Intelligent Fire Detector Detects Fire Within 10 Seconds of Ignition

- 3-cm Flames Detected in the Latent Phase; False Alarms Minimized -

 

The Non-destructive Evaluation Team of the KRISS Safety Measurement Institute and HansunST, a KRISS-affiliated company, developed an intelligent fire detector that distinguishes between fires used by humans and actual fires, and puts out fires within 10 seconds of ignition.

 

The detector is capable of accurately detecting small flames of 3 cm in the latent phase. It is expected to contribute significantly to fire control and evacuation.

 

About 70% to 80% of deaths caused by fires are due to the inhalation of smoke and toxic gases. As such, early detection of fires is essential to prevent damage and minimize casualties.

 

Most fire detectors detect fires in the growing phase, which is 1 minute from ignition. The smoke and flames produced during this phase make fire control and evacuation difficult. Since sprinklers are activated when the indoor temperature exceeds 72 °C, the safety of persons remaining indoors cannot be guaranteed.

 

▲ Comparison of fire-detecting times of fire detectors

 

The intelligent fire detector jointly developed by KRISS and HansunST detects fires in the latent phase, and alerts users through its alarm system and smartphone app. Flame coordinates are displayed, and automatic fire extinguishing is possible by synchronizing extinguishers with the app.

 

The false alarm rate of the intelligent fire detector falls within 3%, which makes it much more reliable than existing smoke or heat detectors with error rates in the range of 34% to 50%.

 

The team achieved such outstanding performance by combining infrared light sensing and infrared thermal sensing technologies, thereby enhancing the flame detection rate. They utilized the key advantage of infrared sensors, which detect flames rapidly based on the wavelength bands of CO2.

 

▲ Basic specifications of the intelligent fire detector

 

 

In addition, AI technology was developed and applied to track human activities in areas where the fire detector is installed, so as to distinguish between fires used for practical tasks and actual fires.

 

▲ Example of a system integrated with the intelligent fire detector

 

 

 

? Existing fire detectors tend to detect fires used by humans as actual fires. To resolve this issue, HansunST used the patent titled “Security camera tracking and monitoring system and method using thermal image coordinates,” transferred by KRISS in 2021.

 

? Using an infrared thermal imaging camera, the fire detector can detect humans and flames at the same time. If both are present, it applies an algorithm that distinguishes between practical fires and actual fires.

 

 

▲ User/administrator interface of intelligent fire detector app

 

 

Man-yong Choi, principal research scientist of the KRISS Safety Measurement Institute, said, “Our fire detection technology addresses the known issues of fire detectors and detects actual fires in a significantly shorter time. We expect to relieve people’s anxiety by producing more reliable fire detectors, and to allow early extinguishing of fires.”

 

HansunST, a KRISS-affiliated company established in December 2015, will complete prototype development in 2021, and pursue commercialization.

 

Su-Un Kim, CEO of HansunST, said, “Currently, we are in the product verification stage, and awaiting fire safety certification. We will cooperate with local governments, security companies, and construction companies, so that our product can be installed in factories, warehouses, restaurants, and offices, as well as general households.”

 

Building upon these achievements, the KRISS research team plans to develop intelligent fire detectors capable of learning and making judgments in various spaces.

 

 

▲ Man-yong Choi (left), principal research scientist of KRISS, and Su-Un Kim (right), CEO of HansunST, are conducting a qualitative reliability test of the intelligent fire detector.

 

 

▲ Su-Un Kim (left), CEO of HansunST, and Man-yong Choi (right), principal research scientist of KRISS