Search

The purpose of this paper is to propose one analytical method to make stochastic floor response spectra in order to easily estimate the dynamic characteristics of secondary system such as piping, equipment and so on, under the stochastic non-stationary excitation and model uncertainty. Applying the transition matrix method with state space to the secondary system mounted on main structure, the non-stationary standard deviation of absolute acceleration response relating to the secondary system is obtained under the assumption of no structural uncertain properties. Then, finally, in order to obtain the statistical results of response in consideration of model uncertainty, through combining the extreme value theory, FOSM(First Order Second Moment) method and ACO(Ant Colony Optimization) algorithm, the floor response spectra can be stochastically derived. This proposed floor response spectra may be called the Stochastic Floor Response Spectra. The analytical results are compared with those by Monte Carlo simulation, and show the accuracy and efficiency.

The regulatory guide for seismic design of nuclear power plant has been vastly revised at several fields. Main items under revision are observed such as the requirement of making appropriate concern for the dynamic coupling effects of horizontal and vertical components of earthquake, and the consideration of probabilistic risk assessment against the uncertainty of earthquake. The purpose of this paper, therefore, is to estimate the stochastic responses of multi-degree-of-freedom structure with closely spaced natural frequencies under the consideration of the correlations between multi components of earthquake. For this target, one mathematical model is developed in order to analytically evaluate the effects of correlation between components of input, and the stochastic vibration analysis is executed based on FEM-based eigen mode analysis and transition matrix method. The analytical results are compared with those by Monte Carlo simulation. Failure probability of structure is also explained as a typical example applying the proposed technique to the probabilistic seismic design.

This paper outlines shaking table tests and their results for an active seismic response control system that uses fuzzy optimal logic (FOL). The test results confirmed the validity of the vibration control effect of this seismic response control system. The results of this study lead to two conclusions, that the effectiveness of this FOL control system can be increased by modifying the membership function, and that the results of seismic response control tests can be qualitatively evaluated by simple simulation methods.

A method to estimate the stochastic response of hysteretic structures under combined random loads regarded as Gaussian white noise is investigated. The response statistical moments of structures, with hysteretic characteristics approximated by Wen's mathematical model, are obtained by using the Fokker-Planck equation method with a cumulant truncation technique. This paper deals with the formulation for the multi-degree-of-freedom system, and demonstrates that resulting variances of the response for a three-degree-of-freedom structure, as a numerical example, agree well with a Monte Carlo simulation.

Recent research related to probability-based structural design has become more active, especially in the field of the load and resistance factor design (LRFD) technique. This paper also describes some contributions to the improvement of the LRFD method. In this paper, a rational and analytical method to determine the required code format and load factors, which were determined experimentally in many former LRFD studies, considering the dynamic characteristics of a structure is newly proposed on the basis of the dynamic reliability technique with the load combination theory. The usefulness of this proposed method is described in an application to the selection problem of the required code format and load factors against a typical structure that is flexible and excited due to multiple random loads.

The authors have attempted to develop an analytical method that estimates the variability of structural response by treating the variations in the physical properties of structures by means of the stochastic finite-element method and by treating the random variation in earthquake excitations by means of the random vibration theory, and to further encode the method of analysis. This article describes the formulation of variability of response used in the analytical method, and then shows the results of benchmark analysis with the Monte Carlo method to confirm the accuracy of the analytical method. The floor response spectrum of a reactor building was also analyzed as an example in order to demonstrate the validity of this method.

A finite element formulation for a class of nonlinear viscoelastic continua subjected to stochastic excitation is developed by means of a stochastic equivalent linearization technique. The nonlinear viscoelastic properties of the continua are modeled in terms of the constitutive equation which linearly involves the hysteretic tensor and in terms of the auxiliary equation which describes a nonlinear relationship among the strain and hysteretic tensors and their time derivatives. This auxiliary equation is linearized with the aid of a stochastic linearization technique which minimizes the expected value of the square of the difference between the nonlinear and linearized auxiliary equations. The linearized boundary value problem then involves a set of two linearization coefficients which are functions of space and time. With the aid of the weighted residuals method, the boundary value problem developed for the continuum is transformed into that in the finite element formulation. In integrating the equation ...

This paper presents a new method to gain the structural reliability of a nonlinear single degree of freedom system with the material uncertainty and deteriorating characteristics due to a nonzero mean random excitation, by utilizing the Fokker-Planck equation approach (F-P method). Since the obtained moment equations of nonlinear systems by F-P method are not closed form, some approximations should be taken to get the closed form. In the presented paper, the response moments can be obtained to analyze the closed moment equations which are derived from the approximation of higher-order moments by lower-order moments with the aid of a truncation technique of higher-order cumulants. This paper shows the analysis of a Duffing oscillator subjected to a stationary Gaussian white noise as a typical example, and demonstrates that the resulting statistical values of the response agree well with the Monte Carlo simulation.