Journal of Engineering Mechanics

Seismic isolation is a widely adopted technology for the protection of building infrastructure from earthquakes. As one of the most effective seismic protection devices, curved surface slider (CSS) systems have been studied extensively in the last few decades and are nowadays widely used all over the world. However, due to their relatively recent adoption for the seismic isolation of buildings and bridges, critical aspects of their behavior and possible causes for their malfunctioning are still under thorough experimental and analytical investigation. After recent severe earthquakes, such as Kobe (1995) and Tohoku (2011), considerable attention has been devoted to the effects of near-fault and long-period earthquakes, with specific reference to tall buildings and base-isolated structures. However, no reference has been made in the literature to the resonant effects of nearly periodic ground motions on CSS systems. For the low friction coefficient values used in seismic isolation, the linear theory currently used for the analysis of the response of friction-based systems leads to an unbounded resonant response. It is the focus of this work to investigate the effects of geometric nonlinearity on the response of such systems and to assess the performance of the CSS system under critical ground motions. Dynamic analyses are carried out under harmonic, near-fault, and long-period ground motions. The results are presented in the form of displacement response spectra for different values of the friction coefficient.

Using potential flow theory, oblique wave scattering and trapping by a partial porous breakwater having finite width are studied under the assumption of small-amplitude water wave theory in the presence of seabed undulation. This study is performed for regular incident waves that approach from deeper water depth and impinge on the porous breakwater at lower water depth. The two constant levels of deeper and lower bottoms are linked by a smooth arbitrary bottom topography, which may be regarded as a continental slope. In the lower-depth water, two types of porous breakwaters, namely, bottom-standing and surface-piercing breakwaters, are considered with and without a seawall. The analytical method of eigenfunction expansion is adopted for constant water depth, whereas an approximate method of the modified mild-slope equation is employed for varying water depth. Different bottom configurations are applied in this study. Reflection and transmission coefficients of waves and wave force on seawall are calculated for various physical parameters related to porous breakwater and waves. The role of seabed undulation on wave scattering and trapping is analyzed. The effectiveness of the breakwater is studied from the reflection and transmission coefficients. Results show a major difference in breakwater’s performance in scattering and dissipating waves under the condition of varying water depth compared to the case of constant depth of water. The tranquillity zone is obtained by optimizing various physical parameters. This study provides some theoretical knowledge related to wave interaction with partial porous structures under varying water-depth conditions. Moreover, it finds engineering applications concerned with the creation of tranquility regions and the protection of seawall.

Among the most challenging problems in the field of damage detection and condition assessment in large structures is the ability to reliably detect, locate, and quantify relatively small changes in their dynamic response, based on vibration signal analysis. In this study, a substructuring approach, which uses a nonparametric identification method, was applied to simulated damage data from a high-fidelity and validated three-dimensional (3D) finite element model of a 52-story high-rise office building, located in downtown Los Angeles. Results of this study indicate that the approach not only yields identification results that match well-known global (linear) system identification methods, such as NExT/ERA, but it also provides additional benefits that global identification approaches suffer from. These benefits include: (1) enhanced sensitivity to small structural parameter changes, (2) ability to provide location information about the region in the large structure in which damage has occurred, and (3) not assuming that the underlying structure is linear. Thus, the approach is capable of detecting, quantifying, and classifying changes, when they do occur, if the actual building is subjected to strong earthquake ground motion.

Since the linear elastic layered theory was developed in 1940s, it has been widely used to evaluate asphalt pavement soundness, and the mechanical responses have been used to analyze the bearing capacity of pavement structures and top-down fatigue cracking. However, there is poor convergence of mechanical responses on the surface of the top layer. Therefore, this article proposes an approach to quickly and accurately determine the mechanical response of asphalt pavement surfaces. This approach separates the equations for calculating the mechanical response into two parts, and then the two parts are solved according to the convergence of integrals of surface mechanical responses. The result shows that the surface point calculation method cannot only ensure the accuracy of calculation, but also greatly improve the efficiency of calculation. This approach provides an effective tool to determine mechanical responses.

Water entry is a complex transient process involving free interface, and the physical mechanism of the cavity dynamics and kinematic characteristics under various initial motion states still are not clear. In this study, the unsteady characteristics of water entry of a hydrophobic sphere were investigated numerically and experimentally, focusing on the effects of various entry speeds and angles. A three-dimensional numerical method with six degrees of freedom was developed for the sphere water-entry process using the shear stress transport (SST) turbulence model based on volume of fluid (VOF) and overlapping grid technology, in conjunction with the corresponding experiments to provide the validation data. The results show that the variation in the entry speeds and angles affects not only the closure mode of the cavity but also the shape evolution and typical moments of the cavitation shape and jet, as well as the changes in force and motion characteristics.

