A Study on Building a Ships Escape Management System Using Neural Network
| Autor: | 張惠華 |
|---|---|
| Rok vydání: | 2007 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 95 The official records of ship escape from Hua-Lien harbor during typhoon period each year were collected from 1996 to 2005 in this research. Criteria of ship escape are classified into five groups depending on the resulting factors, such as typhoon’s scale, path, wind speed, position of typhoon’s center and the distance between the typhoon and the interest point. The relationship between the criteria of ship escape and the resulting factors is established by artificial neural network (NN) which provides an alert model of ship escape. The accuracy of the proposed NN model is examined using recorded events of ship escape in three typhoons. Fair agreements in calibrating and verifying stages show that the proposed model can be applicable for real operation in the future.Another object of this research is to remedy the unavailable data of waves and tides that commonly occur during typhoon periods due to lost or failure of equipments. An NN model with high accuracy was developed to simulate typhoon’s wave heights. Some typhoons that occur in this year are examined to evaluate the forecasting capacity of this model. In the future, the proposed model can be used to simulate wave heights for all typhoons possibly approaching to Taiwan. Storm surge deviations are separated from tidal data by traditional harmonic method. Some separated storm surge deviation is related to wind speed, distance between the interest point and typhoon’s eye. An initial NN model is developed to simulate storm surge deviation. The model of predicting storm surge will be improved by using more data input in the research next year. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|