Pressure Vessel Fabrication KPIs: Benchmarks for Welding, Radiography, Fit-Up and Throughput

Use a benchmark as a reference point for spotting a gap worth investigating, not as a target. Your own trend, segmented by process and vessel complexity and measured the same way every month, says more about your shop than any published range.

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

At a glance

Weld deposition rate and arc-on ratio

Weld deposition rate is the throughput backbone. Measured as deposited pounds per arc-hour, SMAW typically runs 3 to 5 lb/hr, FCAW 8 to 12 lb/hr, and SAW 15 to 25 lb/hr on positioned girth seams¹.

Those bands sit inside the indicative ranges TWI publishes for welding continuously, without breaks for electrode changes or deslagging: 0.4 to 5.5 kg/h for manual metal arc (SMAW), 1.0 to 15 kg/h for FCAW and 3 to 16 kg/h for single-wire SAW¹. World-class shops push SAW toward 30 lb/hr with tandem heads no cited source.

The deeper lever is arc-on ratio, arc time divided by clock time. Typical shops run 20 to 30 percent², and best-in-class shops run 35 to 45 percent no cited source.

TWI's duty cycles run 15 to 30 percent for manual metal arc welding and 25 to 45 percent for FCAW, while mechanized submerged arc welding runs 80 to 95 percent². Those figures exclude set-up and assembly time, so a whole-shop ratio sits at or below them, and they explain why positioned SAW girth seams carry the throughput.

For example, at fixed arc hours clock time scales with 1 / arc-on ratio, so raising arc-on from 25 to 40 percent shortens the weld schedule by 1 - 25/40, or 37.5 percent, a little more than a third.

How to measure

• Deposition: deposited pounds per arc-hour, reported by process.
• Arc-on ratio: arc time divided by clock time, for each welder or station.

What moves it

• Positioners, so more seams run flat under SAW.
• Prestaged consumables.
• Fewer setup interruptions.
• Tandem SAW heads on long girth seams.

Nozzle welding is where arc-on time is hardest to hold; price it with the Nozzle Labor calculator, which carries a productive-time factor.

Radiographic rejects and weld repair rate

Radiographic (RT) reject rate signals both quality and hidden cost. A public reference point is TWI's 2011 industry survey of weld repair rates, which counts repaired welds over total welds: pressure vessels averaged 2.0 percent across all material grades, 2.8 percent in mild steel and 1.1 percent in stainless steel³.

In the same survey, piping systems averaged 3.0 percent, and average repair rates in the oil and gas and power sectors ran 1 to 3 percent, with peaks up to 25 percent in specific locations³. World-class shops hold under 2 percent on carbon steel no cited source.

For example, each rejected joint means grind-out, re-weld and re-shoot, so a shop at 8 percent on 148 ft of weld carries rework hours the schedule never planned.

Moving from 8 to 3 percent reject typically recovers 4 to 7 percent of total weld labor no cited source. TWI's survey names welder skill, weld accessibility and poor fit-up among the major factors behind repair rates³.

How to measure

• Divide joints rejected on the first radiograph by joints radiographed.
• Split rejects by welder, process and position to find the driver.

What moves it

• Root-pass training, since out-of-position root passes are a frequent driver.
• Dry flux, since moisture in flux causes rejects.
• Better fit-up before the root pass goes in.

Rank joints for inspection effort by severity, likelihood and detection difficulty with the Weld Length Estimate calculator, a weighted weld-risk index.

Fit-up first-pass yield

Fit-up first-pass yield gauges assembly discipline. It is the share of shell courses, heads and nozzles that pass dimensional and alignment checks (peaking, ovality, weld gap) on the first attempt.

Typical is 80 to 88 percent, and world-class exceeds 95 percent no cited source. ASME limits like shell out-of-roundness (UG-80) and joint alignment (UW-33) are pass/fail no cited source, so re-fits burn crane and fitter time.

How to measure

• Divide components passing dimensional and alignment checks on the first attempt by components checked.
• Record which check failed: peaking, ovality or weld gap.

What moves it

• Laser trackers for large-diameter roundness.
• Peaking checks on rolled courses before tacking.
• Dedicated fit-up fixtures rather than chain and come-along.

Feed measured hours back into your fit-up and nozzle labor estimates so future quotes stay honest; the Nozzle Labor calculator covers the nozzle side.

