ACORN - AI-powered concrete recipe generator

Important results from the project

The project has successfully developed the first beta version of ACORN, an agent-based AI that optimizes concrete mix designs based on strength, Co2 impact, and cost. We collected over 45,000 recipes, met our accuracy targets, and built a robust architecture with organizational-level access. Interest has been shown from stakeholders in Sweden, the EU, and India, and we have established important collaborations. ACORN demonstrates clear potential to deliver significant positive climate impact.

Expected long term effects

The project has shown that ACORN has the potential to become a powerful tool for sustainable concrete design, enabling optimization of mixes based on strength, cost, and environmental impact. Over time, this can significantly reduce CO₂ emissions and support climate neutrality in Sweden, EU, and beyond. Not only that, but the agent-based architecture offers opportunities to extend similar solutions to other applications, positioning ACORN as both a climate solution and an AI innovation.

Approach and implementation

The project was built on several critical phases: data collection, AI decoding, the establishment of an agent-based AI architecture, and real laboratory testing. Close collaboration between Ecometrix and RISE made it possible to address various challenges that arose during the course of the project. Staff changes and failed lab tests required us to adjust the project organization and adapt activities and timelines, but the collaboration itself was not significantly affected.