The Pain Point: Cracks That Appear After Hydrotest

A flare looks perfect on the bench, passes visual inspection, and then leaks during hydrostatic test or, worse, fails quietly after several weeks of cyclic service. The defect was never a hole to begin with; it was a micro-crack seeded during forming, held closed by residual compressive stress on the surface, and opened by tensile load once pressure, thermal swing, or vibration entered the system. In shell-and-tube exchangers, the tube end lives in a zone of permanent mechanical conflict: thermal expansion pushes it axially, pressure pushes it radially, and the tubesheet restrains it. If the flare carries hidden tensile residual stress at the inner lip, that stress couples with corrosive media to produce stress-corrosion cracking. Post-mortem analysis of failed joints consistently shows that the maximum residual tensile stress sits near the transition just behind the tubesheet, exactly where flaring strain concentrates if the process is uncontrolled. The frustrating part is that the part looked fine. Cold flaring does not forgive a poor strain path, because the damage is subsurface and directional.

The Principle: Hoop Stretch, Bending, and Springback

Flaring is biaxial cold plastic deformation. The mandrel imposes radial displacement at the tube end while the clamp holds the body fixed, so the material experiences circumferential stretch, axial restraint, and local bending at the flare root. On the outer surface of the flare, tangential strain is highest. At the inner corner, the strain state is a mix of bend plus shear. At the throat, the wall thins as material flows outward. The forming limit is governed by the material's uniform elongation: once local strain passes that point, necking begins and micro-voids coalesce into cracks.

Springback is the elastic recovery after the mandrel retracts. It scales with Young's modulus and the strain-hardening exponent. High-strength alloys spring back more, so the final angle is always smaller than the tool angle unless the stroke is over-driven. Rotary flaring reduces peak force by contacting the tube incrementally, spreading strain over time and lowering the risk of shock cracking. Ram flaring, by contrast, delivers a single axial push; it is faster but less forgiving on hard or thin material. The engineering goal is to keep equivalent plastic strain in the 0.2–0.6 range in the active zone, below the damage threshold on the forming limit diagram, while keeping wall thinning inside the allowable band.

A useful way to think about it is strain budget. Cutting, deburring, straightening, and handling all consume ductility before flaring starts. A tube with a work-hardened rim from a bad cut has less budget left for forming. That is why flaring defects often trace back to preparation, not the flare stroke itself.

Equipment Classification

Rigid punch presses are high-force machines suited to carbon steel and thicker walls, but they concentrate load and demand tight material control. Rotary or stagger-spin flaring heads use rollers that gradually push the wall outward; they suit thin-wall stainless, copper, and aluminum where crack resistance matters more than cycle time. Servo-electric incremental formers control stroke, speed, and retraction with feedback, giving the best springback compensation and repeatability for aerospace or nuclear work. Warm-flare units add local induction heating for titanium, high-nickel alloys, and precipitation-hardened materials that are too brittle cold. Multi-action cells combine flare, bead, face, and chamfer in one clamping, reducing handling strain. Orbital forming variants use low-load cyclic contact to stay below ultimate stress while still achieving permanent deformation. Each machine type defines a different strain path, and the strain path decides whether the lip stays sound.

Tooling matters as much as the machine. Sharp mandrel corners create stress risers; generous fillets of 2–3 mm radius distribute strain. Carbide and PVD-coated cores reduce galling on sticky metals. Die surface finish should be around Ra 0.8 μm or better to avoid orange-peel and scratch marks that become crack nuclei.

Practical Operation Guide

Start with material verification. Check hardness, grain direction, ferrite content for stainless, and wall tolerance. Anneal suspect lots; for stainless this may mean softening to improve uniform elongation. Use a mandrel with a large entry radius and confirm concentricity within 0.05 mm. Set the half-flare angle conservatively: 20–30° for most metals, narrower for high-strength alloys. Limit the flare ratio, final OD over initial OD, to about 1.2–1.4 on thin wall. Use staged feeding: pre-center, light contact, 10% expansion steps rather than one full stroke, especially above 1.5 ratio.

Control speed. Soft metals can take 3–5 mm/s; hard metals should run 1–2 mm/s. Apply clean, compatible lubricant in a thin film; chlorine-free for stainless, light oil for copper and brass, PTFE or molybdenum disulfide where galling is likely. After forming, measure flare angle, lip roundness, throat wall thickness, and total indicated runout. Section one part per batch and inspect the inner corner under 10–20× magnification. For critical service, add a low-temperature stress-relief bake to relax residual tension without softening the base tube.

Never flare across a weld seam without aligning the seam to a known low-stress zone, because the weld has different flow behavior and often less ductility. If force curves are available, reject parts with abnormal peaks or premature load drop.

Field Experience

In an aerospace hydraulic line, 5.1 mm stainless tubes with a 37° flare kept cracking at the inner lip. The original setup used a rigid punch, full stroke, no anneal. Crack rate was about 12%. The fix was three changes together: anneal the lot, switch to rotary flaring, and reduce half-angle to 22°. Crack rate dropped below 0.5%. The lesson was that no single change would have solved it; the strain path, material condition, and geometry had to move together.

In a petrochemical exchanger, carbon-steel tubes cracked after hydrotest because they had been stored outdoors and work-hardened at the ends. Normalizing the material before flaring removed the issue. In another case, a high-nickel alloy flared cold showed micro-cracks only visible under dye penetrant; moving to warm-flare at 250–300°C eliminated them. These cases confirm that cold flaring is a system, not a setting.

Summary

Flaring is a strain-budget problem. Stay inside the material's uniform elongation, control the strain path, and manage springback, and the joint stays crack-free. Exceed the budget, even by a few percent locally, and the failure appears later under load.

Outlook

The next step is simulation-first flaring. FEA and forming-limit diagrams will predict thinning, springback, and residual stress before the first part is cut. Shops will "virtually flare" every new alloy, then transfer the validated stroke to a closed-loop servo machine. Defects will be designed out instead of inspected out.