New AI concrete solution could reverse climate damage
New AI concrete solution could reverse climate damage
—————————–
Scientists at the USC Viterbi School of Engineering have developed an AI-powered model, Allegro-FM, that could revolutionize concrete production by making it stronger, longer-lasting, and capable of capturing carbon dioxide (CO₂), Earth.com reported. Concrete, a vital building material, currently accounts for about 8% of global CO₂ emissions and typically deteriorates after around 100 years.
Allegro-FM simulates billions of atoms simultaneously, allowing researchers to digitally test various concrete formulations quickly and efficiently. Running on the Aurora supercomputer, the model achieves 97.5% efficiency while simulating over four billion atoms—about 1,000 times larger than previous models.
This AI-driven approach focuses on CO₂ sequestration, embedding carbon within the concrete to create a carbon-neutral material. The trapped CO₂ forms a “carbonate layer” that enhances the concrete’s strength, potentially extending its lifespan to rival ancient Roman structures that have lasted over 2,000 years. Additionally, the model can design concrete that withstands extreme heat, a critical feature for fire-prone regions like Los Angeles.
By combining machine learning with materials science, Allegro-FM accelerates simulations with quantum accuracy but lower computational costs. This breakthrough opens the door to cleaner, more durable infrastructure that contributes to climate solutions rather than exacerbating the problem. The study is published in The Journal of Physical Chemistry Letters.
