Random Function&Spectral Representation Model Of Strong Ground Motion And Its Application

Engineering structures usually receive random dynamic action, such as earthquakes,strong winds, waves in their useful life. In the random dynamic analysis of engineeringstructures, the most critical, the first and the most difficult steps must to describe therandom dynamic process simply and reasonably and effectively. Classical spectralrepresentation which was widely used during the analysis of stochastic dynamic role areoften requiring up to hundreds of thousands of random variables in order to ensure therequired accuracy, which greatly increased the difficulty of the analysis and computationaleffort.In this paper, the spectral representation orthonormal random variables represent foronly1to2basic random variables orthogonal form of a random function, the randomfunction can be constructed orthogonal to the non-Gaussian random variables and mutuallyindependent Gaussian random variables. Represented by random function randomvariables, it need to re-random arrangement of these variables in the probability space.Then, combined with the classical spectral representation, engineering stochasticdynamical effect of random functions-spectral representation model. Smooth vibrationpower spectral density function of the acceleration process, the establishment of groundmotion random function-spectral representation model, used the model to generate steadyand non-Gaussian process with a given probability, stationary Gaussian process, thenon-stationary non-Gaussian and non-stationary Gaussian process ground motion samplefunction。After the comparative analysis of the target power spectrum, the use of randomfunctions-spectral representation model of ground motion sample value in thesecond-order statistics and sample collection power spectrum which can be in line with thetarget power spectrum is a good vibration sample. Mutually independent Gaussian randomvariables is highly non-linear transformation from the standard orthogonal randomvariables, so the use of the Gaussian process model samples with non-Gaussian samplewhich compared to the second-order numerical statistics and sample collection powerspectrum with the target power spectrum Relevance is slightly worse.Vibration random function-vibration sample spectral representation generated by themodel with a given probability of non-stationary non-Gaussian probability densityevolution method, ground motion along the bridge to the input ground motion transverse tothe input ground motion along the cross-bridge to the input of the Nanjing Yangtze RiverBridge finite element model of stochastic seismic response and reliability analysis, in order to achieve the engineering application of the model. The seismic excitation input adirection in the horizontal plane, the large cross-bridges only in the corresponding directionon the reaction; the reaction is little influence on the direction within a horizontal planeperpendicular to the seismic excitation input.