Nuclear Manufacturing KPIs: Benchmark Ranges for Quality, Delay and Traceability

Seven KPIs cover quality, delay and traceability in nuclear and critical infrastructure manufacturing: first-pass documentation yield, hold point wait time, nonconformance rate, traceability completeness, supplier audit performance, documentation burden ratio, and schedule adherence. Use the ranges below as reference points for spotting a gap, not as targets, and read them as a linked set, because these KPIs trade against each other.

Every figure shows its basis. A superscript number links to a public source in the Sources list, and the sentence gives that source's scope. A figure marked no cited source is kept exactly as originally published because no public source stating it was found, so treat it as illustrative.

At a glance

First-pass documentation yield

Track first-pass documentation yield before anything else, because rejected data packages stop shipment even when the hardware is perfect. Typical shops sit at 70 to 82 percent; world-class programs under the ASME NQA-1 quality assurance standard hold 92 to 97 percent no cited source.

The lever is a structured pre-submittal review against a package checklist, which usually lifts first-pass yield 10 to 15 points in two quarters no cited source. Every failed package adds a full review cycle.

For example, if every rejected package needs another full review and resubmittals can fail again, reviews per accepted package equal 1 / first-pass yield. Moving from 80 to 94 percent takes that from 1 / 0.80 = 1.25 to 1 / 0.94 = 1.064, which cuts package review labor by about 15 percent; if a rejected package needs exactly one extra review, the figures are 1.20 and 1.06, about 12 percent.

How to measure

• Divide packages accepted on first review by packages submitted.
• Count a package as failed if it needs any resubmittal.
• Record the rejection reason with every failed package.

What moves it

• A structured pre-submittal review against a package checklist.
• Templated data packages that leave less to assemble by hand.
• Fixing the most common rejection reason first.

Size package preparation and review effort with the Nuclear Documentation Burden calculator.

Hold point wait time

Hold point wait time is the clearest schedule KPI and the easiest to game if you measure it wrong. Track median hours from readiness to disposition, not the average, since a few long waits distort the mean.

Typical fabrication runs 5 to 9 hours per hold point; disciplined shops that batch inspections and schedule the Authorized Nuclear Inspector on fixed windows reach 1.5 to 3 hours no cited source.

For example, with 9 hold points per traveler, cutting median wait from 6.5 to 2.5 hours recovers about 36 hours of work-in-process time per unit, because 9 x (6.5 - 2.5) = 36.

How to measure

• Log the time each hold point is ready and the time it is dispositioned.
• Report the median wait per hold point, not the average.
• Count hold points per traveler so waits can be added up per unit.

What moves it

• Batching inspections.
• Scheduling the Authorized Nuclear Inspector on fixed windows.

Price the inspections behind each hold point with the Safety-Class Inspection Cost calculator.

Nonconformance rate

Nonconformance rate belongs on the dashboard as nonconformance reports (NCRs) per 100 units and by disposition mix. Commercial-minded shops entering nuclear work often run 18 to 30 NCRs per 100 units early on; mature programs hold 4 to 9 no cited source.

Watch the disposition split too. World-class operations keep scrap and rework under 40 percent of dispositions, using engineering-justified use-as-is and repair for the rest no cited source.

Regulatory clocks run beside your internal ones. Under 10 CFR 21.21, a deviation that could be a substantial safety hazard must be evaluated within 60 days of discovery, or covered by a written interim report within 60 days¹.

Once a director or responsible officer receives information identifying a defect, the NRC must get initial notification within two days and written notification within 30 days². A supplier that cannot evaluate a deviation itself must inform its purchasers or affected licensees within five working days³. These are reporting deadlines, not performance benchmarks.

How to measure

• Count NCRs per 100 units every month.
• Split dispositions into scrap, rework, repair and use-as-is.
• Flag repeat characteristics that recur across lots.

What moves it

• Root-cause discipline on repeat characteristics, which typically removes 30 to 50 percent of recurring NCRs within a year no cited source.
• Engineering-justified use-as-is and repair dispositions.

Size disposition and corrective action effort with the Nonconformance Disposition Time calculator.

Traceability completeness

Traceability completeness must run at or near 100 percent no cited source, so measure the gap, not the level. Track missing or mismatched records per 1,000 required records; world-class is under 2, typical is 8 to 20 no cited source.

