Jtam-A4.dvi JOURNAL OF THEORETICAL SHORT RESEARCH COMMUNICATION AND APPLIED MECHANICS 55, 4, pp. 1443-1448, Warsaw 2017 DOI: 10.15632/jtam-pl.55.4.1443 EXPERIMENTAL ANALYSIS AND MODELLING OF FATIGUE CRACK INITIATION MECHANISMS Aneta Ustrzycka, Zenon Mróz, Zbigniew L. Kowalewski Institute of Fundamental Technological Research, Polish Academy of Sciences, Warsaw, Poland e-mail: austrzyc@ippt.pan.pl; zmroz@ippt.pan.pl; zkowalew@ippt.pan.pl Thepresentwork is devoted to simulation of fatigue crack initiation for cyclic loadingwithin the nominal elastic regime. It is assumed that damage growth occurs due to action of mean stress and its fluctuations induced by crystalline grain inhomogeneity and the free boundary effect. The macrocrack initiation corresponds to a critical value of accumulated damage. Themodelling of damage growth is supported byElectronic Speckle Pattern Interferometry (ESPI) apparatus using coherent laser light. Keywords: fatigue crack initiation, damage evolution, optical methods 1. Introduction The process of fatigue damage development and structural degradation is of local nature. The mechanism responsible for damage accumulation during cyclic loading below the yield point remains elusive. The present paper concerns the fatigue crack initiation and evolution inmetals subjected to loading at the stress level below the conventional yield strength.During the process of cyclic loading due to an inhomogeneous grain structure, the micro plastic effects develop and can be observed on the macro-scale. Sangid (2013) proposed a physically-based model for prediction of fatigue crack initiation based on the material microstructure. Using the potential offered by the novel experimental techniques, it is possible to identify physical phenomena and to describe mechanisms of degradation and fatigue damage development in modern structural materials. Their identification involves usage of damagedetectionmethods, bothdestructive and non-destructive to evaluate material behaviour under different loads (Kowalewski et al., 2008). Electronic SpecklePattern Interferometry (ESPI) is awidely used technique tomeasure full-field deformation on surfaces ofmany kinds of objects. The shielding effect on fatigue crack growth at constant amplitude loading and during application of overloads was investigated using ESPI by Vasco-Olmo et al., (2016). The analysis of the plastic processes that governs crack propagation was analysed usingESPImethod byFerretti et al., (2011). However, its application to the study of fatigue damage mechanisms has not been yet explored extensively. Therefore, in this study, ESPI is applied to investigate the distribution of strain fluctuations. 2. Applications of ESPI method Nowadays, theESPImethod seems to bevery attractive in capturing damage growth. Primarily, optical methods are used to analyze strain fields and strain localization on the surface of a specimen. For instance, strain distribution map using ESPI for nickel alloys (C – 0.09%, Cr – 8.8%,Mn – 0.1%, Si – 0.25%,W– 9.7%, Co – 9.5%, Al – 5.7%,Ta+Ti+HF– 5.5%) is presented for an increasing number of load cycles in Fig. 1. 1444 A. Ustrzycka et al. Fig. 1. Strain distributionmaps on the plane specimen surface using ESPI for different stages of the fatigue process During the process of pulsating cyclic loading the material deforms heterogeneously and numerous strain concentration spots are visible. It is expected that the microstructure of the materials plays an important role in strain localization. The strain concentration occurs especial- ly at grain boundaries. It can be expected that cracks start to nucleate in the zones containing large strain accumulation during cyclic loading. Apart from the density and distribution of de- fects in the volume of tested specimen, the specimen size and location of individual defects are important factors of fatigue damage initiation and its further evolution. The lateral profiles of maximal cross-sectional strain distribution for the increasing number of cycles are presented in Fig. 2. The results are shown for the nickel alloy. Fig. 2. The lateral profiles of maximal cross-sectional strain evolution The distribution of strain across the sample at the place of crack initiation and its evolution during the cyclic loading process is shown in Fig. 2. Using ESPI, the strain distributions for the first and last cycle (at the moment of rupture) are included. The strain values evolve faster at Short Research Communication – Experimental analysis and modelling of fatigue crack... 1445 the edge of the sample. Results will serve to verify and calibrate the mathematical description. The results of fatigue tests in the formof stress-strain diagrams exhibiting the permanent strain, stiffness modulus variation and hysteresis loops during selected cycles of fatigue are presented in Fig. 3. Fig. 3. Development of the fatigue hysteresis loop during selected cycles The fatigue hysteresis loops are notmeasurable during the cyclic process indicating that the material globally deforms in the elastic state, and themacroplastic strain developed in the final phase of process is of order 0.1%. The microdamage can evolve within the local zones of strain concentration. The understanding of mechanical properties of materials plays an important role during the fatigue process. 