台灣地區波浪預警系統之應用研究

Abstract

本研究是以類神經網路建立臺灣海海域的颱風波浪推算模式，並分析颱風的規模與運動行為對於台灣海域波浪波高及週期的影響，使颱風波浪推算模式達到預警及模擬災害的目的。此模式的推算核心是由類神經網路聯結所組成，利用具有不同功能的類神經網路分工推算處理以提高模式的推算品質，並合理正確地推算颱風波浪之波高及週期。

Neural network will be applied to construct a forecasting model for typhoon waves around Taiwan. Wave height and period resulted from typhoons were statistically analyzed. The predicted typhoon wave from typhoon information can provide reference to the alarming system or the input of potential disaster model. The model was comprised of a series of neural network that has different function. The present model has high capability of accurately forecasting wave height and period of typhoon waves.