Prediction Of Strength Properties Of Ultra High-Performance Concrete By Using Artificial Intelligence And Machine Learning Techniques

Abstract

Super elite execution concrete (UHPC) is an as of late evolved material that has drawn in impressive consideration in structural designing because of its extraordinary qualities. One vital figure substantial plan is the compressive strength (CS) of UHPC. As a strong device in man-made reasoning (computer-based intelligence), AI (ML) can precisely foresee cement's mechanical properties. Hyperparameter tuning is urgent for guaranteeing the expectation model's dependability, however it is mind boggling. This study means to advance the CS expectation technique for UHPC. Three ML techniques — irregular woods (RF), support vector machine (SVM), and k-closest neighbor (KNN) — are chosen to anticipate the CS of UHPC. The RF model shows predominant prescient precision, with a R2 of 0.8506 on the testing dataset. Moreover, three meta-heuristic improvement calculations — molecule swarm streamlining (PSO), scarab radio wire search (BAS), and snake enhancement (SO) — are utilized to upgrade the forecast model hyperparameters. The R2 values for the testing dataset of SO-RF, PSO-RF, and BAS-RF are 0.9147, 0.8529, and 0.8607, individually. That's what the outcomes demonstrate SO-RF displays the most elevated prescient presentation. Besides, the significance of information boundaries is assessed, affirming the attainability of the SO-RF model. This exploration improves the forecast technique for the CS of UHPC