3. Mathematical modelling of fatigue damage evolution, numerical implementation and comparison with ESPI results In the present paper, the mathematical description of fatigue crack initiation and evolution is formulated. The problem of damage evolution for metals under cyclic loading inducing fatigue crack initiation and propagationwithin the elastic regime is discussed. The condition of damage accumulation is formulated after Mróz et al., (2004). It is assumed that when the critical stress condition is reached on the material plane, the damage zone Ω is generated. Afterwards, the growthof thedamagezonecanbedescribed.Themathematicalmodel is applied to studydamage evolution under cyclic tension and the predictions are compared with experimental data. The profile of normal strain ε(x) and stress σ(x) along the damage zone Ω is expressed as a sum of the mean (ε,σ) and fluctuation (ε̃(x), σ̃(x)) components ε(x) = ε+ ε̃(x) σ(x)=σ+ σ̃(x) (3.1) The material is assumed to be linearly elastic, but exhibiting a damage process at the strain concentration zones. In order to illustrate the problem (Fig. 4), the potential damage zoneΩ is selected with the largest strain and stress fluctuations. Damage evolution rule (3.2) was originally formulated by Mróz et al. (2004) for brittle materials dD=A (σ−σ∗0 σc−σ0 )n dσ σ∗c −σ ∗ 0 (3.2) whereA, n and p denotematerial parameters, σ0 is the damage initiation threshold, σc denotes the failure stress in tension for the damagedmaterial, σ∗0 and σ ∗ c are the threshold values for the undamagedmaterial andD denotes the scalar measure of damage (0¬D¬ 1). 1446 A. Ustrzycka et al. Fig. 4. Damage zoneΩ with the major stress fluctuation The stress value σ increases but the values of σ0 and σc decrease. Both σ0 and σc depend on the damage state according to the formula σc−σ0 =(σ ∗ c −σ ∗ 0)(1−D) p (3.3) The process of cyclic loading is described by the time variation of stress in the following form σ(t)= 1 2 σa[1+sin(ωt)] ω= 2π T (3.4) where σa is the stress amplitude. Substituting σ from Eq. (3.4) into Eq. (3.2) and integrating, the equation for the damage increase in one cycle∆Ds =Ds+1−Ds is expressed as follows Ds+∆Ds∫ Ds (1−D)np dD= Aωσa 2(σ∗c −σ ∗ 0) n+1 T∫ 0 (1 2 σa(1+sin(ωt))−σ ∗ 0 )n cos(ωt) dt (3.5) Finally, the damage evolution law for one cycle reads ∆Ds = 1 (1−Ds)np A (σ∗c −σ0) n+1 1 n+1 [(1 2 σa(1+sin(ωT))−σ ∗ 0 )n+1 − (1 2 σa−σ ∗ 0 )n+1 ] (3.6) The free edges of the sample due to surface irregularities act as a kind of stress concentrators. The influence of edge defects on the damage evolution and crack propagation is significant (see Fig. 2). Inorder toaccount for the edge effect, the stressfluctuation function (seeFig. 4) is introduced and the total stress expressed as follows σ(x)=σ+ σ̃(x)=σ+σy(x)=σ ( 1+α+β ∣∣∣ x a ∣∣∣ m) y(x)=α+β ∣∣∣ x a ∣∣∣ m (3.7) where y(x) denotes the fluctuation function, a is width of the sample, β and m are material parameters. The integration of stress fluctuation on [0,a] makes it possible to establish the parameter α a∫ 0 σ̃(x) dx=0 → a∫ 0 ( α+β ∣∣∣ x a ∣∣∣ m) dx=0 → α=− β m+1 (3.8) Short Research Communication – Experimental analysis and modelling of fatigue crack... 1447 The boundary condition at the external edge allows one to specify parameter β, thus σ̃(x= a)=σb → β= σb σ m+1 m (3.9) The value of σb is assumed to correspond to themeasured boundary fluctuation, hereσb =1.1σ. Finally, the stress distribution is expressed in the following form σ(x,t) =σ ( 1+α+β ∣∣∣ x a ∣∣∣ m)(1 2 + 1 2 sin(ωt) ) (3.10) According to themathematical description, numerical analysis of damage evolution (3.2) under mechanical loads (3.10) in elastic-plastic solids has beenmade. The evolution of damage during the growing number of cycles is shown in Fig. 5. Themacro-crack initiation is assumed to occur at the critical value of damage Dc =0.3. Fig. 5. Damage evolution related to the number of cycles for different values of the parameter n 4. Conclusions The present paper concerns the damage evolution in elastic-plastic materials subjected to cyclic loading. Themathematical description of crack initiation and propagation under cyclic loading is presented. Thismodel is supported by opticalmethods of stress and strainmonitoring (ESPI) for early detection, localization and monitoring of damage in materials under fatigue loading. The numerical prediction and the experimental results indicate a good correlation. Acknowledgement This work has been supported by the National Science Centre through the Grant No. 2014/15/B/ST8/04368. 1448 A. Ustrzycka et al. References 1. Ferretti D.,RossiM., Royer-CarfagniG., 2011,AnESPI experimental study on the pheno- menon of fracture in glass. Is it brittle or plastic? Journal of the Mechanics and Physics of Solids, 59, 1338-1354 2. KowalewskiZ.L.,Mackiewicz S., SzelążekJ., 2008,Evaluationofdamage in steels subjected to exploitation loading–destructive andnon-destructivemethods, International Journal ofModern Physics B, 22, 5533-5538 3. Mróz Z., Seweryn A., Tomczyk A., 2004, Fatigue crack growth prediction accounting for the damage zone,Fatigue and Fracture of Engineering Materials and Structures, 28, 61-71 4. Sangid M.D., 2013, The physics of fatigue crack initiation, International Journal of Fatigue, 57, 58-72 5. Vasco-Olmo J.M., Daz F.A., Patterson E.A., 2016, Experimental evaluation of shielding effect on growing fatigue cracks under overloads using ESPI, International Journal of Fatigue, 83, 117-126 Manuscript received July 18, 2017; accepted for print July 20, 2017