Stochastic Modelling of Wind Speed at Udaipur

Abstract

Stochastic modelling for wind speed was done using 26 years (1978-2003) data. The maximum mean weekly wind speed 9.60 kmph occurs in the 25th week (June 18-24) and minimum mean weekly wind speed 2.37 kmph occurs in the 50th week (Dec 10-16) with 4.92 kmph mean weekly wind speed. Statistical tests indicated that the series of the weekly wind speed data is trend free. The periodic component of weekly wind speed can be represented by second harmonic expression. The stochastic components of the mean weekly wind speed follow second order Markov model. Validation of generated mean weekly wind speed series was done by comparison of generated mean weekly wind speed series and measured mean weekly wind speed series. The correlation coefficient between generated and measured mean weekly wind speed series was 0.9974. The correlation was tested by t-test and found to be highly significant 1 per cent level. The standard error (0.148 mrn) is quite low. The regression equation is very near to I: I line. Therefore, developed model can be used for future prediction of weekly wind speed.