Neural Networks with Backpropagation in Engineering

Resumo

Nowadays, with the amount of data increasing continuously in online networks there is the possibility
to use artificial intelligence to analyze large databases. The Artificial Neural Network is a very utilized tool to
find patterns in data to optimize the time and efficiency of analysis. Concrete is the most used structural material
in Brazil in civil construction. Its resistance depends on the proportion of the materials used in its fabrication. To
quantify each material to reach the required resistance for a particular need of a structure is a challenge found by
engineering. This article aims to elicit how an Artificial Neural Network is capable of processing data and
understanding patterns that are necessary to develop a coherent formula for the concrete based on obtained data
in mechanical tests. The program is capable of making interactions among the data in a process of training and
learning and it is also capable of finding acceptable solutions. It used a feedforward neural network with
backpropagation algorithm to find the resistance of the concrete. The input data were the material used in a 1m3
construction of adensed concrete with different types of traits and tests and the output data, its resistance.