A single missing heat number or certified material test report (CMTR) link can hold an entire serialized assembly.

The lever is barcode or RFID capture at the operation, rather than end-of-lot transcription. It drives transcription error from the 1 to 3 percent range down under 0.2 percent and shrinks record labor at the same time no cited source.

How to measure

• Count missing or mismatched records per 1,000 required records.
• Check heat numbers and CMTR links against each serialized assembly.
• Measure the transcription error rate as its own figure.

What moves it

• Barcode or RFID capture at the operation.
• Retiring end-of-lot transcription.

Size record capture and verification effort with the Traceability Record Workload calculator.

Supplier audit performance

Supplier audit performance is a leading indicator of downstream quality. Track major findings per audit and corrective action closure time. Typical supplier audits surface 4 to 8 majors with 45 to 90 day closure; strong programs see 1 to 3 majors and close inside 30 days no cited source.

Requalification interval is the companion metric. NRC Regulatory Guide 1.28 (Revision 6, 2023), which endorses ASME NQA-1, says an applicant or licensee should audit a supplier's QA program on a triennial basis or arrange for such an audit⁴. The 3-year cycle is the standard interval, not a reward. Licensees should also evaluate suppliers every year, reviewing supplier records such as certificates of conformance, nonconformance notices and corrective actions⁵.

Marginal suppliers that stay on annual audits no cited source cost more to oversee. Push suppliers toward fewer findings and you keep them on the standard interval, which directly lowers amortized audit cost.

For example, one audit a year instead of one every three years is three times the audit count, roughly tripling audit cost per unit.

Your own program is audited too. Internal QA audits should cover applicable program elements at least once each year, or once during the life of a shorter activity⁶.

How to measure

• Count major findings per supplier audit.
• Time each corrective action from finding to verified closure.
• Track each supplier's audit and evaluation dates.

What moves it

• Pushing suppliers toward fewer findings.
• Fast, verified closure of corrective actions.

Price an audit program with the Qualified Supplier Audit Cost calculator.

Documentation burden ratio

Documentation burden ratio benchmarks how much non-touch effort each part carries: documentation plus inspection hours divided by touch labor hours. Commercial parts sit near 0.2 to 0.5; safety-class nuclear parts commonly run 1.0 to 2.5, meaning paper and inspection equal or exceed machining no cited source.

World-class shops do not eliminate this; they make it efficient, targeting the lower end of their part's band through templated data packages and reusable qualification records. A ratio drifting above 2.5 signals redundant signoffs or duplicated records worth attacking no cited source.

How to measure

• Add documentation and inspection hours for each part.
• Divide by touch labor hours for the same part.
• Compare within a part's band, commercial against commercial and safety-class against safety-class.

What moves it

• Templated data packages.
• Reusable qualification records.
• Removing redundant signoffs and duplicated records.

Estimate the documentation side with the Nuclear Documentation Burden calculator and touch labor for clean assemblies with the Clean Assembly Labor calculator.

Schedule adherence and cycle time

Schedule adherence and cycle time round out the operational picture. Typical safety-class fabrication shows 60 to 75 percent milestone adherence because delay stacks up unpredictably; disciplined programs reach 85 to 92 percent no cited source.

Cutting the idle fraction of cycle time, often 25 to 40 percent of the total, is usually a bigger win than speeding any single machining operation no cited source.

How to measure

• Measure on-time completion against contract milestones.
• Measure total cycle time, including all hold points and package review.
• Split cycle time into working time and idle time.

What moves it

• Front-loading inspection scheduling.
• Assembling data packages in parallel with machining.

How to build your own baseline

• Set targets as a linked scorecard, not isolated numbers.
• Review the whole set every month.
• When one metric improves at another's expense, confirm the net effect on cycle time and total quality cost before locking the change in.

The KPIs trade against each other. Chasing zero NCRs by over-inspecting will inflate your documentation burden ratio and hold point wait.

A balanced world-class profile looks like 94 percent first-pass package yield, under 3 hour median hold point wait, 6 NCRs per 100 units, under 2 traceability gaps per 1,000 records, and 88 percent milestone adherence no cited source.