Models for solid waste and its management in Stockholm metropolitan area

Abstract: In the transition from a linear economy to a sustainable circular economy, waste management is critical. This thesis approaches the management of Municipal Solid Waste (MSW) from the perspective of a mathematical modelling. Within the scope of the thesis, three mathematical models were developed. The first is a probability theory-based model that makes reliable projections of MSW amounts and composition from a relatively large number of small datasets using regressional estimates, by treating the regression parameters of individual municipalities (small sets) as samples from a single distribution. A tailored two-step method for outlier detection is also used. This model shows that the amount of mixed waste is decreasing rapidly, as more waste is separated in the households. It also shows that less food waste is being produced but more is being sorted separately, resulting in a rapid increase in available sorted organic waste. The predictions of this model are used as input data for the two other models. These area multi-objective linear programming model which provides Pareto optimal solutions for the Stockholm MSW management system and a stochastic model which simulates the current MSW management system from the perspective of the procurement process. The models take both Economic Benefits (EB) and emissions of Greenouse Gases (GHG) into account, where the latteris measured in CO2 equivalent. Therefore, a set of Pareto optimal solutions are obtained, ranging from maximized EB to minimized emissions of GHG with compromise solutions in between, rather than just one solution (e.g. minimum emissions). Each Pareto optimal solution is characterized by where waste flows are allocated, as well as which facilities should be operating and which facilities should be closed. One solution was subjectively chosen as the best compromise, as it saves Stockholm around 150 million SEK per year, while slightly decreasing emissions to a level very close to the minimum, as compared to the simulation of currenttreatment. This solution indicated that one of the MSW treatments, waste compression, is not economically viable, and should be removed from the system. This result was also seen in the procurement simulation, which was run both with compression facilities active and inactive. The cost for running these facilities had to be decreased to 1/10 of the estimated value before their economic impact became net zero for the simulation, but even at this level remained unfeasible in the optimization model. Another major indication of the models was that Anaerobic Digestion(AD) plant capacity will at the very latest be surpassed by the amount of separated organicwaste in 2024, so to make use of the benefits this process has compared to incineration, AD capacity should be expanded.