The importance of industrial activities in Sfax (Tunisia) lies not only in the local economy boost, but also in the impact on the urban and natural environments. In fact, Sfax is characterized by two major industrial sectors: Phosphate and petroleum industries, presenting therefore numerous possible hazards caused by pipeline accidental leaks and incidents, tanks or eruptive wells explosions. The risk management in industrial areas such as phosphate and petroleum industries aims to reduce hazards as low as possible in order to protect the environment, the staff and the inhabitants in and around sites. In the proposed study, the areal location of hazardous atmospheres (ALOHA) models have been selected for its simplicity to simulate and assess key hazards like toxicity, flammability, thermal radiation and explosion. Results are used to estimate the individual and societal risk inside and outside the industrial area. These results are useful as operational tools for industrialists and risk experts to assess the vulnerability of the areas surrounding the industry. Finally, the maps obtained from simulating different scenarios can be publicly accessible via google drive to apply emergency measures plan when needed.
Air pollution is among the major threats to human well-being, highlighting the critical need for air quality monitoring, especially in urban areas. Whereas the development of low-cost pollution sensors has facilitated a widespread monitoring, a reliable anomaly detection system is required to properly characterize data for the end-users. In this paper, we propose an enhanced deep learning approach based on the A3T-GCN (Attention Temporal Graph Convolutional Network) model that accurately forecasts particulate matter PM 2. 5 concentrations using real past measurements from a deployed sensor network. Our proposed Enhanced-A3T-GCN embeds all the available spatial and temporal correlations within the sensor network, along with additional information regarding the sensor environment in a graph. It is shown to achieve significant performance improvement with respect to other deep learning forecasting methods, emphasizing the importance of exploiting the sensor environment-based information. Further, the achieved accurate forecasting makes it possible to detect anomalies injected at both single and multiple sensor levels.