Showing papers on "Vibration fatigue published in 1991"

Fatigue-life distributions of concrete for various stress levels

Abstract: The paper presents a study of fatigue-life distributions of concrete for various levels of applied stresses. An experiment was conducted to obtain the fatigue lives of concrete at various given stress levels. The concrete beam specimens were tested under 4-point flexural loading. Three different levels of applied fatigue stresses were considered. The present study indicates that the probabilistic distributions of fatigue life of concrete depend on the level of applied fatigue stress. The shapes of the Weibull distribution to describe the fatigue life of concrete differ for various levels of applied stresses. This effect must be taken into account in realistic fatigue reliablity analysis to secure adequate fatigue resistance of concrete structures. The graphnical method, the method of moment, and the method of maximum likelihood estimation are shown to yield very similar values of distribution parameters for the present fatigue data. The shape parameter of the Weibull distribution for the fatigue life of concrete ranges from 2.0 to 4.0, depending on the level of applied fatigue stress.

Energy dissipation: the leading factor of fatigue

Abstract: By dynamic four point bending tests under controlled strain with a non-sinusoidal, composite signal it is shown that the fatigue life of asphalt beams is not only determined by the height of the strain amplitudes, but that the average energy, dissipated per load cycle describes the fatigue test much better. With the dissipated energy, a predictive model for the fatigue life had been composed, enabling calculation of the fatigue damages for, in principle, all types of load signals.

Cumulative Damage Theory of Concrete Under Variable-Amplitude Fatigue Loadings

Abstract: A nonlinear cumulative damage theory that can model the effects of the magnitude and sequence of variable-amplitude fatigue loadings is proposed. Concrete beam specimens are prepared and tested in 4-point flexural loading conditions. Variable-amplitude fatigue loadings in 2 and 3 stages are considered. The present experimental study indicates that the fatigue failure of concrete is greatly influenced by the magnitude and sequence of applied variable-amplitude fatigue loadings. It is seen that the linear damage theory proposed by Palmgren and Miner is not directly applicable to concrete under such loading cases. The sum of the cumulative damage is found to be greater than 1 when the magnitude of fatigue loading is gradually increased and less than 1 when the magnitude of fatigue loading is gradually decreased. The proposed nonlinear damage theory, which includes the effects of the magnitude and sequence of applied fatigue loadings, allows more realistic fatigue analysis of concrete structures.

Monitoring fatigue cracks in gears

Abstract: Vibration analysis is the most common means of gear monitoring and diagnostics. Gear vibration is affected by faults but the signal is usually picked up at the case, where it is also affected by the structural response. An appropriate filtering function is therefore proposed to recover the torsional gear vibration from the case vibration signal. The restored gear vibration can then be used with greater confidence than case vibration both for particular diagnostics purposes like crack detection and for more general objectives. This technique and its possible advantages in fatigue crack detection are illustrated in the paper.

Variational finite element-tensor formulation for the large deflection random vibration of composite plates

Abstract: A numerical integration routine is derived from a set of unified single step integration algorithms using a weighted satisfaction of the equilibrium equation governing the large deflection random response of laminated composite plates. The equilibrium equation is derived using a constant matrix large deflection finite-element formulation. In-plane inertia terms are considered in the formulation, however, rotary inertia terms are assumed negligible. Probability density, spectral density and autocorrelation functions of the maximum displacement and strain responses are presented for three acoustic excitation levels. Classical thin plate boundary conditions and pseudo white noise excitation are used in this investigation.

A Damage Integral Methodology for Thermal and Mechanical Fatigue of Solder Joints

Abstract: The differential thermal expansion of the connected members of a soldered assembly during temperature cycles produces mechanical displacements in solder joints. The resulting cyclic stresses are the driving force for damage processes that eventually cause fatigue failures. Conventional approaches to solder joint fatigue rely on correlations between fatigue life of the solder joint and some measurable or calculable parameter characteristic of the fatigue loading. These approaches often require extensive fatigue data to establish such correlations. Furthermore, their validity is often limited to the range over which the parameters in the correlation are uniquely related to the stress in the solder joint. The plastic strain range is an example of such a parameter. For a given plastic strain range, the stress in the solder joint may vary, however, depending on the displacement rate, such that the correlation between the plastic strain range and the number of cycles to failure is not unique.

Fatigue reliability investigation for marine structures using a response surface method

Abstract: A method is presented to assess the fatigue reliability of marine structures. Hydrodynamic loads due to motions in waves are calculated using a 3D-diffraction theory. Stresses due to dead and service loads and stress spectra due to loads are calculated at locations with extreme stress concentration (hot-spots) using the Finite Element method. Material deterioration is considered due to crack propagation. Load and resistance parameters are assumed to be uncertain. Modern reliability methods are applied to calculate efficiently failure probabilities. The Finite Element analysis and the reliability calculation is linked through a response surface program using Hermite polynomials. The method is illustrated with a fatigue life estimation of the hatch corner of a third generation containership.

