Efficient algorithms for calculation of shear stress amplitude and amplitude of the second invariant of the stress deviator in fatigue criteria applications
Introduction
Many of the existing high cycle fatigue criteria can be divided into two groups [1]. The first group includes the criteria based on the critical plane concept. The second group comprises the criteria based on the stress invariants. Critical plane type criteria usually require the evaluation of the amplitude of the shear stress τa acting tangentially to the critical plane. Stress invariants based criteria typically call for the determination of the amplitude of the second invariant of the stress deviator J2,a. From this, can be found, which has dimension of stress.
With the exception of proportional paths, the evaluation of τa and is not a trivial problem. A coherent definition of the shear stress amplitude has been given by Papadopoulos [2]. It is based on the construction of the minimum circle circumscribed to the path described by the shear stress vector on the critical plane. The radius of the minimum circumscribed circle is defined to be the shear stress amplitude. This definition can be extended to the case of . It is then required to construct the minimum hyper-sphere which encompasses the load path in the stress deviator space. Its radius is equal to .
Algorithms for constructing the minimum circumscribed circle or hyper-sphere have already been proposed by Papadopoulos [2] and Dang Van et al. [3]. However, these methods often necessitate a large amount of computational time. The purpose of the present paper is to introduce an efficient algorithm for the rapid calculation of τa and of . The proposed method makes use of numerical optimization routines currently available as free Fortran or C source codes or as part of commercial mathematical software.
Section snippets
The critical plane approach
One of the first fatigue criteria based on the critical plane concept first appeared in the paper of Findley [4]. On a given elementary plane Δ, identified by its normal vector n, the stress tensor σ combined with n defines the normal stress vector σNand the shear stress vector τFollowing Findley, the most damaged plane is the one experiencing the greatest sum of the amplitude of the shear stress vector and of the normal stress. Thus the criterion is written as
Criteria based on the stress invariants
Concerning the stress invariant based criteria, the critical quantity is usually the second invariant of the stress deviator, calculated taking into consideration only the alternating components. This quantity is thus defined as J2,a. Among these criteria let us recall the Sines criterionThe term J2,a was originally intended as J2 calculated over the alternating components, i.e. over σij,a if σij(t)=σij,m+σij,af(t).
Later Fuchs, as reported by Garud [14] and Bannantine et al. [15],
Shear stress amplitude calculation
Supposing the shear stress acting on Δ being defined by its components in a Cartesian reference (u,v) fixed on Δthe shear stress tip follows the curve shown in Fig. 4.
The centre of the circumscribed circle and its radius have been calculated with the Points Combination Algorithm (being affected by roundoff errors only, its results can be considered the reference solution), the Incremental Method and fminimax, all implemented in a Matlab [10] script, and yielded the
Concluding remarks
Critical plane type criteria as well as second invariant based criteria applied to non proportional load paths require to find the centre of respectively the circle and the hyper-sphere circumscribing the load path.
The Points Combination Algorithm allows to solve this problem with high accuracy but it makes calculation time too long if the load path is defined with a very fine time step.
Optimization routines give the same results with a number of function evaluations almost independent of the
Acknowledgements
The author expresses his gratitude to Dr. K. Dang Van and Dr. H. Maitournam of LMS, Ecole Polytechnique, France, for their valuable advice and encouragement and for providing him with the stress data and the Point Combination Algorithm results shown in the second example. The author also wishes to thank Dr. I.V. Papadopoulos of JRC, Ispra, Italy, for useful suggestions and comments on the contents of this paper.
References (15)
- et al.
A comparative study of multiaxial high-cycle fatigue criteria for metals
Int. J. Fatigue
(1997) Critical plane approaches in high-cycle fatigue: on the definition of the amplitude and mean value of the shear stress acting on the critical plane
Fatigue Fract. Engng Mater. Struct.
(1998)- et al.
On a new multiaxial fatigue limit criterion: theory and applications
A theory for the effect of mean stress on fatigue of metals under combined torsion and axial load or bending
J. Engng Ind. Trans. ASME
(1959)An explanation on fatigue limit under combined stress
Bull. JSME
(1977)A general criterion for high cycle multiaxial fatigue failure
Fatigue Fract. Engng Mater. Struct.
(1991)- et al.
A theory for fatigue under multiaxial stress–strain conditions
Proc. Inst. Mech. Engrs
(1973)
Cited by (55)
Fatigue reliability prediction of shape memory alloy parts based on multi-scale high cycle fatigue criterion
2023, Reliability Engineering and System SafetyValidation of fatigue criteria under fretting fatigue conditions
2022, International Journal of FatigueCrack initiation and life estimation for 316 and 430 stainless steel specimens by means of a critical plane approach
2020, International Journal of FatigueA numerical investigation on multiaxial fatigue assessment of Nitinol peripheral endovascular devices with emphasis on load non-proportionality effects
2019, Engineering Fracture MechanicsCitation Excerpt :The clear difference between FS and BM criteria predictions, when reversal or MCC methods are used for calculating the shear strain amplitude, underlines the importance of evaluating the effects of all the fatigue cycle when non-proportional local responses are activated. However, due to the high computational cost of the MCC approach implementation, when the fatigue assessment have to be performed on the whole device (no specific areas of high deformation can be recognized), a more efficient algorithm could be employed to allow faster analyses [44,45]. The presence of stress concentrations, combined with high three-dimensional stress/strain gradients, affects the material response, leading to a reduced number of elements experiencing RF > 1: operating the fatigue strength assessment then requires a particular attention in choosing the model mesh size, which may affect the results.
Multiaxial fatigue damage estimation of structures under random vibrations using Matsubara's criterion
2019, International Journal of Fatigue