Fire and rescue teams know that many fires don’t just happen. They are instead due to patterns of human behaviour which, if identified in the first instance, could have seen interventions planned and the fires, some of which result in fatalities, could have been prevented from even happening.
There has never been more pressure on housing providers to protect their residents. With more legislative changes occurring in our sector in the past five years than in the previous two decades, it’s a topic I have regularly touched on in recent presentations at fire safety-focused conferences and events.
As the cost-of-living crisis pushed many households into fuel poverty, we are now also seeing additional fire risks through an increase in alternative, unsafe heating and cooking practices happening behind closed doors, stretching fire and rescue services and housing teams’ limited resources even further.
Reinforcing current fire safety strategies with AI
Effective fire strategies have their foundations in passive fire-safety measures (the first strategic pillar) which aim to prevent the spread of fires. These include many of the post-Grenfell recommendations in the UK, with billions invested in improvements to unsafe cladding and communal fire detection.
The second strategic pillar is active fire-safety measures, which focus on fire detection and evacuation, with the introduction of ‘waking watches’, digital fire logbooks and the conversation around PEEPs all having supported improvements in active fire safety.
But neither passive nor active fire strategies aim to prevent the fires happening in the first place; this is where harnessing the power of AI predictive technologies comes into its own.
Identifying patterns of behaviour
In 2010, as an industry first, FireAngel introduced a battery-powered smoke alarm with diagnostic capabilities which stored alarm events in its internal memory. Close to 10 million of these alarms have since been installed during home safety visits completed by fire and rescue services across the UK.
If any of these devices were in properties where a serious fire took place, they were sent to FireAngel’s head office for forensic investigation. The retrieved data soon highlighted that in many cases, there were multiple activations in the weeks or sometimes days before the larger, and in many cases fatal, fires had occurred.
During these precursor occurrences when a smoke alarm was activated, the device would log events, including the duration and frequency of all recent alarm events. These indications reinforced both national statistics and anecdotal messages from fire and rescue services that many fires are caused by repeated behaviours.
Fire Officers are also aware that reduced mobility, hoarding or inebriation are all factors that can impact the time taken to silence a sounding alarm, while devices are regularly removed by residents who smoke or abuse drugs.
At FireAngel, we used the data collected from these activations or ‘near misses’ to develop our AI fire risk tool Predict, which uses a unique patented algorithm.
Building the final pillar of fire safety
Predict provides a step-change in the fire industry that protects residents, properties and communities from preventable fire risks using unrivalled insights. The fire-risk tool provides a real-time view of the active risk in a property in two simple outputs: either low risk with no further action needed or high risk which requires urgent interventions to prevent the probability of a future fire.
This is the only tool available that can identify high-risk behaviours behind closed doors. Predict is built as standard into FireAngel’s Connected smoke and heat alarms which, when installed and connected to the cloud via a gateway, provide ongoing risk-mapping for all properties without manual trawls through spreadsheets or in-depth data analysis.
Predict can integrate alerts into housing providers’ current asset management systems with an option to send alerts to third parties to enable prioritised interventions for those at a critical risk of a fire, thereby reducing the risk to the wider community.
A step-change in keeping communities safe
The trend in fire deaths since 2000 reveals a plateau in recent years. Yet although we’ve seen improved fire-safety measures, fires are still occurring due to human behaviour.
It’s only through a combination of preventative and predictive fire safety, using the insights that AI-powered tools can give us, that we’ll ever achieve our goal of zero deaths caused by fire. FireAngel Predict is unique in its ability to support this goal. For more information, please visit fireangel.co.uk/predict.
Nick Rutter is the co-founder and chief product officer at FireAngel.
