Abstract
This manuscript introduces a computational approach to micro-damage problems using deep learning for the prediction of loading deflection curves. The location of applied forces, dimensions of the specimen and material parameters are used as inputs of the process. The micro-damage is modelled with a gradient-enhanced damage model which ensures the well-posedness of the boundary value and yields mesh-independent results in computational methods such as FEM. We employ the Adam optimizer and Rectified linear unit activation function for training processes and research into the deep neural network architecture. The performance of our approach is demonstrated through some numerical examples including the three-point bending specimen, shear bending on L-shaped specimen and different failure mechanisms.
Highlights
Neural networks (NN) have been used for numerous applications in different areas including computational mechanics
Passing through the three-point bending as an illustrative example, we suggest some possible architectures of the deep neural networks based on the Adam optimizer
Vk where the parameter γ ≈ 0.9 governs the updating of the iterations within the stochastic gradient descent method (SGDM)
Summary
Neural networks (NN) have been used for numerous applications in different areas including computational mechanics. The vanishing gradient problem for instance has been significantly alleviated thanks to the RELU activation function and cross entropy-driven learning techniques. Certain issues such as over-fitting still remains a challenge in deep neural networks. Lee et al [22] has applied deep learning algorithms to structural analysis. Passing through the three-point bending as an illustrative example, we suggest some possible architectures of the deep neural networks based on the Adam optimizer Such findings can open a new branch of research that may prove beneficial to the fourth industrial revolution, where deep learning algorithms play a major role in big data analysis of structural engineering
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
