U Bend Tubes in Heat Exchangers: Materials, Bending, and Sourcing From a Single Mill

In a shell-and-tube heat exchanger, the U-bend is the most forgiving part of the bundle and the most unforgiving to get wrong. Get the material, the bend radius, and the post-bend heat treatment right, and the exchanger runs for decades. Get them wrong, and the bundle cracks within the first two heating cycles. This guide walks through how the bend is actually made, where it tends to fail, and how a mill-level supplier simplifies the whole chain.

Why the U Bend Exists at All

A U-tube heat exchanger has only one tubesheet. Both legs of every tube are fixed to that single plate, and the tube loops back on itself inside the shell. That geometry is the entire point: the free leg can slide, expand, and contract as the metal heats and cools, which removes the thermal-stress problem that plagues fixed-tubesheet designs. The trade-off is that the inside of the bend is now the most heavily worked zone in the entire tube, and it must remain ductile, dimensionally stable, and free of wrinkles for the life of the unit.

In practice, the same logic shows up wherever thermal cycling is severe — in feedwater heaters, condensers, ethylene cracking coolers, and the cold side of LNG units. In every one of these, the bend is the controlled weak point, and the controlled weak point is only safe when the manufacturing is controlled.

Material Selection: Where Most Projects Start

Material choice is driven by three forces acting on the tube: the shell-side fluid, the tube-side fluid, and the peak metal temperature. Carbon and carbon-manganese steels cover most refinery and power duties; austenitic stainless grades handle corrosive condensers and food-grade service; copper-nickel and nickel alloys take over when seawater, sour service, or high-temperature creep are in play. Many process plants pull tubes from more than one material family within a single bundle, and that is precisely where working with a mill that produces stainless steel pipe and carbon steel pipe on the same line pays off — one heat of steel, one set of test reports, one delivery window.

How the Bend Is Actually Made

Two processes dominate. Cold rotary-draw bending with an internal mandrel is the standard route for tube OD up to about 50–60 mm; the mandrel sits just behind the tangent point and prevents the classic failure modes — wrinkles on the intrados, thinning on the extrados, and ovality at the tangent. For larger diameters or thicker walls, induction hot-bending is used: a narrow ring is heated to the austenitizing range, the tube is pushed through a fixed radius die, and the bend is formed in a single pass. Hot bending is slower and more expensive, but it preserves the metallurgical structure of heavy-wall tubes that would work-harden and crack if bent cold.

The minimum bend radius is the next design decision. A common rule of thumb is 1.5×OD for annealed stainless and 2×OD to 3×OD for harder alloys, but the right number depends on wall thickness, ovality limits, and whether the tube will be solution-annealed after bending. Tightening the radius below the safe minimum looks attractive on a layout drawing — and it is the single most common source of in-service cracks at the tangent line.

Post-Bend Treatment and Inspection

Bending work-hardens the outer wall and leaves residual stress in the springback zone. For austenitic stainless tubes, a full solution anneal followed by a rapid quench restores the corrosion resistance that cold work has degraded; for carbon and low-alloy tubes, a stress-relief soak at 595–705 °C is usually enough. Skipping this step is the most expensive shortcut in the industry: the exchanger passes hydrotest, ships to site, and then develops stress-corrosion cracking within the first year.

Inspection has to verify three things: geometry, integrity, and material identity. Geometry is checked with optical or laser scanners against the bend-radius and straight-leg tolerances. Integrity is confirmed by hydrostatic test on every tube, plus nondestructive testing (eddy current or ultrasonic) on the bend zone. Material identity is locked in by the mill test report that travels with the tube from the steelmaker through bending, heat treatment, and final shipment — the kind of traceability that an integrated industrial tube producer can hand over as a single document instead of a chain of partial certificates.

Why the Mill Matters for U-Tube Bundles

Most U-tube projects are quoted as three separate purchases: the straight tube, the bending subcontract, and the heat-treatment subcontract. Each handoff introduces a delay, a documentation gap, and a quality risk. A mill that controls all three steps on one site — drawing or pilgering the tube, bending it in-house, then solution-annealing and testing it before shipment — collapses that chain. Lead times come down, the mill test report covers the finished bent tube rather than just the raw stock, and the engineering team is dealing with one technical contact instead of three.

This is also where the broader product range of a full-line mill becomes useful. A project that needs a U-bundle for the high-pressure side of a power plant often needs matching pipe fittings and heat efficiency tubes for the economizer and air-preheater sections, all on the same delivery schedule. Bundling the order through one producer — and getting one consolidated documentation package — is usually cheaper than chasing each item separately, even before the schedule savings are counted.

Common Failure Modes and How to Avoid Them

Five Issues Seen in the Field

• Cracking at the tangent: Almost always a bend radius that was tightened to fit a layout. Push the radius back to 1.5–3×OD, or switch to a deeper shell.
• Stress-corrosion cracking in the bend zone: Cold work was not removed by post-bend annealing. Add a full solution anneal and verify with a hardness map.
• Wrinkles on the intrados: Mandrel position was wrong, or the wall is too thin for the radius. Move to induction bending or specify a heavier wall.
• Ovality outside tubesheet tolerance: Bend speed too high, or lubricant starved. Slow the bend, check mandrel lubrication, and re-verify against ASME B&PV Section VIII tolerances.
• Bundle vibration in service: Baffle spacing too wide for the unsupported span. Cross-check the bundle against the TEMA vibration chart, not just the static pressure ｄｒｏｐ.