Narrow Scatter Research Articles

Machine learning for fatigue lifetime predictions in 3D-printed polylactic acid biomaterials based on interpretable extreme gradient boosting model

Modeling of the fatigue lifetimes in 3D-printed biomaterials of Polylactic acid (PLA) is presented in this article based on machine learning (ML) techniques of interpretable extreme gradient boosting (XGBoost) and Shapley additive explanations. For this objective, standard testing samples were additive-manufactured from PLA under different 3D printing parameters. Then, the fatigue experiments were performed on specimens under various stress levels. Based on these data, three ML methods were utilized for modeling the PLA fatigue lifetimes, including XGBoost, random forest, and support vector regression, besides a common nonlinear regression analysis. The obtained results indicated that XGBoost had superior modeling results, compared to other ML techniques and the regression analysis. The coefficient of determination was 97.66 % with a scatter-band of ±1.3, which was a narrow scatter in fatigue modeling.

The structural strain method for fatigue evaluation of welded components: closed-form solutions

A consistent and unified analysis framework is proposed in this paper for extending the early traction-based structural stress method to low-cycle fatigue regimes in fatigue evaluation of welded structures through a novel structural strain method. Instead of local strain or notch strain, the structural strain here refers to a through-thickness strain distribution that satisfies plane-remains-plan conditions under elastic–plastic deformation conditions. This paper provides closed-form solutions to structural strain distributions corresponding to loading and unloading, structural strain range, an equivalent structural strain range, and the structural stress–strain relation by considering elastic-perfectly plastic material behavior. Both finite element method (FEA) simulations and in-house modeling tools confirm the accuracy of these closed-form analytical solutions. Although these solutions are derived based on elastic-perfectly plastic material assumption, the resulting structural strain range, as a unique fatigue damage parameter for welded components, is shown essentially the same as those obtained numerically by considering a modified Ramberg-Osgood material model under pulsating loading condition, i.e., stress ratio equals zero. Good data correlation between the equivalent structural strain range predicted results and the test results obtained for a welded steel structure under the same loading condition is achieved. Finally, the structural strain range-based parameter is proven effective in correlating a large number of both low-cycle and high-cycle fatigue test data of welded structures made of several structural materials into a narrow scatter band demonstrating the universality of the structural strain method.

Experimental investigation on stress state and deformation in bolts through tightening of perfect flange connections

AbstractThe fatigue loads of bolts in ring flange connections are significantly influenced by the pressure body between the compressed flanges, which in turn depends on two main scattering parameters ‐ geometric flange imperfections and the preload level. Depending on the tightening method and geometric boundary conditions, shear and bending stresses occur in addition to the axial stress in the bolt during prestressing of the connection. This multiaxial stress condition leads to different plastic deformations in the threads and thereby influence the performance respective the maximum achievable preload of tightening methods. In accordance with the objective, the authors identify the stress state in HV‐bolts during installation for different tightening procedures through experimental testing. Based on a comparison of the strain measurements with the 3D scan of the bolt geometry, this study evaluate to what extent the initial imperfections corresponds to bolt bending. Derived from the scope of a joint national research project, this contribution point out that the yield controlled tightening ensure prestresses close to the individual yield strength of the bolt within a narrow scatter.

A Numerically Efficient Method to Assess the Elastic–Plastic Strain Energy Density of Notched and Imperfective Cast Steel Components

The fatigue strength of cast steel components is severely affected by manufacturing process-based bulk and surface imperfections. As these defect structures possess an arbitrary spatial shape, the utilization of local assessment methods is encouraged to design for service strength. This work applies the elastic–plastic strain energy density concept to study the fatigue strength properties of a high-strength cast steel alloy G12MnMo7-4+QT. A fatigue design limit curve is derived based on non-linear finite element analyses which merges experimental high-cycle fatigue results of unnotched and notched small-scale specimens tested at three different stress ratios into a unique narrow scatter band characterized by a scatter index of 1:TΔW¯(t)=2.43. A comparison to the linear–elastic assessment conducted in a preceding study reveals a significant improvement in prediction accuracy which is assigned to the consideration of the elastic–plastic material behaviour. In order to reduce computational effort, a novel approximation is presented which facilitates the calculation of the elastic–plastic strain energy density based on linear–elastic finite element results and Neuber’s concept. Validation of the assessment framework reveals a satisfying agreement to non-linear simulation results, showing an average root mean square deviation of only approximately eight percent in terms of total strain energy density. In order to study the effect of bulk and surface imperfections on the fatigue strength of cast steel components, defect-afflicted large-scale specimens are assessed by the presented elastic–plastic framework, yielding fatigue strength results which merge into the scatter band of the derived design limit curve. As the conducted fatigue assessment is based solely on linear–elastic two-dimensional simulations, the computational effort is substantially decreased. Within the present study, a reduction of approximately 400 times in computation time is observed. Hence, the established assessment framework presents an engineering-feasible method to evaluate the fatigue life of imperfective cast steel components based on rapid total strain energy density calculations.

