Research On Probabilistic Design Spectra Based On The Variability Of Soil Nonlinear Parameters

From seismic damage investigation and analyses of the seismic responses of soil layers, it is discovered that soil dynamic modulus ratio and damping ratio are two important factors influencing the results of seismic sites responses. The impacts of soil nonlinear parameters on layered soil responses have been widely testified. However, soil is a kind of material with dramatically uncertain mechanical properties. With its long history of formation, together with the influences of both natural and artificial factors, properties of soil are extremely complex. Therefore, soil parameters obtained from experiments cannot avoid being discrete. Consequently, many uncertainties in geotechnical engineering and earthquake engineering arise.The standard design spectra in Design Criteria for Seismic Resistance of Buildings represent the average responses of various sites. Therefore, they show enormously uncertain characteristics, partly owning to the variability of soil nonlinear dynamic parameters. In order to conduct the seismic risk design of structures with different probabilities, probabilistic design spectra are needed that consider the variability of soil dynamic parameters.The thesis studies the method to compute probabilistic design spectra and provides advices to improve the standard design spectra based on probability theory that take into the variability of soil nonlinear parameters consideration. In addition, the thesis develops theory about the influences of soil nonlinear parameters on design spectra and, further provides references for the risk design of earthquake engineering. Moreover, in order to explore the relationship between characteristics of soil nonlinear and those of site responses, the relationship is established between probabilities of soil nonlinear parameters and those of site responses. Based on the relationship, soil nonlinear parameters can be selected when risk design of engineering is conducted.The main works of the thesis are as follows:1. The basic theory for uncertain analysis is introduced. The notions of normal distribution test, percentile, reference range and their computing methods are discussed with detail. In addition, the calculating method for probabilistic design spectra is advanced that take into the variability of soil dynamic modulus ratio and damping ratio consideration. Moreover, the calculation method is also programmed.2. Based on the calculation method advanced above, it is presented that probabilistic design spectra of four different seismic fortification levels at typical siteâ…¡. Probabilistic design spectra take into the variability of nonlinear dynamic parameters of both cohesive soil and non-cohesive soil consideration. In addition, the impacts of soil parameters on the probabilistic design spectra are also analyzed. The analysis results show that, the variability of soil dynamic modulus ratio and damping ratio dramatically influences probabilistic design spectra on the whole. However, the influences are determined by probabilities of parameters and also influenced by different frequency bands of probabilistic design spectra. Analysis results also reveal that discrete probabilistic design spectra follow partial distribution. Further analysis also shows that the influences of the variability of non-cohesive soil on design spectra are far greater than those of cohesive soil. In addition, the sensitivity of the influences of soil nonlinear dynamic parameters on peak ground acceleration increases with the increment of soil dynamic parameters probabilities.3. Based on the typical sitesâ…¡, the relationship is established between probabilities of soil (non-cohesive soil and cohesive soil) nonlinear parameters and those of ground responses by numerical methods.Analysis results show that, the relationship is different at different frequencies between probabilities of soil dynamic modulus ratio and those of design spectra; the relationship presents linear characteristic, independent of frequencies between probabilities of damping ratio and those of design spectra. Further analysis results also show that linear relationship is quite good between probabilities of soil nonlinear dynamic parameters and those of peak ground acceleration.