Modeling of Concrete Slump Workability and Compressive Strength in a Normal Concrete with waste Ceramic Tiles Using Artificial Neural Network

In this study, there were two (2) derived models which are the compressive strength and slump workability of concrete with waste ceramic tile without adding any additives using an Artificial Neural Network (ANN) model based on five (5) different input parameters which are the Amount of Fine Aggregate (FA), Amount of Coarse Aggregate (CA), Cement Dosage (C), Water-cement ratio (W/C) and the Amount of Waste Ceramic (CW) respectively while concrete slump and compressive strength test result as an output on the model. The two (2) derived models have satisfactory accuracy where the regression values are 0.98007 and 0.99643 and the mean square error of 10.218 and 1.4927, respectively. All models show excellent accuracy has a maximum error of 19.07% and average error of 2.2%. for slump workability, maximum error of 9.26% and average error of 1.81% for compressive strength model. Parametric study was used to describe the behavior of the derived models, the addition of ceramic waste improves the mechanical properties of the concrete, specifically its compressive strength, while the value of slump workability decreases. The study also performs the relative importance calculation, and based on the results, water to cement ratio (w/c) is the main contributing factor for the slump workability and compressive strength model among other parameters, having the most contributing relative importance value of 28.35% on slump model and 27.47% on compressive strength model.