A Measure-Correlation-Predict Method for Turbulence Intensity Distribution Assessment

A method based on a multiple correlation analysis is proposed in this work that is able to offer a long-term assessment of turbulence intensity (TI) distribution at a local meteorological mast using a reference wind dataset. The MCP analysis uses a joint probability distribution function which is derived from wind speed, wind direction and TI data which are simultaneously recorded at the two measurement sites. Validation has been carried out using two met-masts which have a high level of correlation in terms of wind speeds. The algorithm properly corrects the local wind dataset, by estimating the “per sector” and “per speed” long-term distribution of TI; this was proved using different short-term data samples. Other tests, opportunely arranged using artificial wind data, help to understand the score of the method when multiple variables are used, highlighting that fine assessment can be always achieved when all the used variables are dependent on each other.