Abstract
This chapter quantifies the effect of uncertainty in natural frequencies of laminated composite plates based on neural network–based approach coupled with finite element analysis. An exhaustive comparative investigation on the performance of artificial neural network and polynomial neural network is carried out from the viewpoint of accuracy and computational efficiency. The stochastic system parameters are modeled following a layer-wise random variable–based approach, where the random system properties are considered to be different at different layers of the laminate for a particular realization of Monte Carlo simulation (MCS). Both individual and combined variations of stochastic input parameters are considered to address the aspect of low and high dimensional input parameter spaces, respectively. The convergence of the proposed neural network–based algorithm is verified and validated with original finite element method and direct MCS.
Original language | English |
---|---|
Title of host publication | Handbook of Probabilistic Models |
Editors | Pijush Samui, Dieu Tien Bui, Subrata Chakraborty, Ravinesh C. Deo |
Place of Publication | Oxford |
Publisher | Butterworth-Heinemann |
Chapter | 22 |
Pages | 527-547 |
Number of pages | 20 |
ISBN (Electronic) | 978-0-12-816514-0 |
ISBN (Print) | 9780128165461 |
DOIs | |
Publication status | Published - 2020 |
Keywords
- Artificial neural network
- Comparative stochastic natural frequency analysis
- Low-frequency vibration
- Polynomial neural network
- Surrogate-based monte carlo simulation