REPRESNTATION OF NON-STATIONARY INPUT GROUND MOTION MODEL BASED ON SHORT-TIME FOURIER TRANSFORM AND RANDOM RESPONSE ANALYSIS OF ELASTIC SDOF SYSTEM

This study proposes a new representation method of an input ground motion model as a non-stationary stochastic process based on the short-time Fourier transform (STFT). The inverse STFT of the product of a deterministic amplitude function concerning time and frequency and the Fourier transform of a Gaussian white noise gives the input ground motion model. This definition implements the ground motion model as an evolutionary process and delivers the expression of the non-stationary power spectral density function. The sample waveform functions generated numerically from the stochastic model keep expected well-designed characteristics. Furthermore, a method of random response analysis of an elastic SDOF system excited by the input ground motion model is carried out and used to estimate its maximum response. Some numerical simulations demonstrate the effectiveness and accuracy of the proposed method.