Cooling water return pipe pressure

[TKTogs]

I currently have a cooling tower cell with a pump delivering cooling water (at 540 m3/hr) to a header tank 35 m high from the pump. The water is then distributed to several condensers, and routed back to the cooling tower return line at a level of 10 m above the cooling tower. The water then flows back to the cooling tower by gravity. After the installation of additional condensers, which take water from the cooling line to the before it reaches the header tank, there has been an enormous heat load on the cooling tower, resulting in the new condensers failing to cool properly. I now intend to supply water to these new condensers through another cooling tower and separate pump,delivering 300 m3/hr of water at 43 m head. Both these systems will share one common return line to keep the costs at minimal. This will however increase the pipe velocity in this return pipe. Preliminary calculations show that the velocity almost doubles by increasing the total flow in the return pipe (500 mm diameter). My question is, what is the acceptable pressure of the returning cooling water as it enters the cooling tower? Will such an increase in velocity, and hence pressure drop, result in a high pressure drop that would upset the water distribution in the tower?

Is such an arrangement sensible?

 

[Drazen]

it is hard to figure it our, so your header tank is above condensers that are above cooling towers?! do you have any balancing valve installed at tower branches? are they evaporative towers, closed or open loop towers?

many things come in play, but you would need to recalculate your existing network, verify calculations by real-life readings, than you will be able to comprehend whole possible range of modifications.

 

[KiwiMace]

With everything seemingly above the level of the cooling tower... is this a closed or open tower design?
Open towers return to atmospheric pressure at the top, they don't need pressure exactly, just the right system pressure to meet the flow. If you are supplying the tower from higher level you have to manage the supply and return flow well to prevent the tower flooding. It is not difficult in steady state, one tower/one pump, but when you start and stop condenser pumps on multiple tower/pumps you can quickly run into problems. This will get more complex in your common return scenario - how do you ensure that a stopped pump on tower A will not cause tower B to flood?

A closed tower is completely different, see the vendor information for head requirements.

 

[katmar]

Returning cooling water from such a height can be problematic. If the return line is not vented to atmosphere it is possible to draw a vacuum in the return line, and even in the condensers. This can cause the cooling water to boil (vaporise) and then re-condense - just like cavitation in a centrifugal pump. Apart from the obvious danger of imploding the condensers, it can cause the lines to vibrate.

But it is not as simple as just opening the return line to atmosphere. Once this is done you have to look at the risk of sucking air down the return pipe and causing maldistribution in the towers. Your return line of 500 mm would not be large enough to be self venting with 840 m3/h. It is already borderline at 540 m3/h. The alternative is to install a control system on the return line that ensures the pressure in the return line (at the condenser outlets) never goes below atmospheric.

I have never found a system where the pressure in a gravity fed return line was insufficient to drive the nozzles in the tower, but your tower vendor should be able to provide you with the flow vs head characteristic curve for the nozzles.

In all the towers that I have worked with which have had multiple return lines there have been balancing lines between the tower basins. This overcomes the problem highlighted by KiwiMace of when a pump stops on one of the towers.


Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[TKTogs]

Thanks for all the responses. Just to clarify to Drazen and KiwiMace, the cooling tower is indeed an open one, with the water evaporating in the atmosphere. All condensers and return piping are above the cooling tower - find the attached sketch illustrating the relative positions of the header tank, condensers and cooling tower..

Katmar, is there any supporting literature and calculations I can use in my write up to propose the installation of a control system to regulate the cooling pipe return pressure not to fall below atmospheric (showing that a 500mm diameter piping will be insufficient for self breathing at 840 m3/hr)?

 

[katmar]

The only document describing this problem that comes to mind immediately is an article by Larry Simpson in the July 17, 1968 edition of Chemical Engineering. These old documents can be hard to get hold of these days, unless you have access to a University Library. There are online sources where you can order copies, but I don't remember the names. Sorry. Have a look at thread378-87442 where a similar topic was discussed, and reference is also made to an excellent article by Hills. Perry's Chemical Engineering Handbook probably has something on self venting flow too.

However, all the theory in the world is trumped by practical experience and if your installation has not had problems with cavitation etc yet it might be OK. Your sketch seems to indicate that the new condenser return joins the vertical line at the 10m level and the combined line discharges into the cooling towers at the 7m level. If this is correct then the new (additional) load will not have much impact on the self venting characteristic of the return line. If anything, the additional back pressure from the extra flow will help back up the level in the return and this should be an improvement.

At a flow of 840 m³/h in a 500 mm ID line the velocity is only 1.2 m/s and the pressure drop is about 210 mm H₂O per 100 m length.

Your sketch does not show any open vents. Is the whole system sealed - apart from the header tank of course?

The control system does not have to be elaborate. Simply throttling the valves on the return to the cooling towers would probably be adequate if the system does vibrate a bit.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Drazen]

as far as i can comprehend, in the old system header tank established zero pressure point, but i do not see such zero point in new supply piping. if you imagined that new pump pressure creates balance with height difference between old and new condensers (new condenser is drawn at lower altitude than the old ones), it looks quite complex to me.

such situation could be very interesting for some faculty study, but i myself would not dare spending client's money in someone that not passed practice test. not to mention that it could also create some interesting hydraulic hammers - very interesting to faculty professors, but not at all interesting to clients.

even if you would install condensers at the same level and create zero pressure point, yet you would not be able to control flow to cooling towers without balancing valves, and that should be dynamic balancing valves.

if that alone would not offset saving in return piping, than risk of facing the unknown troubles certainly would

 

[katmar]

Drazen, I agree with you that it is a system that could go badly wrong, but I would not regard it as something of interest to academics only. I recently upgraded a system that was very similar to TKTogs setup (i.e. with some condensers fed from a header tank and others directly), but we had the added complexity of it all being done with one pump. I made the vertical return pipe self venting with an open top and it all works very smoothly with no vibration or mal-distribution.

The balance between the header tank and the other condensers was done by simply ensuring a small overflow from the header tank into the vertical return pipe. For a bit more money this could be automated but the system was steady enough for us to do it manually.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"