Enhanced Vibration Analysis of Composite Shells Using Artificial Neural Networks and Optimization Algorithms

In this study, an Artificial Neural Network (ANN)-based approach is proposed for the vibration analysis of fiber-reinforced composite shells. Traditional Finite Element Method (FEM) simulations, while accurate, are computationally expensive, making them unsuitable for real-time applications. To overcome this limitation, we developed an ANN model trained on FEM-generated datasets to predict natural frequencies of composite structures with high accuracy. The model accounts for key parameters such as fiber orientation, material properties, boundary conditions, and shell geometry. The results demonstrate that the ANN model achieves a Mean Absolute Error (MAE) of ۲.۱ Hz and an R² score of ۰.۹۸, significantly reducing computational time from ۶۰۰ seconds (FEM) to just ۵ seconds (ANN). Sensitivity analysis confirms that the model remains robust despite variations in material and geometric parameters. Compared to traditional AI models such as Support Vector Machines (SVM) and Random Forest (RF), the proposed ANN approach exhibits superior predictive accuracy and efficiency. This research highlights the potential of AI-driven structural vibration analysis, offering a powerful alternative to conventional numerical methods. Future work includes expanding the model to complex geometries, integrating ANN with FEM for hybrid modeling, and validating predictions through experimental testing. The proposed framework paves the way for real-time structural health monitoring (SHM) and smart composite materials design in aerospace, automotive, and civil engineering applications.

کلیدواژه ها:

Artificial Neural Network (ANN) ، Vibration Analysis ، Composite Shells ، Natural Frequency Prediction ، Finite Element Method (FEM)