Tadeusz Łagoda

A critical plane approach based on energy concepts: application to biaxial random tension-compression high-cycle fatigue regime

Abstract In this paper the energy parameter, defined for random loadings, is analysed. Under uniaxial loading this parameter distinguishes between the strain energy density for tension (positive) and the strain energy density for compression (negative). As a consequence, if there is no mean component in the random loading, we obtain a random history of strain (elastic and plastic) energy density with zero expected value. Under multiaxial loadings the normal strain energy density in the critical plane (i.e. the plane of the maximum damage) is understood as the energy parameter. The history of strain energy density is schematized with use of the rain-flow algorithm. Fatigue damage is accumulated according to Palmgren-Miner hypothesis and the standard fatigue characteristic of the material, rescaled with use of the considered energy parameter. The proposed parameter was verified during fatigue tests of cruciform specimens made of 10HNAP steel, subjected to biaxial non-proportional random tension-compresion. The calculated fatigue lives are included in the scatter band of the experimental data of a factor of 3. Thus, the normal strain energy density in the critical plane seems to be an efficient fatigue parameter under random non-proportional loadings for high-cycle fatigue.

Assessment of multiaxial fatigue behaviour of welded joints under combined bending and torsion by application of a fictitious notch radius

Abstract A major problem in the assessment of the multiaxial fatigue behaviour in welded joints is the determination of local notch stresses, since the real local radii in most cases are not known. This can be circumvented by introducing a fictitious notch radius, according to Neuber and Radaj. The calculations of local equivalent stress amplitudes for welded joints (tube–tube and flange–tube) under pure bending, pure torsion and combined in- and out-of-phase bending with torsion can then be carried out by stress-based and strain energy density-based criteria. Among the investigated methods, the most reliable results are obtained with the effective equivalent stress hypothesis, an integral interference plane oriented method, which transfers the results of complex loadings onto the safe side of the reference scatter band for pure bending. However, the best fit with the reference scatter band is realized by the application of the maximum shear and normal strain energy density parameter in the critical plane.

The influence of the mean stress on fatigue life of 10HNAP steel under random loading

Abstract This paper contains the results of uniaxial random load fatigue tests with different mean values performed on 10HNAP steel specimens. The experimental fatigue life was compared with the life determined according to different methods of amplitude transformation of the cycles. Cycles were counted with the rain flow algorithm and damage was cumulated with the Palmgren–Miner hypothesis. It has been shown that in the case of “symmetric” stress histories with zero expected values, the mean values of selected cycles do not strongly influence the calculated fatigue life and transformation of the cycle amplitudes; hence, no mean stress correction for cycle mean values different from zero is required. As for the “asymmetric” histories, the algorithm of transformation of the stress history (because of its global expected value different from zero) gives similar results to the algorithm of amplitude transformation of cycles. A suitable choice of the transformation relationship is very important. In the case of 10HNAP steel, stress amplitude transformation according to the Goodman relationship gives good results.

Estimation of Fatigue Life of Materials with Out-of-Parallel Fatigue Characteristics under Block Loading

The paper presents the algorithm of fatigue life estimation for materials with out-of-parallel fatigue characteristics under block loading. Brass CuZn40Pb2, medium-alloy steel 30CrNiMo8 and high-alloy steel 35NCD16 belong to such materials. Brass CuZn40Pb2 was used for analysis. The experimental results were compared with those calculated according to the assumed model, and satisfactory results were obtained.

Fatigue Life Estimation under Cyclic Loading Including Out-of-Parallelism of the Characteristics

The paper presents an algorithm of fatigue life determination for materials with no parallel fatigue characteristics under pure bending and pure torsion. The presented model uses the iteration method, and the applied fatigue criterion is function of the ratio of normal and shear stresses coming from bending and torsion, respectively. Three materials were applied for analysis: CuZn40Pb2 brass, 30CrNiMo8 medium-alloy steel and 35NCD16 high-alloy steel.

Fatigue Life According to Cyclic Strain Characteristics Determined from Variable-Amplitude Loading*

Abstract The paper includes the analysis of influence of cyclic stress-strain curve parameters on the fatigue life estimation. The parameters of cyclic stress-strain curve, described by the Ramberg-Osgood relation, were determined by a new method from the variable-amplitude loading histories. The variable-amplitude stresses were used to calculate strain courses based on different values of cyclic stress-strain curve parameters. The calculated and experimental strain courses were applied to calculate fatigue life with use of damage accumulation hypothesis by Serensen-Kogayev.

The use of spectral method for fatigue life assessment for non-gaussian random loads

Abstract The well-known problem with the fatigue lifetime assessment of non-Gaussian loading signals with the use of spectral method has been presented in the paper. A correction factors that transform the non-Gaussian signal into an equivalent Gaussian signal proposed by Bracessi et al. (2009) has been used for the purpose of lifetime calculations together with Palmgren-Miner Hypothesis. The calculations have been performed for the 10HNAP steel under random non-Gaussian load with four dominating frequencies. The signal has been generated on the test stand SHM250 for random tension-compression tests. The results with zero and non-zero mean stresses have been used to calculate the fatigue life with the frequency domain method based on Dirlik’s model and with a time domain method with the use of the rainflow cycle counting algorithm. The obtained calculation results have been compared with experimental results.

The Fictitious Radius as a Tool for Fatigue Life Estimation of Notched Elements

In this paper, the fictitious radius - according to Neuber’s method for determination of stresses at the notch root was used. Next, the fatigue lives of elements of the ring notches were calculated, and then compared with results of experimental tests of S235JR steel samples. However, the obtained fatigue lives did not bring satisfactory results. It has been demonstrated that the fictitious radius strongly depends on the expected fatigue life

Low cycle fatigue of steel in strain controled cyclic bending

Abstract The paper presents a comparison of the fatigue life curves based on test of 15Mo3 steel under cyclic, pendulum bending and tension-compression. These studies were analyzed in terms of a large and small number of cycles where strain amplitude is dependent on the fatigue life. It has been shown that commonly used Manson-Coffin-Basquin model cannot be used for tests under cyclic bending due to the impossibility of separating elastic and plastic strains. For this purpose, some well-known models of Langer and Kandil and one new model of authors, where strain amplitude is dependent on the number of cycles, were proposed. Comparing the results of bending with tension-compression it was shown that for smaller strain amplitudes the fatigue life for both test methods were similar, for higher strain amplitudes fatigue life for bending tests was greater than for tension-compression.

Concept of Fatigue for Determining Characteristics of Materials with Strengthening

The paper presents the concept of division of the total strain amplitudes. Simulations were performed for high-alloy steel X6NiCr3220 for proposing a new curve of cyclic strain based on the best fit to the experimental points and plotted the hysteresis loop. Proposed allocation of the total strain amplitude in three parts: the plastic strain amplitude, strain amplitude and the amplitude of the linearly elastic strain coupled. In order to preserve the forms of popular formula Ramberg-Osgoda and Manson-Coffin-Basquin modified them in such a way that added to their member responsible for the description of the coupled strain. Inclusion of additional term leading to closer representation of the actual material properties.