Why Underground Mine Fires Keep Happening Despite Falling Numbers

A hydraulic fluid leak is not a dramatic mechanical failure. It is one of the most routine warning events an underground operator encounters, the kind that triggers a pressure alert and a slowdown, not an evacuation.

Yet at a New South Wales underground metal mine, that ordinary leak produced an open fire behind a load haul dump vehicle in seconds. The gap between routine and catastrophic turned out to be uncomfortably narrow.

That gap is why this incident is worth analysing beyond its own facts. The NSW Resources Regulator recorded 179 mobile plant fires in FY25, the lowest in five years, and still describes every one of them as entirely avoidable. Underground metalliferous mines accounted for 15% of that total.

The load haul dump (LHD) fire illustrates precisely the failure mode the regulator has been flagging since Safety Bulletin SB13-05 in 2013: ignition sources and fuel sources that are meant to be separated, but that operational reality allows to converge.

What follows in this analysis gives you a working picture of underground mine fire safety as it actually breaks down: where existing controls tend to fail, what the regulator has specifically asked operators to fix, and what the vehicle stability dimension adds to a risk picture that is already more layered than most fleet reviews assume.

How a low-pressure warning became a fire: the LHD incident mechanics

The sequence started with something entirely normal. The operator released a load from the bucket, and moments later a low hydraulic oil pressure warning triggered.

Then the pressure reading dropped to zero. Flames appeared at the rear of the machine. The operator engaged the onboard automatic suppression system, and when that left a residual fire still burning, finished the job with a portable hand-held extinguisher.

Reconstructed step by step, the chain looks like this:

1. The operator releases the bucket load, a routine manoeuvre.
2. A low hydraulic oil pressure warning triggers.
3. Pressure drops to zero as the leak accelerates.
4. Hydraulic fluid reaches a heated engine component and ignites.
5. Flames become visible at the rear of the machine.
6. The automatic suppression system engages, and a portable extinguisher is required to fully control the fire.

There is nothing exotic in that sequence. The failure mode was hydraulic fluid contacting a hot surface, a scenario the NSW Resources Regulator has directly connected to exhaust systems and turbochargers as established ignition sources.

NSW Resources Regulator Failures in oil and fuel lines near exhaust systems and turbochargers are established causes of underground fires.

The most revealing part of this incident is not that it happened, but how it ended. The two-stage suppression response tells you something the marketing materials for suppression systems tend not to.

An automatic system engaged, and a fire still remained. A person had to physically finish the job with a hand-held extinguisher.

That is a finding about suppression adequacy under real conditions, not a footnote. It means the automatic system was a first response, not a resolution guarantee. Operators who treat their installed suppression as the primary control are accepting a residual exposure the regulator explicitly does not endorse.

This is not the first serious mobile plant fire NSW has investigated. The Cadia East underground loader fire (IIR21-14, incident 25 October 2021) left a worker with serious burns, a reminder that these events carry human consequences well before they reach the statistics.

For an operator, the value of walking the sequence at this granularity is precise. It lets you locate the exact decision points and proximity conditions in your own fleet that replicate the failure mode, rather than filing the LHD fire away as an isolated anomaly that could not happen on your site.

What the statewide data reveals about systemic failure patterns

The headline number reads like good news. The NSW Resources Regulator’s FY25 report, published 23 October 2025, records 179 mobile plant fires, a 14% decrease on the 206 incidents in FY24 and the lowest count in five years.

Read that again with the regulator’s own framing attached. A 14% reduction in a category described as entirely avoidable is not proof the controls are working. It is a signal that the drivers behind these fires are structural, not random, because a category that could in principle be zeroed out is instead settling into a persistent annual baseline.

The distribution matters more than the total. Surface coal mines accounted for 79% of FY25 fires, underground metals mines for 15%. That places the LHD incident inside a smaller but stubborn fire population defined by confinement, limited escape routes, and ventilation contamination risk.

For an operator, that 15% is not a comfort figure. It means fires occurring in the most consequential physical environment there is, where a blocked decline or contaminated ventilation can turn a contained machine fire into a site-wide emergency.

