Artificial neural network model for predicting nonlinear response of uniformly loaded fixed plates

Abstract EN

An artificial neural network (ANN) model has been developed for the prediction of nonlinear response for plates with built-in edges and different sizes, thickness and uniform loads.

The model is based on a six-layer neural network with back propagation learning algorithm.

The learning data were performed using a nonlinear finite element program, the set of 1500x16 represent the deflection response of load.

Incremental stages of the nonlinear finite element analysis was generated by using 25 schemes of built-in rectangular plates with different thickness and uniform distributed loads.

The neural network model has four input nodes representing the uniform distributed load, thickness, length of plate and length to width ratio, four hidden layers and sixteen output nodes representing the deflection response.

Regression analysis between finite element results and values predicted by the neural network model shows the least error.

This approach helps in the reduction of the effort and time required determining the load-deflection response of plate as the FE methods usually deal with only a single problem for each run while ANN methods can solve simultaneously for a patch of problems.