Why the U-Bend Deserves Its Own Specification

Most exchanger tube failures are not in the straight legs. They are in the bend, and they are caused by a small set of recurring issues: insufficient wall thickness after bending, residual tensile stress that cracks in chloride-bearing water, ovality that prevents proper rolling into the tubesheet, and a base-material chemistry that cannot survive the shellside or tubeside service. Each of these can be designed out. The reason most are not is that the U-bend specification is buried in a generic "furnish per ASME SB-111" or "furnish per ASTM A688" line, and the bending is left to the supplier's interpretation.

In our experience as a mill supplying heat efficiency tubes into refineries, chemical plants, desalination trains, and power stations since 1994, the projects with the lowest lifetime cost treat the U-bend as its own engineering line, not as an afterthought to the straight tube order. The four-line spec that follows is what we have used on those projects.

Matching the Base Tube to the Service

The base tube choice depends on what the tubeside and shellside fluids actually are, at the operating temperature, and at the chloride, sulfide, or ammonia level that will be present over the exchanger's life. The four families below cover the majority of industrial U-bend service. Going outside these four is sometimes justified (titanium, super-duplex, alloy 825) but each step away from a standard family increases lead time, cost, and welding complexity.

Copper and copper-nickel — the default for seawater and brackish cooling

For tubeside seawater at velocities up to about 3.0 m/s, and for most desalination and shipboard cooling service, copper-nickel 90/10 (C70600, UNS C7060X) and 70/30 (C71500) remain the default. Both are covered under ASTM B395/B395M for U-bend geometry, which is the standard most mills will reference for procurement. The 90/10 grade is the workhorse: easy to bend, good resistance to seawater impingement, and a thermal expansion that pairs well with aluminum-brass or titanium tubesheets. The 70/30 grade is selected where sulfide or ammonia contamination is expected, or where the cooling water velocity is at the upper end of the design window.

Our copper nickel alloy range ships in both grades, and we typically supply the U-bend bundle with matching copper nickel flanges so the tubeside metallurgy stays consistent through the channel and into the channel cover.

Austenitic stainless — the default for chemical, food, and clean utility service

For most chemical, food, pharmaceutical, and clean-utility service, TP304/304L and TP316/316L remain the standard U-bend materials. They are covered under ASTM A249, A269, and A688, and the L grades are mandatory where the exchanger will see a post-fabrication heat-treat that could otherwise sensitize the grain boundaries. For chloride-bearing service above about 60°C, the boundary moves toward higher-nickel grades (TP321, TP347) or duplex stainless, with the actual choice driven by ASTM A923 testing and the design chloride concentration.

On stainless steel pipe bundles where the U-bend is part of a larger piping package, the safest move is to order the U-bend tubes, the tubesheet-quality straight tubes, and the connecting pipe from a single heat-number pool. This eliminates the galvanic and weld-procedure variables that show up when three different suppliers are involved in one exchanger.

Carbon and carbon-moly — the default for high-pressure steam and hydrocracker service

Above about 400°C, in refinery hydrocracker and power-plant steam service, the choice moves to carbon-moly (T1, T11, T12) or Cr-Mo (T22, T91) grades under ASTM A213. These grades are bendable on a cold rotary draw bender in sizes up to about 50.8 mm OD, but they require a mandatory post-bend stress relief at 620–705°C depending on grade. Skip the stress relief and you get a bend that work-hardens in service and cracks at the extrados within the first operating year.

Titanium, duplex, and the alloy grades — when service forces you off the standard

Chloride service above 80°C in refinery overhead condensers, sour hydrocarbon service with NACE MR0175 limits, and seawater service at high velocity or with biofouling control all push the specification outside the four "standard" families. Titanium Grade 2 and Grade 7 (ASTM B338) are common for refinery overheads. Duplex 2205 (ASTM A789) and super-duplex 2507 cover offshore and high-chloride seawater. Alloy 825 and alloy 625 cover sour hydrocarbon. Each of these is bendable, but the bend procedure must be qualified on the actual tube lot — generic bend procedures from a different material cannot be carried across.

The Wall-Thinning Math — Why "Bend to 1.5 × OD" Is Not Enough

The standard shorthand is "bend centerline radius of 1.5 × OD" for a long-radius U-bend, and 1.0 × OD for a tight-radius return bend. What the shorthand does not tell you is the actual wall thickness on the extrados (the outside of the bend) after the tube has been pulled around the bend die. This is the number that determines whether the tube will hold design pressure for the next 20 years or fail at the next hydrostatic test.

The empirical formula used by ASTM B395 and most exchanger fabricators is straightforward. For a tube with nominal wall thickness t, outside diameter D, and centerline bend radius R:

The procurement implication is simple: if the exchanger design requires 1.2 mm minimum post-bend wall, the nominal wall on the PO must be at least 1.65 mm, and the order must include a "minimum post-bend wall" line, not just a "nominal wall" line. For tight-radius bends (1.0 × OD) and for service where any further reduction is unacceptable, ASTM B395 allows the supplier to use dual-gage (double-wall) tubing — the central bend section is supplied at the next heavier BWG increment, and only the straight legs are machined back to the nominal wall. This is standard practice for copper and copper-nickel U-bends; specifying it on the PO closes the wall-thinning loophole before it becomes a field failure.

