<![CDATA[Gas leaks represent one of the most critical triggers of residential and industrial fire accidents, particularly in environments reliant on Liquefied Petroleum Gas (LPG). Conventional threshold-based detection systems exhibit fundamental limitations in interpreting the gradual and simultaneous dynamics of multiple hazardous parameters, rendering them prone to delayed responses or false alarms. This study proposes an intelligent early warning system for fire detection based on gas leakage, employing the Mamdani Fuzzy Inference System (FIS) implemented on an Arduino Uno microcontroller. The system integrates three environmental input variables Gas Concentration (MQ-2 sensor), Flame Distance (IR flame sensor), and Ambient Temperature (DHT22 sensor) to determine the proportional speed output of a DC exhaust fan via Pulse Width Modulation (PWM) control. A rule base of 27 IF-THEN rules governs the inference process. The system was validated through MATLAB Fuzzy Logic Toolbox simulation and direct hardware implementation, yielding an average error rate of 0.22% between simulated, hardware computation, and actual outputs. The results demonstrate that the proposed multi-parameter Mamdani fuzzy system provides a significantly more adaptive and precise hazard assessment compared to conventional single-threshold approaches, offering a robust foundation for smart safety system deployment.]]>
Copyrights © 2026
