Harnessing ANN and advanced isoconversional (AIC) models for enhanced energy efficiency in thermochemical conversion of medical plastic waste

This study investigates thermodynamic parameters during valorization via thermochemical conversion of single-use medical plastic waste and introduces a new idea by investigating the synergism integrated with ANN and the advanced isoconversional (AIC) method. Using thermo-gravimetric analysis (TGA), the kinetic behavior of face masks (WM), gloves (WNG), face shields (WFS), PPE coat (WPK), and their mixture (MCPW) was studied in an inert atmosphere in a temperature range of 25–700 °C at heating rates of 10, 20, and 30 °C per min. The AIC was utilized to obtain the kinetic triplets [activation energy (Eₐ), pre-exponential factor (Aₐ), and reaction mechanism (f(α))], and the Master Plot technique was used for reaction mechanism identification. A key investigation of this study revealed that the co-pyrolysis of mixed medical plastics (MCPW) significantly reduces the activation energy by 24–54 % compared to single components, indicating strong synergistic interactions. This reduction implies less energy input and better conversion, and hence, MCPW pyrolysis is more scalable and sustainable. Thermodynamic properties (ΔH, ΔG, ΔS) were also considered to assess the process viability and spontaneity. Moreover, a model for predicting thermal degradation behavior using an artificial neural network (ANN) was developed based on single data and validated against MCPW as an unknown input. The ANN model demonstrated high predictive accuracy (R 2 ≥ 0.95) for all heating rates, indicating its usefulness in minimizing experimental effort and enabling the design of a smart pyrolysis reactor for heterogeneous plastic waste streams. • Thermochemical valorization of medical plastic waste studied via TGA. • Co-pyrolysis of mixed plastics reduced activation energy by 24–54 %. • Kinetic triplets and reaction mechanisms determined using AIC and Master Plot. • Thermodynamic parameters (ΔH, ΔG, ΔS) confirm process viability. • ANN accurately predicted degradation (R 2 ≥ 0.95), aiding smart reactor design.