IoT-Enabled Smart Safety System for Real-Time Fire Detection and Suppression in Firecracker Manufacturing Industries

Abstract

Firecracker manufacturing industries are in danger as these involves explosive powder, flammable chemicals and heavy dust. Accordingly, fires can easily break out in such industries. An IoT smart safety system capable of real-time detection of hazards, warning the people beforehand and suppressing the fire incident with least human intervention will be presented further. The device has many environment sensors including temperature (DS18B20, DHT22) smoke (MQ-2) gas sensor (MQ-135) and flame sensor using ESP32 microprocessor. The information received from sensors is sent to a cloud server using wi-fi or low power communication (ZigBee/LoRa). This tool utilizes various machine learning algorithms to perform risk assessments. These algorithms include Random Forest, KNN, and lightweight neural networks. The system quickly responds to the fire in the locality with the alarm, water mist spray, exhaust fans, and circuit shut down. Supervisors can also monitor environmental trends on mobile or web dashboard. Smoke, gas, flame and temperature detection has been done in a laboratory scale fire cracker unit. The AI-powered method successfully detected fires with an accuracy of 82%, recall of 88% and F1-score of 0.80. The proposed system is reliable, scalable, and intelligent and is highly efficient in the safe and fire prevention in fire cracker.