An assembly that has pulled is longer on one side than the other. Heat straightening corrects it by making the long side shorter, using the same mechanism that caused the problem in the first place: local heating, local expansion, and contraction on cooling that leaves the area permanently smaller.
The counter-intuitive part is that the shrink happens on cooling rather than on heating. The heated spot expands against the surrounding cold metal, which resists it, so the spot yields in compression while hot. When it cools it contracts from a yielded state and cannot return to its original size. The result is a small, permanent shortening, and a series of them is what pulls a bowed part straight.
A spot is a roughly circular heated area, used to pull a local area back. A wedge is a triangular pattern with the apex pointing toward the area that needs to move least, used to correct an angle. A line is a heated band across a member, used to shorten a section and take out a bow.
The choice of pattern is a choice of where the shrink goes. All three rely on the same effect: the metal that was heated is smaller afterwards, so the part moves toward it. Getting the pattern wrong on a part with several distortions can fix one and make another worse, which is why straightening is usually done in stages with a check between them rather than in one pass.
Heat alone does not straighten anything. The shrink comes from the resistance of the cold metal around the heated area, so a part that is free to expand will simply expand and cool back to where it started. This is why straightening is often done with the part clamped or weighted, and why the same torch pass on a clamped part and a free part gives different results.
It also means that a part with no stiffness in the direction of the correction cannot be straightened. A thin, unsupported panel will bow out of plane instead of shrinking, and the heat is wasted. Adding a temporary stiffener or using a jig to hold the shape being aimed for is part of the method, not an extra.
Austenitic stainless is the friendliest of the common materials for this work, because it can be heated and cooled without a phase change and because it work hardens predictably. But it is not free of consequences. The heated area loses its passive film and will show heat tint, which on a visible part is a finishing problem even when the geometry is correct.
There is also a sensitisation risk if the part is held in the wrong temperature band for long enough, and a risk of distorting a nearby weld by heating close to it. Straightening near a finished weld is the case to think hardest about, because the weld and the heat-affected zone respond differently from the parent metal and the joint can crack or move.
Titanium cannot be straightened with an air torch. Heated in air it absorbs oxygen and nitrogen, forming a hard, brittle surface layer, and that layer destroys the property the part was chosen for. A titanium part that has been heated in air without shielding has been damaged in a way that no amount of cleaning will undo.
Even with shielding, heat straightening on titanium is limited by the fact that its properties depend on the thermal history and on the atmosphere it was processed in. A shop that welds titanium with a purge and a trailing shield can protect a joint, but correcting a pulled titanium assembly by heating it is a different operation and usually belongs back at the welding stage, where the sequence can be changed instead.
The titanium assemblies we build are handled on that basis, and the parts are shown under titanium fabrication components and TIG titanium welding parts.
Aluminium can be straightened, and heat-treatable alloys are the difficult case. Heating them above a certain range starts to overage the material, so the area that was heated is softer than the rest of the part. The geometry improves and the part gets weaker, and the weakening is invisible.
That trade is worth making on a bracket and not worth making on a structural member. Where the alloy is a heat-treatable grade and the part carries load, a full heat treatment after straightening is the only way to recover the properties, and that is a different process with its own distortion.
There is a difference between bringing a part back to the shape it was designed to have and bending it until it fits. The second is common and it works until the part is under load, at which point the built-in stress does what built-in stress does.
A part that has been straightened carries residual stress from the operation. Where that stress sits at a weld toe or a change of section, it adds to the service stress rather than opposing it. The check is whether the part is straight when free and still straight after a light load, or whether it springs back to a curve when the clamps come off.
Straightening has a limit, and past it the honest answer is to cut the joint that pulled and re-fit it. That is more work, and it produces a part whose stress state is the one the design intended rather than one that has been argued into shape.
The decision usually turns on how much correction is needed, how visible the part is, and whether the material tolerates the heat. A small correction on a stainless tube is routine. A large correction on a heat-treated aluminium bracket, or any correction on titanium in air, belongs in the other category.
Straightening is a corrective operation, and the useful question is whether a part should need it at all.
Muffler and tube assemblies of the kind where this work arises are grouped under stainless steel exhaust system and exhaust pipe, with one example of the finished article being our titanium universal car exhaust muffler. The welding practice behind them is described under TIG stainless steel welding.
