Journal of Engineering Mechanics

Measurement of the compressive strength of parallel strand lumber (PSL) is conducted on specimens of varying size with nominally identical mesostructure. The mean of the compressive strength is found to vary inversely with the specimen size, and the coefficient of variation of the strength is found to decrease with increasing specimen size, and to be smaller than the coefficient of variation of strength for solid lumber. The correlation length of the compressive strength is approximately 0.5 m, and this correlation length leads to significant specimen-to-specimen variation in mean strength. A computational model is developed that includes the following properties of the PSL mesostructure: the strand length, the grain angle, the elastic constants, and the parameters of the Tsai-Hill failure surface. The computational model predicts the mean strength and coefficient of variation reasonably well, and predicts the correct form of correlation decay, but overpredicts the correlation length for compressive strength, likely because of sensitivity to the distribution of strand length used in the model. The estimates of the statistics of the PSL compressive strength are useful for reliability analysis of PSL structures, and the computational model, although still in need of further development, can be used in evaluating the effect of mesostructural parameters on PSL compressive strength.

An interface deformable piezoelectric bilayer beam model is proposed to study the electromechanical responses and interface stress distributions in an intelligent layered structure. Like most of current approaches in the literature, the layerwise approximation of electric potential is employed. While in contrast to the linear approximation where the induced electric field is ignored, the present model takes a quadratic variation of the potentials across the thickness, thus warranting an efficient and accurate modeling of the electric field. Completely different from the widely used equivalent single layer model, in which the whole laminate is assumed to deform as a single layer and thus has a smooth variation of the displacement field over the thickness, the present model considers each sublayer as a single linearly elastic Timoshenko beam perfectly bonded together and therefore with individual deformations. To ensure the continuity of deformations of two adjacent sublayers along the interface, two interface compliance coefficients are introduced, by which both the longitudinal and vertical displacement components along the interface of two sublayers due to the interface shear and normal stresses are taken into account. To assess the performance of the present model, a number of benchmark tests are performed for a piezoelectric bimorph and a piezoelectric-elastic bilayer beam subjected to (1) a force density normal to the upper face and (2) an electric potential applied to the top and bottom faces. A remarkable agreement achieved between the present solution and the finite element computations illustrates the validity of the present study. The present model not only predicts well the global responses (displacement, electric charge, etc.), but also provides excellent estimates of the local responses (through-thickness variations of electromechanical state, interface stress distributions, etc.) of the piezoelectric layered structures. The novel mechanics model of electroelastic layered structures presented can be used to efficiently and effectively characterize hybrid smart devices and develop/optimize new multifunctional materials.

The tensile stiffness and deformation model of rubber bearings are systemically researched in the paper, a series of basic concepts, such as origin tensile longitudinal elastic modulus, tensile stiffness factor, origin tensile stiffness, offset tension origin stiffness, etc. are suggested corresponding to the theoretical formula and the experimental estimation method. Based on the origin tensile stiffness and offset tension origin stiffness, the deformation calculating theories related to pure tensile state and tension-shear state of isolated bearing are established. The double stiffness model and original stiffness model are proposed to describe the tension properties of rubber isolators after the bearings yield. The vertical tensile stiffness, offset tension origin stiffness, and deformation tests are performed with nature rubber bearings and lead plug rubber bearings to verify the new concepts and computation model of rubber isolators. All test results show that the theories established in the paper are suitable to analyze the tensile stiffness and tensile deformation for rubber isolators.

In this study, a new technique for detecting ungrouted cells in concrete block masonry constructions was developed. The concept, based on detecting the local dielectric permittivity variations, was employed to design coplanar capacitance sensors with high sensitivities to detect such construction defects. An analytical model and finite element simulations were used to assess the influence of the sensor geometrical parameters on the sensor signals and to optimize the sensor design. To experimentally verify the model, the dielectric properties of various materials involved in concrete masonry walls were measured. In addition, a masonry wall containing predetermined grouted and ungrouted cells was constructed and inspected using the developed sensors in a laboratory setting. Moreover, different capacitance sensors were designed and compared with respect to their sensitivity, signal-to-noise ratio, and coefficient of variation of the inspected measurements. Excellent agreements were found between the experimental capacitance signal response parameters and those predicted by the analytical and finite element models. The proposed sensor design, coupled with a commercially available portable capacitance meter, would facilitate employing this technique in the field for rapid inspection of masonry structures without the need for sophisticated data analyses usually required by other more expensive and time consuming methods.