Failure accidents of transmission lines due to excessive wind-induced displacements pose a serious threat to the operational stability and security of the electric power grid. A Bayesian approach is proposed to assess the discharge failure probability of overhead transmission lines under typhoon hazards. The Bayesian approach is capable of combining the prior information in the design stage and the full-scale measurement data in the operational stage together to determine the real performance of the tower-line system. By integrating the procedures of the model updating, modal identification, and multiscale wind field simulation in a unified Bayesian framework, the uncertainties of the structural parameters and wind load conditions are reduced and the fragility results are refined according to the updated data and parameters. Based on full-scale measurement data collected during the two strong typhoons Talim (1718) and Kong-rey (1825), the wind-induced fragility of a full-scale transmission line is updated sequentially using the proposed approach. The fragility results show that the failure probability may be underestimated in the design stage, which introduces potential risks in the long-term operation of transmission lines.

Wind effects can be critical for the design of lifelines, such as long-span bridges. The existence of a significant number of aerodynamic force models, used to assess the performance of bridges, poses an important question regarding their comparison and validation. This study utilizes a unified set of metrics for a quantitative comparison of time-histories in bridge aerodynamics with a host of characteristics. Accordingly, nine comparison metrics are included to quantify the discrepancies in local and global signal features, such as phase, time-varying frequency and magnitude content, probability density, nonstationarity, and nonlinearity. Among these, seven metrics available in the literature are introduced after recasting them for time-histories associated with bridge aerodynamics. Two additional metrics are established to overcome the shortcomings of the existing metrics. The performance of the comparison metrics is first assessed using generic signals with prescribed signal features. Subsequently, the metrics are applied to a practical example from bridge aerodynamics to quantify the discrepancies in the aerodynamic forces and response based on numerical and semianalytical aerodynamic models. In this context, it is demonstrated how a discussion based on the set of comparison metrics presented in this study can aid a model evaluation by offering deeper insight. The outcome of the study is intended to provide a framework for quantitative comparison and validation of aerodynamic models based on the underlying physics of fluid-structure interaction. Immediate further applications are expected for the comparison of time-histories that are simulated by data-driven approaches.

The frequency dependence modeling of viscoelastic (VE) materials faces the challenge of high modeling accuracy over a wide frequency range and double-objective optimization in model parameter identification. To address these two issues, a gradient chain structure model and a model parameter identification method are proposed. Inspired by chain structure models, two gradient functions are introduced to update the roles of a chain network and free chains in the mechanical behavior of VE materials. A two-step parameter identification method is proposed to obtain the model parameters and gradient functions. Then a dynamic property test of VE dampers is conducted to study the frequency effect on the equivalent stiffness and loss factor. Also, the test results are used to verify the proposed gradient model. Further, a comparison of the gradient model, a 5-parameter rheological model, the fractional derivative Zener model, and 10 sets of experimental data are conducted to illustrate the feasibility and superiority of the proposed gradient model for VE materials with different kinds of properties. The numerical data from the gradient chain structure model fit well with the experimental data over a wide frequency range, and the proposed model performs much better than the other two typical models in terms of both modeling accuracy and stability.

The two-dimensional (2D) hydraulic analysis of water flow down a slope is crucial. Contrary to past studies, not only the anisotropy of soil but also the vertical component of flow velocity are considered in this study. The flow field is divided into two regions, namely, the water layer and the anisotropic soil layer. By regarding soil as an anisotropic and permeable porous medium, the viscous flow theory and poroelastic theory are employed to govern the surface flow and subsurface flow, respectively. When considering the anisotropy of soil, hydraulic conductivity is supposed to be a second order tensor, and the momentum equations for the porous medium flow in the soil layer should be derived through a new method. Based on the presumption of velocity type, the distributions of horizontal and vertical velocity and pressure could be acquired by an analytical approach by the differential transform method. The results show that both the horizontal and vertical components of flow velocity increase with the anisotropy degree, whereas the pore water pressure decreases with the anisotropy degree. The streamlines in surface water flow depend on higher anisotropy degree. The infiltration rate for sand estimated by the present solution is feasible when compared with previous research result.

