應用WRF與LES耦合模型分析真實颱風邊界層中之風場特性

Abstract

台灣鄰近太平洋且位於易受颱風侵擾的中緯度地區，每年至少受到3至4次的颱風侵襲，其所造成的極端風負載往往是離岸風機結構損壞的主因之一，像是2015年的蘇迪勒颱風以及2016的梅姬颱風都有造成風機嚴重毀損甚至倒塌的案例，因此如何準確評估颱風邊界層中之風場特性為台灣發展離岸風電的重要課題。 本研究應用了中尺度的天氣研究預報模式(Weather Research and Forecasting, WRF)與微尺度的大渦流模擬(large-eddy simulation, LES)，透過先進的耦合方法結合兩者優點，將中尺度大氣參數以及微尺度的紊流變化皆包含於模型之中，以求得更高解析度和更貼近真實風場的模擬數據。 中尺度模擬使用的WRF模式是一種數值天氣預報(numerical weather prediction, NWP)系統，其使用大氣和海洋的數學模型，透過現有的大氣科學物理原理，模擬運算三維空間中的天氣變化，得到數小時或是數天的天氣預報。微尺度風場生成則是使用美國國家再生能源實驗室(National Renewable Energy Laboratory, NREL)所開發的開放原始碼程式SOWFA進行大渦流模擬，能考量大氣邊界層中的溫度對流效應以捕捉紊流的實際特性。 藉由日本氣象廳(JMA)及台灣中央氣象局(CWB)提供的颱風資料，本文分析了近年有經過台灣且風速相對較強勁的颱風，最終挑選了2017年過境台灣北部的尼莎(Nesat)颱風做為分析對象，並將颱風過境台灣的過程分成前、中、後三個時間段，觀察颱風在不同階段所表現的風場變化，依所設計的流程模擬後得出風場結果，並與台灣離岸架設之海氣象觀測塔所得實測資料進行比對與探討。 模擬結果顯示中尺度的颱風風場能描述出尼莎颱風在不同階段大致的特性，模擬路徑也能與JMA公布之最佳路徑吻合，但WRF模擬的颱風過境之風速曲線相對於實測資料較為平緩，無法描述風速急降至颱風眼低速區的數值變化，且WRF模擬略為低估了颱風極端風速數值，最大差距低於實測資料約8 m/s。微尺度的耦合風場也能合理地描述紊流特性與結構，模擬之紊流強度在非風眼時段與實測資料相比平均誤差約在20 %以內，但模擬風場的紊流能量與風速標準差在颱風過境後則明顯較小於量測值。 總結而言，以中尺度與微尺度的耦合模型評估颱風邊界層具有一定程度的準確度與可行性，且透過縮小空間尺度以觀察風機數十公尺範圍內的紊流變化也具有顯著的結果，此耦合後的風場預計可用於評估離岸風機於颱風極端風速下的負載情況。

Taiwan is located in the mid-latitude region of the western Pacific Ocean, a favorable area of typhoon impacts. With an average of 3 to 4 typhoons each year, the extreme loads caused by typhoons could lead to severe damage to the structures of offshore wind turbines. Notable examples include the devastating impacts of typhoon Soudelor in 2015 and typhoon Meranti in 2016, which resulted in severe damage and even the collapse of wind turbine structures. Therefore, how to accurately estimate the wind characteristics of typhoon boundary layers is of importance for the development of offshore wind energy in Taiwan. In this study, we apply the Weather Research and Forecasting (WRF) meso-scale model coupled with micro-scale large-eddy simulation (LES) to consider the atmospheric parameters and generate high-resolution fields which account for more details of small-scale turbulences. WRF model is a numerical weather prediction (NWP) model that utilizes physical models for solving problems of the atmosphere and ocean. It simulates the evolution of 3-D atmospheric fields based on atmospheric dynamics and physics and thus can provide weather forecasts for several days to weeks. The micro-scale simulations are performed using the simulator for wind farm applications (SOWFA) program, which is an open-source software developed by the National Renewable Energy Laboratory (NREL) in the United States. SOWFA employs LES to simulate turbulent flows, considering the effects of temperature convection in the atmospheric boundary layer. Typhoon Nesat in 2017, which passed the northern coast of Taiwan, is selected based on the typhoon data provided by the Japan Meteorological Agency (JMA) and the Central Weather Bureau (CWB) of Taiwan. We divide the typhoon passage in Taiwan into three time periods (pre-typhoon, during typhoon, and post-typhoon), and observe the changes of the wind field during different periods. The wind fields are simulated using a proposed framework and then compared to the observational data of the met mast in the Taiwan’s offshore wind farm for further discussions. The simulation results show that the meso-scale typhoon wind fields can capture the general characteristics of typhoon Nesat during different periods, and the simulation track is also similar to the typhoon best track from JMA. However, the wind speed curve of the typhoon passage simulated by WRF appears relatively smooth compared to the measured data, lacking the drastic drop of wind speed due to the storm eye. Additionally, the WRF simulation slightly underestimates the values of extreme wind speeds associated with the typhoon, with a maximum difference of up to 8 m/s compared to the measurements. The coupled micro-scale model successfully generates wind speed data that matches the observed trends and reasonably describes the characteristics and structure of turbulence. Except for the time period of storm eye, the mean error between the simulated and observational turbulence intensity values is less than 20 %. However, the simulated turbulence kinetic energy and the standard deviation of wind speed are seen smaller than the measured data. In sum, the research results demonstrate that the coupled model of meso- and micro-scale provides a certain level of accuracy and feasibility in evaluating the typhoon boundary layer. Moreover, significant findings have been observed by narrowing down the spatial scale to observe turbulence variations within a range of several tens of meters around wind turbines. The coupled wind field is expected to be useful for assessing the load of offshore wind turbines under extreme wind speeds associated with typhoons.