Asphalt sand is described as a viscoelastoplastic materials. To involve elastic, viscoelastic, and viscoplastic strain components, a modified Schapery’s model is proposed through adding a viscoplastic term to the Schapery’s model with elastic and viscoelastic terms, and a strain-hardening model is used to describe viscoplastic behavior. A set of uniaxial compression tests, repeated and simple creep-recovery tests in different stress levels are performed to determine the stress-dependent parameters in the model. Finally, the model is validated by comparison with the results from the uniaxial creep experiments. Better agreement than the Schapery’s model indicates that the modified model is practicable to describe the mechanical behavior of asphalt sand in uniaxial compression.

An excavation process is a nonlinear dynamic problem that includes geometrical, material, and contact nonlinearities. The simulation of ground excavation has to face contact interaction in a changing geometry composed by several solid domains. The particle finite-element method (PFEM) is based on a Lagrangian description for modeling the motion of a continuum medium. The PFEM is particularly suitable for modeling a fluid motion with free surfaces. The application of the PFEM in ground excavation includes the use of the remeshing process, $\alpha$ -shape concepts for detecting the domain boundary, contact mechanics laws, material constitutive models, and surface wear models. Everything is correctly matched to quantify the excavation and the corresponding damage caused to the ground. The erosion and wear parameters of the soil/rock material govern the evolution of the excavation process. The preliminary results presented in this paper show that the PFEM it is a very suitable tool for the simulation of ground excavation processes.

Submarine pipelines are deemed as thin-walled structures in which relative external pressure may be created in some cases of fluid transmission. The certain effect of this type of loading is local buckling and its propagation along the considerable length of the line. In this study, an experimental program has been performed, in which the influence of ring stiffeners on the buckling strength of pipelines is investigated. In the tests, only hydrostatic pressure is considered as the major loading case, and the effect of further loads is neglected. The modes of initial buckling, buckling propagation, postbuckling, and development of yield lines and the final collapse of the pipeline have been closely appraised. It is verified that the buckling threshold highly hikes up by attaching some light ring stiffeners. By decreasing the ring spacing, the difference between buckling and failure loads is diminished and torsion-type yield lines at failure mode occur on the pipe skin.

In this paper, the static stability of the variable cross section columns, subjected to distributed axial force, is considered. The presented solution is based on the singular perturbation method of Wentzel-Kramers-Brillouin and the column is modeled using Euler-Bernoulli beam theory. Closed-form solutions are obtained for calculation of buckling loads and the corresponding mode shapes. The obtained results are compared with the results in the literature to verify the present approach. Using numerous examples, it is shown that the represented solution has a very good convergence and accuracy for determination of the instability condition.

The present paper examines the convexity of the yield surface in the directional distortional hardening models by Feigenbaum and Dafalias. In these models anisotropy develops through kinematic and directional distortional hardening, supplemented by the classical isotropic hardening, and the associative flow rule is used. However, the issue of convexity, which naturally arises due to the distortion of the yield surface, was not fully addressed. The present paper derives the necessary and sufficient conditions to ensure convexity of the yield surface for the simpler Feigenbaum and Dafalias models, but it is not as straightforward to derive corresponding conditions for convexity of the Feigenbaum and Dafalias model version which contains an evolving fourth-order tensor. In this case convexity will be addressed first in general and then at the limit state for which simple restrictions on the material constants to ensure convexity are derived. Numerical examples will show that some of the yield surfaces simulated in the original Feigenbaum and Dafalias publication will not stay convex if loaded beyond what was done in these publications. Therefore the material constants for these cases are recalibrated based on the derived relations for satisfaction of the convexity requirement, and the fitting of the yield surfaces is repeated with the new set of constants and compared with the previous case.

Direct integration algorithms are typically used to solve temporally discretized equations of motion to evaluate the performance of structures under dynamic loading. The stability of these direct integration algorithms are usually investigated for linear elastic structures. However integration algorithms are often applied to structures with nonlinear behavior. This paper presents a procedure based on discrete control theory to investigate the stability of direct integration algorithms applied to multidegree-of-freedom (MDOF) nonlinear structures. The discrete root locus approach is used to investigate properties of the poles of the discrete transfer function matrix representing the nonlinear structural dynamics and to assess the stability of the integration algorithm. The procedure is illustrated using a nonlinear shear building MDOF system to investigate the stability of popular direct integration algorithms, including the Newmark family of integration algorithms, the Hilber-Hughes-Taylor $\alpha$ -method, and two newly developed explicit integration algorithms. Stability limits are derived for the direct integration algorithms that are found to be conditionally stable.

