BANDEIRA, M. M.; http://lattes.cnpq.br/1080069316722665; BANDEIRA, Maria Marle.
Abstract:
The daily and monthly wind regime in the Northeast of Brazil (NEB), to the latitude of 12 ° South, was analyzed. The available data were the hourly velocities at 10m height, from 1977 to 1981, for 66 climatological stations of the National Department of Meteorology (DNMET) were grouped into four daily periods, each lasting six hours. The biparametric Weibull model (in a hybrid version) was adjusted to the empirical frequency distribution of the wind. For the estimation of the Weibull parameters, three methods were tested: Moments, Maximum Likelihood and Least Squares. The adjustment tests indicated the maximum likelihood as the best fit generator; however, the momentum method provides adjustments with similar deviations from the empirical distribution, and was adopted because it is simpler to apply. The least squares method showed appreciable deviations, and is not recommended. Weibull parameters tables were constructed, characterizing the regime for each daily period and each month of each of the stations. Based on them, the average power available and that available for wind turbines can be estimated for each site studied. The shape parameter (c) allows us to define a correction coefficient g c for average power approximations based on the climatological mean wind <v>. The adjustment obtained was not always satisfactory, since some NEB stations appear to have a multimodal distribution of winds. A regionalization of the chi2 parameter associated to the adjustment allows to highlight the regions that best fit the proposed model. The distribution of the scale parameter (a) evidence the availability of wind power, better than the distribution of the mean <v>. Also, the distribution of the shape parameter (c) provides information on the asymmetry in the occurrence of winds and on the correction coefficient gc for each location. Finally, the comparison between a and c distribution charts allows the detection of subregions with the same available wind power and regime. Specific comments are developed for the month of February. Tables with the results of a and c allow us to construct similar analyzes for other months and periods of the day. The results should be considered valid for the period 1977-81, since the wind regime may vary considerably. On the other hand, its usefulness is restricted to the synoptic scale, consistent with the spatial distribution of climatological seasons. They are, therefore, indicative for the primary detection of favorable subregions for wind farms. Conclusions in meso and microscale require specific experiment.