Prediction of surface crack growth life for AA7075-T6 under nonproportional loading

The growth direction and growth rate of fatigue cracks initiated from the surface of an aluminum alloy 7075-T6 plate were investigated under six sets of nonproportional loadings produced by combined axial-torsion loading. A digital image correlation method was simultaneously applied to measure the crack opening displacements around the crack tip during the fatigue tests. The effective stress intensity factor ranges were determined on the basis of the measured crack opening displacements and applied to arrange the crack growth rate. Results showed that fatigue crack tended to propagate orthogonally to the direction in which the resultant normal stress in the nonproportional loading cycle was maximized, suggesting that the maximum resultant normal stress dominates crack growth under nonproportional loadings. When the effective stress intensity factor range based on the maximum resultant normal stress was used to arrange the crack growth rates, all data collapsed into a single narrow scatter band. The growth lifetime until the surface crack penetrates the plate could be estimated within a relative error of about ± 20 % based on a modified Paris-type equation through numerical integration.

Very-high-cycle fatigue responses of stainless steel for compressor valve plates

Very High Cycle Fatigue (VHCF) responses of stainless steel used for air conditioning compressor valve plates under different loading conditions were investigated. Firstly, the material mechanical responses considering the specimen extracted directions were conducted and compared for predicting the fatigue life. The stress distribution in the simple specimen was examined for these specimens. Furthermore, the fatigue experiments were conducted to explore the variations in calculation methods of materials for the compressor. According to the calculated results under the static and dynamic loading conditions, the results demonstrates that the fatigue life results by the static stress approach is superior to that of the dynamic stress approach with a narrow scatter band. The computed strategy for the high cycle fatigue data could further estimate the fatigue life of structures of air conditioning compressor valve plates.

The Master S-N curve approach for fatigue assessment of welded bridge structural details

The verification of the Master S-N curve approach using an equivalent structural stress parameter, has been carried out for welded joints with different geometries, thickness and loading modes, mainly representative of pressure vessels and piping components. The application of the Master S-N curve to a new field requires the reassessment of the relevant fatigue data, due to different manufacturing conditions, detailing techniques across industries and typical loading modes. In this paper, the referred verification will be further extended by means of available fatigue strength data from large-scale tests of typical bridge structural details, mainly based on girder-type specimens. The reassessment can allow the consideration of loading modes and stress conditions closer to those which develop in actual girder steel bridges. An approach based on experimental S-N data is also useful because it can take into account residual stresses, construction misalignments, local notch and weld bead geometry effects in the estimated fatigue lives. In this manner, the aim is to reassess the robustness and effectiveness of the equivalent-structural stress technique combined with the Master S-N curve approach through the finite element numerical modelling of large-scale tests of welded bridge details subjected to in-plane and out-of-plane loading. The approach is applied to four fatigue test setup programs representative of welded bridge details, which were used as the basis for the establishment of the current nominal S-N curves available in American codes. A performance comparison with the nominal and the hot-spot stress method is carried out. The method proved to collapse nearly 300 datapoints of fatigue failures of welded bridge details into a narrow scatter band.

Distribution of Si ii λ6355 velocities of Type Ia supernovae and implications for asymmetric explosions

