Journal of Engineering Mechanics

This study seeks to bridge the gap between nonlinear system identification and nonlinear dynamic finite-element analysis. Motivated by the needs in earthquake simulation, it is first investigated under which conditions and to what degree the prediction of maximum lateral drift and base shear requires accurate nonlinear hysteretic moment-rotation joint models. A series of simulations is carried out using a simple but typical steel frame under two different earthquake ground motion time histories scaled up to various levels. As one of the two major classes of models in nonlinear system identification, nonparametric models are proposed to be implemented into OpenSees. A methodology with details is provided to effectively implement feedforward neural networks with one hidden layer as a new one-dimensional nonlinear smooth material model directly from a MATLAB environment to OpenSees. The same methodology can be applied to benefit the implementation of other parametric and nonparametric models with linear parameterization. Numerical examples are provided. Challenges are discussed and future work is identified.

An implementation for the pseudodynamic testing of a one-story frame subject to bidirectional loading was proposed and its feasibility was experimentally verified. Three generalized degrees of freedom at the center of mass of the roof diaphragm are often used in formulating the equations of motion for the one-story frame in a nonplanar test. In this work, three loading points of actuators oriented in the original directions of these actuators were selected to formulate equations of motion. Since loading-point degrees of freedom instead of generalized degrees of freedom were used for a pseudodynamic test there is no need to construct the transformation between these two coordinate systems. Thus, there is no error propagation regarding these transformations. External displacement transducers were applied to measure the displacement components along the original directions of the actuators although the directions of these actuators may vary after the specimen deforms. This displacement measurement is easy to implement and saves a large space in the test setup. A series of pseudodynamic tests were performed and the responses were compared to shaking table tests. These test results attested to that a bidirectional pseudodynamic test can be conducted using the loading-point degrees of freedom and this technique is applicable to the three-dimensional testing of structures under multiple components of earthquake ground motion.

The focus of this paper is on the development of an analytical damage model for predicting the deterioration of the mechanical properties of polyester (PET) ropes subjected to static tension loading. Experimental data on small PET ropes are used to estimate the evolution of damage using the effective stress concept and the principle of strain equivalence. The proposed damage model relies on an empirically based cumulative scalar damage function, which is founded on the assumption that the strain range experienced by rope elements is the main source of damage under static tension loading conditions. In this particular study, the evolution of the damage function is represented by both power law and polynomial forms. Based on experimental observations, softening behavior is developed by rope elements after reaching their maximum load-carrying capacities. This softening behavior is captured by the damage function through an asymptotic expansion technique (perturbation method). Comparisons between predicted rope responses and experimental data are provided to illustrate the use of the proposed damage model to estimate PET rope response.

This note deals with three aspects of nonproportional damping in linear damped vibrating systems in which the stiffness and damping matrices are not restricted to being symmetric and positive definite. First, we give results on approximating a general damping matrix by one that commutes with the stiffness matrix when the stiffness matrix is a general diagonalizable matrix, and the damping and stiffness matrices do not commute. The criterion we use for carrying out this approximation is closeness in Euclidean norm between the actual damping matrix and its approximant. When the eigenvalues of the stiffness matrix are all distinct, the best approximant provides justification for the usual practice in structural analysis of disregarding the off-diagonal terms in the transformed damping matrix. However, when the eigenvalues of the stiffness matrix are not distinct, the best approximant to a general damping matrix turns out to be related to a block diagonal matrix, and the aforementioned approximation cannot be justified on the basis of the criterion used here. In this case, even when the damping and stiffness matrices commute, decoupling of the modes is not guaranteed. We show that for general matrices, even for symmetric ones, the response of the approximate system and the actual system can be widely different, in fact qualitatively so. Examples illustrating our results are provided. Second, we present some results related to the difficulty in handling general, nonproportionally damped systems, in which the damping matrix may be indefinite, by considering a simple example of a two degrees-of-freedom system. Last, we use this example to point out the nonintuitive response behavior of general nonproportionally damped systems when the damping matrix is indefinite. Our results point to the need for great caution in approximating nonproportionally damped systems by damping matrices that commute with the stiffness matrix, especially when considering general damping matrices. Such approximations could lead to qualitatively differing responses between the actual system and its proportionally damped approximation.

This paper presents a model for shear response of asphaltic concrete, taking into account of strain-rate and temperature effects. The model employs rate-dependent hyperplasticity theory, which is based on a thermomechanical framework. A principle of the theory is that the entire constitutive behavior can be defined by two scalar potentials: an energy potential and a flow potential. The viscous behavior of the model corresponds to the results of rate process theory and defines the strain-rate and time dependent behavior. The initial modulus and shear strength are each assumed to be exponential functions of the inverse of temperature. The model is verified and calibrated against the unconfined compression test data for asphaltic concrete at different strain rates and temperatures. A viscoelastic damage model is also addressed to make a comparison with the model developed here. Comparison between the test data, the predictions of the new model, and the predictions of the viscoelastic damage model are discussed.

The objective of this study is to characterize the creep behavior of hot-mix asphalt (HMA) at intermediate $\left(20°\text{C}\right)$ and at high temperatures $\left(40°\text{C}\right)$ . To accomplish this objective, a nonlinear time-hardening creep model, characterized through laboratory testing, was incorporated into a three-dimensional (3D) finite-element (FE) model, which was used to calculate permanent creep strains after applying repetitive vehicular loading cycles at the pavement surface. Two different tire configurations were simulated representing a typical dual-tire assembly and a newly introduced wide-base tire (dual-tire: 275/80R22.5 and wide-base tire: 455/55R22.5). Results of the 3D FE model were successfully verified against pavement response measurements in the field at the Virginia Smart Road. While the elastic or linear viscoelastic FE model may not simulate permanent deformation or shear creep strains after repetitions of vehicular loading, a nonlinear time-hardening creep model could predict primary rutting damage in HMA and shear creep strains at the edge of the tire imprint caused by different tire configurations.

This paper presents results of multiaxial strength tests on various high-performance concretes (HPCs) with uniaxial compressive strength from 58 to $94\mathrm{MPa}$ . A large number of stress ratios of biaxial compression-tension, compression-compression, and triaxial compression tests were investigated. Experiments were performed using a servohydraulic jack with variable stress ratios, whereas steel brushes were used to transfer the load to the specimen. With increasing uniaxial strength the obtained normalized multiaxial strength decreases. The results, along with mathematical approximations, are discussed for the description of failure. The formulated failure criteria for HPC can be used, e.g., in finite-element simulations. The increase in strength by acting a multiaxial compressive stress state or the decrease in strength by acting a compressive-tensile stress state related to the uniaxial compressive strength can be considered and enables a better material exploitation.

The porochemoelastic analytical models and solutions have been used to describe the response of chemically active saturated porous media such as clays, shales, and biological tissues. To date, all existing solutions are only applicable to single-porosity and single-permeability model, which could fall short when the porous material exhibits multiporosity and/or multipermeability characteristics, such as secondary porosity or fractures. This work summarizes the general linear dual-porosity and dual-permeability porochemoelastic formulation and presents the solution of an inclined wellbore drilled in a fluid-saturated chemically active fractured formation, such as fractured shale, subjected to a three-dimensional in situ state of stress. The analytical solution to this geometry incorporates coupled matrix-fracture deformation, simultaneous fluid flows, solute transports and interporosity exchanges induced by the combined influences of stress variation, fluid pressure and solute chemical salinity gradients under isothermal conditions. The fracture system is modeled as a secondary porosity porous continuum following Biot’s formulation while using mixture theory and the pore fluid is a binary solution comprised of a solvent and a solute. Results for the transient stresses and dual pore pressure distributions due to the coupled fracture and hydrochemical effects are plotted in the vicinity of the inclined wellbore and compared with the classical porochemoelastic and poroelastic counterparts. Finally, wellbore stability analyses are carried out to demonstrate applications of the solutions to field drilling operations.

In this study, an analytic solution of wave interaction with a rigid porous medium above a poro-elastic sandy bottom is derived to investigate the attenuation of the surface wave and the wave-induced soil response. In the model, both inertial and damping effects of the flow are considered in the rigid porous region using the potential theory, while the consolidation theory is adopted in the sand region. A new complex dispersion relation involving parameters of the rigid porous and the poro-elastic medium is obtained. The analytic solutions are verified by some special cases, such as wave interaction with a porous structure over an impermeable bottom or wave interaction with a poro-elastic medium only. Numerical results indicate that the wave attenuation is highly dependent upon the thickness of the rigid porous layer, the soil stiffness, and their respective coefficients of permeability. Increasing the thickness of the rigid porous layer will shorten the wavelength of the surface wave regardless of the sand coarseness. The pore pressure in fine-sand is larger than in coarse sand, with both decaying with wave progression. It is also found that increasing the thickness of the rigid porous medium will effectively reduce the pore pressure in the sand. For the applications, an extended hyperbolic mild-slope equation is finally obtained, based on the basic analytic solutions. Examples of the wave height transformation over submerged permeable breakwaters on a slope sandy seabed are given. The simulated results show that the wave decay of the coarse sand seabed is larger than those of fine-sand and impermeable seabeds when waves pass after the submerged porous breakwater. The wave damping versus the friction factor for various height of the submerged breakwater is discussed.

Hysteretic damping is often modeled by means of linear viscoelastic approaches such as “nearly constant attenuation” models (e.g. NCQ model). These models do not take into account nonlinear effects either on the stiffness or on the damping, which are well known features of soil dynamic behavior. The aim of this paper is to propose a mechanical model involving nonlinear viscoelastic behavior for isotropic materials under dynamic excitations. This model simultaneously takes into account nonlinear elasticity and nonlinear damping. On one hand, the shear modulus is a function of the excitation level; on the other, the description of viscosity is based on a generalized Maxwell body involving nonlinearity. This formulation (X-NCQ) is implemented into a one-dimensional finite-element approach for a dry soil. The validation of the model shows its ability to retrieve low amplitude seismic ground motion. For larger excitation levels, the analysis of seismic wave propagation in a nonlinear soil layer over an elastic bedrock leads to results which are physically satisfactory (lower amplitudes, larger time delays, higher frequency content).

Despite many successes in concrete creep studies, its effect on the mechanical behavior of concrete members is far from a thorough case-specific understanding. For the members that have been subjected to a long-term load, the classical stress-strain models describing the short-term behavior of either confined or unconfined concrete are unsuitable. In order to investigate this creep-effect, an experiment on eight concrete cylindrical columns confined by fiber-reinforced polymer (FRP) is carried out. Based on the theory of plasticity for concrete, a constitutive model that takes into account the effect of creep on mechanical behavior of concrete confined by FRP is presented. In the model, the creep law inspired in the microprestress-solidification theory is generalized to triaxial stress condition for the calculation of the creep of the concrete columns confined by FRP. The predictions of the model agree well with the experimental results. The present study indicates that the creep increases the elastic modulus, slightly decreases the compressive strength, and degrades the deformation capability of the concrete confined by FRP.

This paper first examined the effect of wood strand orientation on the flexural properties of oriented wood strand composites (OSB) under different engineering stress modes, and second, investigated the applicability of the rule of mixtures in conjunction with the theory of elasticity for predicting the properties of OSB. The results showed that the performance of OSB was closely related to the loading directions and stress modes: The flexural strength and stiffness under both flat and edgewise bending loads consistently decreased with increasing angles between the applied load and longitudinal direction of orientation of strands, but that under flat bending being much more significant. Panel shear at 45° loading angle resulted in higher strength compared to other loading angles tested, indicating an occurrence of diagonal shear stresses. In conjunction with the numerical results from image analysis of the structure of OSB, and the oriented elasticity and stress algorithms, the models for theoretically predicting the strength and stiffness of OSB under various loading angles were derived with a good estimate under bending and panel shear loads, implying that OSB can be treated as a composite and its properties may be modeled by the rule of mixtures by suitable development, even though OSB has a very high volume fraction of wood strands (0.979) and correspondingly very low volume fraction of resin (0.021).

This paper deals with the study of dynamic or parametric instability behavior of laminated composite stiffened plates with step-uniform and concentrated in-plane harmonic edge loading. The eight-noded isoparametric degenerated shell element and a compatible three-noded curved beam element are used to model the plate and the stiffeners, respectively. The method of Hill’s infinite determinant is applied to analyze the dynamic instability regions. Numerical results are presented through convergence and comparison with the published results from the literature. The effects of parameters like loading type, stiffening scheme, lamination scheme, dynamic load factor, and boundary conditions are considered in the dynamic instability analysis of laminated composite stiffened plate. It has been shown that the type of loading and the width of loading have remarkable effect on the dynamic instability characteristics of the stiffened plate.