< Back to previous page

Project

Optimized Plastics Recycling from Waste Electric and Electronic Equipment (WEEE) through Product Clustering

Waste electric and electronic equipment (WEEE) is a large and increasing waste stream. To attain recycling targets set by the European Union, recycling of WEEE plastics is a necessity. The complexity of the polymeric composition of the stream and the possible presence of hazardous substances are substantial challenges in WEEE plastics recycling. 

In this dissertation, the use of product clusters for improved recycling of disassembled plastic components of WEEE is introduced. Clusters are collections of WEEE product types that are suitable for joint treatment in a recycling facility. The derivation of promising clustering strategies relies on a combination of several research methods.

First, data regarding WEEE composition are gathered. These data cast light on the composition of WEEE from a wide categorical level down to the level of specific polymers and additives. To obtain reliable data, proprietary experimental results are combined with existing literature. A novel data integration method is introduced to combine results in the most reliable manner possible. A key insight is that the waste category of  small household appliances and consumer equipment, while complex, may be more manageable than previously believed.

Second, information is collected regarding waste treatment for WEEE plastics specifically. This involves identifying the core sorting technologies used in recycling facilities in practice. It also includes estimating their efficiencies in targeting particular polymers and the investments required to implement them. Novel indicators for assessing these quantities are introduced. 

Third, the collected information is used in a model to describe how material recycling facilities process given input streams. This model can be used in an optimization routine to determine the optimal sequence of recycling technologies and corresponding parameters. A novelty in this model is the inclusion of a physically precise modeling mechanism for density-based sorting, incorporated in large-scale recycling facility optimization. Objective functions that more accurately reflect the European plastic recyclates market are also novel. The efficient computation of an optimal order for multiple types of separation technologies improves older methods for single technology types.

Fourth, a method for computing promising product clustering strategies for improved WEEE plastics recycling is proposed. The core of the method is in a novel feature engineering technique. This technique uses a classical problem from graph theory for the derivation of features that model the ease with which non-compatible polymers in product types can be sorted. The integration of a statistical clustering model with the designed optimization routine is also noteworthy. Particularly, the clusters obtained in a clustering strategy are used as input in the optimization routine. In consequence, the ecological and economic performance of each strategy can be evaluated. This helps identify a clustering strategy that is not only statistically relevant, but can also be shown to improve recycling rates and recyclers' profits.

Fifth, the gathered data and designed methodologies are combined to identify practically relevant optimal clustering strategies. A clustering strategy with six distinct clusters is proposed. Adoption of this strategy could lead to a recuperation of almost half of all WEEE plastics collected in the European Union and boost the profitability of recycling facilities. The optimality could further be increased by separate treatment of specific product components. The optimal clustering strategy was robust against fluctuations in proportions of collected WEEE categories, product types and plastic proportions. However, there was unpredictability involved when the specific polymer compositions of product types varied. This underscores the need for continuous data collection and integration, for instance through the proposed integration methods, as such compositions may change over time.

Date:17 Jan 2018 →  8 Mar 2023
Keywords:plastic sorting
Disciplines:Business administration and accounting, Management
Project type:PhD project