Predicting the thermo-mechanical behaviour of heterogeneous material systems, derived from recycled polymers, bio-fillers, and hybrid polymer blends, is crucial when we are aiming to achieve a transition toward circularity in the automotive plastics supply chain. However, significant variability in feedstock composition and microstructure poses challenges to the prediction process by hampering material qualification and process optimisation in extrusion and injection moulding applications.
The EcoPlast (Empowering Circular Operations in the Automotive Plastics Value Chain) project consortium partners recently published a conference paper, titled “AI-Enhanced Multiscale Predictive Modelling for Characterising Recycled, Bio-Based, and Virgin Plastic Materials”, in the Proceedings of the 22nd European Conference on Composite Materials, Europe’s leading conference in composite materials and structures, to showcase how they are addressing this issue while promoting circularity in the automotive sector. Organised by ASEMlab and under the patronage of the European Society for Composite Materials (ESCM), the conference was held at the Norwegian University of Science and Technology (NTNU) from June 21 to 25, 2026.
In collaboration with the project Coordinator, Technovative Solutions Ltd (TVS), the paper was published by Brunel Composites Centre at Brunel University of London. Michail Kazilas, Director of the Centre, presented the publication at the conference on June 22, 2026, as six authors from these two partner organisations collaboratively prepared the paper, as listed below.
Authors from Brunel Composites Centre:
- Nithin Amirth Jayasree, Head of the Centre
- Michail Kazilas, Director of the Centre
- Dr Sadik Omairey, Senior Research Fellow
- Faranak Bahrami, Composites Processes Technology Leader
Authors from Technovative Solutions Ltd (TVS):
- Mohammad Anis Uz Zaman, Senior Electrical Engineer and Coordinator of the Project
- Mir Nafis Sharear Shopnil, AI Engineer
In the paper, they discussed a physics-informed predictive modelling framework that integrates Representative Volume Element (RVE)-based numerical homogenisation into AI-compatible virtual dataset generation for recycled and bio-based polymer systems. To estimate homogenised thermo-mechanical properties of materials containing recycled polymers, recycled glass fibres, talc fillers, and bio-fillers, the framework employs the EasyPBC environment. The estimating properties include elastic behaviour, Poisson's ratios, density, specific heat capacity, and coefficient of thermal expansion (CTE).
The conference paper is now available online, please click here to read it!
EcoPlast is funded by the Horizon Europe Framework Programme (HORIZON) under Grant Agreement No 101182147.
