Stochastic process model of ground motion to meet the needs of reliability analyses of nuclear power plants

For the safety of nuclear power plants and their critical components, necessity is emphasized to analyze them three dimensionally, taking into account specifically the effect of vertical component of ground motion. The effect of nonlinearity of the foundation and structure is also of vital importance for the economical design. To meet the above requirements, the stochastic model of ground motion in reliability analyses must be multi-dimensional and nonstationary in the amplitude and frequency. In this direction, this paper proposes as the mathematical means a multidimensional nonstationary model and the related simulation theory. First, the role of Fourier amplitude and phase of a stochastic process is discussed to lead to the nonstationary process model. Then, two simulation methods by use of trigonometric series and autoregressive series are presented. A methodology of preparing the model is also discussed. (orig.)