The viscoelastic continuum damage model, developed based on Schapery’s correspondence principle and the continuum damage mechanics, has received a great deal of attention because of its mathematical soundness and effectiveness in describing damage growth in viscoelastic media and has been used to make reliable estimations on the fatigue lives of asphalt mixtures. Its applications to field mixtures, however, have been limited because the model requires performing the uniaxial tension test. As an alternative, this study developed an analytical methodology for applying the model to the indirect tension test, which has been successfully used in testing both laboratory-made and field-cored mixtures. From the results of the indirect tension tests conducted on asphalt mixture at three different crosshead-controlled rates, it was found that the stress-pseudostrain curves could be superimposed onto one equality line in the linear viscoelastic range of the given mixture, and its rate dependency was successfully eliminated in the $C_1$ and $S_1$ plots. This indicates that the methodology developed for the indirect tension test has a capability of evaluating damage development in asphalt mixtures through the viscoelastic continuum damage model. It would be potentially of great benefit to pavement engineers who want to estimate the remaining lives of field mixtures.

In this paper, a preexisting higher-order depth-integrated wave propagation model is extended to include a moving seabed. As a result, the extended model can be applied to both wave propagation and the dynamic process of wave generation by a seabed disturbance such as a submarine landslide. The model has the linear dispersion relation in a form of (4,4) Padè approximant, and approximates the water velocity profiles along the water depth with a fourth-order polynomial of the vertical coordinates. The fourth-order model is aimed at extending the validity of the lower-order depth-integrated models from long waves to both long and shorter waves, as well as improving the approximation of the velocity field from the second order to the fourth order. Laboratory experiments are carried out in a wave flume to study wave generation by a submerged landslide model. Both water waves and water velocities are measured by using resistance-type wave gauges and a particle image velocimetry. The experimental data are then compared with the predicted wave height and water current based on the new model and two existing lower-order Boussinesq-type models. The results clearly show that the new model predicts the fluid velocity more accurately and is also able to predict the shorter trailing waves very well where the traditional Boussinesq model may be inadequate, thus validating the improvement provided by the fourth-order model.

This paper presents the head reconstruction method (HRM), a new technique that can be used within the finite volume framework to make shallow water models well balanced, i.e., to correct the imbalance that exists between flux and source terms in the equation discretization in the case of irregular bathymetry thus providing unphysical solutions. This technique, based on considering, within each computational cell, the total head of the flow (i.e., the sum of the elevation, pressure and kinetic energies per unit weight of the fluid) as an equilibrium variable, enables the preservation of dynamic equilibria under subcritical, transcritical, and supercritical flow conditions. The new technique is applied to the one-dimensional total variation diminishing (TVD) MUSCL-Hancock scheme and the conservation property is then proven mathematically for this scheme under static equilibrium conditions. Furthermore, the effectiveness of the HRM is tested and compared with two other well balancing techniques based on considering the water elevation as an equilibrium variable in various steady flow case studies. In the end the robustness of the HRM is tested in the simulation of dam-break flow over irregular bathymetry.

This paper presents a robust analytical model for a moment connection of concrete-filled fiber reinforced-polymer (FRP) tubes (CFFTs) to concrete footings. The CFFT connection is based on a simple approach of direct embedment into the footing, thereby eliminating the need for connection rebar or mechanical devices. The CFFT is externally subjected to lateral and axial loads, resembling practical applications such as piles affixed to pile caps, bridge columns, or utility poles. The model adopts the concepts of equilibrium, deformations compatibility, and nonlinear concrete stress-strain behavior. It also employs a “bond stress-slip” relation that can be obtained from simple push-through tests on some of the commercially used tubes. The model can predict the critical embedment length $X_{\text{cr}}$ , which is the minimum length required to achieve material failure of the CFFT outside the footing, and bond failure inside the footing, simultaneously. If the actual embedment length is less than $X_{\text{cr}}$ , bond failure occurs prematurely at a lower strength that can also be predicted by the model. The model was verified using experimental data and showed that $X_{\text{cr}}$ was only 0.7 of the diameter for that case. A sensitivity parametric analysis was carried out that led to some approximations. Based on which, a simple closed-form expression was established for $X_{\text{cr}}$ in the case of lateral loading only.

An infinite beam attached to an elastic foundation is buckled by an axial force. The beam is weakened by one or more joints or partial cracks. The governing equations are solved analytically and an exact nonlinear characteristic equation gives the buckling criterion. It is found that the buckling force depends on the foundation stiffness and the rotational resistance of the joints. The buckling modes are complex, and may be either antisymmetric or symmetric.