A repaired container goes back into a fleet and is then handled, stacked and loaded as if it were sound. Where the repair was not sound, it fails in service, which costs more than the repair saved and occasionally drops a load. Two checks at the point of return – one specific to the repair and one on the container as a whole – catch the ones that would fail, and both take seconds.
Check one: the repair itself
Look at the bond line. On a welded repair the material either side should be fused, not merely adjacent, and a bead that can be lifted at an edge was not fused. On a mechanical repair the fasteners should be seated and the plate should not flex away from the surface. Flex the repaired area by hand: any movement means the repair has not taken.
Check two: the container as a whole
A repair addresses one place, and the container failed there because something was happening to it. Check that the container still sits flat, that the walls are square, that the stacking features seat, and that nothing else in the same area is cracked. A container with a repaired corner and a cracked base joint is not repaired, it is half repaired – and the second crack is the one that will fail.
Load test for structural repairs
Where the repair is to a load-bearing area – a corner, a base-to-wall joint, anything that carries stack load – a static test is worth doing: load the container to its working weight, leave it for a day, and check that the repair has not opened or deformed. That is a day of floor space per container, which is affordable for a sample and not for all of them.
How many to test
Testing every repair is expensive and testing none is how a bad repair method goes unnoticed for months. A sensible approach is to test every repair in the first batch from a new method or a new operator, and then a sample of one in ten routinely. Where the sample fails, the method or the operator needs attention rather than the container.
Marking repaired containers
A repaired container should be identifiable, because it may belong on a manual lane rather than an automated one and because its history matters when it next fails. A durable mark – a date code, a colour dot, a tag – takes seconds to apply and makes the repair programme measurable. Without it, a repaired container that fails again looks like a new failure.
What to do with a failed repair
Record it, and look at why. A repair that fails repeatedly at the same place usually means the repair method is wrong for that damage, or the damage was not repairable in the first place. Both are specification problems in the repair programme rather than workmanship problems, and both are fixable once identified. The sorting rule is in which damage is worth repairing.
Where repaired containers should go
Back into the same duty if the repair restored full strength, and into a less demanding duty if it did not. A container with a welded corner is fine for manual handling and questionable in an ASRS slot, where the tolerance is tight and the consequences of a failure are a stopped aisle. Routing repaired containers deliberately is cheaper than repairing them to a higher standard.
What the record should say
Container identity, damage, repair method, date, and the result of the check. Five fields, and they make the programme auditable – which matters when a customer asks how containers are maintained, and it matters internally when the repair budget is reviewed.