IoT and AI-Based Approaches for Monitoring and Preventing Hazards from Inflammable Materials

Abstract

Inflammable materials pose significant risks in industrial, residential, and transportation environments, often leading to catastrophic fire accidents and economic losses. Traditional hazard detection and safety management systems rely on manual monitoring or delayed response mechanisms, which limit their effectiveness. This research proposes an integrated approach using the Internet of Things (IoT) and Artificial Intelligence (AI) for real-time monitoring, prediction, and prevention of hazards related to inflammable materials. IoT-based sensor networks were deployed to continuously track temperature, gas concentration, and environmental parameters, while AI algorithms analyzed the data to identify early warning signs of ignition risks. A hybrid machine learning model combining Support Vector Machines and Deep Neural Networks achieved high accuracy in hazard prediction, significantly reducing false alarms. Experimental simulations demonstrated that the proposed system improves response time and enhances overall safety. The findings suggest that the integration of IoT and AI can serve as a scalable, cost-effective solution for proactive fire prevention and risk management in industries handling inflammable substances.

Introduction

The text presents a hybrid Internet of Things (IoT) and Artificial Intelligence (AI) framework for early detection and prevention of hazards associated with inflammable materials. The primary objective is to improve industrial safety by continuously monitoring environmental conditions, predicting potential fire risks, and providing timely alerts before hazardous situations escalate.

Traditional fire detection systems, such as smoke detectors, thermal sensors, and manual inspections, often respond only after significant environmental changes or ignition have occurred. The proposed system addresses these limitations by integrating IoT-based sensors with machine learning models to enable real-time monitoring and proactive hazard detection.

The system consists of three main layers: the perception layer for collecting sensor data, the network layer for transmitting and storing information, and the application layer for AI-based analysis. Sensors, including the DHT22 temperature and humidity sensor, MQ-2/MQ-7 gas sensors, and a flame sensor, are connected to an ESP32 or Arduino microcontroller with Wi-Fi connectivity. Data are collected every five seconds and transmitted to a cloud platform using the MQTT protocol.

The collected data undergo preprocessing, including noise removal, normalization, and classification into safe and hazardous conditions. Two machine learning models are employed: Support Vector Machine (SVM) for binary classification and Deep Neural Network (DNN) for more complex hazard prediction. The system provides local alerts through LEDs and buzzers and remote notifications through mobile applications, email, or SMS when potential hazards are detected.

Experimental evaluation involved a dataset of 10,000 sensor readings, divided into 80% training data and 20% testing data. The reported results are as follows:

• DNN: 96.7% accuracy, 95.5% precision, 97.2% recall, and 96.3% F1-score.

• SVM: 91.4% accuracy, 89.8% precision, 90.2% recall, and 90.0% F1-score.

• Traditional threshold-based method: 82.6% accuracy, 80.1% precision, 78.9% recall, and 79.5% F1-score.

These reported results indicate that the AI-based approaches performed better than the conventional threshold-based method on the evaluated dataset, with the DNN achieving the highest scores across the listed metrics.

Conclusion

This study presented an integrated IoT–AI framework for monitoring and preventing hazards from inflammable materials. The IoT-based sensor network successfully captured real-time environmental parameters, while AI algorithms, particularly the Deep Neural Network (DNN), demonstrated superior performance in predicting hazardous conditions with an accuracy of 96.7%. Compared to traditional threshold-based systems, the proposed approach significantly reduced false alarms and improved hazard detection response time from over 12 seconds to under 5 seconds. The results underscore the potential of IoT–AI integration as a scalable, intelligent, and cost-effective solution for enhancing fire safety in industries, residential buildings, and transportation systems handling inflammable substances. By shifting from reactive to predictive safety management, the proposed system contributes to minimizing human, environmental, and economic losses caused by fire accidents.

References

[1] Drysdale D. An Introduction to Fire Dynamics. 3rd ed. Chichester: Wiley; 2011. [2] Rasbash DJ, Ramachandran G, Kandola B, Watts JM, Law M. Evaluation of Fire Safety. 2nd ed. London: Routledge; 2004. [3] Ko B, Ham Y, Nam J. Fire detection based on vision sensor and support vector machines. Fire Saf J. 2009;44(3):322-9. [4] Lloret J, Garcia M, Bri D, Sendra S. A wireless sensor network deployment for rural and forest fire detection and verification. Sensors. 2009;9(11):8722-47. [5] Muhammad K, Ahmad J, Baik SW. Early fire detection using convolutional neural networks during surveillance for effective disaster management. Neurocomputing. 2018;288:30-42. [6] Sharma A, Kumar R. A novel machine learning approach for fire prediction using environmental parameters. J Ambient Intell Humaniz Comput. 2020;11(3):1075-86. [7] Al-Turjman F. Smart-city IoT applications for fire prevention and safety: Trends and challenges. Measurement. 2019;129:589-98. [8] Chen X, Xu Y, Gao Z, Shi H. An IoT-based gas leakage detection system. IEEE Access. 2019;7:63690-8. [9] Chouhan N, Yadav SK, Sinha GR. IoT based smart fire detection and alarm system. Procedia Comput Sci. 2020;167:1545-54. [10] Patel K, Doshi A, Patel J, Patel K. Fire monitoring and control system using IoT. Int J Eng Res Technol. 2018;6(11):1-4. [11] Gaurav A, Singh P, Rajan R. Hybrid IoT and AI-based fire monitoring system for industrial safety. J Ind Inf Integr. 2021;24:100220. [12] Gupta S, Kumar S, Rana A. AI-enabled IoT framework for fire hazard prediction in smart cities. Sustain Cities Soc. 2022;76:103456

Copyright

Copyright © 2026 Ajeet Pratap Singh, Shishupal Singh, Yastuti Rao Gautam. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.