A Lifecycle‐Aware Optimization Framework for Low‐Carbon Material Selection in Engineering Product Design

Achieving carbon neutrality in engineering design requires effective emission reduction strategies throughout the entire product lifecycle. Material selection is a key element in this process, as it spans multiple stages and has a decisive impact on total carbon emissions. However, existing material selection methods often rely on static rules or single‐stage assessments, lacking adaptability to complex process design scenarios and the ability to capture cross‐stage interactions. As a result, they fall short of meeting the dynamic and accuracy requirements of real‐world applications. To address this, we propose a lifecycle‐oriented framework, Low‐Carbon Industrial Design Material Selection (LIDMS), for intelligent material selection in engineering applications. The framework integrates a deep learning‐based prediction module, the Multi‐Branch Fusion Module (MFM), with a constraint‐aware combinatorial optimization engine. The MFM captures complex relationships between material properties and design parameters across four lifecycle stages using a stage‐aware attention mechanism. These predictions are embedded into a differentiable optimization model that identifies material combinations minimizing total lifecycle emissions while meeting engineering constraints such as strength, cost, and recyclability. Experimental validation using synthetic data and a real‐world case study (a handheld electric sander casing) demonstrates the method's performance and practical applicability in reducing emissions under realistic constraints. The proposed LIDMS framework offers a scalable solution for sustainable material decision‐making in industrial product and process design. Our code is available at https://github.com/XueFuqiang/Low‐Carbon‐Industrial‐Design‐Material‐Selection.git .