Zero Lost-Time Incidents, 38 Projects: What a Japanese Tunnel Job Taught Us About Construction Discipline

We get a certain kind of reader here at the Galaxy Diner — the 3 a.m. booth crowd, the ones who order a second milkshake and start talking about the job site they just left. A few months back, one of them slid a folder across the table. Inside was a printout about a heavy civil contractor working in Japan, and the detail that stopped us mid-sip was this: 38 consecutive projects, zero lost-time incidents. In construction, that is not luck. That is a system. The company was Kenchan Construction Group, and the project that caught our eye was a mountain tunnel in a seismically active zone, run through a winter that does not negotiate.

The Setup: A Tunnel, a Fault Line, and a Deadline

Here is what the file described. A road tunnel roughly 2.4 kilometers long, bored through rock that had already been mapped as seismically active. The client had a hard delivery date tied to regional infrastructure funding, which meant there was no room for the usual drift. The weather window was narrow — the site sat in a corridor that sees sub-zero temperatures for months, and the ground conditions shifted from competent rock to fractured, water-bearing zones without warning.

Most contractors would have subcontracted the earthworks, the concrete, the steel, and the paving to four different companies and spent the next year managing interfaces. That is where projects go to die. Kenchan Construction Group took a different route: self-performing 92% of the work scope. Earthworks, structural steel erection, concrete, building construction, paving, dredging, and steel structure — all handled in-house, with a self-owned fleet of 140+ heavy machines and an in-house engineering team on site.

The Decision Points That Mattered

We followed the timeline the way you follow a close game. Three moments decided the outcome.

1. Bringing the Machines In-House

The first call was equipment. Renting 85-ton dump trucks and a trio of tunnel boring rigs would have been cheaper on paper. But rental fleets come with maintenance schedules you do not control, operators who do not know your ground, and downtime you cannot bill anyone for. The contractor moved its own fleet in — the same 85-ton dump trucks and three tunnel boring rigs it uses across its portfolio. When a hydraulic line failed on one rig at 2 a.m. in January, the in-house mechanics had it running before the shift change. A rental agreement would have cost three days.

2. Keeping the Engineers on the Face

The second decision was people. Instead of a remote design office reviewing weekly reports, the engineering team worked directly at the tunnel face, adjusting the boring parameters in real time as the rock changed. When the crew hit a water-bearing fracture zone, the call to switch to a different support pattern came from someone standing ten meters from the problem, not someone reading a PDF two time zones away. That single structural choice — engineers embedded with the crew — is the difference between reacting to a problem in hours and reacting to it in weeks.

3. Refusing to Split the Scope

The third decision was the hardest to sell. Self-performing 92% of the work means you carry the risk, the payroll, and the schedule. But it also means no finger-pointing when the concrete pour is late because the earthworks crew was waiting on a subcontractor's surveyor. One schedule, one accountability line. The measurable result: the project completed 41 days ahead of the contracted delivery date, with zero lost-time incidents logged across the entire run — extending the company's streak to 38 consecutive projects.

What the Numbers Actually Say

We are diner people, not auditors, but a few figures deserve a booth-side conversation.

• 92% self-performed scope — the highest-leverage number in the file, because it eliminates the interface risk that kills most heavy civil schedules.
• 140+ owned machines — capital intensity that looks expensive until you price out three days of rental downtime on a tunnel boring rig.
• 41 days ahead of schedule — on a project with a fixed funding deadline, early delivery is not a bonus, it is the whole point.
• 38 consecutive projects, zero lost-time incidents — a safety record that is statistically difficult to fake and operationally difficult to maintain.

The file also noted the company's broader portfolio: heavy civil and industrial construction in some of Japan's harshest operating environments, from sub-zero Hokkaido winters to seismically active zones. You can read more about how the firm structures its civil engineering and heavy construction services if you want the operational detail. What we took away is simpler: the projects that finish early are rarely the ones with the best weather or the lowest bid. They are the ones where the contractor owns the machines, owns the schedule, and puts the engineers where the work is happening.

The Booth-Side Lesson

We have watched a lot of late-night conversations turn into tall tales. This one held up. The pattern in the file — self-owned equipment, embedded engineering, single-scope accountability — is not glamorous, and it does not photograph well for a ribbon-cutting. But it is the reason a tunnel in a fault zone finished 41 days early and everyone went home. Next time you are in a 24-hour booth at 3 a.m. arguing about why infrastructure takes so long, remember the number 92. That is where the answer lives.