Modélisation et optimisation d'un procédé de traitement mécanique pour la valorisation d'un courant hétérogène de matières résiduelles solides

The design and optimization of a mechanical treatment plant for mixed solid waste brings many challenges. A refuse derived fuel (RDF) production plant must be designed to meet final product requirements depending on the downstream recovery route and it must be optimized to minimize its cost. Each product flow resulting from a mechanical treatment plant is characterized by specific physicochemical properties that can be modified and controlled along the mechanical treatment sequence. Equipment efficiency depends on operation and design parameters as well as physicochemical properties of the material feed. Modelling of the mechanical treatment unit operations in waste processing plants provides a useful analysis tool to predict plant performance and profitability. The main objective of this work is to develop a tool for the optimization of waste mechanical treatment. The model serves as an analysis tool to study the production of RDF from a waste stream collected at a recycling sorting plant located in Montreal. The tool provides the optimal mechanical sequence i.e. the most cost-effective solution that allows the production of RDF that conforms to the quality requirements for waste co-processing in a cement plant (CRH Canada, Joliette). Mathematical models for mechanical unit operations (hammer mill, trommel, magnetic separator, eddy current separator, air classifier, optical sorter) based on a wide variety of operating parameters and material properties are developed from literature. The waste stream properties (size, shape, density, heating value) and composition are defined based on characterization studies. Modelling results are compared to experimental results to prove the validity of the models. Extensive work on the hammer mill allowed to develop a more flexible model able to simulate the size reduction of a waste stream of any composition. The results obtained offer a better understanding of the kinetic parameters that control the size reduction. An economic tool is developed to evaluate the cost performance of a given plant configuration. Lastly, the tool is used to solve the case study previously described. The results show that the alternative of producing a RDF instead of disposing the waste to landfill is not economically feasible with the actual quality constraints. The chlorine limit concentration value (0,05%) is indeed too strict. For a higher chlorine limit concentration value (0,3%), the optimal configuration consists of manual sorting, two eddy current separators and one near-infrared optical sorter for the removal of PVC. The estimated cost (75 $\$$/t) competes with the landfill tipping fee (65 $\$$/t). The tool has certain limitations, particularly in terms of the models used. In the future, it would be interesting to carry out experimental work to parameterize the models in more detail and thus consider the influence of other characteristics of the waste material (water content, density, etc.). To evaluate performance on an environmental basis, it would be interesting to combine mechanical treatment models with recovery process models and to link the tool to life cycle analysis and impact assessment models.