Machine Learning-Based Prediction of Strength and Environmental-Economic Optimization of HSC Concrete Incorporating Pozzolanic Ash and Nano-Eggshells

Abstract By 2030, global egg production is expected to reach 90 million tons, generating substantial eggshell waste. This study looks into the use of nano-eggshell powders (NEP) as a portion substitute for cement in HSC (high-strength concrete) mixed with pozzolanic ashes (OA). Seven machine learning (ML) methods were used to forecast the 28-day compressive (CS) and flexural strength (FS) of waste-based HSC. Random Forest Regression (RFR) had the highest reliability (R 2 > 0.97). Model interpretation with SHapley Additive exPlanations (SHAP) revealed that water (negative), cement (positive), and NEP (combined) played the most important roles. For considering CO2 emissions, strength, and cost in addition to mechanical performance, a modern sustainability indicator (SCER) was established. Multi-objective optimization was allowed by applying an integrated approach of RFR and Differential Evolution (DE), causing environmentally friendly high-strength blends. The novelty of this study is the achievement of an equilibrium in economic, environmental, and mechanical performance of concrete by a combination of NEP, pozzolanic ashes, and optimization by ML. This conduct has not been investigated previously in the literature. Graphic Abstract