Temper, Heat Treatment, and the Hidden Stress-Corrosion Risk

The second most common U-bend failure is stress-corrosion cracking at the extrados, driven by residual tensile stress from cold bending. The fix is metallurgical, and it depends on the alloy family.

The cost of post-bend heat treatment is small relative to the cost of a single exchanger failure — typically 4–8% of the U-bend bundle price. The cost of skipping it on a service that demands it is the entire bundle plus a forced outage. On a stainless U-bend, cold work above 15% without a re-anneal leaves the extrados in a sensitized condition that cracks in chloride-bearing water within 12 to 36 months. On a carbon-moly bend, missing the post-bend stress relief leaves the bend in a hard, brittle condition that cracks at the first thermal cycle.

Inspection and Quality Documents — What the RFQ Should Demand

A U-bend bundle is not a commodity product, and the inspection regime should match that reality. The minimum documentation set that any serious procurement should demand, irrespective of alloy family, looks like this:

• Mill Test Certificate to EN 10204 3.1 for every heat used, with full chemistry, mechanical properties, and grain size.
• Eddy current test (ECT) report on every straight tube before bending, per ASTM E243 or equivalent. The bend cannot be Eddy-tested reliably after forming, so the straight-leg inspection is the only one that counts.
• Hydrostatic test on every U-bend at the pressure required to produce 48 MPa fiber stress (per ASTM B395) or the equivalent ASME code stress, whichever is higher. Pneumatic testing is acceptable for small-diameter copper-nickel but is rarely used on stainless or alloy bends.
• Dimensional report covering leg length, leg-to-leg centerline distance, bend angle, ovality at the bend apex, and measured wall thickness at the extrados.
• Heat treatment certificate (when applicable) with actual furnace temperature and time-at-temperature data, not just a "stress relieved per spec" statement.

A common procurement error is to accept the supplier's standard certificate pack and assume the inspection is covered. It usually is, for the straight tube, but rarely covers the bend-specific checks. Build the bend-specific documentation into the PO, and the supplier will quote to it. Leave it off, and the supplier will ship to whatever default they happen to run that month.

A Practical RFQ Walkthrough for the Next U-Bend Order

The U-bend RFQ is short, but every line in it does work. A clean RFQ for a 200-tube bundle of 25.4 mm OD × 1.65 mm wall copper-nickel 90/10 U-bends to 38.1 mm centerline radius would read approximately as follows. The line-by-line notes are what we wish every inquiry included:

• Line 1 — Quantity and base tube standard: 200 pieces, ASTM B395/B395M, UNS C70600, temper O61 (annealed).
• Line 2 — Dimensions: OD 25.4 mm, nominal wall 1.65 mm, leg length 6000 mm, centerline bend radius 38.1 mm (1.5 × OD).
• Line 3 — Post-bend minimum wall: 1.20 mm minimum at the extrados, verified by ultrasonic thickness measurement at the bend apex and 90° on either side.
• Line 4 — Heat treatment: stress relief at 300°C ± 10°C for 1 hour minimum, then air cool.
• Line 5 — Inspection: ECT per ASTM E243 on straight tube before bending; hydrostatic test to 48 MPa fiber stress on every bend; dimensional report per Line 2; MTC to EN 10204 3.1.
• Line 6 — Packaging and delivery: individual plastic sleeve on each bend, foam separator between layers, wooden crate seaworthy for 60-day ocean transit; delivery CIF main seaport within 30 days of PO.

A clean RFQ like this typically comes back with a quote inside five working days, with a price that is realistic for the actual specification. A vague RFQ comes back with a low-ball quote that the supplier will later modify once the gaps are filled in — usually at the buyer's expense in the form of change-order fees. The two-hour investment in writing the RFQ properly saves a far larger sum in claim avoidance downstream.

Bundling the U-Bend With the Rest of the Heat Exchanger Package

The cheapest way to source a U-bend bundle is rarely to source it on its own. The tubesheet-quality straight tubes, the U-bend tubes, the tubesheet forgings, the channel and channel cover, the pipe flanges, the stud bolts and gaskets, and the nozzle pipe fittings are all part of one metallurgical and dimensional system. Sourcing them from a single integrated mill produces four measurable benefits on real projects: harmonized MTCs, matched heat numbers, single-point accountability for any field issue, and shorter overall RFQ-to-delivery time because the mill is running one production plan rather than chasing four.

On a typical 200-tube U-bend bundle, the savings on the bundled package are not on the unit price of the tubes themselves — that price is broadly market. The savings show up on the schedule (fewer clarification rounds, fewer receiving-day surprises), on the documentation (one MTC chain rather than four), and on the field issues (one phone call to one supplier when a question comes up at tubesheet rolling). For a refinery turnaround, where every day of delay is a six-figure cost, the bundled package is almost always the right commercial answer as well as the right engineering one.