Prediction of Wind Power Density Using Machine Learning Algorithms
This paper concentrates on spatial prediction of the wind power density in the complex mountainous region of Swiss Alps. The data-driven non-parametric Machine Learning modeling methods are applied to this task. Digital Elevation Model is extensively used in the modeling. Terrain characteristics calculated from DEM are used as the inputs of the mapping methods. The relevant information on the wind speed in complex alpine terrain is learned from data, and the use of numerous heuristics is avoided. The available meteorological information is obtained from 110 stations of Swiss meteorological monitoring network. Multi-Layer Perceptron and Support Vector Machine are used for the modeling. The basic problem which are considered in the paper are: spatial classification of the zones where the mean annual wind speed exceeds 5 m/s, spatial regression of the mean annual wind speed. Another case study deals with simultaneous modeling of the mean daily wind speed and direction, resulting in a data-driven wind field model.
Alexei Pozdnoukhov Mikhail Kanevski Vadim Timonin
Institute of Geomatics and Analysis of Risk (IGAR), University of Lausanne, Amphipole, 1015 Lausanne, Switzerland
国际会议
The 12th Conference of the International Association for Mathematical Geology(第12届国际数学地质大会)
北京
英文
620-623
2007-08-26（万方平台首次上网日期，不代表论文的发表时间）