(237c) A Decision Tree-Based Model for the Techno-Economic Assessment and Environmental Impact of Aromatic Production from Polystyrene Pyrolysis

Polystyrene (PS) is a lightweight, petroleum-derived plastic known for its strength and heat resistance accounting for about 10 wt.% of accumulated plastic waste generated annually in the past 10 years [1-2]. These properties have driven its global demand for plastic, increasing its production by approximately 4% annually for applications in packaging, insulation, construction, and household products [1]. However, its widespread use has led to a significant accumulation of plastic waste, highlighting the need for effective disposal and recycling methods [3]. Global plastic production is projected to reach around 500 million tons by 2025 [1]. In 2021, PS production was about 20.7 million metric tons (MT), representing 5.3% of the global plastics market [4]. Due to its durability, PS can persist in the environment for thousands of years, creating major waste management and sustainability challenges. Pyrolysis has gained attention as a viable method for converting PS waste into valuable chemicals and fuels; products with highly valued aromatic compounds [5]. Despite its potential, challenges related to economic feasibility and environmental impact continue to hinder large-scale adoption [6].

This study employs decision-tree machine learning (ML) algorithms to develop predictive models for product yields from continuous, non-catalytic pyrolysis of waste plastics, with a strong focus on PS [7]. The ML model considers various plastic types, including PS, high-density polyethylene (HDPE), and low-density polyethylene (LDPE), alongside process parameters such as reaction temperature (°C) and vapor residence time (s) as input variables. The target outputs are BTX (wt.%) and styrene (wt.%). To enhance model performance, the dataset was split into 80% for training and 20% for validation. The ML model was integrated into a BioSTEAM framework to simulate PS pyrolysis, perform a techno-economic assessment (TEA) and life-cycle assessment (LCA) of the PS pyrolysis.

Model performance was evaluated using key metrics, including the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE). Results indicate strong predictive capability, with RMSE values of 9.99×10⁻¹⁶ for training data and 2.3 for test data. The MAE was 0.82, while the R² values for the training and validation datasets were 1.0 and 0.95, respectively. The model predicted BTX and styrene yields of 26.43 wt.% and 27.33 wt.% at 600°C, supporting previous findings that higher reactor temperatures enhance BTX production due to increased secondary reaction rates. The TEA results show that there is an increase in the minimum selling price (MSP) from $861.20 to 863.43 per MT as the pyrolysis reaction temperature increases from approximately 477 to 577. The Global Warming Potential (GWP) impact of recycling PS is a negative -3.18 kg CO2 equivalent per kg of PS recycled which indicate that the process reduces the emission of Greenhouse gases (GHG).

References

1. Gonzalez-Aguilar, A. M., Pérez-García, V., and Riesco-Ávila, J. M. (2023), A Thermo-Catalytic Pyrolysis of Polystyrene Waste Review: A Systematic, Statistical, and Bibliometric Approach. https://doi.org/10.3390/polym15061582
2. Aljabri, N. M., Lai, Z., Hadjichristidis, N., & Huang, K.-W. (2017). Renewable aromatics from the degradation of polystyrene under mild conditions. Journal of Saudi Chemical Society, 21(8), 983–989. https://doi.org/10.1016/j.jscs.2017.05.005
3. Nayanathara Thathsarani Pilapitiya, P.G.C., and Ratnayake, A. S. (2024), The World of Plastic Waste: A Review. https://doi.org/10.1016/j.clema.2024.100220
4. Royuela, D., Veses, A., Martínez, J. D., Callén, M. S., López, J. M., García, T., & Murillo, R. (2024). Thermochemical recycling of polystyrene waste by pyrolysis using a pilot-scale auger reactor: Process demonstration in a relevant environment. Resources, Conservation and Recycling, 211, 107869. https://doi.org/10.1016/j.resconrec.2024.107869
5. Dement’ev, K. I., Bedenko, S. P., Minina, Y. D., Mukusheva, A. A., Alekseeva, O. A., and Palankoev, T. A. (2023), Catalytic Pyrolysis of Polystyrene Waste in Hydrocarbon Medium https://doi.org/10.3390/polym15020290
6. Valizadeh, S., Valizadeh, B., Won Seo, M., Choi, Y. J., Lee, J. Chen, W. Lin, K. A., and Park, Y. (2024), Recent Advances in Liquid Fuel Production from Plastic Waste via Pyrolysis: Emphasis on Polyolefins and Polystyrene. https://doi.org/10.1016/j.envres.2024.118154
7. Cheng, Y., Ekici, E., Yildiz, G., Yang, Y., Coward, B., and Wang, J. (2023), Applied Machine Learning for Prediction of Waste Plastic Pyrolysis Towards Valuable Fuel and Chemicals Production. https://doi.org/10.1016/j.jaap.2023.105857