Prediction of Concrete Compressive Strength Using Artificial Neural Network

Abstract

Concrete is the most widely used material by humans after water. Rapid growth in the construction industry means concrete will continue to be the dominant material in the future. Concrete is a composite material consisting of aggregates, water, and admixtures. Destructive testing of concrete to determine its strength after mix design is an expensive and time-consuming process. With recent advances in soft computing techniques like artificial intelligence, these results can be predicted by feeding the algorithm with a large number of available data to obtain the desired results. In the present research work, it is proposed to use artificial neural networks to predict the strength of different types of concrete. A Multilayer Perceptron has input and output layers, and one or more hidden layers with many neurons stacked together. Data capturing will be done regarding different types of concrete, and artificial neural networks are preliminarily trained with various inputs to solve problems with data applicable to obtain the desired results. This ANN, with captured data, helps minimize the repetitive process and tests involved to obtain the results through experimental procedures, which is time, material, and money-consuming with practical difficulties. The advantage of Python is that designers can create a customized program for interactive design. Python also improves the analytical skills of students, and programs can be converted into executable software. Concrete cubes are cast to validate the predicted result of the software.