AutoGrünBeton - Auto-adaptive learning process for optimizing recycled concrete production

The AutoGrünBeton project aims to improve sustainability in the concrete industry. It integrates up to 90% recycled material from construction and demolition waste and reduces cement consumption by up to 70% in order to reduce_{CO2 emissions} and the consumption of natural resources. The focus is on the development of a machine learning system that ensures the quality of the recycled concrete during production, optimizes the manufacturing process and dynamically adjusts the recipe. The system helps to increase production efficiency by predicting concrete quality, identifying production deviations at an early stage and adapting to varying properties of the recycled materials.

Focus of the work:

• Characterization of the recycled raw materials
• Conducting laboratory and field tests
• Development of test plans using active learning
• Development of an automatic camera-based slump flow test system.
• Increasing production efficiency by predicting concrete quality, detecting production errors at an early stage and adapting to the properties of the recycled materials.

M.Sc. Farzad Rezazadeh Pilehdarboni

January 2024 - September 2024

State of Hesse LOEWE 3