%0 Journal Article %A Gomes, Guilherme Ferreira %A Ancelotti, Jr., Antonio Carlos %A da Cunha, Jr., Sebastião Simões %D 2018 %I Begell House %K artificial neural networks, porous carbon fiber, fatigue test, residual stress %N 1 %P 27-40 %R 10.1615/CompMechComputApplIntJ.v9.i1.30 %T RESIDUAL STRESS PREDICTION IN POROUS CFRP USING ARTIFICIAL NEURAL NETWORKS %U https://www.dl.begellhouse.com/journals/36ff4a142dec9609,2072c4b7696dc891,66bb275c73801c34.html %V 9 %X The use of composite materials, especially the ones made of carbon fiber/epoxy, has considerably increased for structural applications in the aerospace industry. One of the most common defects related to composite processing refers to void formation or porosity. In general, porosity causes reduction of the mechanical properties of composites and therefore it is important to evaluate the behavior of this material in the presence of this type of defect. The porosity level was taken as the input of the network. Four fatigue test data groups were used in this work, three for the training state and one set of data for validation. The ultimate strength prediction was performed with an artificial neural network backpropagation algorithm. The neural network results showed that the application of the Levenberg–Marquardt learning algorithm leads to a high predictive ultimate strength quality. %8 2018-03-27