Plate utilization

Plate utilization protects material margin. Nesting yield is finished part area divided by purchased plate area.

It benchmarks at 78 to 85 percent typical and 88 to 92 percent with nesting software and common-line cutting no cited source.

For example, on a vessel buying 12 to 22 percent of its plate as drop, moving yield up five points on 20,000 lb of SA-516 at $1.20/lb saves 20,000 lb x 5% x $1.20, or about $1,200 per vessel.

How to measure

• Divide finished part area by purchased plate area.
• Compare purchased plate weight with finished weight for each vessel.
• Review the highest-drop nests monthly.

What moves it

• Nesting software.
• Common-line cutting.
• Reusing remnants from the drop bin.

Work out the share of purchased plate that ends up in shells, heads and nozzles. Plate Yield

Fabrication hours per finished ton

Labor productivity per vessel ton normalizes across job sizes. Express it as fabrication man-hours per finished ton.

Simple carbon steel vessels run 40 to 80 hr/ton, while complex or exotic-alloy, heavily nozzled vessels can exceed 150 hr/ton no cited source. Best-in-class repeat production hits the low end of each band no cited source.

Watch the trend, not the absolute, because nozzle count and code class shift the baseline. A vessel with twelve nozzles versus four can add 20 to 40 hr/ton on its own no cited source.

How to measure

• Divide fabrication man-hours by finished vessel tons.
• Segment by complexity class before comparing shops or quarters.

What moves it

• Nozzle count and code class, which set the baseline.
• Repeat production, which pulls hours toward the low end of the band.

Estimate how nozzle count drives hours with the Nozzle Labor calculator.

Hydrotest success and schedule milestones

Schedule and hydrotest KPIs catch flow problems. Track first-pass hydrotest success rate, benchmark above 95 percent no cited source.

A failed test means locating the leak, re-welding, in some cases post-weld heat treatment (PWHT) re-qualification, and a full re-test, often a lost week. One failed hydrotest can consume the slack of three on-time vessels.

How to measure

• Divide vessels passing hydrotest on the first attempt by vessels tested.
• Track on-time completion of shell, heads, nozzles and NDE milestones against the plan.

What moves it

• Catching fit-up and weld defects before the vessel reaches test.
• Scheduling test-bay fill and drain windows rather than improvising them.

Size test-bay throughput from vessels per cycle, available cycles, uptime and first-pass hydrotest yield. Hydrotest Capacity

Rework as a share of direct labor

Rework and scrap rate is the summary quality metric. Target total rework under 3 percent of direct labor hours world-class, versus 6 to 10 percent typical no cited source.

Pair it with cost of poor quality: rework hours times loaded rate, plus wasted consumables and NDE. Shops that make rework visible and review it weekly usually cut it in half within two quarters no cited source, mostly by fixing fit-up and root-pass technique rather than chasing final-pass cosmetics.

TWI's survey points the same way: peak repair rates typically occur in root runs, fillet welds and areas with limited access³.

How to measure

• Capture every grind-out, re-fit, re-weld and re-shoot as rework hours, not as normal production, or the number lies.
• Divide rework hours by direct labor hours.

What moves it

• Fit-up discipline.
• Root-pass technique.
• A weekly rework review that keeps the number visible.

A deep grind-out can eat into wall thickness; check the remaining margin against the code minimum with the Shell Thickness Allowance calculator.

How to build your own baseline

Instrument the front of the line, where errors are cheapest. A peaking or ovality miss caught at rolled-course inspection costs minutes; the same defect found at hydrotest costs a week.

• Write each KPI's definition down and keep it fixed.
• Segment by process and complexity class before comparing periods.
• Prioritize levers by payback: positioners and arc-on discipline for throughput, fit-up fixtures and laser roundness for first-pass yield, root-pass training and dry flux for radiographic rejects.
• Set one target per KPI per quarter and feed measured actuals back into your cost and labor estimates.

For example, one quarter's target might be an arc-on ratio of 25 to 32 percent.

Safety is one KPI with a public industry average. In 2024, U.S. heavy-gauge metal tank manufacturing (NAICS 33242) recorded 3.9 total recordable injury and illness cases per 100 full-time workers, with 2.5 days-away, restricted or transferred (DART) cases, against 2.7 total recordable cases for all U.S. manufacturing⁴.