Investigation of multiaxial fatigue in the context of turboengine disc applications
Introduction
A critical concern related to fatigue of materials and structures is multiaxial fatigue. Many engineering components that undergo fatigue loading experience multiaxial stress states. Rotating parts in turboengines, like turbine or compressor discs, are typical examples. They involve the interaction of low cycle fatigue (ground-to-ground cycles) and high cycle fatigue (vibratory) in regions with high biaxialities and high mean-stresses. Advances in material testing equipments during last 30 years have enabled to develop multiaxial testing facilities at room and high temperatures. These devices allow us to study the behaviour of materials and structural components by applying loads representative of the service life. The cruciform specimen and the thin-walled tubular specimen are usually used in multiaxial testing facilities. The tubular specimen is a versatile geometry. It can be tested either under axial–torsional loadings, which are the most commonly used multiaxial conditions, or under axial–internal/external pressure loadings. Some reviews of fatigue life investigations using axial–torsional or axial–internal pressure devices can be found in the literature [1], [2], [3], [4]. Testing facilities allowing the combination of axial load, differential pressure and torsion have been also reported in the literature [5], [6], [7], [8]. For the cruciform specimens, the tension–compression loads applied to its each pair of arms can produce a large range of biaxiality values in the central region. However, the stress distribution is usually evaluated using finite element analyses, and depends on several parameters (specimen design, configuration of the loading frame, specimen attachment and actuation method) leading to various testing facilities. A review of the cruciform specimen testing facilities is proposed by Makinde et al. [9], Boehler et al. [10], Hannon and Tiernan [11], and Abu-Fahra et al. [12].
Regarding the theoretical aspects of the multiaxial fatigue, many multiaxial criteria can be found in the literature nowadays. In spite of the number of proposed criteria, no universally accepted approach yet exists. Several good reviews of multiaxial fatigue criteria are available in the literature [13], [14], [15], [16], [17], [18]. These approaches can be divided into three main categories. One popular approach has extended the static yield criteria to fatigue by combining octahedral shear stress amplitude with hydrostatic pressure. By similarity, several strain based criteria have been proposed. Another approach uses energy as a correlating parameter for the multiaxial data. Energy criteria have been proposed and tested by several researchers (namely Garud [19], Ellyin [20], [21] and Radakrishan [22] and more recently Atzori et al. [23]) but present some difficulties for quasi elastic cycles. A third method for multiaxial fatigue life evaluation is to use the critical plane approach [24], [25], [26], [27], [28]. In all cases, as is underlined in the present paper, the largest difficulty with any of such fatigue criteria is to take simultaneously into account both the mean-stress effects (in uniaxial conditions) and the multiaxiality effects.
This paper presents the results of a large study on multiaxial fatigue in the context of compressor or turbine disc applications, involving both experimental and theoretical aspects. The objective of the experimental investigation is to study, under a large range of biaxial stress states, two typical turboengine disc materials: a titanium alloy (TA6V) and a nickel-based alloy (INCO718DA). In addition to standard LCF fatigue test under strain control with various values of strain ratio and HCF tests under stress control, the experimental program includes a significant number of complex multiaxial tests by using alternatively: axial–torsional tests, axial –internal pressure tests and biaxial fatigue tests on cruciform specimens.
Several existing multiaxial fatigue criteria (namely Crossland [29], Sines [30], Smith Watson Topper [24], Brown–Miller [25], Fatemi–Socie [26] and Gonçalvès et al. [31]) are then employed to determine their suitability at correlating the multiaxial fatigue data generated in the experimental program. Regarding the observed limitations, a new criterion is then proposed to estimate shear and equibiaxial fatigue conditions together with mean-stress effects.
Section snippets
Experimental devices to study multiaxiality effects
An extensive test program has been conducted on the facilities of multiaxial fatigue from ONERA and CEAT with two kinds of geometries: tubular specimens and cross-like specimens. Table 1 gathers the different testing facilities of both partners and their mechanical and thermal possibilities. Fig. 1 shows in broad outline the stress domain explored by the experimental devices. We can observe the good complementarity of tests that allow us to investigate a large domain, almost complete except
Cyclic inelastic analysis
The first step of the lifetime analysis is to find the stabilised operating conditions of the structure. The experimental program presented here includes three types of multiaxial tests: tension/torsion in-phase and out-of-phase conditions, tension/internal pressure in-phase conditions, biaxial fatigue loadings. In order to obtain the local stress–strain fields, all the multiaxial conditions have been calculated by various methods, strength of materials rules, (thermo-)elastic finite element
Multiaxial fatigue models
The objective of the present work is to identify reliable multiaxial fatigue models in the context of turbine disc applications. The existing models evaluated here can be classified in two categories: equivalent effective stress models and critical plane models.
Applications and discussion
The presented multiaxial models are examined to evaluate their ability to correlate both uniaxial data and multiaxial data. All the models have been identified through the same uniaxial data base, namely strain and stress fatigue tests at several loading ratio.
Conclusion
Some multiaxial fatigue test facilities have been presented here, involving the characterisation and the identification of multiaxiality effects for fatigue crack initiation in the prospect of turbine disc applications. Associated experimental procedures have been also described, taking into account thermal effects. All these multiaxial test conditions have been evaluated by strength of materials rules and thermo-elastic finite element analysis. Obviously the full cyclic inelastic finite
Acknowledgements
Elisabeth Ostoja Kuczynski, Francois Vogel (from Turbomeca) and Hacène Cherouali (from Snecma) are gratefully acknowledged for their constant collaboration on this research program. The financial support of DGAC (Délégation Générale à l’Aviation Civile) is also gratefully acknowledged.
References (44)
- et al.
A critical review on multiaxial fatigue assessments of metals
Int J Fatigue
(1996) - et al.
A comparative study of multiaxial high-cycle fatigue criteria for metals
Int J Fatigue
(1997) - et al.
Evaluation and comparison of several multiaxial fatigue criteria
Int J Fatigue
(2004) A criterion for fatigue under multiaxial states of stress
Mech Res Commun
(1974)- et al.
Multiaxial fatigue of severely notched carbon steel
Int J Fatigue
(2006) - et al.
Multiaxial high cycle fatigue criterion for hard metals
Int J Fatigue
(2001) - et al.
Multiaxial fatigue life estimation in welded joints using the critical plane approach
Int J Fatigue
(2009) - et al.
Multiaxial fatigue: a stress based criterion for hard metals
Int J Fatigue
(2005) - et al.
In-phase and out-of-phase axial–torsional fatigue behavior of Haynes 188 superalloy at 760 °C
- et al.
Testing materials under multi-axial loading
Materialprufung
(1995)