Show Less
Restricted access

Estimation of Uncertainty of Wind Energy Predictions

With Application to Weather Routing and Wind Power Generation


David Zastrau

Currently, a new generation of fuel-efficient ships, which use wind force in addition to conventional propulsion technology, is being developed. This study describes a mathematical method for a probabilistic estimate of the wind propulsion force on a ship route. The method is based on quantile regression, which makes it suitable for various ship routes with variable weather conditions. Furthermore, the author takes different macro weather situations into account for the calculation of the statistical distributions. He validates the results for a multi-purpose carrier, a ship route in the North Atlantic Ocean and archived weather forecasts. It showed that the wind force can be estimated more accurately if the macro weather situation is taken into account properly.

Show Summary Details
Restricted access

5. The Quantile Regression Model


Abstract: In chapter 4 quantile regression (QR) was selected as a method to estimate the uncertainty of predictions with prediction intervals. This chapter describes the quantile regression model. Furthermore, methods are described to reduce a weather prediction to a small set of variables to be used as regressors in the QR model (in addition to the variables which have been identified in sec. 3.3.5). In this context the North Atlantic Oscillation index (NAOI)—which summarizes the air pressure difference in the North Atlantic Ocean—is introduced in this chapter. Finally, two new regressors (Local Energy Distribution Moments (LEDM)) are introduced which are calculated by the mean and variance of the predicted energy in a region of interest. At the end of the chapter a short discussion compares the NAOI with the LEDM.


5.1. Linear quantile regression (QR)

You are not authenticated to view the full text of this chapter or article.

This site requires a subscription or purchase to access the full text of books or journals.

Do you have any questions? Contact us.

Or login to access all content.