Research Of Wind Farm Power Resource Assessment Based On Support Vector Machine Method

Booming global Scientific and technological progress always come with giganticenergy consumption. Coal, fossil oil, nature gas and other traditional resources arecalled non-renewable resources, because they can not or take quite a long periodtime to regenerate. Man realized that, not only these energies are nonrenewable andran out of rapidly, but also cause serious environmental pollution problems. In recentyears, people have devoted more efforts looking for cleaner, pollution-freealternative energy resources. As one of the potential clean, pollution-free andrenewable energy resources, people realized that the wind power can be easilytransferred to electricity and help to fulfill the increasingly energy demanding wefaced today. Therefore, a worth studying and challengeable researching field is howto develop existing wind power resources and put them to use efficiently.In recent decades, more and more countries and organizations focused onresearches of how to develop wind power energies. For his society sustainabledevelopment, Chinese government introduced many preferential policies toencourage the development and use of the wind resources. Along with the massivedevelopment and construction of the wind farms, people faced many technical issues;one of the huge challenges is how to design the wind mill distribution in a givenregion. Traditional methods for wind resource assessment are proved to be notprecise enough in recent project practices. An overvalued assessment causes themismatch of transmission network planning and the expected electricity output goalcannot reach, similarly an undervalued assessment causes wind turbines to operate atlower capacities and cannot generate the prospective earnings.Above all, the researches of finding a more accurate wind power resourceassessment method are worthy and necessary. Based on a precise assessment method,a forecast for accurate annually equivalent electrical energy will be likely to becomereality. The thesis discusses the controlling strategies of the double fed asynchronousmachine, the traditional methods of the wind power distribution, introduces thealgorithm of the support Vector Machine (SVM) and compares the traditionalmethods with SVM. We analyzed related data about the wind farm in Shanghairegion from planning to in operation. Comparing and analyzing the calculatedresults of the traditional methods against the improve SVM by the real wind farmand testify the methods of the improved SVM are more accurately.