New analytical model for delamination of CFRP during drilling

https://doi.org/10.1016/j.jmatprotec.2016.10.007Get rights and content

Abstract

Drilling of composite material structure is a widely used process in aeronautical assembly due to its best manufacturing value (ratio quality/cost). However, delamination which can be generated at the hole exit is still a problem and is directly related to the drill axial force, the composite material mechanical characteristics and the load distribution along the cutting and chisel edges. This paper analyzes the distribution of the load along the drill edges and points out a triangular distribution associated to an additional load located at the centre of the drill. This additional load depends strongly on the feed rate. An orthotropic analytical model is then proposed in order to determine the delamination critical thrust during drilling for different combined loadings. This new model is compared with the existing ones in the literature and with experimental results. This analysis highlights that delamination occurs in a mixed mode and not in mode I only. The use of the B-K criterion for the determination of the equivalent critical energy release rate GC and a cutting edge combined loading (triangular distribution associated to an additional load) allow a close correlation between experimental measurements and the prediction of this new model.

Introduction

The development of large civil aircrafts (Airbus A380, Airbus A350, Boeing 787…) is associated with an increased use of composite materials to reduce the weight of the plane while maintaining high specific characteristics. The widespread use of these materials in aircraft structures has become a necessity and must be achieved within certain requirements such as adequate load transfer from one structural element to another via joining. The latter, usually bolted or riveted, presents performances that depend heavily on the quality and dimensional accuracy of the drilled holes. Persson et al. (1997) pointed out the effects of drilling defects on strength and fatigue life of composite laminates. The results of the static testing showed that damaged specimens yielded significantly lower strengths (about 11%) than effect-free specimens. At l06 cycles, damaged specimens yielded substantially lower strengths (19%).

Although non-conventional means have been developed in the recent years (laser drilling, water jet drilling, ultrasonic…), drilling using cutting tool is still the most common process to achieve the best manufacturing value (ratio quality/cost).

Industrial challenges of drilling composite materials are now relatively clear. As defined by Airbus, Bombardier or Embraer in their Aerospace Process Specifications, they consist in several requirements: (i) a dimensional accuracy of IT grade 9 (for high load transfer holes) with a roughness Ra of less than 125 μm, (ii) delamination of one ply over 2.54 mm around the hole whatever its diameter, (iii) tool lifetime sufficiently long without improving the price per hole compared with aluminum or steel, (iv) in stacks made of only CFRP or hybrid materials (CFRP/metallic assemblies).

Drilling holes in composite structures is significantly different than drilling holes in metallic parts. Most of the problems encountered are associated with the drilling quality. The different types of damage induced by the drill are as follows.

Thermal damage (matrix softening or burning…) is due to the friction between the tool and the CFRP material and depends primarily on the margin width in contact with the borehole wall, associated with inadequate cutting conditions (high cutting speed and low feed rate). Consequently, the matrix is removed locally and the fiber reinforcement uncovered.

Brinksmeier et al. (2011) investigated different process parameters and their influence on surface integrity of the borehole. This was accomplished by measuring the cutting temperatures and forces. The investigations revealed that the use of high cutting speeds leads to increasing borehole surface layer damage in the CFRP material.

Jawahir et al. (2011) showed that this phenomenon is easily manageable controlling the spindle rotation speed (or tool cutting speed).

Park et al. (2013) developed and used empirical equations to model and analyze thermal damage and chip removal during the drilling process.

Montoya (2013) demonstrated that the temperature field near the tool/chip interface influences significantly the tool life. This temperature field is difficult to reach by experimentation, but can be obtained using numerical simulation of the workpiece thermal solicitations. The model developed allowed, by the use of an inverse method, to determine the hole wall temperature. Temperatures between 60 and 120 °C can be reached depending on the cutting conditions.

Mechanical damage (top and bottom delamination, fiber bundle tearing off the machined surface, spalling, uncut fibers, uncut resin, cracking…) results in the generation of defects, of which the most problematic one is delamination (at entry or exit of the drill within the material) and fiber bundle tearing off as pointed out by Ho-Cheng and Dharan (1990) or Jain and Yang (1993). This residual damage, depending on its size and occurrence, leads to lower mechanical properties for the CFRP material and in particular reduces the strength and the tensile fatigue endurance limit as shown by Brinksmeier et al. (2011) or Saleem et al. (2013).

