台灣地區風速機率分布特性及年最大風速之模擬

機構典藏 > College of Engineering > Graduate Institute & Department of Civil Engineering > Thesis > Item 987654321/102474

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/102474

Title:	台灣地區風速機率分布特性及年最大風速之模擬
Other Titles:	The simulation of annual maximum wind and probability distribution of Taiwan
Authors:	李軒誠;Li, Hsuan-Chang
Contributors:	淡江大學土木工程學系碩士班羅元隆;Lo, Yuan-Lung
Keywords:	極值分佈;年最大風速;非高斯轉換;風工程;統計特性;Extreme value distribution;Non-Gaussian simulation;Annual maximum wind speed;Hermite-based transformation
Date:	2014
Issue Date:	2015-05-04 09:56:45 (UTC+8)
Abstract:	台灣地理位置處於西太平洋，一年四季飽受強風之侵襲如夏秋兩季之颱風與冬季有強烈的大陸冷氣團，由於飽受強風的侵襲，所以不論在做結構耐風設計，氣象預測，等相關從業人員都會需要取得當地的相關風力資訊，如此一來便可以減少強風所帶來的災害以及損失，由於觀測風速的資料量不足並不能準確的求取長期回歸期風速，因此需要使用風速模型來模擬產生人造風速。本文所使用之方法為年最大值法以及颱風風速法，選取資料為全台灣29個人工氣象觀測站所有年份之風速風向觀測資料，針對所選取之資料進行統計特性計算，並藉由計算逐年統計特性並觀察其逐年統計特性間之相關性及各測站逐年最大風速之趨勢，接著進行區域統計特性計算，並利用逐年統計特性及區域統計特性所得之結果做為基礎並進行年最大風速之模擬。本文使用GEVD極值分佈模型，並討論Gumbel、Frechet、Weibull三種極值分佈模型，給予各測站合適之極值分佈模型，並利用Hermite-based transformation 將高斯隨機變數轉換成非高斯隨機變數，並將前述之統計特性帶入Hermite-based transformation進行模擬，並選取最大值，視為模擬之年最大風速。短期觀測資料部分將所有風速風向歷時資料進行分類，將資料分為5年歷時資料及10年歷時資料並進行逐年統計特性計算及區域統計特性計算步驟並利用Hermite-based transformation模擬方法模擬短期觀測資料在年最大風速模擬之可行性。此外在颱風極值分佈的部分選出宜蘭、阿里山、新竹、高雄測站做為研究對象並搜尋其觀測年間所量測颱風風速，並將颱風風速之極值分佈與年最大風速之極值分佈進行比較。 Taiwan is an island located in the western Pacific Ocean and suffered from strong winds, such as tropical cyclones or typhoons in the summer season and monsoons in the winter season. To give a wind resistant design of a structure, specialists and structural engineers need to understand the local design wind speed in detail. However, due to the lack of observed wind speed data, numerical models are developed to simulate design wind speed. In this study, analysis on annual maximum wind speed and the peak wind speed during a typhoon event is executed. 29 meteorological stations in Taiwan are collected to understand the basic statistical parameters of 10-minute mean wind speeds. Further, the trend of annual maximum wind speed and the highly correlated third and fourth moments are found. To tell the extreme distribution of 29 stations, the GEVD extreme value distribution model is applied. Then a non-Gaussian simulation technique called Hermite-based transform is adopted for simulation of annual maximum wind speed for each station by inputting their statistical parameters. Results show that even the simulation is based on short term period record, the non-Gaussian simulation method to simulation the annual maximum wind speed provides fairly good agreement with the full-year observation. Finally in this study, the typhoon event is especially picked up among 4 stations, Yilan , Aliahsn , Hsinchu , Kaohsiung, to show the inconsistence with the distribution with annual maximum wind speeds. The GEVD model is also applied for discussion.
Appears in Collections:	[Graduate Institute & Department of Civil Engineering] Thesis

Files in This Item:

File	Size	Format
index.html	0Kb	HTML	233	View/Open

Loading...