Multi-scale optimization of PC/ABS polymer blends: Microstructural design for superior toughness, strength, and weight efficiency

The PC/ABS polymer blend is widely used in automotive and consumer electronics due to its balanced combination of thermal, mechanical, and processing properties. Its behavior depends on deformation mechanisms such as rubber particle cavitation and debonding at the PC/ABS interface, which vary with loading conditions and morphology. Modeling and optimizing the PC/ABS microstructure is a complex challenge. This work proposes a multi-scale framework to model and optimize different PC/ABS blends, based on: (i) efficient generation of representative volume elements, (ii) accurate constitutive models for the blend phases, and (iii) an unsupervised optimization process for microstructural design. The optimization considers ABS content in the blend, rubber fraction in ABS, and ABS particle orientation to maximize toughness and strength while minimizing cost and weight — a challenging task due to the negative correlation between toughness and strength. To handle the high-dimensional solution space, Multi-Criteria Decision Making techniques are employed to select optimal solutions within the Pareto front. Two case studies are explored: (i) a lightweight, high-toughness application and (ii) a high-strength application. Additionally, the framework is tested in a functionally graded material optimization problem, where a Cook’s membrane is discretized into PC/ABS layers, with the ABS fraction in each layer adjusted to simultaneously minimize maximum displacement and structural weight. The numerical results validate the proposed design framework for PC/ABS while demonstrating its flexibility for other materials and structural applications. • Characterization of PC/ABS polymer blend and its technological relevance. • Development of a multi-scale optimization-based design framework for PC/ABS. • Evaluation of the performance and robustness of the design framework. • Use of Multi-Criteria Decision Making techniques to select the best solution within a high-dimensional Pareto front. • Application of the framework to a functionally graded material optimization problem.