ABSTRACT The ejecta velocity is a very important parameter in studying the structure and properties of Type Ia supernovae (SNe Ia) and is a candidate key parameter in improving the utility of SNe Ia for cosmological distance determinations. Here, we study the velocity distribution of a sample of 311 SNe Ia from the kaepora data base. The velocities are derived from the Si ii λ6355 absorption line in optical spectra measured at (or extrapolated to) the time of peak brightness. We statistically show that the observed velocity has a bimodal Gaussian distribution (population ratio 201:110 or 65 per cent:35 per cent) consisting of two groups of SNe Ia: Group I with a lower but narrower scatter ($11\, 000 \pm 700\, \mathrm{km\, s}^{-1}$), and Group II with a higher but broader scatter ($12\, 300 \pm 1800\, \mathrm{km\, s}^{-1}$). The true origin of the two components is unknown. Naturally, there could exist two intrinsic velocity distributions observed. However, we try to use asymmetric geometric models through statistical simulations to reproduce the observed distribution assuming that all SNe Ia share the same intrinsic distribution. In the two cases we consider, 35 per cent of SNe Ia are considered to be asymmetric in Case 1, and all SNe Ia are asymmetric in Case 2. Simulations for both cases can reproduce the observed velocity distribution but require a significantly large portion ($\gt 35{{\ \rm per\ cent}}$) of SNe Ia to be asymmetric. In addition, the Case 1 result is consistent with recent SNe Ia polarization observations that higher Si ii λ6355 velocities tend to be more polarized.

Modified strip model for punching-shear strength of FRP-reinforced concrete edge–column slab connections

The use of fiber-reinforced polymer (FRP) bars as internal reinforcement for concrete structures in aggressive environments has emerged as an innovative solution to eliminate corrosion problems. Glass FRP (GFRP) reinforcing bars have recently gained wide acceptance as a viable construction material for sustainable new constructions due to their costing less than other FRP types. This paper reports the test results for full-sized edge–column slab connections (ECS –connections) reinforced with sand coated GFRP bars. These connections were cast using normal-strength concrete (NSC) and high strength concrete (HSC). The effect of concrete strength (NSC and HSC) and moment-to-shear force ratio (M/V) (300 mm, 600 mm) on the performance of the tested connections were examined. In addition, a new mechanical model, called Modified Strip Model, is presented for punching-shear strength predictions of FRP reinforced concrete (RC) ECS –connections. The test results reveal that the use of HSC directly augmented the punching-shear strength and improved the load–deflection response of the tested connections, However, Increasing the M/V ratio produced significant shear stresses, thus decreasing the punching-shear strength for the NSC and HSC connections, respectively. The mechanical model presented herein takes into account all possible mechanisms of loads transfer in FRP-RC ECS-connections. Hence, a complete shear–moment (V-M) interaction diagram can be plotted to predict the punching strength of such connections. Moreover, the model revealed higher accuracy and narrower scatter for punching-shear strength predictions compared to other available equations in FRP design codes and guides. The EN 1992–1-1–05 punching-shear equation was also examined for FRP-RC ECS–connections and adapted to consider the tensile properties of the FRP bars instead of steel.

Rib loading effects on weld root fatigue failure modes at rib‐to‐deck welded joint

AbstractThe service life of orthotropic steel decks is dependent on the fatigue resistance of rib‐to‐deck welded joints, which is often tested using two kinds of experimental models in terms of the rib loading condition. Different weld root fatigue failure modes have been observed in the different models, but the role of rib loading remains unclear. This paper aims to clarify the effect of rib loadings on the weld root fatigue failure modes at rib‐to‐deck welded joints. The loadings are decomposed into the deck loadings and rib loadings according to the principle of superposition. Formulae of the weld root notch stress intensity factors and T‐stress under rib loadings are developed by multiparameter regression analysis and subsequently used for the local stress analysis. The fatigue failure modes determined from the local stress field agree well with the experimental results. The results reveal that the weld root failure modes depend on the rib loadings but are independent of the weld geometries. The averaged strain energy density (SED) that can capture both weld geometry and loading condition effects is used to correlate the fatigue test data of different weld root failure modes. The SED is capable of evaluating the fatigue strength of the rib‐to‐deck welded joint failed by different weld root failure modes with a narrow scatter band.

Polyacrylamide hydrogels. I. Network imperfection

A real polymer network is never perfect, but the quantification of network imperfection has been elusive. Here we quantify the network imperfection of a polyacrylamide hydrogel through its mechanical properties. We find that the ultimate properties—strength, extensibility, and work of fracture—have narrow scatter, comparable to the scatter of modulus. Despite the narrow scatter, the work of fracture is about four orders of magnitude lower than that of the perfect network. This reduction in work of fracture is a measure of network imperfection. The toughness of the hydrogel is about two orders of magnitude higher than that of the perfect network. When the effects of other inelastic processes are minimized, this amplification in toughness is likely due to distributed chain scission, and is another measure of network imperfection. We find that the ultimate properties of the hydrogel are insensitive to cuts up to about 1 mm. We trace this cut insensitivity to network imperfection.

