Journal of Engineering Mechanics

A novel method of Bayesian learning with automatic relevance determination prior is presented that provides a powerful approach to problems of classification based on data features, for example, classifying soil liquefaction potential based on soil and seismic shaking parameters, automatically classifying the damage states of a structure after severe loading based on features of its dynamic response, and real-time classification of earthquakes based on seismic signals. After introduction of the theory, the method is illustrated by applying it to an earthquake record dataset from nine earthquakes to build an efficient real-time algorithm for near-source versus far-source classification of incoming seismic ground motion signals. This classification is needed in the development of early warning systems for large earthquakes. It is shown that the proposed methodology is promising since it provides a classifier with higher correct classification rates and better generalization performance than a previous Bayesian learning method with a fixed prior distribution that was applied to the same classification problem.

The Venturi effect refers to the increase in fluid speed due to a decrease of the flow section in confined flows. The wind speed conditions in converging and diverging passages between perpendicular buildings are studied with computational fluid dynamics to investigate the extent to which the so-called Venturi effect is present in the passages. Model validation is performed by comparing the numerical results with wind tunnel measurements. The validated model is employed for a detailed investigation of the wind speed and the flow rate in the passages for a wide range of passage widths. The simulations show an increase in wind speed near ground level, but a decrease of horizontal wind speed in the upper part of the converging passages. The reason is the wind-blocking effect, which causes a large part of the oncoming wind to flow over and around the buildings, rather than being forced through the passage. Due to this effect, the flow rates through the converging passages are consistently lower than the free-field flow rate, implying that the term Venturi effect is less applicable for such building configurations.

Uncertainties need to be taken into account for credible predictions of the dynamic response of complex structural systems in the high and medium frequency ranges of vibration. Such uncertainties should include uncertainties in the system parameters and those arising due to the modeling of a complex system. For most practical systems, the detailed and complete information regarding these two types of uncertainties is not available. In this paper, the Wishart random matrix model is proposed to quantify the total uncertainty in the mass, stiffness, and damping matrices when such detailed information regarding uncertainty is unavailable. Using two approaches, namely, (a) the maximum entropy approach; and (b) a matrix factorization approach, it is shown that the Wishart random matrix model is the simplest possible random matrix model for uncertainty quantification in discrete linear dynamical systems. Four possible approaches for identifying the parameters of the Wishart distribution are proposed and compared. It is shown that out of the four parameter choices, the best approach is when the mean of the inverse of the random matrices is same as the inverse of the mean of the corresponding matrix. A simple simulation algorithm is developed to implement the Wishart random matrix model in conjunction with the conventional finite-element method. The method is applied vibration of a cantilever plate with two different types of uncertainties across the frequency range. Statistics of dynamic responses obtained using the suggested Wishart random matrix model agree well with the results obtained from the direct Monte Carlo simulation.

A micromechanics-based model is proposed to determine the nonlinear stress-strain relations of cement-matrix composites at different concentrations of inclusions (aggregates). We first conducted some experiments to uncover the stress-strain behavior of the cement paste with a water-to-cement ratio of 0.45, and those of the mortar with the same cement paste but at three different volume concentrations of aggregates. The behavior of the cement paste is then simulated by Burgers’ rheological model. In the development of the composite model, we extend the linear elastic response to the nonlinear one through the replacement of elastic moduli by the corresponding secant moduli. The nonlinear stress-strain curves of the cement-matrix composite are then determined from those of the cement paste and inclusions. It is shown that the predicted stress-strain curves of the mortar are in close agreement with the experimental curves up to an aggregate volume fraction of 49% or $60\mathrm{wt}\%$ .

The main objective of this paper is to develop three-dimensional (3D) microstructure-based discrete element models of asphalt mixtures to study the dynamic modulus from the stress-strain response under compressive loads. The 3D microstructure of the asphalt mixture was obtained from a number of two-dimensional (2D) images. In the 2D discrete element model, the aggregate and mastic were simulated with the captured aggregate and mastic images. The 3D models were reconstructed with a number of 2D models. This stress-strain response of the 3D model was computed under the loading cycles. The stress-strain response was used to predict the asphalt mixture’s stiffness (modulus) by using the aggregate and mastic stiffness. The moduli of the 3D models were compared with the experimental measurements. It was found that the 3D discrete element models were able to predict the mixture moduli across a range of temperatures and loading frequencies. The 3D model prediction was found to be better than that of the 2D model. In addition, the effects of different air void percentages and aggregate moduli to the mixture moduli were investigated and discussed.

