Data-Driven Functional Modeling of Corroded Bolted Joints: A Framework for Remanufacturing

The circular economy, emphasizing resource efficiency and sustainability, highlights remanufacturing as a key aspect. However, ensuring the quality and reliability of remanufactured products is challenging due to uncertainties from prior product usage, especially for bolted joints affected by corrosion. Current methodologies, including analytical, numerical, and empirical models, often fail to meet accuracy and computation time criteria. Addressing this research gap, this paper presents a data-driven framework that couples empirical methods with artificial intelligence techniques to develop a predictive functional model. This model aims to accurately predict key variables such as load-bearing capacity and thread friction coefficient for the remanufactured bolted joints. The proposed framework comprises three main components: development of a test rig and experimental analysis, identification of key variables, and development of a predictive model. The model is based on empirical data from various Design of Experiments (DoE) tests and is refined using machine learning techniques. By embracing a data-driven approach, this research seeks to fill the gaps left by existing modeling techniques, enhancing the reliability of modeling techniques for corroded bolted joints. ... mehrThis framework contributes to the circular economy, product quality, reliability, and sustainability and can eventually be applied beyond bolted joints, addressing uncertainty across various products.