Fiber bundle tearing off occurs on the surface of the hole for a particular orientation of the laminate ply with respect to the feed rate of the tool (−45°) associated with a loss of tool edge sharpness. It results in fiber bundles which are torn off the surface of the hole, this phenomenon being repeated for each ply in the same angular orientation between the tool and the fibers. To date no model has been developed to reflect the occurrence of this phenomenon.

Infeed side delamination (or peel-up delamination) is a damage which does not appear systematically. It is associated with the combined action of the helix angle, the drilling torque and the shape of the flute as demonstrated by Ho-Cheng and Dharan (1990) or Lachaud et al. (2001). It occurs by sliding the first plies up the flutes of the drill similar to the action of a corkscrew. This vertical action separates the first plies of the material and creates a delaminated area by fracture in Mode III loading (Fig. 1a). It can also be caused by unfavorable cutting conditions upon which the fibers are not cut sufficiently. Fiber fringes are being pulled up and cause delamination also according to mode I.

Hole exit delamination (or push-out delamination) occurs when the composite plate is not supported in its bottom part and when the drill is ready to exit the material as pointed out by Ho-Cheng and Dharan (1990) or Jain and Yang, 1993, Jain and Yang, 1994. When there are few plies left to drill, the rigidity reduction of the material makes the deformation easier and in particular the bending of this area of the plate. The penetration force of the drill may then be sufficient to cause the rupture of the interlayers and to allow the generation of a delaminated area when the tool exits the part (Fig. 1b). The fracture of the bottom surface plies occurs by both mode I and mode II fracture.

This type of delamination is more problematic since it is systematically repaired and therefore has an added manufacturing cost. De Zarate-Knorr (2014) pointed out that although this type of defect represents only 6% of the total damages associated with the drilling process, compared with 27% for oval holes, repairing the delaminated area takes 5 to 6 h per hole versus 10 to 15 min for an oval hole.

This is why it is important to control this phenomenon and to be able to predict when the drill starts to generate such default due to its wear and/or to inadequate cutting conditions. It has been pointed out by various authors that the axial or thrust force is a good indicator for controlling this damage. In addition, the optimal thrust force for no delamination can be used to control the drilling machine with thrust force feedback for maximizing productivity. The axial load can be controlled in real time and the feed rate decreased when the recorded axial load approaches the critical value.

The study presented below has therefore focused on carrying out (i) correlations between cutting forces and damage, mainly delamination at the exit of the hole, by defining different analytical models which take into account the loading on the drill cutting edge, (ii) experimental evaluation of the cutting edge loading as a function of the feed rate, and (iii) comparison of the models predictions with experimental data. The originality of this work is based on (i) an analytical approach which takes into account the coupling between bending and stretching commonly observed in CFRPs with various lay-ups, and (ii) an experimental approach which identifies the thrust force distribution along the cutting edge for different feed rate of the tool.

Section snippets

Models review for delamination prediction

Among the different existing models, the most interesting ones are presented thereafter. They can be classified according to the hypothesis made on the CFRP remaining to be drilled.

Experimental evaluation of the cutting edge loading

To determine analytically the critical thrust force, it is important to have an idea of the distribution of the cutting force on the tool cutting edge as a function of the drilling conditions, in particular the feed rate.

Tests have been performed using a Sandvik CoroDrill R846 twist drill which is suitable for titanium-based alloys and CFRP drilling. The material used in the experiments is a CFRP composite material made of intermediate modulus T800 carbon fibers and 924C epoxy resin. It

Point loading

In this case, it is assumed that the thrust force is exerted mainly by the chisel edge of the tool and therefore the loading will be concentrated in the center of the plate (Fig. 6a). Q is related to the axial force FZ distributed over the plate as follows:Q2πr=FzQ=Fz2πr

Substituting Q in the general formula (11), integrating and applying the boundary conditions of a clamped plate on its periphery, the deflection of the plate subjected to a concentrated or point load in its center can be

Conclusions

This study has focused on three important aspects, (i) the distribution of the load along the drill edges, (ii) the development of a combined loading model with mixed mode delamination propagation and (iii) the experimental validation of this new model.

In the case of a Sandvik R846 twist drill, it has been demonstrated a triangular distribution of the pressure along the cutting and chisel edges associated to an additional load located at the center of the drill. This additional load depends

Acknowledgments

Part of this work has been supported by the Unit for Training and Research in Mechanical Engineering of the University of the Basque Country (UFI11/29). The authors also wish to thank Ugaitz Orrantia, Begoña de Zarate-Knörr and Paula Arbe for their help in carrying out the experimental tests.

References (35)

Cited by (0)

View full text