Integrated AI Frameworks For Structural Monitoring, Mechanical Performance, And Engineering Optimization

Abstract

This study develops an integrated artificial intelligence solution that simultaneously addresses structural health monitoring, mechanical-performance prognosis, and engineering optimization by leveraging open-sourced Indian experimental case bases. Structural monitoring assessment was based on the six-storey IIT Patna shear-building benchmark with fourth-storey damage, whereas mechanical behavior predictions were based on an innovative 188-sample experimental database of recycled-aggregate concrete mixes from IIT Bhubaneswar. The functioning of the automated XGBoost regressor was validated on the independent test set, achieving R²=0.943 , RMSE=3.99 MPa, and MAE=2.90 MPa for the mechanical performance prognosis task. The five-fold cross-validation procedure scored R²=0.946 ±0.009. The observed-data Pareto front analysis revealed that an observed 28-day strength of 59.0 MPa was reached with 41.2% lower cement content while still retaining 83.2% of the maximum compressive strength. This finding highlights the potential of the proposed framework to enable transparent decision-making in engineering by integrating sensing, mechanical-performance models, and design optimization in a unified data-driven paradigm.