Stochastic differential calculus for wind-exposed structures with autoregressive continuous (ARC) filters

Abstract

In this paper, an alternative method to represent Gaussian stationary processes describing wind velocity fluctuations is introduced. The technique may be considered the extension to a time continuous description of the well-known discrete-time autoregressive model to generate Gaussian processes. Digital simulation of Gaussian random processes with assigned auto-correlation function is provided by means of a stochastic differential equation with time delayed terms forced by Gaussian white noise. Solution of the differential equation is a specific sample of the target Gaussian wind process, and in this paper it describes a digitally obtained record of the wind turbolence. The representation of wind fluctuations with the proposed model is suitable for the use of stochastic differential calculus in wind-engineering applications. Some numerical applications dealing with structural models in presence of the wind fluctuations have been reported to challenge the robustness of the proposed method in the representation of stationary random process of wind-turbolence and its accuracy for engineering analysis.