Publication:
The influence of laminate stacking sequence on ballistic limit using a combined Experimental/FEM/Artificial Neural Networks (ANN) methodology

Loading...
Thumbnail Image
Identifiers
Publication date
2018-01-01
Defense date
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Elsevier
Impact
Google Scholar
Export
Research Projects
Organizational Units
Journal Issue
Abstract
Composite laminates subjected to high velocity impacts are usually studied by means of experimental or numerical approaches. Nevertheless, these techniques are not appropriate to analyze the wide range of possibilities in the design of laminates (a great amount of time and economic resources are required); therefore, more efficient methods would be desirable. This work presents the capability of an ANN approach to predict the change of the ballistic limit with the laminate stacking sequence, and hence to find the optimum laminate combination. In order to obtain a refined ANN tool, a combined methodology of experimental and finite element method has been used. The results of the experimentally validated FEM model, are used to provide the data to the ANN. Once trained, the ANN is able to predict accurately the ballistic limit of composite laminates studied. The ANN allows studying very efficiently the whole possibilities of laminate stacking sequence using the common orientations, in symmetric 12 plies laminates (4096 cases). In addition, a deeper comprehension of composite plates when subjected to high velocity impact has been achieved by means of the analysis of the results. Conclusions obtained can be used by composite design engineers to improve ballistic performance of composite plates. (C) 2017 Elsevier Ltd. All rights reserved.
Description
Keywords
Finite element method, Experimental tests, Artificial neural network, Multilayer perceptron, High velocity impacts, Composite laminates
Bibliographic citation
Artero-Guerrero, J. A., Pernas-Sánchez, J., Martín-Montal, J., Varas, D., & López-Puente, J. (2018). The influence of laminate stacking sequence on ballistic limit using a combined experimental/FEM/artificial neural networks (ANN) methodology. Composite Structures, 183, 299-308.