01Executive summary
Two machines that move in opposite directions and solve the same class of problem: how to make ground and height usable without disturbing what is already there.
Henri-Joseph Maus built a mechanised tunnelling machine in 1845 for the Alpine crossing between France and Italy — a large percussive rig, driven from outside the tunnel, intended to replace hand drilling. It was superseded before it could prove itself, but the concept it established was correct: excavate mechanically and continuously rather than by drill-and-blast cycle. Sixteen years later Elisha Otis patented a steam passenger lift, and in 1880 Werner von Siemens demonstrated an electric one. In both cases the enabling invention was not the motion but the safety of the motion.
02Tunnel boring machines
A modern tunnel boring machine is best understood not as a drill but as a mobile factory that manufactures a lined tunnel continuously. Excavation is only one of the things it does, and arguably not the hardest.
Cut
A rotating cutterhead carries disc cutters or drag picks across the full face. In rock, disc cutters induce tensile failure and chip material away rather than grinding it; the machine is exploiting rock’s weakness in tension.
Support the face
In soft or water-bearing ground the face must be actively supported. Earth pressure balance machines use the excavated spoil itself, held under controlled pressure in the chamber; slurry machines use pressurised bentonite. Losing face support means settlement at the surface.
Remove spoil
Cuttings pass through openings in the head to a screw conveyor or slurry circuit, then to belt or rail haulage. Spoil handling rate frequently limits advance rate, not cutting rate.
Line and grout
Behind the shield an erector places precast concrete segments into a complete ring. Annular grout fills the gap between the ring and excavated ground, controlling settlement and locking the ring into position.
Advance
Hydraulic rams thrust against the last completed ring to push the head forward. The tunnel lining is therefore also the machine’s reaction structure — the tunnel pushes itself forward against what it has already built.
Steer and survey
Differential ram pressure articulates the shield. Continuous laser guidance against a survey control network holds line and level to tolerances measured in tens of millimetres over kilometres.
The undisturbed surface is the whole point
Cut-and-cover works are cheaper per metre in shallow ground, but they consume the surface: roads close, services divert, buildings are underpinned or demolished, and disruption runs for the full length of the works. Bored tunnelling transfers that cost into the machine and the lining. In dense urban environments the surface disruption avoided is usually worth far more than the additional construction cost — which is why every major Australian city has moved to bored solutions for its recent metro and motorway projects.
Ground loss at the face or around the shield produces a settlement trough at the surface, and structures within that trough experience differential settlement, distortion and cracking. The controls are face pressure management, prompt and complete annular grouting, and continuous monitoring with defined trigger levels and response plans. A tunnelling project’s risk register is largely a settlement risk register.
Australian tunnelling is dominated by Sydney’s Hawkesbury sandstone and Melbourne’s mixed-face conditions, and these behave very differently. Sydney sandstone is generally competent and self-supporting, which favours roadheaders and open-mode operation; Melbourne’s Coode Island Silt and variable basalt drive closed-mode machines and much tighter face pressure control. Work health and safety obligations for tunnelling sit under the model WHS Regulations with specific provisions for underground work, atmospheric monitoring and emergency egress. Underground services location before any excavation is a statutory obligation, not a courtesy — strikes on high-pressure gas and high-voltage assets remain a recurring cause of serious incidents.
03The lift: selling arrest, not ascent
Hoists predate Otis by centuries. Ropes, drums and counterweights were entirely familiar technology. What did not exist was any reason for a person to get in one. A rope is a single point of failure with a fatal consequence, and no amount of rope quality changes the fact that the failure mode is unarrested free fall.
Otis’s safety gear inverted the logic. Tension in the hoist rope holds a spring-loaded mechanism clear of the guide rails. If tension is lost, the spring drives pawls or wedges into the rails and arrests the car. The failure of the primary system is what triggers the protection. This is fail-safe design in its purest form, and it is the reason the invention is remembered: it converted a fatal failure mode into a stopped lift.
| Layer | Protects against | Principle |
|---|---|---|
| Multiple suspension ropes | Single rope failure | Redundancy — remaining ropes carry the car |
| Overspeed governor | Descent above rated speed | Independent speed sensing, mechanically linked to the safety gear |
| Safety gear | Free fall or overspeed | Wedges grip the guide rails; energy absorbed progressively |
| Terminal limit switches | Overtravel at top or bottom | Independent of the normal control system |
| Pit and headroom buffers | Residual impact energy | Controlled deceleration at the extremes of travel |
| Interlocked landing doors | Shaft access when no car is present | Car cannot move unless every door is closed and locked |
Why the counterweight matters more than it looks
A traction lift balances the car against a counterweight equal to the car mass plus roughly half the rated load. The motor then only has to accelerate and overcome the imbalance, not lift the whole car. This reduces installed motor power by a large factor, reduces energy consumption, and means that at about half load the system is nearly balanced. It is the same principle as a bascule bridge or a sash window, applied to make a building habitable.
People will reliably walk four or five storeys and no more. Until the lift was trusted, that constraint capped building height regardless of what structural engineering could achieve. Steel frames made tall buildings possible; safe lifts made them usable. Neither would have produced the modern city alone, and it is worth noticing how often a structural breakthrough waits on an apparently unrelated services technology before it delivers any value.
Lifts and escalators are designed, installed and maintained to the AS 1735 series,
with registration, inspection and maintenance obligations for plant under the model WHS
Regulations. Two design items are commonly under-considered at concept stage: fire and emergency
lift provisions under the National Construction Code, and accessibility requirements under the
Disability Discrimination Act and AS 1428, which affect car dimensions,
control heights and door timing. Both are far cheaper to accommodate in the shaft layout than to
retrofit.
04Takeaways
Fail-safe beats fail-strong
Design so that loss of the primary function triggers protection, rather than relying on the primary function not failing.
The machine builds its own reaction
A TBM thrusts against the lining it has just placed. Look for structures that serve construction and service life both.
Balance the load before powering it
Counterweighting reduces installed power more effectively than any efficiency improvement to the drive.
Trust is an engineering deliverable
Neither machine was adopted because it worked. Both were adopted because people believed they would keep working.