This paper focuses on the calculation of the structural vibration–induced hydrodynamic added mass of the rectangular pile cap submerged in water. A total of 150 pile caps with different physical parameters are sampled by Latin hypercube sampling (LHS). The added mass coefficients of all the pile cap samples are computed using a three-dimensional (3D) potential-based numerical method. Based on the database of physical parameters and added mass coefficient results of the pile cap samples, the analytical model with Bayesian updating for the hydrodynamic added mass of the rectangular pile cap is proposed. The proposed model consists of a deterministic model and the additional correction terms accounting for the model inherent bias and model error. The formulation of the model is expressed as a function of a series of physical parameters of the pile cap. The unknown parameters in the proposed model are updated based on the produced database, and the formulation of the correction terms is optimized through a stepwise deletion process. The proposed model is then applied to calculate the hydrodynamic added mass according to the experimental modal testing setup of the isolated pile cap specimens and the group-pile cap foundation specimens submerged in water, respectively. Through the comparison between the modal results from the experiments and the proposed model, it is concluded that the proposed model can efficiently estimate the hydrodynamic added mass of a rectangular pile cap submerged in water with acceptable prediction error.

This paper focuses on an inverse problem involving a commonly used material model in structural steel applications—the Voce-Chaboche model. In particular, the paper proposes an approach focused on obtaining a consistent set of material parameters that accurately represents a wide range of mechanical uniaxial cyclic load histories. The main focus of the methodology presented here is its application to structural steels in the field of earthquake engineering. A set of load protocols representative of seismic loading are proposed for this purpose. It is shown that the calibration procedure optimally leverages Voce-Chaboche’s ability to describe the material response to an arbitrary load history. Parameters for prevailing American, European, and Japanese structural steels are also proposed.

Electromagnetic (EM) energy harvesters have been proposed in civil engineering for simultaneous energy scavenging and structural vibration mitigation. One important component governing the performance of EM energy harvester is the energy harvesting circuit. These circuits consist of electronic components that could introduce nonlinearity to the electromechanically coupled harvester–structure system. The presence of this circuit nonlinearity complicates the performance analysis of the harvester–structure system as well as the design of the energy harvesting circuit. In this study, the performance of a structure equipped with an EM energy harvester connected to a representative energy harvesting circuit under Gaussian white noise force excitation is presented. The examined circuit, termed the standard energy harvesting circuit, consists of a full-wave bridge rectifier and a capacitor connected in parallel with a resistor. A statistical linearization technique is adopted to estimate the system’s stationary response. The accuracy of the technique is validated both numerically and experimentally. The results show that neglecting the circuit nonlinearity due to diodes in the full-wave bridge rectifier can result in overestimating both the damping capability of the harvester and the scavenged output power. When the blockage effect of diodes is prevented effectively, the performance of vibration mitigation and energy harvesting can be enhanced simultaneously. The results of this study indicate that neglecting circuit nonlinearity in the analysis could lead to nonoptimal circuit design and could affect the efficiency of energy harvesting.

A three-node triangular finite element is developed for the static theorem of limit analysis to discretize plane strain and stress problems. The element does not provide rigorous lower-bound solutions because the equilibrium equations and mechanical boundary conditions are satisfied on average, but exhibits a remarkable performance as illustrated by standard benchmark tests. Nowhere does the generated stress field violate the traction continuity across the element interfaces and the yield criterion. The stated nonlinear convex optimization problem is cast as second-order cone programming before its solution by a robust interior-point algorithm implemented in the dedicated solver MOSEK.

Permanent deformation and cracking of asphalt mixtures are common distresses of pavement structures. These two damage processes are usually coupled and occur simultaneously. The evolution speed of each process across the temperature spectrum determines which one is more obvious. To more realistically represent this phenomenon, a new approach that combines the energy-based mechanics and kinetics is proposed to separate and model viscoplastic deformation and cracking of asphalt mixtures in this study. First, based on the energy-based mechanistic (EBM) approach, the criterion for crack initiation of asphalt mixtures is formulated, and the dissipated pseudostrain energy for viscoplastic deformation and that for cracking (${DPSE}_p$ and ${DPSE}_c$) are separated at different temperatures. Then, an energy-based kinetics (EBK) approach is proposed with a dimensionless logarithmic rate for an energy changing process and the Arrhenius equation. The kinetic parameters include the activation energies and preexponential factors, which are determined from creep tests for different types of asphalt mixtures under different loading modes and aging conditions. Besides, the viscoelastic–viscoplastic–viscofracture model containing the kinetic models and parameters is established and the weak-form partial differential equations (PDEs) are implemented into a finite element program, COMSOL Multiphysics. The accuracy of the model is verified by the experiment results of asphalt mixtures. In addition, the numerical simulation using the models and parameters from the EBK approach is performed for a typical pavement structure. It successfully evaluates the distributions of viscoplastic strains in the surface layer.