Estimating the importance of cyclic thermal loads in thermo-mechanical fatigue

Abstract: A method for evaluating the effect of cyclic thermal loading on crack tip stress fields is developed. In its development, advantage is taken of the periodic nature of fatigue loading and only harmonic loadings are evaluated. Formulating the problem in this way permits the extraction of time as an explicit variable and replaces its role with a dependence on the frequency of the thermal loading. The means for evaluating the effect of periodic loadings on crack tip stress fields is the stress intensity factor which is calculated from numerically defined stress and displacement fields using a path independent integral. Results obtained indicate that stress intensity factors of cracked components exposed to thermal fatigue conditions have a significant dependence on the frequency of the thermal cycle and the crack geometry. Numerical estimates for mode I thermal stress intensity have been obtained using thermal fatigue test data for a titanium alloy and can be as high as 25 percent of the critical mode I mechanical stress intensity.

Statistical Evaluation of the Distribution of Crack Propagation Fatigue Life by Simulating the Crack Growth Process

Abstract: It is widely recognized that the fatigue crack propagation is fundamentally a random process which can be predicted only in terms of probability. The primary source of statistical variation of fatigue crack propagation is material inhomogeneity. To explain its effects, the authors proposed in the previous paper a new stochastic model which treats the material’s resistance against fatigue crack growth as a spatial stochastic process along the path of the crack. This paper investigates the influence of the parameter variation on the results using Monte-Carlo simulations for the proposed model in the case of the constant load amplitude. It is shown that the statistical properties of random crack propagation resistance has great influence on the distribution of the crack propagation fatigue life. The results are also compared with well-known experimental data sets and satisfactory agreements are obtained.

Application of Closure Concept to Estimation of Lifetime of Fatigue Crack Propagating Under Random Load

Abstract: A stochastic approach describing fatigue crack growth under random loads and incorporating the closure concept for assessment of load interaction effects is presented. Two extreme loading histories are considered: (a) the superposition of overloads, random in time and magnitude, on base-line constant—amplitude cyclic loading, and (b) Random sequence of load peaks. In the first case the model is based on presentation of the delay time due to retardation effects associated with the overloads as a purely discontinuous Markov process. A numerical procedure developed is based on application of Kolmogarov—Feller integrodifferential equation. An expression for failure probability is derived. In the second case the load peaks are considered as a sequence of equally distributed noncorrelated random variables The crack opening stress, determined for an arbitrary i t h cycle, is estimated on the basis of the highest maximum and the lowest trough observed in η preceding cycles. It is shown that in this case some characteristic integral of crack length is normally distributed. Numerical procedure developed accounts for stochastic nature of both the load and the material. Examples illustrate the application of the proposed procedures.

Detection of Fatigue Damage in a Steel Member

Abstract: In this paper the possibilities of detection of crack extension in a steel beam by observation of changes in the dynamical response are investigated. Sytem changes are observed by frequency domain and time domain techniques. The position and the size of the crack by finite element calculations. The estimated values are compared to the real values observed in the experiment.

Probabilistic design of a ship type floating production vessel

Abstract: Probabilistic analyses of a ship type floating production vessel is performed with respect to the ultimate limit state and the fatigue limit state. Variability in wave induced and still water bending moments are included in the analysis. Load and material coefficients have been derived for various target reliabilities for the ultimate limit state. The fatigue design has been based on S-N data and fracture mechanics analysis. Requirements to in-service inspection have been evaluated. Based on reliability analysis the optimal fatigue design life factor has been calculated based on minimum life-time costs due to design, construction, inspection and possible repair and shut-down. The effect of material yield strength on fatigue reliability has been quantified.

Experimental verification of random load fatigue endurance

Abstract: — This paper is concerned with various problems related to laboratory experimental verification of fatigue endurance under random loading. Based on statistical characteristics of stationary and non-stationary operating random processes (probability density function and/or power spectral density, transition probability densities) it is shown how to design real time simulation algorithms that are further used to create inputs of computer controlled electrohydraulic loading systems. The corresponding fatigue tests (strain or stress constrolled) produce endurance curves that can be plotted in various coordinates (peak or RMS values vs fictitious number of cycles or time to failure) offering non-identical conclusions and qualitative interpretations. Some examples of the influence of simulated random process probability density function and/or power spectral density on the resulting fatigue endurance are presented and also certain suggestions concerning prospects for random load fatigue tests are summarized.

A finite element random response analysis of a complex panel with fluid-conveying pipes

Abstract: A finite element approach is developed for a complex panel with fluid-conveying pipes undergoing large deflections subjected to random loadings. The influence of fluid velocity on the random response is investigated. The root-mean-square (rms) deflections and frequencies for different sound spectrum level value are studied. All four edges of the panel are considered to have the same conditions and restrictions. Either simply supported of clamped boundary condition with respect to the transverse deflection of the panel is considered. The prediction of fatigue life is then based on obtained rms stress. This analytical investigation will help to broaden the basic understanding of the role of fluid flow within structures subjected to acoustic loading.