How AI Is Helping Address Daily Threat of Industrial Fires

In recent years, there has been major technological progress in the safety and security of large, complex premises. For too long, life-safety systems integrators have faced the challenges of gaps in coverage and how to bridge them.

This applies to all aspects of on-site safety, but in an array of industrial and commercial environments, none pose a greater risk of harming onsite personnel, damaging property and disrupting business continuity than fires and combustion.

Frequent reports of devastating fires worldwide continue to demonstrate that standard detection, monitoring and alarm equipment is not effective or fast enough to deal with the rapid onset and spread of fires in hazardous settings, like factories or waste processing plants.

Artificial intelligence (AI) and Industry 4.0 — the fourth revolution that has occurred in manufacturing — in combination with video-based surveillance, has brought fire-safety provision to a whole new era.

Inadequate Protection

Before data analysts and technical developers could design and implement such cutting-edge solutions, they first assessed all issues and challenges that needed to be addressed and overcome.

A global exchange of professional knowledge and experience in the fire-safety industry outlined the pain points and hazards identified over past years.

For instance, classic smoke detectors fixed onto ceilings require smoke or vapor to reach the sensors in their immediate detection field before any alarm is sounded. A delayed response means precious time is lost while a fire takes hold and spreads across an area before the alarm is triggered.

Regular surveillance cameras often produce grainy imagery if they pick up movement in low-light conditions or overnight. Outdoors it is even worse. Raindrops and frost can blur or obscure a lens and prevent it from capturing any distinguishable footage.

Intruders leave a site unidentified, and even continual live surveillance projected over screens will unlikely alert security personnel of any heat buildup or imminent fires. So, whereas many industrial safety managers and systems integrators may have planned and equipped a facility for decent fire-safety coverage, the efficacy of these measures are severely lacking in the level of security they should provide.

Standard equipment can also be hypersensitive to any sign of fire and signal frequent false alarms that disrupt business continuity with site shutdowns. This can even create a dangerous “cry wolf ” effect, so real threats of fire are not taken seriously.

Specific Fire-Safety Challenges

This combination of risk factors and insufficient measures and responses has made radical developments in industrial fire prevention essential. In addition, new technology must be flexible enough for tailored solutions, as a one-size-fits-all approach does not work for the unique requirements of end customers and specific premises.

In the warehouse and logistics industry, large storage depots typically feature high ceilings where smoke detectors are fitted, along with a sprinkler system. Goods and materials being stored change constantly, so the exact fire load or combustion risk at any moment is unknown.

Movement by robotics and forklifts creates hazardous conditions of potential collisions. By the time smoldering vapor travels to smoke sensors, damage to the warehouse is already well under way.

Waste management and recycling plants hold a dangerous cocktail of flammable or explosive chemicals and materials, including lithium batteries, discarded devices, aerosol cans and cleaning products, such as antibacterial waste with a high percentage of alcohol.

Once a fire takes hold at sites containing these and other hazardous items, it can spread rapidly with catastrophic results. Excavators, conveyor belts, and incinerators generate heat and movement which exacerbate the dangers.

The pulp and paper industry produces and stocks huge quantities of raw materials for packaging and transportation of goods. These materials provide perfect kindling. If a flame catches a bale of cardboard, the blaze needs to be stopped within seconds for the storage area to be saved.

Stock is often stored in sheltered constructions, open to outdoor weather conditions such as strong sunlight and flame-fanning winds. Overnight they are vulnerable to intruders, including arsonists and vandals. A deliberate fire on such sites usually results in complete devastation.  Yet standard surveillance and smoke detectors offer no reliable protection.

Speed and Precision of AI Every environment, from industrial sites to office complexes, shopping centers, depots and residential tower blocks needs efficient fire detection and monitoring, and rapid alerts and reactions.

Video-based fire detection with specially developed algorithms recognizes the unique shapes and movement of smoldering, smoke and flames at their sources, before they take hold to cause significant damage. An alarm can be sounded within seconds when necessary.