Evaluation of the crystallographic fatigue crack growth rate in a single-crystal nickel-base superalloy

Cracks in single-crystal nickel-base superalloys have been observed to switch cracking mode from Mode I to crystallographic cracking. The crack propagation rate is usually higher on the crystallographic planes compared to Mode I, which is important to account for in crack growth life predictions. In this paper, a method to evaluate the crystallographic fatigue crack growth rate, based on a previously developed crystallographic crack driving force parameter, is presented. The crystallographic crack growth rate was determined by evaluating heat tints on the fracture surfaces of the test specimens from the experiments. Complicated crack geometries including multiple crystallographic crack fronts were modelled in a three dimensional finite element context. The data points of the crystallographic fatigue crack growth rate collapse on a narrow scatter band for the crystallographic cracks indicating a correlation with the previously developed crystallographic crack driving force.

Effects of porosity and pore distribution on static strength properties of porous zirconia

AbstractBulk porosity, along with size and spatial distribution of pores, play key roles in strength of porous ceramics, as reported in a study on porous alumina. Hence, a fracture mechanics procedure was proposed to evaluate their strength by presuming that behavior of pore distribution is equivalent to that of crack distribution, and each pore is surrounded by virtual crack. In contrast to alumina, zirconia has distinct spherical‐shaped pores. Moreover, its strength properties vary with stabilizing additives. In this research, strength properties of yttria‐stabilized zirconia ceramics were studied to verify applicability of the procedure proposed for simulating strength of porous ceramics. The effect of pore characteristics on static strength properties was determined experimentally and confirmed by Monte‐Carlo simulations. It was revealed that simulated strength coincided with experimental results within a narrow scatter band, ie, factor of 21/2. Therefore, the proposed procedure was found to be appropriate for estimating strength of porous zirconia.

Assessment of tensile and fatigue behavior of PEEK in a physiologically relevant environment

In the last decades the necessity of implant devices is continuously increasing. The researchers have thus focused their attentions on the development of new biocompatible materials, in particular polymers. Among them, polyetheretherketone (PEEK) has gained wide interest in load-bearing applications such as spinal cages due to its yielding behavior and its superior corrosion resistance. Since such applications are characterized by notches and other stress concentrators weakening the implant resistance, a design tool for assessing their tensile and fatigue behavior in the presence of such discontinuities is highly claimed. To this aim, tensile and fatigue data available in literature of neat and differently notched PEEK samples (circumferentially cracked and U-notched specimens with different notch radii) experimentally tested in a phosphate-buffered saline (PBS) at 37 °C have been analyzed using the strain energy density (SED) approach. The method is shown to provide accurate results regardless of the different notch geometries, both for tensile and fatigue data. Concerning the former, the tensile strength was in fact estimated with an error lower than ±10%, regardless of the strain rate, while for the latter the SED approach was able to summarize the fatigue data with a single narrow scatter band independently from the notch geometry

Creep-fatigue crack growth behavior of G115 steel under different hold time conditions

The creep-fatigue crack growth behavior and fracture mechanisms of G115 steel at 650 °C were studied using compact tension specimens in a constant load control mode. The crack growth length and load-line displacement analyses reveal that G115 steel is a creep-ductile material and the crack growth life and initiation time increase with increasing dwell time. Meanwhile, the parameter ΔK was used to characterize the crack growth behavior and the effect of dwell time on da/dN depends on ΔK. These data were also compared with those getting from P91 steel under 625 °C and P92 steel at 630 °C. When the crack growth rate, da/dtavg, is represented by Ctavg, all the data seem to collapse within a single narrow scatter band. Further, the approximate Nikbin-Smith-Webster (NSW) crack growth model was utilized to predict the creep-fatigue crack propagation behavior of G115 steel. And the microstructures of the cross-sections along the crack growth direction were studied through scanning electron microscopy to reveal the crack growth mechanisms of G115 steel.

Effect of different environmental conditions on surface crack growth in aluminum alloys

Fatigue surface crack growth is studied through experiments and computations for aluminum alloys D16T and B95AT (analogue of 2024 and 7075 aluminum). Subjects for studies are cylindrical hollow specimens with external semi-elliptical surface crack. The variation of fatigue crack growth rate and surface crack paths behavior was studied under cyclic loading for different environmental conditions. Uniaxial tension tests were carried out at low (-60°C), room (+23°C) and high (+250°C) temperature. For the same specimen configuration and the different crack front position as a function of cyclic loading and temperatures conditions the distributions of governing parameter of the elastic-plastic stress fields in the form of In-factor along various crack fronts was determined from numerical calculations. This governing parameter was used as the foundation of the elastic-plastic stress intensity factor (SIF). Both elastic and plastic SIF approach was applied to the fatigue crack growth rate interpretation. It is found that there is a steady relationship between the crack growth rate and the plastic SIF in the form of general curve within a relatively narrow scatter band for all tested specimens at different temperatures.

