Estimation of Strain Values of Aluminum Based Aerospace Materials By Using Artificial Neural Networks

Nowadays, artificial intelligence and machine learning methods are increasingly used for applications in engineering disciplines. As a result, methods provide solutions to technical problems. These algorithms also provide an alternative to the methods and techniques used so far. Artificial neural networks (ANN) are one of these artificial intelligence methods. ANN covers many processes, such as information acquisition, digitization, and improvement. ANNs are also widely used in aerospace engineering. Various studies have been conducted, especially in the field of the metal engineering and material science. In this study, conventional test methods for determining the tensile strength of metals and alloys used in the aerospace industry were investigated. In this context, a series of the analytical modeling by digitization with the Levenberg-Marquardt algorithm was performed to investigate the performance of ANNs as an alternative testing method for determining the tensile strength of metals used in the aerospace industry. An analytical model was created in an Excel spreadsheet that was digitized using artificial neural networks. The success of the analytical modeling was determined by comparing it to the experimental data. As a conclusion, a high accuracy rate was obtained by the application. The correlation between ANNs and the Levenberg-Marquardt algorithm, the limitations of its application in materials and structural engineering, and the advantages of the method are discussed.