Analysis of early indicators for fire detection using a multi-sensor system

Details

Supervisors

Abreu Araujo, Flavio

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil électricien, à finalité spécialisée

Abstract

This thesis investigates the potential of a multi-sensor system for enhancing early fire detection by monitoring key emissions during the early stages of combustion. Traditional fire detectors primarily focus on smoke particles, often leading to delayed alarms in smouldering fires where toxic gases accumulate to dangerous levels. This research explores whether the combination of sensors measuring carbon dioxide (CO2), Volatile Organic Compounds (VOCs) and particulate matter (PM) can enhance early fire detection. A prototype sensor node was developed to monitor these emissions in real-time. Controlled burn experiments were conducted using common household materials such as rubber, plastic, wood, and paper, to analyse the behaviour of the indicators. The study also explored the optimal placement of sensors within indoor environments to enhance detection efficiency. A second experiment was realised to assess the effectiveness of the multi-sensor system in enhancing early fire detection through a real-time monitoring system based on the MQTT protocol. The results demonstrate that multi-sensor systems can significantly improve early fire detection compared to conventional smoke detectors. Emissions in CO2, VOCs, and PM were detected within minutes of fire ignition, with higher detection efficiency at elevated sensor heights. The research highlights the potential of these indicators to provide a more comprehensive fire detection system.