Prediction of the Standard Deviation of Wind Speed Turbulence

Abstract

Turbulent wind speed is a vital component of wind speed, corresponding to the small-scale uncertainty of wind speed. The wind power output and aerodynamic loads are significantly influenced by turbulent wind speed. The present paper concentrates on the study of turbulent wind speed prediction, which is hardly addressed in previous studies and is a novel approach to understand turbulent wind speed in advance. In the present study, turbulent wind speed is measured by turbulence standard deviation. Although turbulent wind speed itself is unpredictable, the correlation analysis proves that the turbulence standard deviation is predictable within a proper time horizon. As a result, the turbulent wind speed can be understood ahead of time by predicting turbulence standard deviation. The 10-min ahead, 30-min ahead and 60-min ahead predictions of turbulence standard deviation are provided by Support Vector Regression and Kernel Ridge Regression. Furthermore, the average wind speed is fused into the forecasting model to improve the prediction accuracy of turbulence standard deviation.

Keywords: Wind speed, Turbulence, Uncertainty, Standard deviation, Predictability, Correlation analysis