碩士 / 國立中興大學 / 土木工程學系所 / 98 / The determination of wave overtopping rate is an essential in the design of a coastal structure. An exact mathematical description of the wave overtopping process for coastal dikes or seawalls seems not possible due to the stochastic nature of wave breaking, wave run-up and the various factors influencing on the wave overtopping process. Therefore, wave overtopping rates for coastal dikes or seawalls were mainly determined by empirical formulas derived from experimental or field investigations.
In this paper, the artificial neural network model (ANN) is applied to the prediction of overtopping rates at sloped seawalls. The data of wave overtopping rates of sloped seawalls selected from the database of CLASH are adopted for the learning and testing in the present ANN model. This paper first configures the optimum architecture of the ANN using different combinations of input factors. The results show that the ANN can achieve satisfactory prediction of the dimensionless wave overtopping rate at a sloped seawall from three input parameters; they are the relative freeboard, the surf similarity parameter and the relative water depth at toe. It is found that the accuracy is better if the learning of the ANN is based on the overtopping at the individual slope of seawall, rather than learning from all slopes of seawalls. As comparing with the prediction using empirical formula by Van der Meer et al. (1994), the results show that the present ANN model obtains better prediction.
| Identifer | oai:union.ndltd.org:TW/098NCHU5015091 |
| Date | January 2010 |
| Creators | Yi-Ting Lee, 李宜庭 |
| Contributors | Ching-Piao Tsai, 蔡清標 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 59 |
Page generated in 0.006 seconds