Do you know your weld repair rate?

· Erman Demirel · 2 min read

On a piping job, most teams can answer "how many welds went to repair?" Most cannot answer "what did that cost us?"

According to TWI's industry survey, the average weld repair rate in oil-and-gas and power work is 1-3%. Local peaks reached 25%, exceptionally 55%. The survey was done in 2011, does not state its sample size, and the rate was calculated in two ways (repaired weld length ÷ total length, or number of repaired welds ÷ total number). For piping systems, all materials together 3.0%.

Careful: this is not the same as the first-radiograph rejection rate. In random sampling (B31.3 guidance talks of at least 5% full radiography for normal service) the share of welds examined and the acceptable defect rate also get mixed up. They have different denominators.

On cost we have no solid unit figure. A customer case page from an equipment manufacturer says "USD 800-1,000 per test" and gives an example of about USD 300,000 a year at 6,000 welds and 5% repair, but does not say what the figure covers (test only, or removal and re-welding too, or delay). It is a marketing page, not a measured project. In general construction, in a 359-project study the direct cost of rework was 4.5% of actual construction cost on owner projects and 2.5% on contractor projects; but that covers rework from every cause, not pipe welding.

So we cannot say "the industry average saves this much". What we can do: find your own number. Your annual weld count, your repair rate, what one repair costs you. If these three come from your own records, the calculation takes a minute. If they do not, that is the real finding.

Go to the calculator

Sources: TWI, "Repair statistics of manufacturing and fabrication welding"; Novarc case page (Metropolitan Mechanical Contractors), manufacturer's account; Hwang (2006) rework study; Painter (2023) B31.3 progressive sampling.

Related articles

In material traceability, the error is usually at the installation, not the certificate

The “wrong material rate” does not reduce to one number. The figures found measure different things; the pattern in documented incidents is common.

Sep 30, 2026 · 1 min read

The honest answer to “how many documents are lost at handover?”

We could not find a measured rate for the percentage of documents lost at handover. What we found, what we did not, and a better measure.

Sep 30, 2026 · 1 min read

A test package “coming back” is not the same as a hydrotest “failing”

A package returned before water and a leak under pressure have different causes, costs and denominators. What we found and what to measure.

Sep 30, 2026 · 1 min read

Join the waitlist.

WeldTracker is not sold as an off-the-shelf package; it is set up by integrating with your organization’s systems. Join the list and we will contact you when it is your turn.