Isa Ali Ibrahim

Gas leaks require an early detection system capable of responding quickly without false alarms . This study developed an IoT system based on MQ-5 sensors and Arduino UNO with a hybrid architecture of Rule-Based and Fuzzy Logic. Rule-Based classifies severity based on concentration and rate of change, while Fuzzy generates alarm intensity through defuzzification to control actuators. Simulation results show a false alarm rate of 0.00% under normal conditions. Light leaks were detected at 202.0 seconds (intensity 31.10), while heavy leaks were detected early in 53.0 seconds (intensity 87.19) thanks to the rate of change heuristic. Although the rule-based severity classification oscillated (33 transitions), the Fuzzy component stabilized the final decision, so that ventilation activation achieved 100% accuracy under critical conditions. This combination of response speed and control stability proves that the hybrid approach is effective in suppressing false alarms , detecting early, and optimizing mitigation, making it feasible to be implemented for smart gas safety. which is more stable, accurate, and contextual compared to a pure threshold system.