AMMOSHIELD: AI-Based Smart Ammonia Gas Leak Detection and Industrial Safety Monitoring System

This paper has presented AMMOSHIELD, an AI-based smart ammonia gas leak detection and industrial safety monitoring system intended for deployment in fertilizer, cold-storage, chemical, and food-processing facilities where ammonia handling poses a recognized occupational hazard. The proposed architecture combines ammonia-specific gas sensing with ESP32-based embedded processing, environmental parameter monitoring, and an AI/ML classification layer built around the Random Forest algorithm to produce a graded risk assessment rather than a single fixed-threshold alarm. This classification layer is coupled with IoT/cloud connectivity for real-time remote visualization and historical logging, and with an automated safety-response subsystem that escalates local alarms and exhaust ventilation in proportion to the assessed risk level, while remaining functional independently of network connectivity for critical local safety actions. The hardware implementation, software control algorithm, and evaluation methodology have been described in detail, and illustrative reporting formats for classification and system-response performance have been outlined for use once calibrated experimental trials are conducted. No specific accuracy or performance figures are claimed in this paper, as the system described is a bench-scale design that has not yet undergone the controlled exposure trials and field validation outlined in Sections IX and XVI. Future work will focus on edge-AI and TinyML deployment, multi-gas sensing, sensor fusion, explainable AI, predictive leak forecasting, integration with plant-level SCADA systems, and, ultimately, field validation under real industrial conditions, with the broader goal of contributing a practical, data-driven improvement to industrial ammonia safety practice.  

[1]          G. J. V. Lakshmi, G. Puthilibhai, and S. R. L. Siddarth, "Implementation of Ammonia Gas Leakage Detection and Monitoring System Using Internet of Things," in Proc. 2019 Int. Conf. Communication and Signal Processing (ICCSP), Chennai, India, 2019, pp. 1–6. [2]          A. Rokade, S. Shaikh, S. Nadaf, A. Attar, P. Kandekar, and A. Kaur, "A Survey on Enhanced Safety of IoT Enabled Gas Leakage Detection System," in M. Tuba, S. Akashe, and A. Joshi (Eds.), Springer, 2026. [3]          M. Santiputri, "IoT-Based Gas Leak Detection Device," in Proc. 2018 Int. Conf. Computer, Control, Informatics and its Applications (IC3INA), Tangerang, Indonesia, 2018, pp. 1–4. [4]          R. J. Jadhav, P. Radhakrishnan, S. Mukherjee, R. Goyal, R. Goyal, and S. Thakur, "Internet of Things Enabled Gas Leakage Detection over Industrial Areas Using Powerful MQ Series Sensor and Controller," in Proc. 2024 Int. Conf. Smart Technologies and Systems for Next Generation Computing (ICSTSN), Pune, India, 2024, pp. 1–6. [5]          V. P. Sharma, R. Dugyala, V. Padmavathi, and G. V. R. Gurram, "Gas Leakage Detection System Using IoT and Cloud Technology: A Review," E3S Web of Conferences, vol. 391, art. 01063, 2023. [6]          G. A. Senthil, P. Suganthi, R. Prabha, and S. Sridevi, "An Enhanced Smart Intelligent Detecting and Alerting System for Industrial Gas Leakage Using IoT in Sensor Network," in Proc. 2023 Int. Conf. Intelligent Sustainable Systems (ICISS), Coimbatore, India, 2023, pp. 1–6. [7]          N. Manjunathan, S. Muthulingam, and D. Jaganathan, "IoT Based Gas Leakage Detection System," in Proc. Int. Conf. Sustainable Computing and Data Communication Systems (ICSCDS), 2023, pp. 1–5. [8]          M. Al-Jemeli and M. K. A. Al-Adilee, "Portable Gas Leak Detection System Using IoT and Off-the-Shelf Sensor Node," Indonesian Journal of Electrical Engineering and Computer Science, 2021. [9]          F. Chraim, Y. B. Erol, and K. Pister, "Wireless Gas Leak Detection and Localization," IEEE Trans. Industrial Informatics, 2016. [10]        Bin et al., "Tensor-Based Approach for LPG Leakage Detection from Surveillance Thermal Cameras," IEEE Trans. Industrial Informatics, vol. 17, no. 12, pp. 8296–8305, Dec. 2021.