Penetration

Penetration is not the same measurement as bead width or reinforcement height, both of which are visible on the finished surface. It is the depth the fusion boundary reaches into the parent material, and on a single-sided joint with no back access it is also the dimension an inspector cannot check by looking at the top of the weld — the surface can look completely sound while the root is unfused underneath it.

How much penetration a given pass achieves is set mainly by current density and travel speed, with the process itself setting the envelope: gas tungsten arc welding runs cool and shallow, submerged arc drives deep penetration at high current, and everything else sits between them. Heat input is the summary figure most procedures actually control, but heat input and penetration are not interchangeable — the same energy delivered with a wider arc or a wandering torch spreads sideways into the weld pool instead of down into the joint.

Joint design makes penetration a deliberate choice rather than only an outcome. Full joint penetration fuses the complete thickness of the joint and is specified where the joint must carry the design load through its full depth. Partial joint penetration accepts a calculated, smaller effective depth — sized against the load the joint actually needs to carry rather than the full member thickness — and is a legitimate design choice, not a defect, provided the declared depth is what the weld actually achieves. The comparable calculation on a fillet weld is throat thickness rather than penetration depth, because a fillet weld’s governing dimension is measured from the root to the face rather than into the parent material.

When penetration falls short of what a joint was designed and qualified for, the result is classified as lack of penetration — a planar discontinuity that visual inspection from the top surface will not find, which is why volumetric NDT methods lead detection on joints where full penetration is required and cannot be confirmed by eye.