Municipal Solid Waste Fractions and Their Source Separation: Forecasting for Large Geographical Area and Its Subregions

journal article in Web of Science

en

This paper introduces an approach toward forecasting municipal solid waste and its fractions in a large geographical area divided into subregions. A multi-commodity system, where components overlap between streams of residual waste and separately collected recyclables, is developed to predict composition, future amounts and separation efficiencies. The approach combines a reconciliation-based balancing model with regression analysis and time series analysis. Regression analysis provides models which are later used to get complete information for all nodes of tree-like structure describing the geographical area of interest. Time series analysis proposes initial models on future amounts for all fractions. The balancing model with newly formulated composition constraints corrects initial estimates, which is a key issue especially for short-time series where precise extrapolation models can hardly be secured. The developed approach contributes to analysing rational recovery targets by reflecting the current situation in individual (micro) regions and, at the same time, it exploits examples of good practice from regions with high recovery rates. Here the analogy with rigorous regression models (historical data from one region can serve as one scenario for another region) is utilised. The algorithm is demonstrated through a case study inspired by an extensive project for the Ministry of the Environment of the Czech Republic.

This paper introduces an approach toward forecasting municipal solid waste and its fractions in a large geographical area divided into subregions. A multi-commodity system, where components overlap between streams of residual waste and separately collected recyclables, is developed to predict composition, future amounts and separation efficiencies. The approach combines a reconciliation-based balancing model with regression analysis and time series analysis. Regression analysis provides models which are later used to get complete information for all nodes of tree-like structure describing the geographical area of interest. Time series analysis proposes initial models on future amounts for all fractions. The balancing model with newly formulated composition constraints corrects initial estimates, which is a key issue especially for short-time series where precise extrapolation models can hardly be secured. The developed approach contributes to analysing rational recovery targets by reflecting the current situation in individual (micro) regions and, at the same time, it exploits examples of good practice from regions with high recovery rates. Here the analogy with rigorous regression models (historical data from one region can serve as one scenario for another region) is utilised. The algorithm is demonstrated through a case study inspired by an extensive project for the Ministry of the Environment of the Czech Republic.

Municipal solid waste; Circular economy;Separation efficiency;Separation rate;Network flow model;Paper separation;Plastic separation;Forecasting

31.07.2019

Springer Nature B.V.

1877-265X

11

2

1–18

18