Please use this identifier to cite or link to this item:
Title: 風壓負荷下樹傾斜值的預測技術
Prediction Methods of the Tilt of Trees under Wind Loading
Authors: 林均翰
Lin, Chun-Han
Jiang, Zheng-Quan
Keywords: 風壓負荷
Wind load
Tree tilt
Damping oscillation
Prediction model
Issue Date: 2018
Abstract: 都市內的樹木傾倒會造成居民與財產損害,其中風災是造成樹木傾倒的主因,若能預測樹木在風災中的傾斜情況並進行適當處理,將可有效的降低損害。本論文著重於研發樹木在風壓負荷下所造成傾斜角度的預測技術。我們首先測量實際風壓負荷下的樹木傾斜加速度。接著,透過將加速度資料換為球座標系獲得樹木的傾斜角度與方向。然後,配合氣象站的風速和風向資料,根據樹木擺動的特性,取得風速與相對應的傾斜角度的資訊。隨後,根據此資訊,建立一個決策樹非線性迴歸模型,來預測在各風向的不同風速下樹木的傾斜角度。最後,我們以實際量測資料驗證此預測模型。實驗結果顯示所開發的預測模型在預測傾斜角度的準確率上有良好的效果,並獲得對將來的研究方向與系統實作有助的資訊。
Tilting or rotation of trees in the city will cause damage to residents and property, the main cause of the tilting or rotation of trees was wind disaster. If we can predict the tilting of trees in the wind disaster and carry out appropriate treatment, it will effectively reduce the damage. This paper focuses on the development of technology for predicting the tilting angle of trees under wind load. We first measure the tree's tilting acceleration under actual wind load. Next, the angle and direction of the trees are obtained by converting acceleration data into a sphere coordinate system. Then, according to the wind speed and wind direction information of the weather station, the wind speed and the corresponding tilting angle information are obtained based on the characteristics of the tree oscillation. Subsequently, based on this information, a non-linear regression model of the decision tree was established to predict the tilting angle of the trees at different wind speeds in different directions. Finally, we validate this prediction model with actual measurement data. The experimental results show that the developed predictive model has a good performance in predicting the accuracy of the tilting angle and obtains information that will help future research directions and system implementation.
Other Identifiers: G060447039S
Appears in Collections:學位論文

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.