Applying Weibull Distribution and Discriminant Function Techniques to Predict Damaged Cup Anemometers in the 2011 PHM Competition



Joshua Cassity Christopher Aven Danny Parker


Cup anemometers are frequently employed in the wind power industry for wind resource assessment at prospective wind farm sites. In this paper, we demonstrate a method for identifying faulty three cup anemometers. This method is applicable to cases where data is available from two or more anemometers at equal height and cases where data is available from anemometers at different heights. It is based on examining the Weibull parameters of the distribution generated from the difference between the anemometer’s reported measurements and utilizing a discriminant function technique to separate out the data corresponding to bad cup anemometers. For anemometers at different heights, only data from the same height pair combinations are compared. In addition, various preprocessing techniques are discussed to improve performance of the algorithm. These include removing data that corresponds to poor wind directions for comparing the anemometers and removing data that corresponds to frozen anemometers. These methods are employed on the data from the PHM 2011 Data Competition with results presented.

Abstract 8 | PDF Downloads 15



Anemometer, Linear Discriminant

Duda, R. O, Hart, P. E., Stork, D. G., (2001). Pattern Classification. New York, NY: John Wiley & Sons, Inc
Eadie, W. T., Drijard, D., James, F. E, Roos, M., & Sadoulet, B. (1971). Statistical Methods in Experimental Physics. Amsterdam, North-Holland: Elsevier Science Ltd
Oke, T, R, (1987). Boundary Layer Climates, London, UK: Routledge
Touma, J, S, (1977). Dependence of the wind profile power law on stability for various locations, J. Air Pollution Control Association, Vol. 22, pp. 863-866.
Wolverton, C., Wagner, T. (1969), Asymptotically optimal discriminant functions for pattern classifiers, IEEE Transactions on Information Theory,
Ye, X., Veeramachaneni, K., Yan, Y., & Osadciw, L. A. (2009), Unsupervised Learning and Fusion for Failure Detection in Wind Turbines, Proceedings of 12th International Conference on Information Fusion. July 6-9,Seattle,WA
Technical Papers