Explaining LDPE Pyrolysis Using Thermal Analysis and Machine Learning Approaches

What is it about?

Plastic waste, especially everyday items made from low‑density polyethylene (LDPE), is piling up around the world and causing serious environmental problems. Traditional disposal methods like landfilling and burning create pollution and waste valuable resources. Our study explores a cleaner and more useful alternative: pyrolysis, a process that heats plastic without oxygen to turn it into fuel and other valuable products. We examined how LDPE breaks down when heated and measured the gases, liquids, and chemical compounds it produces. We also studied how much energy is needed for the plastic to decompose. To make these predictions more accurate, we combined laboratory experiments with machine learning models, which can recognize patterns in data and forecast how the material behaves under different conditions. Our results show that LDPE can be converted into useful products—nearly half becomes liquid fuel—and that machine learning can reliably predict the energy required for this transformation. This combination of experimental work and advanced data analysis provides a powerful approach for improving plastic‑to‑fuel technologies.

Featured Image

Photo by SURU HAN on Unsplash

Photo by SURU HAN on Unsplash

Why is it important?

This study stands out because it brings together three elements that are rarely integrated in LDPE pyrolysis research: 1. high‑resolution experimental kinetics, 2. advanced reaction‑mechanism modeling (DAEM), and 3. multiple machine learning models used specifically to predict activation energy. While previous studies have explored LDPE pyrolysis, they typically focus on only one of these aspects—either experimental kinetics, or product analysis, or a single ML model. Your work is the first to combine all three into a unified predictive framework, giving a more complete and accurate picture of how LDPE breaks down and how its behavior can be forecasted. It is also timely. Global plastic waste is rising sharply, and industries are urgently seeking data‑driven, low‑emission technologies to convert waste into useful fuels. By showing that machine learning can reliably predict key kinetic parameters, your study demonstrates a pathway toward faster, more cost‑effective design of pyrolysis systems, reducing the need for repeated experiments. This combination of experimental depth, mechanistic insight, and predictive modeling offers a new direction for researchers and practitioners working on waste‑to‑energy solutions. It helps accelerate the development of cleaner technologies and supports global efforts toward sustainable plastic management.