Top tensioned strings and beams are often used in civil and marine applications. Typically these members have constant cross sections, and a pronounced, usually linear, tension variation, due to the effects of gravity. In this paper simple, approximate formulas for the natural frequency of such strings are derived, based on asymptotic techniques, while for the tensioned beam case approximate closed-form results are developed by the Wentzel–Kramers–Brillouin method. Both derivations are shown in reasonable detail. While similar work is known for a beam with varying axial tension this is believed to be the first time that a single analytic expression is developed for the full length of the beam. A simple example in which the bottom tension is only 9% of the top tension is analyzed for cases with and without bending stiffness, and the solutions have been compared to the exact solution for the string case and to the results from three finite-element programs for the beam case. The accuracy was found to be very good, even in this situation, in which the tension variation is large.

In this work, stability and accuracy of the Newmark method for nonlinear systems are obtained from a linearized analysis. This analysis reveals that an unconditionally stable integration method for linear elastic systems is unconditionally stable for nonlinear systems and a conditionally stable integration method for linear elastic systems remains conditionally stable for nonlinear systems except that its upper stability limit might vary with the step degree of nonlinearity and step degree of convergence. A sufficient condition to have a stable computation for nonlinear systems in a whole step-by-step integration procedure is also developed in this study. Furthermore, it is also found that numerical accuracy in the solution of nonlinear systems is closely related to the step degree of nonlinearity and step degree of convergence although its characteristics are similar to those of the preceding works for linear elastic systems. Since these results are obtained from a linearized analysis, they can be applicable to the nonlinear systems that satisfied the simplifications for the analysis but may not be applicable to general nonlinear systems.

This paper examines the use of data obtained from large eddy simulations in exploring the instantaneous characteristics of boundary shear stress. The probability density functions (PDF) of the shear stress are shown to be non-Gaussian, whereas the PDF of the shear force per unit length is found to be approximately Gaussian. Conditional sampling is used in order to provide an insight into the average behavior of an extreme shear stress event. Finally, proper orthogonal decomposition illustrates that the behavior of the shear stress is complex and cannot necessarily be attributed to a single major flow event.

We probe the origins of nonaffine deformation in discrete element simulations of densely packed granular systems under biaxial compression, using a new local measure of nonaffine micropolar deformation. This measure represents the deviation of particle motion from that dictated by a measure of micropolar strain and curvature, devised on the scale of a particle and its first ring of neighbors. Highly correlated mesoscopic structures emerge and contribute significantly to nonaffine deformation, an aspect of material behavior that is largely neglected in existing constitutive models. Distinct regimes of deformation were observed: a globally affine regime in which particles move in uncorrelated Brownian motion with some “rattlers” present followed by a globally nonaffine regime involving rattlers, microbands, vortices, confined buckling of force chains, and persistent shear band(s). Structural development leading to and during persistent shear banding is elucidated. The highest levels of nonaffine deformation were observed during drops in the macroscopic stress ratio: here, nonaffine deformation is essentially confined to the shear band, where confined buckling of force chains is the intrinsic mechanism. Degrees of correlation between the local measures of nonaffine strain and curvature confirm the key role of particle rotations in shear bands. Probability density functions of local nonaffine deformation exhibit power law behavior cut off by an exponential tail, consistent with experiments. Self-diffusion anisotropy was observed inside the shear band. The degree of diffusion is greater tangential rather than normal to the shear band; particles in buckling force chains exhibit the highest degree of diffusion along the band. Self-diffusion of particles after the peak stress ratio was found to be superdiffusive during each drop in stress ratio, and diffusive during each rise in stress ratio. The implications of these results for constitutive modeling are discussed, with particular attention paid to the significance of confined buckling of force chains.