Synthetic turbulence: A wavelet based simulation

• A new scheme for simulating wind velocity based on the Haar wavelet is proposed.
• The log-Poisson model has been used in the present simulation scheme.
• Effects of intermittency on the wind-induced response of structures are evaluated.

The treatment of wind-induced vibrations is an important consideration in the design of civil structures with increasing span-lengths and heights. The turbulence in the atmospheric boundary layer has been treated as Gaussian in conventional stochastic simulation schemes, wherein higher-order statistics have been disregarded. However, experimental evidence points at turbulence being a typical multifractal process, which suggests that the statistics at different scales of atmospheric turbulence are not strictly self-similar but exhibit stronger non-Gaussianity as the length scale decreases. Intermittency characterized by the occasional bursts in the wind velocity leads to non-Gaussianity. Intermittency and its potential impact on wind-induced response have been neglected though. Recent studies have addressed the multifractal property of turbulence in wind with wavelet-basis representations of the log-Normal or log-Poisson models. These schemes offer new insight into the simulation of turbulent wind, but suffer from several significant drawbacks, such as the inappropriate sampling of the wavelet coefficients. To overcome these shortcomings, a new simulation scheme, based on the Haar wavelet representation, is proposed in this study where the exact relation between the wind velocities and the wavelet coefficients is introduced. In addition, the effects of intermittency on the wind-induced response of structures are evaluated for a number of cases. A quasi-steady theory-based assessment indicates that the intermittency results in amplifying extremes of the wind-induced response and exhibits higher impact on relatively more rigid structures. It is likely that intermittency may invoke flow-structure interaction with possible enhancement in load effects at related scales, which may further amplify the extremes.