| Because of wind speed with volatility and randomness, wind speed probability distribution model is generally used to describe the distribution characteristics, the accuracy of which will directly determine the feasibility of wind farm pre-project, and impact the economy in commissioning phase, as well the stable operation of grid-connected wind farms. Thus, it is of great significance to study wind speed probability distribution models under different weather and terrain conditions for the rational use of wind resources.Influenced by geographic factors (such as latitude, topography) and regional climatic factors (such as temperature, humidity, pressure), the diurnal, monthly, yearly and inter-annual variation of wind speed has less regularity, and there is no standard model. Weibull model is commonly used. The longer the fitting period of time, the better the results, but such principle does not work in a shorter time scale (e.g. weeks, days). To solve these problems, through studying frequency features of near-surface wind speed:(1) The concept of considerable pressure, considerable humidity and temperature is proposed, from the angle of probability distribution, the fitting similarity between wind speed and main meteorological elements is studied.(2) The comprehensive index is applied. By integrating fitting index, the best model is selected;(3) MCMC method to estimate Weibull parameters is proposed to reflect the diurnal variation characteristics of wind speeds.(4) Power spectrum method to study the multi-scale wind distribution is put forward. The time-scale characteristic of wind speed probability distribution have been demonstrated; it is illustrated the fact that sub-period wind speed probability model is subjected to the total period; the validity of the power spectrum to identify the time scale is confirmed; determining statistical cycles and the probability distribution of wind speed in different time scales, and in different weather conditions. |
PDF Full Text Request