Algorithms developed based on recorded video footage can be customized by adjustable sensitivity settings. Definable criteria covering a diverse range of possible scenarios and conditions is also integrated. This provides a strong basis for machine learning (ML) as more footage is recorded and collected over time — and anomalies detected — so only real signs of fire are identified for triggering alerts when essential.

AI within the video-based system analyzes any new data received to help determine if there is a real fire hazard or not. The preconfigured algorithms with precise sensitivity settings can then differentiate between sources of real danger and regular objects or non-hazardous activities, such as between sparks or flickering flames and cigarette smoke, and between bright orange flames and a person rushing by in high-visibility workwear.

Other algorithms recognize the specific forms and characteristics of smoldering and smoke drifting upwards, and the illumination and natural behavior of flames. These capabilities make this type of video monitoring highly efficient and robust against false alarms.

It raises an alert only when the danger is real, and a response is necessary. Smart, video-based methods detect fires directly at their source, rather than relying on heat and smoke to first reach the sensors of classic smoke detectors.

Future generations of the technology will be even more advanced and completely AI-driven, with no need for definable criteria input, thanks to a continual accumulation and sharing of real-life data and experience. IoT technology is thus making fire detection and prevention much faster, more precise — and safer — than ever before.

Mitigating Outdoor Dangers

The fact that the video cameras can be installed where they are needed, at optimal angles just like surveillance cameras, and outside under sheltered outdoor constructions, offers huge advantages. Even after operational hours, when staff are safely not on the premises, financial costs of fires to businesses can be huge.

If a fire ignites and takes hold of outdoor storage shelters containing a high fire load, the entire business stock is at risk of being devoured by flames. A fire will not stop at adjacent buildings or any machinery or vehicles nearby; they too can quickly turn to ash.

Thanks to smart communications, AI indicators in video cameras can instantly alert safety personnel and emergency services before the risk of major damage occurs. Another advantage with video-driven safety technology is that monitoring in low-light conditions — overcast days outside or poorly lit spaces indoors — can be covered by specially fitted infrared lighting.

IR illuminators help the cameras to function efficiently around the clock, not just in bright daylight or well-lit areas. So even in pitch darkness, intruders, vandals and any clear sign of fire or smoke are picked up by the cameras thanks to the artificial lighting. Alarms can then be promptly triggered and smart alerts can be sent within seconds.

Housing around the latest surveillance cameras protect their lenses from dust and condensation, so they are continually effective in all conditions.

Future of Fire Prevention

Ignition, kindling and even combustion cannot be prevented altogether with precautionary measures, due to a whole host of hazardous sources and conditions. However, using the latest fire- and smoke-detection technology, dangers of smoldering vapors turning into billowing toxic smoke and small flames growing into devastating fires that cause total damage can be averted.

Limits of classic fire-prevention equipment have been overcome by state-of-the-art developments. Algorithms providing precise recognition of heat, smoke and fire at their source, with customized anomaly detection, all built into video-based solutions, make personnel, property and local environments much safer than ever before.

Reliable and rapid solutions save businesses huge costs in insurance and damage repairs, and business continuity is improved massively with fewer false alarms and shutdowns. By installing smart solutions, premises are safer and more secure, and business costs related to fire damages and frequent interruptions to business continuity are reduced.

From a business perspective, there are clear advantages of this cutting-edge technology: efficiency, as the flame- and smoke-detection technology can be installed at the same site as existing surveillance equipment; and speed, as by reliably identifying a real fire at its source, these solutions trigger a much more rapid response. And they are unmatched in their ability to monitor for fires under outdoor sheltered constructions, even overnight.

Systems integrators offering the latest technologies to clients can, thus, finally meet their demand for round-the-clock fire monitoring in challenging business environments, such as industrial sites featuring high fire loads or hazardous materials.

The new AI and IoT developments, combined with video surveillance, equips them with unprecedented business opportunities in a highly competitive landscape for fire safety provision.

Opportunities for them to position themselves as a supplier of innovative detection and prevention methods which ensure fire safety and business continuity for clients of all sizes are diverse.

---

Theresa Grunewald is the Global Business Development Manager at Bosch Security and Safety Systems.

The post How AI Is Helping Address Daily Threat of Industrial Fires appeared first on Security Sales & Integration.