1. Start with criticality
Rank how bad a failure of this component would be: safety, the environment, lost production and the cost of the failure itself. A high-criticality component needs more evidence and more margin in the decision. A low-criticality one may justify a simpler repair and a closer watch.
2. Establish the condition, with numbers
"It's badly cracked" is not a condition. Measure it: crack locations, lengths and, where it matters, depths from NDT, remaining thickness from ultrasonic thickness surveys, deformation and wear from measurement or 3D scanning. The decision is only as good as this data.
3. Understand the cause
This is the step most often skipped. If a component cracked through fatigue at a poor weld detail, a like-for-like replacement carries the same detail and can crack again. A well-engineered repair that improves the detail can outlast the original. If the cause is unknown, a short failure investigation is usually cheaper than guessing.
4. Check the repair can actually be built
A repair that looks good on paper still has to be welded. Check:
- Weldability. Castings, quenched and tempered steels and unidentified materials need specific procedures, and sometimes testing, before anyone strikes an arc.
- Access and position. Can the welder reach the joint in a position the procedure allows?
- Restraint and distortion. Repairs in stiff structures carry a higher cracking risk and need a planned sequence.
- Inspection. Can the finished repair be inspected by the method the acceptance criteria assume?
5. Compare whole-of-life cost, including lead time
Compare the options over the life you actually need from the asset, not over the next invoice:
- Repair: engineering and certification, the repair itself, downtime, and any extra inspection afterwards.
- Replace: the new component, freight, installation, downtime and the lead time. A replacement with a long lead time needs a temporary repair anyway.
- Risk: the cost of a failure, weighted by how likely each option makes it over the period.
Lead time often decides it. When a replacement is months away, the question becomes how to repair safely until it arrives.
6. Know when replacement is right
Repair is not always the answer. Replacement is usually the better call when the damage is extensive, when repairs keep failing at the same place, when the material can no longer be welded with confidence, or when the design itself is the problem and the real fix is a redesign.
7. Write the decision down
A defensible decision is a documented one: the options considered, the evidence, the assumptions, the risk and the reason for the choice. If the repaired component will be relied on for structural adequacy, the repair design is independently reviewed and certified by a registered engineer before it goes back into service.
How Veristruct helps
Veristruct provides the engineering behind repair-or-replace decisions: condition assessment, failure investigation, repair design and whole-of-life costing, documented so the decision holds up to review. Veristruct prepares the full technical package and coordinates independent review and certification through registered engineering partners.