Journal of Engineering Mechanics

This paper addresses a problem often encountered in calibrating resistance factors for the ultimate capacities of piles based on load test databases. In practice, many pile load tests are not conducted to failure but only to a multiple (e.g., 2) of the design load. This leads to a difficult situation of incomplete information: for these test results, the ultimate bearing capacities of the test piles are unknown. How can these test results still be used to calibrate resistance factors of piles? A full probabilistic framework is proposed in this research to resolve this problem. A local pile test database of Taipei (Taiwan) is presented for demonstration. The analysis results show that the inclusion of the incomplete pile load test data enhances the stability of the calibrated resistance factors. For a target reliability index of 3, the calibrated resistance factor is in the range of 0.4–0.66 for two design models adopted in the current Taiwan design code. Moreover, it is found that the safety factor adopted in the Taiwan design code corresponds to a reliability index larger than 3.5, which is reasonably conservative.

This paper presents a technique developed on the basis of ultrasonic guided waves to monitor prestress levels in multiwire prestressing strands. The transducer layout identified for stress measurement is composed of an ultrasound excitation provided by a piezoelectric element bonded on a peripheral wire. Ultrasound detection is performed on the central and peripheral wires at the strand’s end. The ultrasonic feature used for stress monitoring is the interwire leakage between the peripheral and the central wire, occurring across the strand anchorage. A semianalytical finite-element analysis is first used to predict modal and forced wave solutions in seven-wire strands as a function of the applied prestress level. The numerical analysis accounts for the changing interwire contact as a function of applied loads and predicts the attenuation occurring in loaded strand when the wave travels across the anchorage. Results of load monitoring in free strands during laboratory tests are then presented. Finally, a statistical approach is used to enhance the sensitivity of the technique to stress level in the strands. The study presented focuses on unbonded tendons. However, the final goal of the research is to monitor prestress loss in bonded tendons that are found in the majority of the bridges built in California.

Reinforcement corrosion owing to chemical attack could lead to premature steel-mortar debonding, concrete cracking, and catastrophic failure of structures if not well attended. In conventional design and maintenance practices, heterogeneous concrete matrix is commonly treated as a homogeneous medium when the evolution of chemical ingress and concrete cracking need to be determined. Such oversimplification has caused significantly inaccurate prediction and evaluation of structural service life. This paper presents a finite-element (FE) model developed to evaluate the service life of reinforced concrete (RC) structures in three key steps: chemical ingress, steel corrosion, and concrete cracking. The mass conservation principle is employed in the first step to model the ingress of multiple chemical species into variably saturated heterogeneous concrete matrix. By using Faraday’s law, steel corrosion and the incurred diametric expansion are then formulated as a transient displacement boundary condition for subsequent analysis of concrete cracking. The cracking pattern of concrete under the expansion force of corrosion products is finally characterized by using a cohesive-fracture approach. The FE model is validated with laboratory experiments.

Concrete-filled fiber-reinforced polymer (FRP) tubes (CFFTs) have been introduced as a new system for piles, columns, and poles. A simple moment connection based on direct embedment of the CFFT into concrete footings or pile caps, without using dowel-bar reinforcement, has been proposed by the authors. Robust analytical models to predict the critical embedment length ($X_{cr}$) were also developed and experimentally validated. In this paper, a comprehensive parametric study is carried out using the models developed earlier along with a newly developed closed-form model for the general case of axial loading, bending, and shear applied to the CFFT member. The parameters studied are the diameter ($D$), thickness ($t$), length outside the footing ($L$), and laminate structure of the FRP tube, as well as the tube-concrete interface bond strength (${\tau }_{\max }$), concrete compressive strength in the CFFT ($f_{ct}^{\prime }$) and footing ($f_c^{\prime }$), and the magnitude and eccentricity of axial compressive or tensile loads. It was shown that increasing $D$, $L/D$, ${\tau }_{\max }$, and $f_c^{\prime }$ of the footing, or the axial compression load, reduces $(X/D{)}_{cr}$, whereas increasing $t$ and $f_{ct}^{\prime }$ of the CFFT, the fraction of longitudinal fibers in the tube, or the axial tension load, increases $X_{cr}$. As the axial load eccentricity increases, $X_{cr}$ reduces for tension loads and increases for compression loads until both cases converge asymptotically to the same $X_{cr}$ value, essentially that of pure bending.

This paper presents a methodology to reduce the computation time for discrete-element (DE) modeling of asphalt-based materials, based on the frequency-temperature superposition principle. Laboratory tests on the dynamic modulus of asphalt sand mastics and asphalt mixtures were conducted at temperatures of $-5$, 4, 13, and 21°C and frequencies of 1, 5, 10, and 25 Hz, respectively. The test results of the asphalt sand mastics were fitted with the Burger’s model to obtain the microparameters for DE models. To reduce the computation time of the DE modeling, the regular loading frequencies were amplified to virtual frequencies. Simultaneously, the Burger’s model parameters (microparameters in DE models) of asphalt sand mastic at regular frequencies were modified to those at virtual frequencies on the basis of the frequency-temperature superposition principle. Because the virtual frequencies were much larger than the regular frequencies, the computation time was significantly reduced by conducting the DE modeling with the virtual frequencies and the corresponding modified Burger’s model parameters. The modeling work, which typically takes several months or years with the traditional methods, only took a few hours or less in this study.

This paper considers a portal frame with rigid members and flexible joints. The stability and postbuckling characteristics are derived analytically. The optimum frame to raise a load to a given height is trapezoidal in shape, with only the top two joints heavily strengthened.

This paper presents new closed-form analytical approximations to the first-passage problem in structural reliability by using the exact closed-form solutions for the spectral characteristics of nonstationary random processes. The first-passage problem applied to a structural system possibly with random parameters and subjected to stochastic loading consists of computing the probability of a response quantity exceeding a deterministic threshold in a given exposure time. This paper also investigates, on the basis of benchmark problems, the absolute and relative accuracy of analytical approximations of the time-variant failure probability, such as Poisson, classical Vanmarcke, and modified Vanmarcke approximations, in the case of nonstationary random vibration. The classical and modified Vanmarcke approximations are expressed as time integrals of the closed forms of the corresponding hazard functions. These closed forms refer to linear elastic systems subjected to stationary and nonstationary base excitation from at-rest initial conditions, and they are obtained using recently developed exact closed-form solutions for the time-variant bandwidth parameter. These closed-form Vanmarcke’s approximate solutions to the first-passage problem are compared with the well known Poisson approximation and accurate simulation results obtained via the importance sampling using elementary events (ISEE) method for two benchmark applications: (1) a set of linear elastic single-degree-of-freedom (SDOF) systems defined by different natural periods and damping ratios, and (2) an idealized, yet realistic three-dimensional asymmetric steel building model. The linear elastic SDOF systems are subjected to white noise base excitation from at-rest initial conditions, while the steel building model is subjected, from at-rest initial conditions, first to white noise and then to a time-modulated colored noise base excitation. The retrofit of this second benchmark structure with viscous dampers is also considered, illustrating (1) the use of the newly available closed-form approximations of the failure probability for nonclassically damped linear elastic systems, and (2) an example of practical use in structural engineering of the presented analytical solutions. The results presented in this study show that, for nonstationary random vibration problems, the two Vanmarcke approximations can improve considerably the estimates of the time-variant failure probability for the first-passage problem when compared with the simpler Poisson approximation.