Liquid flows through granular material are common phenomena in different engineering fields. Under certain conditions, this type of flow can be described by Darcy’s law, which involves the permeability of the porous medium. Experimental characterization of this parameter is then of importance to many practical applications. The falling-head permeability test is regarded as one of the most commonly used methods for that purpose. The required manipulations are easy and rapid, which makes it preferable, especially for field tests. However, due to practical difficulties of carrying out such measurements, it is only applicable to porous materials with permeability values lower than ${10}^{-10}{\mathrm{m}}^2$. To enlarge the test range while keeping its advantages, a modified test procedure is proposed here to measure saturated permeability values two orders of magnitude larger, namely around ${10}^{-8}{\mathrm{m}}^2$. Tests were performed on granular beds containing single-diameter and multiple-diameter beads. Experimental results showed that the saturated permeability can be accurately predicted by the revisited Ergun’s equation with its first empirical constant equal to 180. Two important factors must be considered when performing these tests. The first one is the data-processing method to compute the permeability from experimental data: both the gravity and pressure drop of the setup must be taken into account. The second one is the selection of test fluid. By comparing water and silicone oil, it was shown that the viscosity should be adapted depending on the permeability of the sample to ensure consistent and repeatable results.

Cyclic loading tests were conducted on self-centering reinforced concrete (RC) walls with disc spring devices (SC-RCW-DSD), and results confirmed that these walls had sufficient energy dissipation and satisfactory self-centering capabilities to reduce damage; however, the bearing capacity of such walls was relatively low compared with conventional RC walls. This study further presented a new self-centering RC wall with improved disc spring devices (SC-RCW-IDSD) to achieve a high bearing capacity. The improved disc spring device was designed to exhibit symmetrical flag-shaped hysteretic behavior to provide stable restoring force. A numerical model of the SC-RCW-IDSD was developed and validated by the experimental results. The effects of activation force, initial stiffness, post-activation stiffness, and the ratio of friction force to preloading of the improved disc spring device on the seismic resilient performance of the SC-RCW-IDSD were then analyzed. The SC-RCW-IDSD had stable energy dissipation and excellent self-centering capabilities, and the bearing capacity was greatly improved relative to that of the SC-RCW-DSD. As the activation force of the improved disc spring device increased, the seismic performance of the SC-RCW-IDSD was greatly improved compared with that of the conventional RC wall. The improved disc spring devices dissipated more than 90% of the seismic energy and reduced damage sustained by the RC wall in the SC-RCW-IDSD.

The lateral pile integrity test (LPIT) method, a low-strain nondestructive test technique, is employed in this paper to evaluate the integrity of an extended pile shaft supporting superstructure. An extended pile shaft with various defects is simulated by a Timoshenko beam (TB) model, and the lateral soil resistance is generalized as a beam-on-dynamic-Winkler-foundation (BDWF) model. The pile–soil model and presented solution are verified by rigorous finite-element analysis (FEA). The effects of pile defect and technical aspects of the LPIT method on the pile dynamic response are then investigated, and a practical method is proposed as well to evaluate the potential defect of an extended pile shaft on site. By comparing the actual arrival time of the pile toe reflection with the calibration value, it is possible to determine the type (stiff or weak) and extent (length or total length) of the pile defect.

The use of simplified numerical substructures in hybrid fire simulation is clearly advantageous as long as the resulting simulation accuracy is sufficient. However, excluding geometrical and material nonlinearities from the numerical substructure might make a significant difference in internal force redistribution and reduce the simulation accuracy beyond acceptable levels. Also, materials at a high temperature very often exhibit time-dependent behavior, including strain-rate dependency, high-temperature creep, and stress relaxation, which prohibit the use of extended testing time scales. This standpoint motivated the development of the real-time hybrid fire simulation method presented in this paper. Dynamic relaxation is proposed to solve the static response of the hybrid numerical-experimental fire simulation. As an equivalent dynamic solution method, dynamic relaxation allows for coupling substructure equations of motion by using a partitioned time integration approach. Minimal data exchange between substructures and negligible computational overhead plus ease of reusability of verified finite-element software makes the proposed algorithm suitable for coordinating real-time hybrid fire simulations. The hybrid fire simulation of a virtual steel frame case study is reported as a validation example.

A multiaxial degrading hysteretic model is developed, enabling consistent multiaxial yield/capacity surface evolution with degradations, and is appropriately incorporated in a finite-element framework using hysteretic Timoshenko beam elements. Degradation phenomena are introduced in this model in the form of either symmetric or asymmetric strength degradation, stiffness degradation, pinching functions, and various combinations thereof. More specifically, a new strength degradation function is developed and enhancements in other existing functions are suggested to simulate the physically observed degradation phenomena in structural elements. The degradation functions are then employed in a multiaxial classical damage-plasticity framework to satisfy the consistency criterion of the yield/capacity surface, thereby resulting in a set of new multiaxial hysteretic evolution equations. The proposed evolution equations are specifically formulated so as they could be seamlessly incorporated into a hysteretic finite-element formulation, using appropriate displacement and hysteretic interpolation functions, to satisfy the exact equilibrium conditions and model distributed plasticity characteristics, thereby avoiding any shear locking effects. As such, the proposed hysteretic finite-element model accounts for equilibrium, distributed plasticity, degradations, and multiaxial inelasticity with capacity interactions in a single consistent and unified framework. Constant system matrices are employed that do not require updating throughout the analysis, while the degradations and inelasticity are captured through the suggested multiaxial hysteretic evolution equations. An efficient numerical solution scheme is also devised, where the finite-element model can be expressed explicitly in terms of first order ordinary differential equations (ODEs), rather than a set of complex differential-algebraic equations for quasi-static cases. The resulting system of equations can be then straightforwardly solved using any standard ODE solver, without any required linearization. Numerical illustrations and experimental verifications are provided to demonstrate the performance and utility of the suggested methodology.

In the present study, two new models are proposed to model nonstationary stochastic processes with time-frequency-dependent spectra and coherence. The models are developed based on the discrete orthonormal S-transform and are enhanced by using the time-frequency representation from the S-transform. One of the models is shown to be usable in carrying out the conditional simulation. A parallel is drawn between the proposed models and those based on the spectral representation method using the ordinary Fourier analysis to facilitate an understanding of the proposed models. Unlike the use of the evolutionary spectral theory, no amplitude modulation or frequency modulation functions need to be assigned and identified for the proposed models because they are included explicitly in the time-frequency spectra. Moreover, the use of the S-transform and discrete orthonormal S-transform provides a familiar time-frequency spectral representation rather than the timescale representation obtained from the wavelet transform. The exposition and use of the models are focused on the simulation of ground motions at single and multiple support, although the models can be equally applicable to nonstationary processes, such as the winds and waves. The adequacy of the simulated ground motion records is assessed in terms of the time-frequency spectral representation and response spectrum. Numerical examples, including conditionally simulate ground motions, are also given.

In the dynamic characteristic analysis of high-speed railway bridges, free-vibration responses after a train passes are valuable for modal identification. However, distinguishing between free- and forced-vibration segments requires user intervention, which hinders the reliable and continuous identification of modal parameters, thereby impeding the vibration-based structural state assessment. This paper proposes a free-vibration detection technique whose basis is that the envelope function of each modal component decomposed from the free-vibration data decays exponentially. To extract the modal component adaptively, iterative variational mode decomposition is proposed where the signal is iteratively decomposed into two components until a single-degree-of-freedom component is obtained. Subsequently, the estimated free-vibration data are adopted to identify the modal parameters by the eigensystem realization algorithm with data correlation. A numerical simulation illustrates that the proposed method can provide the optimal free-vibration data for modal analysis. To verify the effectiveness of the proposed method in practice, the accelerations of the railway bridge during the passage of a train are analyzed. The modes can be identified from the estimated free-vibration data but not from the combination of forced- and free-vibration data, which indicates that the separation of forced and free vibration is necessary and can be achieved by the proposed method.

The two-layer model previously established by the author is expanded to well capture the response of back-rotated and shape-distorted piles set with $P-{\mathrm{\Delta }}_o$ effect ($P$ is the vertical load on piles and ${\mathrm{\Delta }}_o$ is the initial lateral displacement) and subjected to lateral spreading. It is underpinned by sets of parameters of modulus ($k_{\mathrm{s}}$), modulus ratio ($m$), rotational stiffness ($k_{\theta }$), and limiting force per unit length ($p_b$), and calibrated using back-rotating four-pile groups ($1g$ model tests). The study shows that the imposed $P{\mathrm{\Delta }}_o$ values reduce the normalized stiffness ${\overline{k}}_{\theta }$ to $-(0.14–2.54)$ or (0.114–0.494). This ${\overline{k}}_{\theta }$ together with $p_b$ at $m=1.5–8$ incur the diverse response of (1) stable I2 and I5 piles for sufficient forward-rotating capacity; (2) stable I9 piles for a high ${\overline{k}}_{\theta }$; (3) distorted I14 piles at stiffness singularity (SS); (4) initially stable through to later distorted (at SS) I4 piles due to $p_b$ reduction; and (5) hinged I6 piles despite the largest $P{\mathrm{\Delta }}_o$ value and back-rotation. The flexible two-layer model should be adopted to design restrained piles subjected to lateral spreading.