Concrete.ai's Mix Design Optimization Platform
Innovative Product or Service
Concrete.ai is on a mission to avoid 500 million tons of CO2 from entering the atmosphere annually through an artificial intelligence driven optimization platform. The foundation of Concrete.ai’s platform is a unique technology developed at the University of California, Los Angeles.
Concrete.ai offers the concrete sector a capital-light, rapidly deployable Software-as-a-Service solution that brings performance and sustainability efficiencies to concrete design by leveraging the power of data. Unlike other technologies, the concrete.ai software can be implemented virtually instantly with no need to utilize new material sources or plant retrofits. Additionally, unlike other technologies that rely on one method or product to reduce the carbon footprint, Concrete.ai’s software utilizes all available technologies to find the optimal combination to produce the best result. In independent trials across the U.S., reductions in concrete material costs and embodied carbon by up to 10% and 70%, respectively we achieved. These reductions are achieved by applying AI/ML-enabled optimization to predict the performance of concrete as a function of its proportions, the characteristics of coarse or fine aggregates, supplementary cementitious materials, plus chemical admixture type and dosage. The result is an optimized, cost-effective concrete that fulfills all performance characteristics such as slump, set time and strength, while utilizing local raw materials to ensure safety, longevity, and code compliance.
Through the course of 2022, Concrete.ai was engaged in pre-commercial platform piloting and subsequently beta testing with Votorantim Cimentos (Prairie Material), one of the largest cement and concrete producers in the world. VCNA Prairie Materials believes that cement, concrete, and aggregate production requires a commitment to improving natural resource allocation, controlling emissions, and reducing waste. With this belief in mind, Prairie Materials saw the value the platform offered to easily deploy a scalable solution across their operations to further reduce the carbon footprint while maintaining the high-quality concrete products demanded by the industry. Prairie Materials thought-leadership and partnership in piloting and beta testing, accelerated the development of the Concrete.ai platform. Prairie Materials is currently implementing the software across their Chicagoland ready mix plants to offer even more sustainable products to their customers.
VNCA Prairie materials has been utilizing the concrete.ai platform in order to reduce the carbon footprint of their concrete products through AI driven mix optimization. Below is an example of the results seen that VCNA Prairie has been able to achieve utilizing the algorithms.
Three commonly utilized mix designs (baselines) were optimized to reduce the embodied carbon while maintaining the required performance parameters.
1. High early 3000 psi at 3 day and 6000 psi at 28 day f’c
2. 4000 psi at 28 day f’c with air entrainment
3. 4000 psi at 28 day f’c without air entrainment
The optimized mix designs met or exceeded the performance requirements and we able to achieve up to 34% CO2 reductions.
The attached images shows results for the comparison of GWP for all three optimized mix designs compared to their baseline. These were all accomplished by utilizing existing raw materials in the plant with no changes to the operation.
The attached images shows results for the predicted and measured compressive strength for the high early strength mixture. The target of the optimization was a f’cr of 7,300 psi at 28 days and a minimum of 3,000 psi at 1 day.
The attached images shows results for the predicted and measured compressive strength for the 4,000 psi mixture without air entrainment. The target of the optimization was a f’cr of 5,200 psi at 28 days. The model predictions vs the actual measured strength shows a high degree of accuracy, especially at 28 days.
The attached images shows results for the predicted and measured compressive strength for the 4,000 psi mixture with air entrainment. The target of the optimization was a f’cr of 5,200 psi at 28 days. The model predictions vs the actual measured strength shows a high degree of accuracy, especially at 28 days.
The attached images shows results for the predicted and measured slump for the 4,000 psi mixture with and without air entrainment. The target of the optimization was for a 3 to 6 inch slump. The model predictions vs the actual measured slump shows a high degree of accuracy, in these trials they were within nearly 0.25 inches.
We have included screenshot of the software Home Screen. We have also included a representative example of a mix optimization within the software that utilizes generic data created for demonstration purposes and is not an actual optimization performed by Prairie Materials. This may be a duplicate submission as the form reset without confirmation