PSI - Issue 18

D.A. Bondarchuk et al. / Procedia Structural Integrity 18 (2019) 353–367 D.A. Bondarchuk, B.N.Fedulov, A.N. Fedorenko/ Structural Integrity Procedia 00 (2019) 000 – 000

354

2

1. Introduction A well-known stress concentration problem - the so-called free-edge effect was investigated by a large number of scientists such as Mittelstedt et al. (2007), Rasuo et al. (2011), Hu et al. (1995), Solis et al. (2018), Amrutharaj et al . (1996), Islam and Prabhakar (2017), Hajikazemi and Paepegem (2018). The effects mainly explained by the mismatch of the elastic material properties between two adjacent dissimilar laminate layers, which causes a concentrated occurrence of three-dimensional out of plane stress components. Hu et al. (1995) carried out a detailed study of the free edge effect on redistribution of residual stresses, radial residual stress reversal and interfacial debonding through theoretical modeling, numerical analysis and experimental investigation for metal matrix composites (MMCs). The consideration of the effect of internal stress phenomena on free edge in carbon-epoxy composites unlike MMCs is not sufficiently studied and covered in research articles up to date. However, the articles devoted to the analysis of the defects in the free edge zone under loading (including thermo-mechanical loading) annually appear in scientific journals. It is worth paying attention to the works conducted by Solis et al. (2018), Amrutharaj et al . (1996), Islam and Prabhakar (2017), Hajikazemi and Paepegem (2018). What is an important, in most researches, residual stress in composites are not taken into account in mathematical models. Fedulov et al. (2016) and Ushakov et al. (2015) showed that residual stresses developing during manufacturing have a direct influence on the product quality and can immediately cause problems with assembly, as it requires the use of additional shims to ensure close contact between construction parts. Moreover, the residual stress can reach values close to the ultimate one and significantly reduce the strength of the final composite structure. The existence of residual stresses combined with mechanical loads can cause several defects in composite laminates and structures such as fiber and tow misalignment, transverse cracking (known as micro-cracking) delamination and warpage. In order to have safe and reliable structures, the designers are forced to use the higher margins of safety that, in turn, leads to «overdesign» of structures. All these uncontrolle d manufacturing effects make researches to study chemical and physical processes accompanying matrix cure. The development of approaches to analyze the thermo-chemical and mechanical aspects of the manufacturing process, such as changes in mechanical properties, stresses, and deformations of thermosetting composites well systematized in works conducted by Baran et al. (2016), Johnston (1997), White and Hahn (1992), Johnston et al. (2001), Chachad et al. (1995), Valliappan et al. (1996). The purpose of this study can be formulated by following steps:  Developing of the modeling technique for prediction of stress-strain state distribution at the composite structure caused by manufacturing process  Determine the value of residual stresses in composite sample during the polymerization process  Determine the change in residual stress states of polymerized composite sample after straight line cut into two parts  Compare damage distribution near the free edge with the presence of manufacturing residual stresses with and without one in case of uniaxial loading of composite material The problem of residual stress formation is complex as it depends on many factors and includes several subproblems. The analysis of sample with [0 0 /90 0 ] n lay-up was selected for consideration in this work in terms of simplicity and as the first step of research. Since there is symmetry, the modeling can be provided in two dimensional plane strain formulation. It is worth to note, effects of shape distortion and fracture effects become more evident for composites with [0 0 /90 0 ] n lay-up, due to maximum difference in anisotropic properties of each individual layer. The present work considers an example of AS4/8552-1 carbon-epoxy composite in view of popularity in engineering practice and availability of data in literature. The data provided in works conducted by Baran et al. (2016), Paramentier et al. (2014), Ersoy et al. (2010), Boyard N. (2016) and others. The study was conducted by means of general-purpose FE package ABAQUS and special constitutive material model that was implemented in special user subroutine-UMAT.