類神經網路的位置因子輸入對颱風波浪推算模式的精度提升探討

Accuracy promotion of ANN typhoon's wave model with an extra position factor

吳岳潤(Yueh-Jun Wu)

指導教授 ： 張憲國

國立交通大學/工學院/土木工程系所/碩士(2014年)

https://doi.org/10.6842/NCTU.2014.00511

摘要

本研究使用位置因子輸入於類神經網路來建立颱風波浪推算模式，主要目的希望能提升往昔的類神經網路模式的預測精準度。本研究使用的風速資料點為臺北港的觀測樁2009至2013年而波浪資料為2000年至2012年。
往昔的基本類神經網路颱風波浪推算模式的主要輸入參數為目標點距離(D)、目標點方位角(θ1)、颱風侵襲角(θ3)、目標點風速(V10)以此參數為基礎建立的模式稱為ANN-B；另外以歸屬函數轉換方位角(λ1)及歸屬函數轉換侵襲角(λ3)取代原本ANN-B的輸入值而建構出ANN-M模式。本文新提出波高空間資訊參數(Ch)及利用回歸公式之修正風速(Vc)或風速類神經模式推算風速(Vann)等輸入參數分別建立ANN-V及ANN-P兩個模式。最後以四個檢定指標及推算波高峰值差異比較4個模式推算未學習的颱風波高來比較模式推算能力，結果以ANN-B與ANN-P的推算結果較差，以ANN-V模式推算結果最佳。
本模式有可接受的精度及快速推算的能力，可應用颱風來襲前的即時波高預測，並可提供相關單位來管理海上活動及防災預浪的決策參考。

關鍵字

類神經網路 ； 颱風波浪

並列摘要

An Neural Network (ANN) model imposed by position factor was developed to estimate better typhoon waves than the original ANN model. Wave data observed by the Harbor and Marine Technology Center from 2000 to 2012 at the Taipei port and wind data from 2009 to 2013 and typhoon data collected by JMA RSMC-Tokyo Center were collected to train the proposed ANN model.
The input parameters of original ANN-B model includes the distance from typhoon center to the interesting point (D), the azimuth between typhoon center and the interesting point (θ1), position angle in the typhoon (θ3), the wind velocity of the interesting point (V). The input parameters θ1 and θ3 in the ANN-B model were transformed by the Gauss member functions into λ1 and λ3. The corresponding model is called ANN-M model. Two new ANN-V and ANN-P models were established by an alternative input parameter, wave height position factor (Ch), instead of original inputsθ1 and θ3 and corrected local wind velocity (Vc) which is obtained by a regression formula between wind velocity and wave heights or calculated wind velocity (Vann) which is obtained by an ANN-W wind model. Model accuracy of estimating wave heights of untrained typhoons is examined by comparing four assessment indexes. The ANN-V model is examined the best among four models. The ANN-B and ANN-M models are worse for forecasting typhoon’s waves than the other two models.
The proposed ANN-V wave model is examined to have high accuracy on real-time calculating typhoon waves. Therefore, the proposed model can be applied to provide wave information of a marine warning system for navigation and marine activities.

並列關鍵字

ANN ； Typhoon's wave

參考文獻

4. 梁乃匡(2010)，「颱風湧浪預報經驗法的探討」，海洋工程學刊 第10卷 第2期 第87-120頁

5. 林賢銘(2011) “結合歸屬函數之類神經網路颱風波浪推算模式”，國立交通大學土木工程研究所碩士論文。

8. Camargo, S. J., A. W. Robertson, S. J. Gaffney, P. Smyth, and M. Ghil, (2007a) “Cluster analysis of typhoon tracks. Part I: General properties.” J. Climate, 20, 3635–3653.

9. Camargo, S. J., A. W. Robertson, S. J. Gaffney, P. Smyth, and M. Ghil (2007b) “Cluster analysis of typhoon tracks. Part II: Large-scale circulation and ENSO,” J. Climate, 20, 3654-3676.

10. Deo, M.C., Sridhar Naidu, C. (1999) “Real time wave forecasting using neural networks,” Ocean Engineering,Vol. 26, No. 3, pp. 191-203.