Predicting the compressive strength of ecological concrete made with PET granules using artificial neural networks (MATLAB)

Abstract

The problem of getting rid of waste, especially plastic waste, has become a problem that worries governments. From this standpoint, the method of getting rid of plastic appeared by inserting it and making it one of the components of concrete, and it became called environmentally friendly concrete.The purpose of this research is to evaluate the performance of long-lasting concrete that has partial volumetric substitution of aggregate with polyethylene terephthalate (PET) granules. mechanical characteristics, such as tensile and compressive strength, This work offers a prediction model-based method for predicting the compressive and tensile strength of environmentally friendly concrete with various kinds of plastic aggregates (PET) using artificial neural networks (ANN). Artificial neurons, which resemble brain neurons in general, are linked units or nodes that make up an ANN. These are linked together by edges that resemble brain synapses. Connected neurons send messages to an artificial neuron. Previous literature collected a data group with five affecting characteristics: water, fine aggregates, cement, coarse aggregates, PET aggregates, and model validation. Tensile and compressive strength were the outcomes. Additionally, a sensitivity analysis was done to confirm the stability and robustness of this model. The testing results demonstrated the excellent performance of the ANN model, which makes it a viable method for predicting the compressive and tensile strength of an environmentally friendly concrete. In addition, the correlation coefficient for compressive strength was found to be 0.9999, and tensile strength is 0.9999 which indicates the high reliability of the model used in the analysis and the accuracy of the results obtained.