Afsaneh Banitalebi Dehkordi - Department of Computer Science, Payam Noor University
Kaveh Kumarci - Sama technical and vocational training college, IslamicAzad University, Shahr-e-Kord branch, Shahr-e-Kord, Iran

The general goal of this research is to determine natural regular frequency of an arch by artificial neural network with various supporting conditions. For the subject of neural network, training or learning algorithms are applied .The most famous of which is back propagation algorithm. In this research, the real frequency of plate is calculated first using ANSYS program and is defined as a goal function for neural network, so that all outputs of the network can be compared to this function and the error can be calculated. Then, a set of inputs including dimensions or specifications of arches are made using MATLAB program. After the determination of algorithm and quantification of the network, the phases of training and testing of the results are carried out and the output of the network is created. It is concluded that the performance of the neural network is optimum, and the errors are less than 7%, so the network trains in different manner. Furthermore the time of frequency calculations in neural network is less than real analysis time that calculated by ANSYS software, and it’s precision is acceptable(less than 10%).

frequency, artificial intelligence, arch, training function, learning function