Analysis of multiaxial low cycle fatigue of notched specimens for type 316L stainless steel under non-proportional loading

The paper analyzes multiaxial low cycle fatigue tests of notched specimens under proportional and non-proportional loading conditions. Strain controlled multiaxial low cycle fatigue tests were carried out using circumferentially notched round-bar specimens of type 316L stainless steel, with different stress concentration factors. The experimental results show that the crack initiation site is shifted from the notch tip. Based on this finding, a new model for life evaluation is proposed by taking into account the strain gradient in the proximity of the notch tip and the effective maximum strain range. The new model allows evaluating the fatigue life in a narrow scatter band.

Fatigue of titanium weldments: S-N testing and analysis for data transferability among different joint types

Fatigue testing of both TIG and MIG welded titanium components were first carried out using fillet welded cruciform specimens. A mesh-insensitive traction structural stress method is then introduced for establishing data transferability from one joint geometry to another and one base plate thickness to another. In addition, available test data from literature are also analyzed using the traction stress parameter proposed. Major findings are:(a) There exist no noticeable differences in S-N curve behaviors between Ti-CP and Ti-6-4 weldments.(b) The proposed traction stress parameter with a thickness correction term can be used to effectively correlate lab specimen test data in the form of a single S-N curve with a narrow scatter band regardless of joint types and base plate thicknesses, which has been demonstrated for both MIG and TIG welded specimens.(c) By using both the S-N curve and traction stress parameter developed, satisfactory fatigue life estimates for a series of full-scale MIG-welded structural component fatigue tests have been achieved.

Multiaxial cyclic behaviour and fatigue modelling of AM30 Mg alloy extrusion

Multiaxial fatigue characteristics of AM30Mg alloy extrusion are studied through fully-reversed strain-controlled cyclic experiments including pure torsional and combined axial-torsional at 0, 45 and 90° phase angle shifts. Under pure torsional cyclic loading, AM30 extrusion is realized to exhibit better fatigue properties than AZ31B and AZ61A extrusions, especially in low-cycle fatigue regime. Under proportional axial-torsional cyclic loading, twinning/de-twinning in axial mode results in asymmetric shear hysteresis loop. The effect of non-proportionality of biaxial loading on various aspects of material response is also examined and observed to be depending on the magnitude of axial strain amplitude. Finally, the life prediction capabilities of two critical plane models, i.e., modified Smith-Watson-Topper (SWT) and Fatemi-Socie (FS), as well as Jahed-Varvani (JV) energy-based approach are assessed, employing fatigue life data of AM30 extrusion. Correlation data between experimental and estimated lives are found to lie within narrow scatter band.

Diaphragm shear strength and stiffness of aluminum roof panel assemblies

A diaphragm is an assembly of planar structural elements interconnected to each other to provide in-plane bracing system and transmit in-plane forces. The diaphragm geometry, supporting frame, type of the panel profiles, the way the panels are attached to each other and to the diaphragm frame govern the overall behavior of diaphragm systems. Full scale testing is the best way to understand the overall behavior and determine the strength and stiffness of diaphragm systems.This paper is a result of experimental and analytical studies on aluminum roof panel assemblies. The main purpose is to check the applicability of Metal Construction Association’s (MCA) “Primer on Diaphragm Design” for the strength and stiffness equations when applied over a range of panel depths, thicknesses and profiles. Five full scale cantilever tests were conducted at Virginia Tech, Blacksburg, Virginia. The tests were conducted in accordance with the AISI “Cantilever Test Method for Cold-Formed Steel Diaphragms”.While the diaphragm shear stiffness development is similar for both MCA and Steel Deck Institute’s (SDI) “Diaphragm Design Manual”, the only difference is the panel edge term (K). The results of this study show remarkably narrow scatter in tested-to-calculated strength ratios. The tested-to-calculated stiffness ratios compare very well, supporting the proposed use of panel edge term (K=2/3) for aluminum diaphragms. The results indicate that the panel thickness limit can be raised safely to 1.27mm (0.050in.) for MCA procedures. The test data further indicate that the MCA strength and stiffness formulations work well for panels with depths through 101.6mm (4in.).