A team at K. R. Mangalam University applies a deep neural network coupled with Bayesian hyperparameter tuning and Multi-Objective Particle Swarm Optimization to develop sustainable concrete mixes that achieve high compressive strength, cut costs, and reduce cement content by up to 25%.
Key points
Developed a DNN surrogate (cvR²=0.936, RMSE=5.71 MPa) for strength prediction.
Employed MOPSO to balance compressive strength, cost, and cement usage under practical constraints.
Achieved mixes exceeding 50 MPa strength with up to 25% cement reduction and 15% cost savings.
Why it matters:
This AI-driven approach streamlines sustainable concrete design, reducing environmental impact while maintaining structural performance.
Q&A
What is Multi-Objective Particle Swarm Optimization?
How does Bayesian hyperparameter tuning work?
Why focus on cement reduction?
Read full article
Academy
Green Concrete Technology
Green concrete refers to concrete mixes designed to reduce environmental impact by optimizing material selection and formulation. Traditional concrete heavily relies on Portland cement, whose manufacturing accounts for approximately 7% of global CO₂ emissions. Green concrete substitutes part of the cement with supplementary cementitious materials (SCMs) such as fly ash, blast furnace slag, silica fume, and natural pozzolans.
Key Materials:
- Cement: The primary binder, providing strength and cohesion.
- Fly Ash: A byproduct of coal combustion that improves workability and durability.
- Blast Furnace Slag: A residual material from steel production that enhances long-term strength.
- Silica Fume: An ultrafine byproduct of silicon metal and ferrosilicon alloy manufacturing, boosting density and strength.
Benefits of Green Concrete:
- Reduced Carbon Footprint: Lower cement content directly cuts CO₂ emissions.
- Cost Savings: SCMs are often cheaper than cement, reducing material costs.
- Enhanced Durability: SCMs can improve resistance to chloride penetration, sulfate attack, and alkali–silica reaction.
- Waste Valorization: Utilizes industrial byproducts, reducing landfill disposal.
Multi-Objective Optimization in Engineering
Multi-objective optimization addresses problems involving two or more conflicting objectives. In engineering, objectives can include maximizing performance (e.g., compressive strength), minimizing cost, and reducing environmental impact. Solutions are evaluated based on Pareto dominance, where no objective can be improved without worsening another.
Pareto Front: The Pareto front represents the set of non-dominated solutions, illustrating optimal trade-offs among objectives. Engineers can select mixes based on project priorities—prioritizing strength, cost reduction, or sustainability.
Algorithms:
- Particle Swarm Optimization (PSO): Simulates flocking behavior, updating candidate solutions based on personal and global bests.
- MOPSO: Extends PSO for multiple objectives by maintaining an archive of Pareto-optimal solutions.
- Genetic Algorithms: Use selection, crossover, and mutation to evolve populations toward Pareto-optimal sets.
- Bayesian Optimization: Efficiently tunes surrogate model parameters by balancing exploration and exploitation.
By applying these techniques, engineers develop concrete mixes that meet structural requirements while optimizing cost and environmental performance. This course overview equips readers to understand the principles behind green concrete formulation and the computational methods used to achieve sustainable design solutions.
