Page 14 - Demo

NEWS14 CPI %u2013 Concrete Plant International | India Edition %u2013 5 | 2025 www.cpi-worldwide.comA team at the Massachusetts Institute of Technology (MIT) has developed an artificial intelligence (AI) framework to help concrete producers identify materials that can partially replace cement, supporting both cost and GHG reduction efforts.Replacing a portion of cement to achieve these savings is already a well-established practice. Byproducts such as fly ash from coal production and slag from steelmaking have long been used as supplementary cementitious materials (SCMs). However, demand for these materials is outpacing supply as the industry looks to reduce climate impacts by expanding their use, making the search for alternatives urgent.AI-driven materials screeningWhile many other potential SCMs exist, the challenge lies in efficiently sorting and evaluating them for use in lasting concrete structures. To address this, the MIT team, led by postdoctoral associate Dr. Soroush Mahjoubi, published an open-access paper in Nature%u2019s Communications Materials outlining their solution. With large language models including chatbots, the team built a machine-learning framework that evaluates and sorts candidate materials based on their physical and chemical properties. In particular, the framework screened materials based on three metrics derived from ASTM International%u00b4s R3 test for evaluating SCM reactivity [1, 2]: heat release during hydration, calcium hydroxide (lime) consumption during hydration, and how water is chemically combined or physically adsorbed onto surfaces. The third metric can serve as a proxy for heat release when isometric calorimeters are not available, making reactivity estimation more accessible. The system processed over one million rock samples and hundreds of thousands of pages of literature, sorting candidates into 19 categories.According to the researchers, many of these materials show potential for practical use in concrete mixes with minimal additional processing, often requiring only grinding before incorporation. This creates new opportunities to use locally available industrial byproducts and waste materials, helping the industry expand its SCM portfolio while enabling a more circular economy. Ceramic waste, such as tiles and pottery fragments, is one example. The study notes that these materials can offer favorable reactivity properties similar to traditional SCMs, echoing techniques seen in ancient Roman concrete structures. Other identified cement substitutes include landfilled nonferrous slags from copper, nickel, or zinc smelters, electric arc furnace slags from scrap steel melting, and harvested coal fly ash.AI and the future of concrete scienceLooking ahead, the MIT team plans to expand the AI framework%u2019s capabilities to evaluate an even broader range of maMIT AI identifies partial cement replacements for more sustainable concreteArtificial intelligence frameworkAn MIT team led by Dr. Soroush Mahjoubi (pictured) built a machine-learning framework that evaluates and sorts partial cement replacements based on physical and chemical properties. The study highlights the potential of materials like discarded ceramics and construction materials to enhance the environmental performance of concrete mixes.