Abstract:

Objective To investigate drone-based toxic gas detection methods suitable for hazardous chemical fires or leakage sites, aiming to improve on-site reconnaissance efficiency and reduce safety risks for rescue personnel.

Methods A simulated scenario was designed in which a fire occurred in a high-rise hazardous chemical warehouse. A quadrotor drone equipped with a toxic gas detector was deployed to the simulated accident site to collect and analyze some kinds of airborne toxic gases. To validate the reliability of drone sampling results, three fixed ground reference monitoring stations were established in accordance with toxicant dispersion patterns and drone sampling locations. At each station, the drone and ground monitors simultaneously collected 30 sets of data, and the correlation between the two datasets was analyzed.

Results Pearson correlation analysis showed that the correlation coefficients of the three paired datasets were all greater than 0.94, indicating that drone-based sampling was accurate and reliable. During sampling, eight gases (VOCs, CO, Cl₂, NO₂, H₂S, CO₂, C_xH_y, O₂) were monitored at relative altitudes ranging from 0 to 54.80 m. The sample acquisition rate reached 4,036 samples per hour, outperforming manual methods. Furthermore, contour and 3D point cloud analyses were used to generate detailed spatial distribution maps of toxic gases.

Conclusions Drone-based toxic gas sampling and detection technology enables safe and remote precise monitoring of toxic gases in fire scenarios. This approach can help reduce risks to rescue personnel and enhances sampling efficiency in emergency response to hazardous chemical fires or leaks.