Carbon Emission Optimization of Ultra-High-Performance Concrete Using Machine Learning Methods.

Area of Science:

• Materials Science
• Civil Engineering
• Sustainable Construction

Background:

• Ultra-high-performance concrete (UHPC) is a key research area due to its superior properties.
• However, UHPC production has high carbon emissions, conflicting with sustainability goals.
• Optimizing UHPC mix design is crucial for reducing environmental impact.

Purpose of the Study:

• To develop a machine learning-based strategy for optimizing UHPC mix proportions.
• To reduce the significant carbon emissions associated with UHPC production.
• To provide a sustainable alternative for UHPC manufacturing.

Main Methods:

• An artificial neural network (ANN) was used to create a predictive model for UHPC compressive strength and slump flow.
• A genetic algorithm (GA) was employed to minimize UHPC carbon emissions under specific constraints.
• Experimental validation was conducted to support the model's predictions and optimization outcomes.

Main Results:

• The ANN model demonstrated high prediction accuracy for UHPC properties, with errors below 10%.
• GA optimization successfully reduced UHPC carbon emissions to 688 kg/m³.
• The study confirmed substantial environmental benefits through the implemented machine learning approach.

Conclusions:

• Machine learning models, specifically ANN and GA, are effective for optimizing UHPC mix proportions.
• This approach significantly lowers the carbon footprint of UHPC, aligning with sustainable development trends.
• The developed ML model provides a robust theoretical framework for future UHPC optimization efforts.