High-Pressure U-Tube
Heat Exchangers

The tube side, nearly always, because a small-bore tube carries pressure far more cheaply than a large-diameter shell. Required wall thickness for a given pressure scales with diameter, so a tube at several thousand psi is still a thin section, while a shell of the same rating becomes a heavy forging. Keeping the high pressure inside the tubes also means the shell can be carbon steel at a modest rating while the tubes, tubesheet and front head take the alloy and the thickness. The exception is a kettle or a condensing duty where the high-pressure stream has to be on the shell side; then the shell rating drives the cost and the tubesheet carries the differential in the other direction.

When the tubesheet forging, not the surface area, is setting the price. Tubesheet thickness grows with shell diameter at a given differential pressure, and the forging cost grows faster still, because the volume of metal rises with the square of the diameter times the thickness. Two shells in series at a smaller diameter each need a thinner tubesheet, a thinner shell and lighter flanges, and the forging material saved can pay for the second set of heads and nozzles. Series flow also gives two tube passes per shell for four effective passes, which helps when tube-side velocity is otherwise low. Ask for both arrangements priced at inquiry when the diameter is heading past what standard plate thickness covers.

Either a relief device sized for the tube rupture case or a low-pressure side designed to withstand it. ASME Section VIII requires protection against internal failure, and the practice described in API 521 treats a low-pressure side designed to at least two-thirds of the high-pressure side design pressure as covered without a dedicated rupture relief. Where the ratio is wider than that, the shell gets a relief valve or rupture disk sized for flow through both open ends of a broken tube, and the discharge has to go somewhere safe. On a U-tube unit the shell is normally the low-pressure side, so this is a shell-rating decision made at design.

Each side separately, at 1.3 times it's design pressure corrected for the material stress ratio at test temperature, per ASME Section VIII Division 1. The shell side is tested first with the bonnet or channel removed, so every tube joint on the tubesheet face is exposed; shell pressure pushes any joint weep out the tube end where it can be seen. The tube side is tested afterward with the head bolted up, which checks the head gasket and the pass partition as well as the tubes. Expect both tests to be witnessed by the Authorized Inspector on a stamped unit and recorded on the U-1 form.

Yes. TEMA sets the expanded length at the smaller of 2 inches or the tubesheet thickness minus 1/8 inch, so on a tubesheet several inches thick most of the hole depth is not gripped by the roll. The unexpanded hole length is a crevice where chloride or sour attack starts, so those services ask for full-depth expansion; strength welds take over as pressure climbs for load, not for the crevice. With a welded joint the weld at the front face carries the load and seals; the expansion behind it only keeps the tube from vibrating in the hole. Ask for the joint detail on the drawing: expanded length, seal or strength weld, and whether the holes are grooved.