Simulating process parameters of sensor-based sorting for plastic packaging waste: Methodology and validation

Sensor-based sorting (SBS) is routinely used to sort plastics according to polymer types in centralized sorting facilities. However, this sorting is currently suboptimal, which contributes to the low recovery and recycling rates for post-consumer plastic packaging in Europe. Existing literature heavily focuses on sensor-based identification, while other parameters that play a key role are not explored. In this paper, we present a methodology to build an accurate SBS unit simulation model, focusing on process parameters that we show to have a significant effect on sorting performance. We performed comprehensive experiments using real-life SBS unit to acquire dataset of sorting performance with different parameters (throughput rate and material composition). We use this data to fit our SBS unit model that was built using principles of discrete event simulation. We employed this model to simulate two realistic sorting scenarios with multiple sorting steps, and we performed these scenarios also empirically. Comparing the simulation to the experiments, the model could very accurately estimate the parameters in each sorting step. This simulation could be used to design different sorting scenarios and even optimize processes in industrial-scale sorting facilities. • Sensor-based sorting (SBS) yields low recovery rates for plastics in Europe. • Sorting performance depends on throughput rate and material composition. • A discrete event simulation model was built and fitted using real SBS unit data. • The model accurately predicted sorting outputs in multi-step real-life scenarios. • Our approach can help design and optimize industrial-scale sorting processes.