Machine learning-driven multi-objective optimization of sustainable engineered cementitious composites: Balancing...

Area of Science:

• Materials Science
• Civil Engineering
• Sustainable Construction Materials

Background:

• Engineered Cementitious Composites (ECC) offer superior mechanical performance but face challenges in environmental sustainability.
• Natural sand, a key component in ECC, has environmental implications and limited availability.
• Municipal Solid Waste Incineration Bottom Ash (MSWIBA) presents a potential supplementary cementitious material for sustainable construction.

Purpose of the Study:

• To investigate the feasibility of replacing natural sand with MSWIBA in ECC mixtures.
• To enhance the mechanical properties and environmental sustainability of ECC through MSWIBA incorporation.
• To develop a Machine Learning (ML)-driven Multi-Objective Optimization (MOO) framework for sustainable ECC mix design.

Main Methods:

• Experimental testing of fifteen ECC mixtures with varying water-to-binder (W/B) ratios and MSWIBA replacement levels (0-100%).
• Development of an ML-driven MOO framework utilizing Support Vector Regression (SVR) and Ant Colony Optimization (ACO).
• Entropy Weight Method (EWM) was employed to assign weights for balancing compressive strength, tensile strength, and CO2 emissions per unit strength.

Main Results:

• MSWIBA inclusion reduced ECC flowability and strength, with compressive strength decreasing from 74.1 MPa (control) to 32.7 MPa (100% replacement).
• Optimal strain-hardening behavior (tensile strain > 2.5%) was observed at W/B=0.27 with 25%-50% MSWIBA.
• MSWIBA incorporation reduced CO2 emissions per cubic meter from 0.9665 kg/m3 to 0.8788 kg/m3.
• The ML-MOO framework identified an optimal mix (W/B=0.2526, MSWIBA=20.00%) predicting compressive strength of 61.09 MPa, tensile strength of 5.56 MPa, and CO2/fc of 0.0154 kg/(m3·MPa).

Conclusions:

• MSWIBA can be utilized as a partial replacement for natural sand in ECC, offering environmental benefits.
• A hybrid approach combining experimental testing and ML-MOO provides an effective strategy for optimizing sustainable ECC mix design.
• The developed framework recommends specific mix proportions within the investigated range for balancing mechanical performance and environmental impact.