Science & Tech · AI · 3 hrs ago
South Korean institute develops AI system to predict underground gas risks
South Korea’s construction research institute developed an AI system to predict hazardous conditions at underground work sites.
It collects real-time readings of harmful gases, oxygen, temperature, humidity and airflow.
The system can estimate gas levels where no sensors are installed and predict where gas is spreading from and where it began.
Earlier safety systems mainly measured conditions and sounded alarms, while this one is designed to help workers anticipate changes.
In tests, readings from two sensors were used to estimate gas levels at 14 locations, and about 30 seconds of data helped identify where gas was coming from.
The goal is to give workers more time to respond to dangers they cannot see.
The institute says it plans to apply the system in tunnels first and then in other underground spaces.
The Korea Institute of Civil Engineering and Building Technology has developed an AI system to predict hazardous gas risks at underground construction sites.
The system collects real-time readings of hazardous gases, temperature, humidity and airflow, then uses AI to analyze them.
It can estimate gas levels in areas without sensors and predict the direction of gas spread and where gas originated.
Tests found that readings from two sensors could estimate gas concentrations at 14 locations.
The institute said the technology could give workers time to respond to underground hazards.
- Who
- The Korea Institute of Civil Engineering and Building Technology developed the system.
- What
- An AI system measures underground environmental conditions and predicts hazardous gas risks.
- When
- The institute announced the development on November 2, 2026.
- Where
- Underground construction sites; testing included a large-diameter conduit and an underground passage-type tunnel in Daegu.
- Why
- To identify hidden underground risks in advance and give workers time to respond.
This story does not have two clearly opposing sides.
Researchers tested the system in indoor experiments and at a large-diameter conduit and an underground passage-type tunnel in Daegu.
The institute announced that it had developed the underground environmental risk management system.
- Developer
- Korea Institute of Civil Engineering and Building Technology
- Measurements
- Hazardous gas concentrations, temperature, humidity and airflow
- Sensor test
- Two sensors were used to estimate gas concentrations at 14 locations
- Gas-source detection
- The system identified the source area using information collected for about 30 seconds after gas appeared
- Testing locations
- A large-diameter conduit and an underground passage-type tunnel in Daegu
1 source
What people are saying
This technology is meaningful because it can identify risks in invisible underground work environments in advance and give workers time to respond.
Starting with tunnels, we will apply it to various underground spaces and contribute to preventing serious accidents at construction sites.