Innovative Optimization of Biochar-Enhanced Concrete Mixes through Meta-Heuristic Algorithms

This study introduces a meta-heuristic algorithm integrating PSO, LSSVM, and NSGAII for optimizing biochar-enhanced concrete mixes, aiming to reduce CO2 emissions and improve mechanical properties. It employs machine learning to tailor mix ratios effectively, enhancing sustainability and design efficiency in construction.

April 13, 2024

2–3 minutes

Chen, et al (2024) Biochar-enhanced concrete mixes: Pioneering multi-objective optimization. Journal of Building Engineering. https://doi.org/10.1016/j.jobe.2024.109263

The recent study introduces a meta-heuristic integration algorithm to optimize BiocharBiochar is a carbon-rich material created from biomass decomposition in low-oxygen conditions. It has important applications in environmental remediation, soil improvement, agriculture, carbon sequestration, energy storage, and sustainable materials, promoting efficiency and reducing waste in various contexts while addressing climate change challenges. More (BC) enhanced concrete mixes, incorporating an innovative triangular model and a sophisticated visualization interface to enhance mix design efficiency. The use of biochar, a carbon-negative material, as a partial cement substitute in concrete formulations presents a sustainable approach to significantly reduce CO2 emissions associated with conventional cement production.

The research employs a hybrid optimization model combining Particle Swarm Optimization (PSO), Least Squares Support Vector Machine (LSSVM), and Non-dominated Sorting Genetic Algorithm II (NSGAII) to predict and optimize the performance and composition of BC-enhanced concrete. This model framework achieves a remarkable generalization performance with an overall accuracy (R2 = 0.95), effectively presenting a complete Pareto front for optimizing concrete mix ratios.

Cement concrete is crucial in the construction industry but notorious for its environmental impact, notably CO2 emissions. Innovative measures like the integration of biochar not only promise reduced emissions but also enhance the mechanical properties and sustainability of the construction materials. Studies have shown that substituting cement with biochar improves the compressive strength of concrete and increases CO2 absorption through improved carbonation processes.

This paper highlights the use of advanced Machine Learning (ML) models for predicting concrete properties and optimizing mix designs. The introduction of LSSVM for performance modeling and XGBoost for feature screening exemplifies the application of data-driven techniques to refine the concrete mix ratios. Moreover, the NSGAII algorithm’s application in multi-objective optimization helps in achieving an optimal balance between mechanical properties, cost, and CO2 emissions.

The integration of meta-heuristic algorithms like PSO enhances the parameter optimization for LSSVM, providing a robust model with high accuracy and reduced computational costs. This novel approach not only streamlines the complex data processing involved but also significantly improves the design efficiency of BC-enhanced concrete mixes.

In conclusion, the study’s innovative use of hybrid meta-heuristic algorithms in developing BC-enhanced concrete formulations addresses the urgent need for sustainable construction materials. By optimizing the mix ratios to tailor to specific requirements, this approach significantly contributes to the environmental goals of reducing carbon footprints and promoting resource efficiency in the construction industry.