Evolution of residual stresses with fatigue crack growth in integral structures with crack retarders

https://doi.org/10.1016/j.msea.2009.06.024Get rights and content

Abstract

Bonded straps are investigated for their ability to retard a growing fatigue crack in metallic structures. The evolution of the residual stresses in the vicinity of the strap with fatigue crack growth has been studied. Cracks were grown in single edge-notched tension (SEN(T)) specimens reinforced with either a titanium or a carbon fibre reinforced plastics (CFRP) strap. The residual stress evolution has been measured in situ during crack growth using neutron diffraction, and modelled with a finite element approach. The peak residual stresses induced by the mismatch of the coefficient of thermal expansion between the strap and plate materials were seen to be fairly constant with crack growth. Good correlation between the experimental and the modelling results was found, except at very long crack lengths for a specimen that exhibited considerable fracture surface roughness at long crack lengths. The difference was attributed to wedging of the fracture surface changing the expected stress state, rather than any effect of the strap.

Introduction

The use of integral structures can potentially reduce the weight and the cost of aerospace assemblies. An inherent inconvenience with integral structures is, however, that there are no natural crack stoppers as there are in riveted structures. Large safety factors therefore have to be applied to integral structures if they are to comply with airworthiness regulations. This then reduces the weight competitiveness of metallic integral structures as compared with other technologies such as fibre reinforced polymer composites.

The research presented here is part of a project to investigate the role of adhesively bonded reinforcements in improving the damage tolerance characteristics of integral structures. This concept has been tested successfully on coupon specimens and a significant life extension was obtained [1].

The reinforcement, which is usually present in the form of an elongated plate or ‘strap’, should have a reasonably high strain to failure, high strength and good fatigue resistance. In order to achieve these properties at minimal additional weight, this implies that a different material than that of which the integral structure is made will be used for the reinforcing strap. As the bonding adhesive used to affix the strap will be cured at an elevated temperature, when the structure subsequently cools residual stresses are induced due to the mismatch of coefficient of thermal expansion between the integral structure and the strap. These residual stresses can be deleterious to fatigue crack initiation and growth in the structure. It has been found that these cure-induced residual stress have a significant effect on the fatigue crack growth rate in selectively reinforced structures [2]. Hence in order to be able to design an optimal structure without having to use severely conservative safety factors, the residual stresses have to be measured to be able to develop and verify predictive models. Measurements of the induced residual stresses in structures reinforced with candidate strap materials have previously been carried out at room temperature [3], [4]. Good correlation was found between the measurement and a finite element simulation of the cure-induced stresses.

In this work the evolution of the residual stress with fatigue crack growth has been measured with neutron diffraction and modelled with a finite element approach.

Section snippets

Specimens

Single edge-notched tension (SEN(T)) specimens were sectioned from a 10 mm-thick aluminium alloy 7085-T7651 plate. The final length was 400 mm and the width was 140 mm (Fig. 1). The assembly was bonded with FM® 94 adhesive supplied by Cytec Ltd. [5], and cured at 120 °C for 1 h with an applied pressure of 0.28 MPa. The samples were prepared at Cranfield University. The material properties used in the following analyses are given in Table 1.

Two specimens were tested, one with a titanium strap and one

Finite element modelling

The temperature at which the thermal residual stresses become zero upon re-heating after cure is referred to as the stress-free temperature (Tsf). The stress-free temperature often coincides with the glass transition temperature (Tg) [16] if the adhesive is cured above Tg. For a system where Tg is above the curing temperature, Tsf is taken to be equal to the curing temperature (Tc) [17]. Hence, in this work the induced thermal residual stresses have been modelled applying the temperature drop

Ti strap

The evolution of the residual stresses in the specimen bonded with a titanium strap can be seen in Fig. 3. Only the stresses which contribute to opening of the crack in mode I (the longitudinal strains as defined in Fig. 1) are presented to save space.

The evolution of the crack front determined as described in Section 2.3 can be seen in Fig. 4.

The evolution of the residual strain along the length of the strap (2.5 mm below the plate surface as indicated in Fig. 1) with crack length, and the

Discussion

In Fig. 3a the initial stress state in the specimen before any load cycles had been applied is shown. In this figure it can be seen that, apart from the higher stresses (compared with most of the other positions along the measured path) at the stress concentration at the initial notch tip, higher stresses are also found at the position under the strap. The peak stresses at the strap arise as the strap constrains the specimen in both the longitudinal and the transverse direction. This is

Conclusions

  • (1)

    The residual stresses associated with crack-retarding straps have been measured using neutron diffraction. The evolution of the residual stresses as a growing fatigue crack passes the strap have been measured experimentally and modelled using the finite element method.

  • (2)

    The strap provides a bridging effect to the fatigue crack wake, although there is progressive debonding of the strap as the crack length increases beyond the strap.

  • (3)

    Good correlation was obtained between the measured and modelled

Acknowledgements

The authors gratefully acknowledge Dr Ed Oliver at ENGIN-X, ISIS and Dr Darren Hughes at SALSA, ILL for help during the experiments. The specimens were prepared at Cranfield University, and Professor Phil Irving is thanked for his part in the collaboration. Alcoa and Airbus UK are gratefully acknowledged for financial support of this project, and Dr Markus Heinimann and Gerry Shepherd are thanked for their comments on the draft of this paper. MEF is supported by a grant through The Open

References (17)

  • X. Zhang et al.

    Engineering Fracture Mechanics

    (2009)
  • C.D.M. Liljedahl et al.

    Composite Structures

    (2008)
  • C.D.M. Liljedahl et al.

    Materials Science and Engineering

    (2008)
  • T. Pirling et al.

    Materials Science and Engineering: A

    (2006)
  • B. Clausen et al.

    Acta Materialia

    (1998)
  • D.-C. Seo et al.

    Composite Structures

    (2002)
  • H. Hosseini-Toudeshky et al.

    Composite Structures

    (2007)
  • P. E. Irving, D. Figueroa-Gordon, Routes to improved damage tolerance; prediction of damage tolerant performance of...
There are more references available in the full text version of this article.

Cited by (0)

1

Current address: Institute of Materials Engineering, ANSTO, PMB1, Menai, NSW 2234, Australia.

View full text