The common thread across incidents sits in maintenance. WorkSafe New Zealand and the NSW Resources Regulator repeatedly attribute fires to component failure and inadequate inspection, and the recurring deficiencies form a recognisable inventory:

• Corroded fittings that fail under pressure
• Pin-hole leaks that spray rather than pool
• Failed hose seals allowing fluid to escape near heat
• Build-up of combustible material close to hot surfaces
• Leaks that operators notice but maintenance never traces back to the underlying corrosion or degraded fitting

That last point is where the systemic pattern becomes undeniable. A leak identified but not resolved is a control that failed twice: once at the component, once at the maintenance response.

Historical data reinforces the mechanism. Analysis of in-service mobile plant fires between September 2014 and May 2017 recorded 203 fire events, and high-temperature diesel exhaust and turbocharger surfaces were the ignition source in 69% of fires during that period. The ignition physics has not changed in a decade.

Neither has the difficulty of embedding fixes. Russell Vale recorded five gas-fire and frictional ignition events in under two years, between April 2022 and 2024, prompting the regulator to sharply increase compliance activity at the site.

Five events at one location tell you the same thing the statewide trend does. Systemic failure can persist across multiple incidents at the same operation, which is why the regulator treats individual fires as indicators of structural breakdown rather than one-off bad luck.

Understanding why hydraulic fluid fires are hard to stop once started

The starting point is the fluid itself. Mineral-oil-based hydraulic fluids are combustible and relatively easy to ignite, and under pressure they behave far more dangerously than a spilled puddle would suggest.

A pressurised leak from a hose, fitting, or pump does not pool. It atomises into a spray or mist with a vastly larger ignition surface area, so contact with a turbocharger or exhaust surface is immediately dangerous, according to guidance from the UK Health and Safety Executive (HSE) and the US National Institute for Occupational Safety and Health (NIOSH).

Then there is the amplification pathway. When engine shutdown does not occur, or occurs too slowly after ignition, the hydraulic pumps keep delivering fluid to the fire.

That is the mechanism that turns a manageable-looking fire into an escalating one. The fire has a fuel supply that continues feeding it until the flow is physically stopped.

Why suppression systems alone are not enough

Commercial suppression systems for underground mobile equipment, including ANSUL A-101, AFEX, and Fogmaker, discharge dry chemical, foam, or water mist to arrest fires early. In Australia these systems are built to AS 5062; the equivalent US benchmarks are NFPA 17 and 17A.

The standard that matters most, though, is integration. A suppression system must be correctly designed, installed, maintained, and tied into engine shutdown to work as intended, and where any of those fails, the fire outlives the discharge.

Coverage is the structural weak point. A WorkSafe Victoria case study of an underground truck fire found the vehicle’s aqueous film-forming foam (AFFF) system ineffective because it did not cover an overheated drive-shaft bearing, and a handheld extinguisher was required to control the blaze.

That is not a product defect. It is a coverage gap, a mismatch between where the nozzles were aimed and where the fire actually started, and it is exactly the pattern that repeated in the NSW LHD incident where a second, manual intervention was needed.

NSW Resources Regulator Suppression and detection systems do not eliminate the risk of a fire occurring and must not be relied upon as the primary control.

Compliance with AS 5062 confirms the system meets the standard. It does not confirm the nozzles cover the drive-shafts, thermal lagging transition points, and other locations where the next fire is most likely to start.

For an operator treating installed suppression as the primary control, the coverage-gap finding is the uncomfortable read. You may be accepting residual exposure at precisely the points your suppression design never accounted for.

That is the foundation for what comes next. Without understanding why suppression falls short as a primary control, the regulator’s layered engineering and maintenance prescriptions can look disproportionate. They are not.

What regulators are asking operators to fix, and what they are not saying

Following the LHD fire, the NSW Resources Regulator issued five preventive recommendations. Each one reads simply, and each carries an operational implication that is easy to miss.

1. Manage thermal lagging joints and transition points. This is not a materials question but an inspection frequency and documentation question: a gap at a lagging joint is invisible unless someone is scheduled to look for it and record what they find.
2. Physically separate fuel sources and ignition points. Separation on the original design drawing means nothing if maintenance or a field repair has since routed a hose closer to an exhaust surface.
3. Ensure fire prevention controls remain active and effective over time. A control installed once and never re-verified is an assumption, not a control, and the burden of proving ongoing effectiveness sits with the operator.
4. Regularly review inspection schedules, housekeeping, and emergency response protocols. The value is in the review cadence: standards drift downward quietly unless something forces a periodic reset against the minimum required.
5. Apply strict oversight and quality assurance to all maintenance and repair work. Maintenance itself introduces risk when a repair creates a new leak path or removes protective shielding, so the quality-assurance layer is what catches the fix that made things worse.

