AI Model Predicted Chloride Resistance in Concrete
Researchers demonstrated an artificial neural network capable of optimizing sustainable concrete mixtures.
Updated on Sept. 30, 2026 in Artificial Intelligence
Live Poll
Do you trust artificial intelligence to improve the safety and sustainability of public infrastructure?
Researchers have published a study in Scientific Reports detailing an AI-assisted method for predicting chloride resistance in recycled aggregate concrete. The research, which remains at the academic study stage, utilizes neural networks to model sustainable material compositions.
Why it matters
This approach aims to accelerate the adoption of sustainable construction materials by minimizing reliance on virgin resources through data-driven mixture design. The study provides a framework for balancing performance and environmental impact in high-density infrastructure projects.
The artificial neural network achieved an R2 coefficient of 0.9935 based on a database of 729 concrete mixtures. Approximately 98% of model predictions for chloride resistance fell within 6% of measured laboratory results.
The players
Scientific Reports
An open-access, peer-reviewed journal from Nature Portfolio covering primary research across all areas of the natural sciences and engineering.
The details
The research employed artificial neural networks — computational models that recognize non-linear patterns within complex datasets — to map the relationship between concrete components and chloride ingress. Researchers also utilized response surface methodology, a mathematical technique for identifying optimal variable interactions, to refine the mix design. The final optimized formulation incorporates 40% ground granulated blast furnace slag and 2.7% waste crumb rubber by binder weight at a water-to-binder ratio of 0.35.
Timeline
September 30, 2026: Study findings were officially published in the journal Scientific Reports.
The Tech Race
This work directly challenges traditional manual trial-and-error methods for optimizing concrete performance in corrosive environments. It marks a shift toward predictive modeling in civil engineering, moving beyond the benchmarks set by standard physical testing protocols like ASTM C1202.
This research provides a digital design tool for engineers seeking to replace virgin aggregates with recycled industrial waste like slag and rubber. Widespread adoption will depend on future validation of the model's accuracy regarding mechanical strength and cost-efficiency in large-scale builds.
The takeaway
The research establishes a new baseline for predicting material longevity using machine learning rather than iterative physical testing. Industry practitioners should monitor upcoming publications from the research team, which will address the remaining variables of mechanical load-bearing capacity and CO2 emissions.
Further reading
For broader trends in machine learning applications for material science, visit Artificial Intelligence.
More information
Read the complete Scientific Reports journal study for the full methodology and data tables.
Source note: This article includes information reported by AZoBuild.
Live Poll
Do you trust artificial intelligence to improve the safety and sustainability of public infrastructure?
