi, in the strict sense you are correct but the new flow is so small compared with the existing flow (20 ton vs 1000 ton) that I would just make sure that the new pump can match the head at the tapping point and allow the pumps to find their own operating points.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[Artisi]

The pressure at the tap-in point for the new pump needs to be above that of the pressure at the tap-in point, otherwise flow won't take place. I would still go with a valve of some sort to ensure the new pump is not overcome by the existing system pressure. Getting the 2 systems to balance by themselves might be a problem.

 

[BigInch]

They will balance by themselves, but the question is, will that balance point be where you want them to balance?

The older pumps will tend to have their suction pressure reduced and their discharge pressure increased when the new pumps are running. They will tend to back up on their curves as they reduce their flow to compensate. When they reduce their flow, the question becomes, will the otal flow in the common loop be equal to the total flowrate you are expecting? Thus, the old pump curves should be examined to ensure that they will produce the lower loop's new flowrate at the lower loop's new required head.

Termo: The new pumps will send the new load's hotter water to the existing cooling tower, perhaps making the inlet temperature to the new pumps higher too, if the tower's cooling capacity is not increased.

"I'm all in favor of keeping dangerous weapons out of the hands of fools. Let's start with typewriters."
- Frank Lloyd Wright (1868-1959)

 

[Artisi]

BigInch - you are of course correct - the additional flow into the return leg from the tap-in point to the discharge will cause a shift of performance of the older pumps anyway as the total head on these pumps will increase due to higher friction loss.

Maybe my first suggestion of a seperate discharge line is worth considering - no reduction of flow for the existing pumps etc.

 

[katmar]

The new pumps have a flow of about 2% of the flow of the old ones, and the new flow only applies to 2 storeys out of 13. If you can mark that change in friction head on your pump curve your pencil is too sharp to be safe ;-)

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[BigInch]

Right u r Kat. Sounds good then.

"I'm all in favor of keeping dangerous weapons out of the hands of fools. Let's start with typewriters."
- Frank Lloyd Wright (1868-1959)

 

[quark]

Separate discharge line is possible only if the cooling tower is an open deck type one. The two lines should join at some place if the return line is directly connected to the spray header. It may be a good option to connect the two line nearer to the spray header.

 

[chicopee]

I have looked at your problem and my approach would be :

a) get pump performance on existing and new equipment; new pump performance should have a churn point matching or closely matching the existing pumps

b) determine friction losses of suction lines and of discharge lines including chillers for a at leat one flow

c) using semilog paper (ordinate of semi log paper labelled head (ft)and absissa labelled as flow (gpm), plot existing pump performance curves, suction line friction curve based on a single flow value, combined discharge line and chiller friction loss all the way up the cooling tower and these curves s/b close to being straight lines in the semilog paper.

d) add the two parallel existing pump performance on the semi log paper to get one equivalent graph for two pumps in operation;
add the friction losses of suction and discharge(including chiller loss) lines to get an equivalent friction loss curve on the same graph. Where the two curves meet s/b a rated point. At that rated point there will be a range acceptable to the operation of the oringinal cooling system.

e) with the new pump, plot its perfromance curve and equivalent pipe friction loss (based on suction and discharge including chiller).
add the new pump performance curve to the equivalent existing pump performance curve ; again treat as a parallel pump arrangement.
add the existing equivalent friction loss curve to the equivalent pipe frictionloss of the new pump.
this intersection of the two new straight line will be another rated point and hopefully should fall within the rated point range in d).

This is the fastest and easiest way to figure out if your proposal will not affect the existing system--at least it will put you in the ball park.