Read together, these five recommendations sit within a long-running regulatory position. Safety Bulletin SB13-05, titled “Too many underground fires”, dates back to 2013, and the “Fires on mobile plant” safety guidance was last updated 8 June 2026, signalling an actively maintained framework rather than a one-off response.

The regulator’s stance is unambiguous: all mobile plant fires are avoidable. That is a zero-tolerance framing, and it sets the standard current practice is measured against.

Here is the harder observation. These recommendations are framed as good-practice guidance, not mandatory compliance instruments, which means accountability for embedding them sits entirely inside the operator’s own safety management system.

The gap between the regulator issuing guidance and an operator making it live in daily maintenance is where most of these fires actually originate. Russell Vale’s five repeated ignitions are the evidence: guidance existed, and it did not reliably translate into sustained operational change.

So the question these recommendations really pose is not whether you have read them. It is whether your quality-assurance system would catch a lagging joint gap or a degraded hose fitting before it becomes an ignition event.

For an operator, that reframes the list. It stops being a checklist to tick once and becomes a self-audit framework you have to keep re-running.

The vehicle stability dimension operators often overlook

Most operators file fluid tank behaviour under load management, not fire safety. That separation is where a real gap opens up.

In a partly filled hydraulic tank on underground mobile equipment, manoeuvre-induced fluid motion shifts the vehicle’s centre of gravity. Dynamic analyses of tank vehicles, including PLOS ONE studies on tanker stability, show that lateral sloshing forces increase rollover torque, while longitudinal sloshing during braking can introduce yaw instability.

Underground, those consequences sharpen. Grades, confined spaces, and limited visibility make a rollover more severe than the same event would be on a surface bench.

The connection to fire is not obvious until you trace the fluid. The same dynamics that make a partly filled tank a stability risk also govern how fast fluid reaches an ignition source once a line is breached.

The engineering answer addresses both problems at once. Internal baffling and compartment separation are usually specified for stability, and research shows they:

• Reduce slosh amplitude
• Suppress peak slosh forces
• Dissipate energy from fluid motion
• Raise rollover thresholds

The same baffling that steadies the vehicle also restricts uncontrolled fluid distribution when a structural leak or breach occurs. One control, two risk registers.

The dual-purpose read In underground confined environments, baffling and compartmentalisation are not single-purpose add-ons. They limit dynamic loads that cause loss of control or rollover, and they restrict how freely fluid spreads after a breach, earning their place as both a stability control and a fire-safety control.

This dimension sits outside the regulator’s five LHD recommendations entirely. That makes it an operator-identified gap rather than a regulator-mandated fix, which is exactly why it gets overlooked.

The implication is direct. An operator who has actioned all five recommendations but never reviewed baffling adequacy across the underground fleet has closed the fire side of the picture while leaving the stability side, and the interaction between the two, unexamined.

Seeing stability and fire risk as connected systems rather than separate workstreams is the perspective that lets you design controls that are genuinely dual-purpose instead of siloed.

Where the risk picture stands and what operators should prioritise first

Three findings converge across this analysis. Suppression is a secondary control, not a primary one. Maintenance quality assurance is the mechanism through which most of the regulator’s recommendations live or die. And baffling adequacy is a gap sitting outside current guidance but operationally tied to fire risk.

For operators with limited review capacity, the sequence matters:

1. Thermal lagging joint inspection. It sits directly in the LHD failure pathway, where a joint gap let fluid reach a hot surface.
2. Suppression system coverage mapping. The coverage gap is what forced a manual extinguisher in both the NSW and WorkSafe Victoria cases, so verifying nozzle placement against real fire zones comes next.
3. Baffling review. Important and dual-purpose, but one step removed from the immediate LHD failure chain, so it follows the first two.

What the FY25 data does not resolve is the bigger point. 179 fires in a category the regulator calls entirely avoidable means the structural drivers are not yet consistently controlled, and with the “Fires on mobile plant” guidance last updated 8 June 2026, the framework is one operators should be actively tracking.

The next incident is more likely to follow the LHD pattern than to arrive as something novel.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions.