The Reincarnate project is announcing the publication of the conference paper An Adaptive Upscaling Approach for Assessing Materials’ Circularity Potential with Non-Destructive Testing (NDT) by Zia, Ghezal Ahmad Jan, Moreno Torres, Benjamí, Kruschwitz, Sabine Völker, Christoph (2024, April 10).
Presented at the Spring Convention & Conference on Advanced Construction Materials and Processes for a Carbon Neutral Society (RILEM) in Milan, Italy, from April 10-12, 2024, the paper explores an AI-driven adaptive sampling (AS) method integrated with Non-Destructive Testing (NDT) techniques.
The aim of this research is to assess construction materials’ circularity potential while minimising data collection efforts. NDT is widely used to evaluate the structural integrity of materials without causing damage, and this method enhances its efficiency.
The paper introduces a novel technique where AI-driven adaptive sampling focuses on critical data points to minimise the number of samples required for precise assessments. Typically, traditional sampling methods require a large amount of data, with up to 29 samples needed for accurate assessments. However, this new approach significantly reduces the required number to an average of 7 samples for Logistic Regression models and 8 for Random Forest models.
The research highlights how this method can be applied to detect material degradation mechanisms, such as corrosion in reinforced concrete. By integrating NDT with adaptive sampling, the process becomes more scalable and cost-efficient, particularly in large-scale construction projects. The study also discusses how this approach helps meet sustainability objectives by aligning with the principles of the circular economy, which seeks to reuse resources and minimise waste.
Reincarnate partners involved: Bundesanstalt für Materialforschung und -prüfung and the Technischen Universität Berlin.
Read the paper HERE