Estimation of electricity cost of wind energy using Monte Carlo simulations based on nonparametric and parametric probability density functions

Abstract

It is conventional in wind energy assessment projects to calculate the expected value of the annual energy production (AEP). However, ignoring the effect of wind speed distribution on the distribution of the annual energy production may cause wrong predictions in the cost analysis. For a better estimation of the payback period, it is essential to accurately determine the confidence levels of the electricity cost around the expected value.Wind speed distributions are commonly represented by the Weibull model. Improved functions or nonparametric functions are preferred in the case of multimodal wind speed distributions. As nonparametric probability density function, the optimum spline based functions are implemented and compared to Weibull, Weibull & Weibull for two different wind sites. The results show that the spline based probability density functions produce minimum fitting error for the analyzed cases. Once the wind speed distributions are characterized, random wind speeds are generated to calculate the AEP distributions by Monte Carlo simulations. The cost analysis is then carried out based on the AEP distributions which includes the determination of the confidence levels. It is observed that the confidence levels of the electricity costs which are not falling close to the expected cost region are much greater.