Experimental and machine learning based analysis of pervious concrete enhanced with fly ash and silica fume

Pervious concrete (PC) has quickly gained attention as an eco-friendly solution to urban stormwater management, offering improved drainage performance while decreasing environmental impact. This study explores the use of fly ash (FA) and silica fume (SF) as supplementary cementitious materials (SCMs) to enhance both mechanical and environmental performance of PC concrete. Experimental investigations were performed using FA replacement levels of 20 %, 30 % and 40 %; and SF replacement levels of 7.5 %, 10 % and 15 % to determine optimal mix proportions for compressive strength, flexural strength, split tensile strength evaluation. Results indicated that a blend containing 30 % FA and 10 % SF achieved a 28 day compressive strength of 34 MPa representing a 51.1 % increase over the control mix as well as improved flexural (4.8 MPa) and tensile strength (3.3 MPa), while reducing porosity to 12.4 % and maintaining high permeability. It also resulted in a 4.7 % reduction in CO₂ emissions and 6.19 % lower material costs, supporting its suitability for sustainable infrastructure applications. Machine learning (ML) models were also created in order to predict compressive strength based on mix composition and curing age using Orange Data Mining software version 3.36. Five algorithms: KNN, Support Vector Machine (SVM), Artificial Neural Networks (ANN), Decision Tree (DT), and Random Forest (RF), were trained and evaluated. SVM achieved the highest predictive accuracy (R 2 = 0.98), while KNN and ANN showed lower performance (R 2 = 0.69 and 0.71, respectively). These findings not only validate the synergy of SCMs in enhancing PC performance but also support their practical application in urban infrastructure such as pedestrian pavements, permeable roads, and stormwater infiltration systems , contributing to sustainable and resilient construction.