Modeling and simulation of a wind model using a spectral representation method

In this paper, we give a detailed way to model a horizontal wind profile applied to turbine blades. Wind speed consists of four components, namely the mean wind speed, the gust wind speed, the ramp wind speed and finally, the turbulence. In order to get an accurate wind speed profile, wind models need to be in the form of time series to carry out the simulation. All the first three wind components are given analytically, but not the turbulence; it is the most difficult to model. The objective is to explore a simulation method for stochastic processes and apply it to wind turbulence. This is accomplished by using the spectral representation method (SRM). This technique was developed by Shinozuka and Jan, and produces sample realizations of the process according to the prescribed power spectral density function (PSD). The SRM method is applied on two kinds of wind PSD functions (Kaimal and Von Karman spectrum). An evaluation of the method is carried out by superimposition of the autocorrelation function of each spectrum found analytically, and the one obtained directly by MATLAB processing of the generated signal. Afterwards, an experimental evaluation of the method is achieved using real wind speed data. Simulation results show that time series of the wind speed turbulence are well reflected by the SRM method and that the proposed model can be used to generate synthetic turbulent wind speed data from the power spectral density of a spectrum.