Predicting Wear Properties of Alloy 7090/SiC Nanoparticle Composites Using Artificial Neural Networks for aerospace application

Abstract

In the current study, we have explored the wear characteristics of Al 7090 alloy-silicon carbide nanoparticles composites that can be applied to aerospace. The composite material is prepared very carefully by stir casting in the laboratory, and silicon carbide nanoparticles are incorporated to enhance the strength. Wear testing was done using the Taguchi Design of Experiments methodology, and factors like load, speed, and composition were systematically varied. Finally, the results of the experiment, on the basis on which wear behavior was inferred to extract optimal conditions in terms of Sw and friction force , are reported. Furthermore, an ANN model is developed, calculations are done by the feed-forward backpropagation algorithm, and the Levenberg-Marquardt optimization algorithm is applied to optimize the parameters . A completely representative of 500 readings are undertaken, and 70% are used for training; 15% are used for testing and the remaining 15% are taken for validation. Surprisingly, the trained ANN model prediction accuracy rate reached 100%, a linear regression plot showing a good representation of the model’s accuracy. Conclusively, our results demonstrate that the developed ANN model is the most effective approach for obtaining the optimal prediction of Sw and friction force , thereby the ANN model can be used for predicting the performance of other composites.

Keywords

• composite materials
• wear properties
• artificial neural network
• aerospace applications
• optimization techniques