Combining application of wavelet analysis and genetic algorithm in wind tunnel simulation of unidirectional natural wind field near a sand ground surface

This study proposes a method for the approximate simulation of field gust in an atmospheric boundary layer wind tunnel driven by an axial fan. Wavelet analysis and genetic algorithm have been adopted to convert actual field gust data into a series of simplified and smooth control signals. One frequency converter is used to receive the signals via a programmable logic controller and change the fan speed correspondingly. Results show that the low-frequency energetic part of field gust on top of a saltation layer has been almost reproduced in the wind tunnel with energy fluctuations ±3.26% from the mean compared with the field data in terms of energy. The appropriate filter frequency for field gust velocity history is 1/32 Hz. A unidirectional natural wind field near a sand bed surface can be approximately simulated in the tunnel by adjusting carefully the roughness element arrangement. The time-averaged velocity profile of the simulated wind, which satisfies the logarithmic law, is nearly equivalent to the steady flow with the same average velocity.