Prediction of typhoon waves off the Port Taipei using artificial neural networks / 以類神經網路預測台北港海域的颱風波浪

碩士 / 國立臺灣海洋大學 / 海洋環境資訊學系 / 101 / This study developed predictive models for typhoon waves off the Port of Taipei. The model was based on back propagation neural network method. The present paper collected information of typhoons occurred between 2001 and 2007, and wave height measured near the Port of Taipei, by the Harbor and Marine Technology Center, Institute of Transportation, Ministry of Transportation. Typhoon information was obtained from the Central Weather Bureau and the Japan Meteorological Agency (JMA). After initial screening of all data, it was found that there were seventeen typhoons having sufficient typhoon parameters and wave height for model building. Among the 17 typhoons, 14 of them were used for training and 3 for validation. This study tested three models, whose input consisted of different parameters representing the strength of typhoon as well as the distances between the typhoon and the wave station and the bearings of the typhoons. The latter two were represented by their respective fuzzy membership function. All the input parameters include the data 2-6 hours previous to the time of the prediction. Tests of this study found that the model using the speed of typhoon center relative to the wave station, the distance of the typhoon center and the bearing of the typhoon and the maximum wind speed at the typhoon center as the input parameters has the best result. The validation runs showed that the mean correlation coefficient of the three validation typhoons between the predicted significant wave height and the measured wave height was 0.878, while the mean root mean square error was 52 cm. The percentage errors for the peak wave heights during the typhoon period between the models and measurements for the three typhoons were 4.7%, 3.2%, and 10.3 % respectively. The errors in the time of the peak wave height were 0, 7, and 2 hours respectively. The positive time error means that the peak wave height occurs later than the measured. The validation showed that the model constructed in this study can satisfactorily predict typhoon wave height, if the typhoon track did not pass through Taiwan. For these typhoons their wind fields were not affected by the land and high mountain ranges.

Identiferoai:union.ndltd.org:TW/101NTOU5276001
Date January 2013
CreatorsLo, Tzu-Hua, 駱姿樺
ContributorsTsai, Chen-Han, Tsai, Jen-Chih, 蔡政翰, 蔡仁智
Source SetsNational Digital Library of Theses and Dissertations in Taiwan
Languagezh-TW
Detected LanguageEnglish
Type學位論文 ; thesis
Format109

Page generated in 0.0625 seconds