Sustainable Brick Production Using Ceramic Waste and Coconut Fibers: Performance Prediction Using Machine Learning

The increasing environmental concerns associated with cement-based construction necessitate sustainable alternatives. The increasing environmental issues associated with cement-based construction highlight the need for sustainable alternatives. The construction sector produces a considerable amount of ceramic waste, presenting environmental challenges. Employing this waste as a binder can led to decreased cement usage and lower carbon emissions. Furthermore, coconut fiber enhances crack resistance and overall performance, playing a significant role in sustainable construction. This research investigates the integration of ceramic waste powder (CWP) and the addition of coconut fiber (CF). Replacement of CWP as a binder ranging from 0% to 15%, and CF from 0% to 0.2% by weight of cement in brick production. Experimental investigations evaluated efflorescence, water absorption, compressive strength (CS), rebound hammer (RH), and ultrasonic pulse velocity (UPV) following 7 and 28 days of curing with a gunny bag. The reduction of efflorescence was notable as the content of CWP increased, with a formulation of 15% CWP and 0.2% CF showing no visible salt deposits. The findings show that a combination of 10% CWP and 0.15% CF achieved the maximum CS of 15.7 MPa at 28 days, whereas 15% CWP and 0.2% CF resulted in the lowest water absorption rate of 10.4%, thereby decreasing permeability. Machine learning models, specifically gradient boosting (GB) and Random Forest (RF), were utilized to forecast brick performance. The results indicated that GB surpassed RF, demonstrating reduced errors (MSE: 1.424, MAE: 0.965, RMSE: 1.193) and enhanced predictive accuracy (R2:0.304) concerning water absorption. The results indicate that the integration of CWP and CF improves mechanical properties and supports sustainable construction methods by decreasing reliance on cement and effectively using industrial waste.