In the ever-evolving construction industry, the quest for more efficient and accurate methods of predicting concrete compressive strength has long been a priority, especially as the energy sector increasingly relies on durable, sustainable materials. A groundbreaking study led by Mouhamadou Amar of Univ. Lille IMT Nord Europe, Univ. Artois, and Yncrea Hauts-de-France, France, has taken significant strides in this direction by leveraging the power of artificial intelligence (AI).
The research, published in ‘Cleaner Materials’ (translated from French, it means ‘Cleaner Materials’), focused on the use of AI to predict the compressive strength of concrete, a crucial property for construction projects. Traditional methods of determining concrete properties often involve time-consuming laboratory testing and can be less accurate when dealing with nontraditional materials, such as supplementary cementitious materials (SCMs). These materials, which include fly ash and slag, are increasingly used in the energy sector to reduce the carbon footprint of concrete production.
Amar and his team utilized RapidMiner® software to design and test various machine learning models, including artificial neural networks, decision trees, random forests, support vector machines, and gradient-boosted trees. The study analyzed over 5,373 concrete formulations from 137 literature sources, providing a robust dataset for training and validating the models. “The goal was to find the most accurate model to predict compressive strength without the need for extensive laboratory testing,” Amar explained.
The results were impressive. The gradient-boosted tree model emerged as the most accurate, closely followed by deep learning and random forest models. These models were validated by comparing experimental results to numerical data, showing high accuracy in predicting compressive strength. “The potential impact of this research is significant,” Amar noted. “By reducing the need for laboratory testing, we can streamline the concrete mix design process, making it more efficient and cost-effective. This is particularly important for the energy sector, where the use of SCMs is becoming more prevalent.”
The commercial implications are vast. Energy companies are under increasing pressure to adopt sustainable practices, and the ability to predict concrete properties accurately and efficiently can help them meet these goals. “This research opens up new possibilities for the energy sector,” Amar said. “By integrating AI into the mix design process, we can improve the durability and sustainability of concrete structures, ultimately leading to more efficient and environmentally friendly energy infrastructure.”
The study’s findings are a testament to the potential of AI in revolutionizing the construction industry. As AI continues to advance, its applications in concrete mix design and other areas of construction are likely to become even more sophisticated. This research sets the stage for future developments, paving the way for more innovative and sustainable construction practices.
