Modeling and field validation of the gravimetric composition of municipal solid waste disposed of in landfills

Gravimetric analysis of Municipal Solid Waste (MSW) stands as a fundamental procedure in landfill waste management. The characteristics of MSW are intricately shaped by various factors within a municipality, encompassing economy, climate, culture, education, and degree of urbanization. While the field implementation of gravimetric determination follows a relatively straightforward operational protocol, it remains a labor-intensive and financially demanding procedure. Additionally, it presents potential hazards of contamination to individuals involved in the screening process. Based on the foregoing, this research aims to compare the gravimetric composition of waste within a landfill situated in the semi-arid region of Brazil with its theoretical counterpart, derived from modeling through Artificial Neural Networks (ANN). Field characterization of the waste adhered to established technical standards, complemented by statistical planning for MSW collection and sampling. The assessment of theoretical composition was conducted using ANN models, with socioeconomic data serving as input variables and the gravimetric fractions of waste as outputs across various Brazilian municipalities. Multiple topologies were explored to identify an optimal configuration that yielded appropriate statistical validations. In general, the examination of both the empirical and theoretical gravimetric composition of MSW indicated a notable congruence between the datasets, thus emphasizing the effectiveness of mathematical modeling substantiated by statistical validations. Consequently, the utilization of mathematical modeling with ANN holds significant potential as a methodology for predicting the gravimetric composition of MSW. This approach efficiently mitigates environmental and health hazards while reducing financial expenditures and time constraints inherent in traditional methods.