ID | 119679 |
Title Alternative | ACCURACY IMPROVEMENT OF TYPHOON BUILDING DAMAGE PROBABILITY MODELS USING WEIBULL DISTRIBUTION BY BUILDING SHAPE POLYGON
|
Author |
Oda, Haruo
Kochi university
Mashiko, Wataru
Meteorological Research Institute
|
Keywords | 被害確率
最弱リンクモデル
建築面積率
Damage probability
Weakest link model
Building area ratio
|
Content Type |
Journal Article
|
Description | 最弱リンクモデルを適用した建物被害確率モデルの精度を向上させるため、建物ポリゴンを導入した。建物の密集度の指標として、評価メッシュ面積に対する建物面積の割合である建物面積比(ε)を使用した。また,被害建物ポリゴンを用いて,各メッシュの被害有無の判定を行った.2018年台風21号(T1821)における建物被害について検討した結果、説明変数としてεと最大風速の2乗を用いた被害確率モデルによる推定値は、実際の値とよく一致することがわかった。そして、構築したモデルによる推定値と実際の被害メッシュ分布とを比較すると、概ね対応することが示された。また、T1821の情報を基に構築したモデルを2019年の台風Faxaiの被害に適応して被害確率を推定し、被害の有無と比較したところ、今回提案したモデルが他の台風にも適用可能であることが示唆された。
|
Description Alternative | Building shape polygons were adopted to improve the accuracy of the building damage probability model applying the weakest link model. The building area ratio(ε), which is the ratio of building area to evaluation mesh area, was used as an indicator of building density. Additionally the damage or non-damage was determined for each mesh using the damaged building shape polygons. As results of investigation for building damage by Typhoon Jebi in 2018, it was found that the damage probability estimated from the damage probability model using the ε and the maximum wind speed squared as explanatory variables show good match with the actual values. Moreover, comparison of the estimated values by the constructed model with the actual damage mesh distribution showed that they generally corresponded to each other. In addition, the results of damage probability estimation of building damage by Typhoon Faxai in 2019 with the probability model constructed based on the building damages by Typhoon Jebi suggested that the proposed model was applicable to the other typhoons as well.
|
Journal Title |
Wind Engineering Research
|
ISSN | 24355429
24354384
|
Publisher | 日本風工学会
|
Volume | 27
|
Start Page | 109
|
End Page | 117
|
Published Date | 2022
|
EDB ID | |
DOI (Published Version) | |
URL ( Publisher's Version ) | |
FullText File | |
language |
jpn
|
TextVersion |
Publisher
|
departments |
Science and Technology
|