Common Mistakes

Common Mistakes in TBM Fabrication and Heavy Civil Equipment (and How to Fix Them)

The most expensive mistakes in TBM fabrication, logistics, and field support, each broken down into symptom, root cause, and a fix you can put a number on.

A tunnel boring machine is a one-off product built to a schedule that cannot slip, so mistakes that would cost a job shop a few thousand dollars can cost a TBM fabricator or tunneling contractor millions. A stalled drive burns 15,000 to 50,000 dollars per day in standing time, and a cutterhead rework can add 6 to 10 weeks to delivery. This guide covers the mistakes we see most often in fabrication, logistics, and field support, each with the symptom, the root cause, and a fix you can quantify and check before it reaches the tunnel.

Symptom: the cutter budget says one cutter change per 200 cubic meters excavated, but the crew is changing cutters at half that interval by ring 300. Root cause: wear was estimated from a reference project with a Cerchar Abrasivity Index of 2.0, and the current ground runs 3.5 to 4.5. Wear scales roughly linearly with CAI above 2.0, so consumption can double. The fix is to rebuild the estimate in the Disc Cutter Wear Rate calculator using the geotechnical baseline report values per reach, not a project average, and reforecast against actual changes every 100 rings.

Symptom: the site runs out of 19 inch disc cutters mid-drive and the machine sits three weeks waiting on a shipment. Root cause: spare inventory was sized to average consumption, but wear in abrasive reaches routinely runs at the 80th to 95th percentile of the estimate, and replacement lead times run 8 to 14 weeks from most suppliers. The fix is to size stock to P80 consumption plus lead time demand using the Spare Cutter Buffer calculator. On a typical 6 kilometer drive that means carrying 60 to 90 cutters on site, not the 25 the average suggests.

Symptom: plate cost on the cutterhead structure comes in 18 percent over quote. Root cause: the estimator assumed 85 percent material yield on 80 to 150 millimeter thick plate, but heavy plate nests poorly. Flame cut kerf of 4 to 8 millimeters, 50 millimeter edge margins for UT test coupons, and grain direction requirements on stress carrying members push real yield down to 65 to 72 percent. Run nests through the Steel Plate Yield calculator before quoting and track remnant usage. Every 5 points of yield on a 120 tonne cutterhead is roughly 9 tonnes of plate, worth 25,000 to 40,000 dollars.

Symptom: fabrication is 95 percent complete but the shop misses the ship date by five weeks because NDT is backed up. Root cause: the schedule carried welding hours but treated inspection as free. A cutterhead and main bearing housing can carry 2,000 to 3,000 meters of Class A weld, and ultrasonic testing runs 1 to 2 meters per hour including reporting, so that is 1,500 to 2,500 technician hours before repair and retest cycles at a 3 to 8 percent reject rate. Load the plan with the Weld Inspection Load calculator and staff NDT at roughly 12 to 18 percent of direct weld hours.

Symptom: installed hydraulic power across the thrust, erector, and articulation circuits totals 40 percent more than the site transformer can feed, or breakers trip during simultaneous thrust and segment erection. Root cause: engineers summed nameplate power instead of applying a simultaneity factor. In practice thrust, cutterhead drive, and erection rarely peak together, and a diversity factor of 0.6 to 0.75 is normal. The fix is to model duty cycles per operating mode in the Hydraulic Power Load calculator, then size prime power to the worst real mode plus 15 percent margin rather than the arithmetic sum of nameplates.

Symptom: hairline cracks appear on 3 to 5 percent of segments at the erector, discovered only after rings are built. Root cause: lifting gear and vacuum plates were rated for the static segment mass, typically 8 to 16 tonnes, with no dynamic amplification. Crane and erector motions impose factors of 1.3 to 1.6, and a worn vacuum seal can drop holding force by 25 percent. Check every lift path in the Segment Handling Capacity calculator using a 1.5 dynamic factor and a minimum 2.0 safety factor on vacuum lifts, and proof test plates every six months.

Two logistics mistakes round out the shop floor list. First, designing a cutterhead or shield that exceeds the 4.9 to 5.5 meter road transport envelope in one piece, which forces a late redesign into bolted segments; run the Transport Cost calculator at concept stage, because a superload permit with escort convoy can add 80,000 to 250,000 dollars per move. Second, budgeting site assembly at OEM brochure figures. Real field commissioning of a 6 to 9 meter TBM runs 10 to 16 weeks with a 6 to 10 person crew, and the Field Commissioning Hours calculator will price that honestly before bid day.

Symptom: the shop quotes 9 month deliveries but ships in 12. Root cause: assembly bays are booked at 95 to 100 percent utilization, leaving no buffer for the reject and rework cycles above; queueing behavior roughly doubles lead time between 85 and 95 percent loading. Plan bays at 80 to 85 percent with the Assembly Bay Utilization calculator and defend the margin. The common thread across all of these failures is estimating from averages and brochures instead of measured project data. Recalculate at every phase gate and compare forecast to actuals monthly, and most of these mistakes show up on paper before they show up in the tunnel.

Published 2026-07-02.