Hide/Show Apps

Development of an artificial neural network based analysis method for skin-stringer structures

2017
Cankur, Anıl
The purpose of this work is to develop a tool for the design of skin-stringer structures. The tool to be developed needs to quickly identify the load carrying capacity and weight of these structures. For this purpose, finite element (FE) models of 1440 different skin-stringer structures were created with a script written in Python 2.7, and these models were analyzed using the commercial FE program ABAQUS. The script was used to construct the model, analyze the model, calculate the buckling load, the collapse load and the total mass, and collect the ABAQUS analysis results. The input parameters and analysis results of these 1440 model were used to create an artificial neural network (ANN) in the MATLAB NNTOOL toolbox for fast determination of the buckling load, the collapse load, and the weight of the skin-stringer assemblies that have not yet been created. The performance of the generated ANN was shown in comparison with FE results.