column overhead line to aircooler 4

[roker]

a 30" ovhd line (at ~60 meters above grade) is routed to an aircooler located at 15 meters above grade.
the aircooler has 9 bundels with two 10" inlet nozzles each.
what is the best way to design the piping manifold so the flow to each bundle shall be as equal as possible by symmetrical piping only (the operating conditions are 80 deg C and 2 barg).

 

[MMcLean]

Roker

Keep it simple. Without seeing it I'd say, 30” O/H line teeing into the centre of a 30 “, or smaller, manifold with 10 “ lines coming off it to the nozzles. Consider provision for butterfly valves or orifice plates on the cooler outlet lines to equalise flow if required. Often the pressure drop in the line is small (if sized correctly) compared to that in the cooler, so even flow distribution is not a problem, as long as the piping is symmetrical. Temperature indicators in/out of each bank would be useful to determine that flow to each bank is relatively even. I might be stating the obvious but I’d recommend a minimal slope on the horizontal sections of the 30 “ line to allow any condensate to drain back to the vessel/to the cooler, and obviously no pockets.

 

[25362]

As MMcLean rightly says, the pressure drops through the inlet and outlet headers should be designed very small when compared with that expected to take place through the "consumer" parallel bundles, thus achieving about equal pressure drops for all nine of them.

Even then process fluid flows through these nine parallel bundles may vary due to other, however small, causes of asymmetry.

For example, the circulating cooling air amounts or temperatures may not be equal for all nine air cooled bundles.

When the cooling loads among the nine bundles differ, as can be checked by measuring temperatures, and depending on the fluid, some salts may sublime and deposit on the slightly ("over") cooled "consumers" directly and indirectly creating corrosion or partial plugging, themselves factors for a further maldistribution of heat loads, serving to create a vicious circle.

In some cases, chemical treatments and water washing are sine qua non process-side requirements.

Sorry for playing the role of the devil's advocate. Good luck.

 

[saxon]

roker, There is standard piping layout design that has been used for years to approximate equal flow to nozzles in exchangers. The standard arrangements are based on the number of total inlet nozzles; less than 4, 4 to 6, more than 6.

In your case, 2-10", manifold: 30" with two 10" nozzle take offs, fed from one side or the other. "borrowed from Rip Weaver, Process piping design V-II, Gulf publishing"

Hope this helps.
saxon

 

[roker]

Dear saxon,
thank you for your reply, since the book is not at my possession, could you give a more detailed description: the total no of nozzles are for all the 9 bundels (total of 18) or for one bundle (2)?
what is meant by "fed from one side or the other?
regards,
roker

 

[steveen]

Roker,

I designed a similar system for a crude overhead system. Best to model on xcel spreadsheet using Crane to estimate pressur losses. You can then look at how different arrangements affect flow. I used a 2% difference in flow through each of 8 air fan bundles as a criteria. This took about a week to get right. The original design, even though it looked good in the field, had the flow preferentially through the inner bundles.

Don't forget the outlet piping has to figure into the equation. Use of special gates and reducing elbows were required to get to my 2% target. My phone number is 713-204-3615 if you would like to discuss. If you have a corrosion issue here, I have additioanl recommendations.
Ask for Steve.