Long-term simulation of the mean wind speed

This paper introduces a procedure for the simulation of long-term time series of the mean wind speed. A series of meteorological records are used to describe the probabilistic distribution and the spectrum of the process to be simulated. In particular, a complete description of the macro-meteorological spectrum of the wind speed is obtained, piecing together spectral portions deriving from different records. Then, an advanced simulation algorithm is introduced, based on existing iterative methods for simulating non-Gaussian processes via memoryless transformations, revised in order to obtain random extreme values of the simulated time-histories. The outlined procedure is finally applied for generating sample functions of the mean speed process. The availability of a large number of realizations permits the ability of the new algorithm to produce functions with desired features to be investigated in terms of ensemble statistics. This robust database could be useful for further studies.

► A procedure to simulate long-term time series of the mean wind speed is proposed. ► Records of different sensors are involved to describe the mean wind speed process. ► An advanced algorithm to simulate Weibull Hybrid distributed processes is introduced. ► Outlined procedure is used for generating many samples of the mean wind velocity.