Identification of process-induced residual stresses in 3D woven carbon/epoxy composites by combination of FEA and blind hole drilling

Abstract

Process-induced residual stresses in 3D woven carbon-epoxy composites are studied by blind hole drilling experiments interpreted with finite element (FE) modeling. It is assumed that residual stresses are primarily caused by the difference in thermal expansion coefficients of the constituents which are modelled as temperature-dependent linear elastic solids. The impact of residual stresses is quantified by drilling blind holes in the composite panels and mapping the resulting in-plane surface displacements by electronic speckle pattern interferometry. Mesoscale finite element models are used to correlate these surface displacements with the volumetric distribution of the residual stresses in the composite. This is done by determining the effective temperature drop ΔT^eff that results in the same predictions for the surface displacements as experimentally measured. The effective temperature drop approach allows to use linear elastic models while approximately accounting for various nonlinear effects occurring in the material during processing. The models are also used to establish the sensitivity of the predicted results to the exact location of a hole and its depth.

Introduction

Manufacturing-induced residual stresses in carbon epoxy composites can occur due to the difference between coefficients of thermal expansion (CTE) of the resin and fibers and also chemical shrinkage [1], compaction of the resin [2], nonlinear distribution of temperature and degree of cure throughout the part during curing, etc. It has been shown that the level of residual stresses can substantially affect the quality and performance of the composite parts including their final shape [3], [4] and strength [5], [6].

A significant amount of publications have been devoted to the manufacturing-induced residual stresses in laminated and 2D woven composites. For example, Cowley and Beaumont [3] carried out experimental study on the residual stresses in the laminated fibrous polymers and modeled the stress state using classical lamination theory. Golestanian and El-Gizawy [7] simulated the entire curing and cooling cycle for the woven carbon and fiberglass mats impregnated with epoxy resin using resin transfer molding (RTM). They obtained values for the residual stresses in the selected points of the composite plate. Fiedler et al. [8] investigated the influence of the manufacturing-induced thermal residual stress on the transverse strength of the unidirectional CFRP composite material. Agius et al. [9] performed FE simulations to predict residual stresses in the multidirectional laminates based on the epoxy resin chemical shrinkage and mismatch in CTEs of the epoxy and carbon fiber reinforcement. Benavente et al. [10] simulated macroscopic residual deformation of a laminate composite part based on the temperature-dependent viscoelastic epoxy resin behavior.

In 3D woven composites, presence of additional constraints in the third direction leads to higher residual stresses from mismatch of CTEs of the fiber and matrix. It has been shown in [11] that for certain reinforcement architectures manufacturing-induced residual stresses might lead to significant levels of residual stresses causing microcracking in the resin pockets between the tows.

Determining the residual stress distribution in 3D woven composites experimentally is challenging due to their complex microstructure and the resulting high level of inhomogeneity and anisotropy. One of the approaches to estimate residual stresses is to utilize hole drilling experiments. In such experiments, the residual stresses are estimated based on the displacements around a circular hole drilled in the material, as described in [12], [13]. However, this purely experimental approach developed for homogeneous materials does not allow to obtain a detailed distribution of the residual stress in woven composites. There is a need to numerically interpret the measurement results. For example, Pisarev et al. [14] proposed to use analytical solutions given in [15] to correlate residual stresses with the displacements due to the through-thickness holes drilled in the composite plates. They measured the displacements using electronic speckle pattern interferometry (ESPI) and assumed composite material to be homogeneous and orthotropic. Akbari et al. [16] obtained residual stress in a filament wound laminated carbon/epoxy ring using incremental hole drilling method. In their study, they used a combination of strain gage measurements and finite element simulations to obtain the in-plane residual stresses in the plies assuming the material of each ply to be homogeneous and orthotropic. Wu et al. [17] estimated residual stresses in 2D woven composite utilizing a combination of FE modeling and Moiré interferometry. In their study, they assumed a uniform distribution of stress within the material removed during drilling to produce the values of residual stress components based on the measured displacements at the sample points on the composite plate surface around the drilled hole.

In the present paper, we estimate residual stresses in 3D woven composites using measurements of displacements on the surface of the composite panels caused by drilling of circular blind holes in various locations. The displacements can be measured utilizing either ESPI or digital image correlation (DIC) techniques. In our previous studies we determined that ESPI provided better resolution of the displacement gradients [18] so it was selected for the present work. The displacements are correlated to the residual stress by mesoscale finite element models of the composites. These models assume that the primary mechanism of the residual stress formation is the mismatch in CTE of fiber and matrix as the composite cools from curing to room temperature.

The rest of the paper is organized as follows. Section 2 provides a description of the experimental techniques utilized in this work. Section 3 describes the numerical modeling procedure and also includes information on mechanical properties of the composite phases. Section 4 presents the results of numerical parametric studies related to the hole drilling experiment. Section 5 provides interpretation of the hole drilling experiments by FEA to obtain the full residual stress field in 3D woven composites. Section 6 contains concluding remarks and comments.