Analysis of Ambient Air PM10-Bound Pollutants Surrounding an Industrial Site and Their Prediction Using Artificial Neural Network

Abstract

The 2030 Agenda dictated the Sustainable Development Goals. It states the waste reduction needs through their reuse, i.e., considering them as secondary raw materials (Objective 12.5). Bottom ashes from municipal or industrial incinerators can be reused as partial cement replacement in concrete after preventive physical processes such as ferrous metals removal (magnetic separation) and nonferrous metals removal (Eddy current separation). Net of the principal pollutant containment systems, diffusive emissions of fine particles from these processes, coupled with several screening steps and a final long-time open-air residues stabilization, could impact the surrounding environment due to the chemical composition of the particulate matter itself (inorganic and organic pollutants). Moreover, the particulate may also arise from transporting the raw bottom ashes to the pre-treatment plant (point source). The present work aims to predict the concentration of the PM10-bound organic contaminants that are usually sampled weekly (PCDD/Fs, PCBs, PAHs) from the concentration of the daily analyzed inorganic pollutants in the surrounding area of an municipal solid waste slag treatment plant, using Artificial Neural Networks (ANNs) as a forecasting tool. Moreover, ANNs have also been used as a clustering tool to evaluate the plant’s environmental impact on the surrounding area with respect to other additional emission sources.