A computational framework for predicting onset and crack propagation in composite structures via eXtended Finite Element Method (XFEM)

Autor(es)

Vinicius, Angelo Marcus ^[1] ; Leite, Ribeiro Marcelo ^[2] ; Volnei, Tita ^[2]

Afiliação do(s) autor(es)

^[1] Helicopteros Brasil SA, Itajuba, MG - Brazil

^[2] Univ Sao Paulo, Escola Engn Sao Carlos, Grp Estruturas Aeronaut, Sao Carlos, SP - Brazil

• Tipo de documento: Artigo Científico
• Fonte: LATIN AMERICAN JOURNAL OF SOLIDS AND STRUCTURES; v. 15, n. 11 2018.
• Citações Web of Science: 3

Resumo

Abstract The eXtended Finite Element Method (XFEM) has been reliably used for analyzing crack growth in 3D structural elements over last years. In fact, many researchers have worked in this field, but it is scarce to find scientific contributions about 3D XFEM models applied to the failure of non-standard composite parts, such as tapered structures and thick laminated composites. Thus, a new computational framework is developed, which is based on a new enhanced golden section search algorithm and 3D Puck’s action plane principle in order to define the crack initiation direction. This in-formation is integrated into a XFEM and used to enrich elements, which have failed during analysis. Compared to the traditional algorithm, the new methodology has convergence one order higher than the traditional one; and it is 20 times more efficient computationally. Therefore, if more precision is needed, then higher gains are achieved combined to lower computational cost by using the proposed framework. Moreover, thick laminated composites with layers mainly oriented to 90o were simulated under tension and compression via the computational framework, displaying results as reported in the literature. Also, compact tension tests with 0°, 90° and 45° specimens were evaluated, and numerical results were qualitatively coherent with experimental data. (AU)

• Processo FAPESP: 15/13844-8 - Formulação de elementos finitos e modelos de delaminação para material compósito laminado.
• Beneficiário: Marcelo Leite Ribeiro
• Modalidade de apoio: Auxílio à Pesquisa - Regular