IFSJ Exclusive: Rethinking fire prevention in high-risk sites, with viAct

viAct examines the limits of reactive fire protection, fire prevention and argues for continuous monitoring of fire precursors across complex worksites

Industrial fires are often treated as sudden events.

In practice, they are usually the final outcome of conditions that have been developing for hours or days.

Across heavy industry, fire safety has traditionally relied on response mechanisms.

Alarms, suppression systems and evacuation procedures are designed to activate once smoke or heat reaches a defined threshold.

These systems remain essential for compliance and life safety, but they are triggered late in the sequence of events.

From viAct’s perspective, this reactive model leaves a gap between hazard formation and ignition.

Recent incidents have illustrated how quickly that gap can close.

The November 2025 fire in a high-rise housing and renovation complex in Hong Kong spread rapidly through combustible bamboo scaffolding and construction mesh materials.

Once ignition occurred, the speed of fire spread left little margin for intervention.

In environments such as construction, logistics, manufacturing and energy, viAct argues that fire safety effectiveness depends less on response speed and more on how early risk conditions are identified.

Why Traditional Fire Safety Still Leaves a Gap

Industrial workplaces remain among the higher-risk environments for fire.

According to a recent survey of workplace fires, industrial premises accounted for 24.85 percent of all workplace fires in the UK in 2024–25.

The National Safety Council reports that in the United States, fire departments respond to a fire incident every 23 seconds.

The National Fire Protection Association (NFPA) estimated an average of 4,300 construction fires each year, resulting in around 62 civilian injuries, five civilian fatalities and property damage worth $375 million.

Many established fire safety measures, including periodic inspections, manual audits and alarm-based detection systems, are designed to respond once defined thresholds are reached.

As a result, they primarily capture lagging indicators, recording events after ignition has occurred.

Earlier conditions that can contribute to ignition, such as gradual temperature increases, smouldering equipment, undetected gas leaks or the accumulation of combustible dust, may fall outside routine reporting and monitoring processes.

These conditions are often addressed informally or remain undocumented, despite their potential role in later incidents.

This interval between the presence of a hazard and the onset of a fire represents an area where additional monitoring and analysis methods are increasingly being explored.

Real-time fire intelligence in practice

Advances in computer vision, sensor integration and edge analytics now allow risk conditions to be monitored continuously rather than intermittently.

In practical terms, this can involve identifying unsafe states such as blocked fire exits, combustible materials positioned near heat sources, sparks in restricted zones or early smoke particles that fall below alarm thresholds.

By analysing visual data, air quality, temperature change and historical patterns together, fire risk can be treated as a dynamic condition rather than a binary event.

Manufacturing plants have used machine vision systems to alert supervisors when debris blocks heat vents hours before temperatures reach critical levels.

Large logistics hubs have deployed particle-level smoke analytics to identify smouldering pallets long before sprinklers activate.

“Real-time fire intelligence isn’t about replacing fire marshals or safety officers — it’s about expanding their field of vision,” says Gary Ng, CEO of viAct. “When hidden thermal build-up or risky worker routines surface early, safety teams can intervene before danger turns irrecoverable.”

From detection to foresight

In complex industrial environments, risk is rarely static.

Equipment heats gradually, layouts shift and temporary works introduce new ignition sources.

When data from cameras, sensors, maintenance logs and environmental conditions is analysed together, patterns can emerge that are difficult to detect through manual checks alone.

A motor running slightly hotter than its baseline or debris repeatedly accumulating near a vent may appear insignificant in isolation, but can become more meaningful when tracked over time.

viAct views this as a shift from detection to foresight.

A complementary approach

Real-time fire intelligence does not remove responsibility from duty holders.

It also does not replace alarms, suppression systems or trained personnel.

It requires governance, defined thresholds and integration with existing safety management systems.

Industrial sites are becoming larger, denser and more automated, so continuous monitoring approaches are gaining attention as a way to supplement scheduled checks and conventional detection.

From viAct’s perspective, established protections and continuous risk awareness can work together to support earlier intervention.

Fires are sometimes preceded by signals that can be detected, but identifying them in time depends on what is monitored, how consistently it is tracked and how findings are acted on.