Floating Head Vapor Condensers

Flexibility here means the condenser can be re-tasked without being rebuilt, which is what a refinery or chemical plant values over the life of a unit. The same AES shell can go from a light overhead service to a heavy, fouling one by changing the cleaning interval rather than the exchanger, and the bundle can be retubed in a shop while a spare runs.

The expansion freedom is also wider than a U-tube's, because the floating tubesheet moves as a whole and no tube is asked to flex at a bend. That lets a floating head condenser accept a 150 F metal temperature difference and repeated thermal cycling that would fatigue other designs. The price is roughly a third to a half above fixed tubesheet construction and a bundle that takes longer to pull, which is why it should be chosen on a documented need.

The difference buyers feel is tube count: a pull-through head leaves a ring around the floating cover that carries no tubes, and in a mid-size shell that typically costs 10 to 15 percent of the bundle. The split-ring design fills that ring and recovers the surface, at the price of an internal disassembly before the bundle will move.

Decide by how often the bundle comes out. If an overhead condenser is pulled every turnaround, once every three to five years, the split-ring unit is cheaper and the disassembly is a small part of a planned shutdown. If polymer fouling drives a pull every few months, the pull-through pays for its larger shell in hours saved and in the risk removed from working inside the shell. Specify the shell diameter for the pull-through design accordingly, since the lost tubes have to be made up somewhere.

Pitch decides whether a brush or a lance can reach the outside of every tube, and it is cheap to get right at the drawing stage and impossible to change afterward. A 3/4 in tube on a 1 in square pitch gives a clear 1/4 in lane in two directions; the same tube on a 15/16 in triangular pitch gives no straight lane at all.

The square layout gives up on the order of 10 to 15 percent of the tubes in a given shell, which is why thermal design software proposes triangular by default. On a fouling vapor side that saving is false economy, because the surface you cannot clean is surface you lose within a year. Rotated square keeps the lanes and adds a little turbulence on the shell side, and is the usual compromise on a condenser that fouls but is not a heavy polymer service.

When a weep to atmosphere would be tolerable and the operating conditions stay modest, which in practice means water, glycol or a benign process liquid on the shell side at low pressure and moderate temperature. These rear heads are generally confined to a few hundred psig and a few hundred degrees F, below the ratings of an internal floating head.

They have a real place on a utility condenser where the shell side is cooling water and the vapor is contained in the tubes, since the packing then sees only water and a drip is a maintenance item rather than a release. The moment the shell-side fluid is a solvent, a hydrocarbon or anything that would set off a gas detector, the saving is not worth it. Thermal cycling shortens packing life, so a batch condenser is a poor candidate even with a benign fluid.

Put a number on the fouling first, because the premium is justified by cleaning economics rather than by design elegance. If an overhead condenser loses 20 percent of capacity between cleanings and that loss throttles a column worth thousands of dollars a day, a bundle that can be brushed on both sides in a planned day pays for its premium in the first lost week it prevents.

If the vapor side is clean and only the cooling water fouls, a U-tube gives the shell-side access for less, and if neither fouls appreciably the fixed tubesheet is the right buy. The honest cases for a floating head are refinery and petrochemical overheads, heavy hydrocarbon condensing, and any duty where the plant has written cleaning intervals into its maintenance plan rather than reacting to lost capacity.

Allow the full tube length plus a working margin in line with the front end of the shell, so a unit with 16 ft tubes needs about 18 to 20 ft of clear floor. Add room for a bundle puller or a crane hook above the pull path. The bundle itself is heavy: a 24 in diameter bundle with 16 ft tubes can weigh around 3,000 lb or more before it is drained.

Plan the maintenence layout with that in mind. Pipe racks, cable trays and neighboring exchangers routinely end up in the pull path, and a condenser that cannot be pulled without a shutdown of something else loses the advantage you paid for. On a split-ring unit, add access at the rear for a person to remove the backing ring, which means a removable shell cover and a platform if the unit is elevated. Note the pull direction on the general arrangement drawing and keep it clear.