Simulation of wind-speed time series for wind-energy conversion analysis.

Publisher Website
Google Scholar
Abstract
In order to investigate operating characteristics of a wind energy conversion system it is often desirable to have a sequential record of wind speeds. Sometimes a long enough actual data record is not available at the time an analysis is needed. This may be the case if, e.g., data are recorded three times a day at a candidate wind turbine site, and then the hourly performance of generated power is desired. In such cases it is often possible to use statistical characteristics of the wind speed data to calibrate a stochastic model and then generate a simulated wind speed time series. Any length of record may be simulated by this method, and desired system characteristics may be studied. A simple wind speed simulation model, WEISIM, is developed based on the Weibull probability distribution for wind speeds with a correction based on the lag-one autocorrelation value. The model can simulate at rates from one a second to one an hour, and wind speeds can represent short-term averages (e.g., 1-sec averages) or longer-term averages (e.g., 1-min or 1 hr averages). The validity of the model is verified with PNL data for both histogram characteristics and persistance characteristics.

This publication has 0 references indexed in Scilit: