Evaluation of the Performance of ANN Algorithms with the Bidirectional Functionally Graded Circular Plate Problem

Functionally graded materials (FGMs) are composed of metals and ceramics in which the distribution material components varies according to a particular volumetric function. FGMs often used high-temperature applications. In our study, models were created Artificial Neural Network depending on equivalent stress levels compositional gradient exponent, is most important parameter determining thermo-mechanical behavior circular plates functionally staggered two directions, performances these evaluated. These obtained with four different training algorithms: Levenberg-Marquardt, Backpropagation Algorithm, Resilient Propagation Conjugate Gradient Powell-Beale Restarts To train ANN, by performing numerical analyzes at upper values. The data sets considering largest value levels, smallest value, value. this stages performance values examined interpreted 4 algorithms detail.