應用混合式類神經網路模式推算季節風波浪之研究

Hybrid monsoon-wave calculation related to Artificial Neural Network

翁瑞嘉(Jui-Chia Weng)

指導教授 ： 張憲國

國立交通大學/工學院/土木工程系所/碩士(2009年)

https://doi.org/10.6842/NCTU.2009.00951

摘要

本文使用交通部運輸研究所港灣技術研究中心提供之2004-2006年臺北港測站波浪資料，配合美國國家環境預報中心與國家大氣研究中心所提供之模擬風速及Mike 21模式產生的模擬波高，分析季節風與示性波浪之特性，以瞭解之間的關係，進而應用類神經網路模式，建立季節風波浪推算模式，來推算示性波高。
本文以模擬波高、模擬風速與風向等參數作為架構季節風推算波浪之模式之輸入值，以建立模式季節風波浪推算。其模式所推測波高與實測值之相關係數平方皆在0.55以上，其推測波高之均方根誤差值則為0.38米以下，相對偏差的絕對值約在0.09以下。應用此模式補遺之後進行波浪推算，得到其相關係數平方皆在0.56以上，均方根誤差值則為0.35米左右，相對偏差的絕對值約在0.06以下。模式加入兩小時延時的輸入值，從推估結果顯示，三種統計指標的變化量皆不到0.05，證實延時效應對於模式預測精確性影響不大。整體結果証實本模式在季節風波浪推估具有不錯之精確度，未來可應用提供海上育樂活動的民眾及漁民船隻作業參考應用。

關鍵字

類神經網路 ； 季節風 ； 波浪推算

並列摘要

This paper investigates the relationship between monsoon and corresponding wave data observed by Harbor and Marine Technology Center during year 2004-2006 at Taipei harbor, and sets up Artificial Neural Network (ANN) model to calculate significant wave heights. Alternative wind data calculated by National Centers for Environmental Prediction and National Centers for Atmospheric Research and simulated wave height by SW model in Mike 21 software are used to establish a hybrid artificial neural model for accurate calculating monsoon waves.
Three parameters, simulated wave height, wind velocity and wind direction, are determined to be inputs in the ANN model. The proposed model has high accuracy of calculating waves by a R-square exceeding 0.55 and by a Root Mean Square Error less than 0.38m and by a Bias less than 0.09m. After the missing wave data were remedied by the original ANN model, the evaluation on calculated waves shows that R2 exceeds 0.56 and RMSE and Bias are less than 0.35m and 0.06m, respectively. Time delay having very slight effect on calculation was examined this paper. Therefore, this proposed model is applicable for a reference of marine activity and ships work due to fast and accurately simulating waves.

並列關鍵字

ANN ； monsoon ； wave calculation

參考文獻

6. Browne, M., Castelle, B. Strauss, D., Tomlinson, R., Blumenstein, M., and Lane, C., (2007) “Near-shore swell estimation from a global wind-wave model: Spectral process, linear, and artificial neural network models,” Coastal Engineering, Vol. 54, pp. 445-460.

7. Chang, H.K., and Chien, W.A., (2006a) “Neural network with multi-trend simulating transfer function for forecasting typhoon wave. ” Advances in Engineering Software, 37, 184–194.

8. Chang, H.K. and Chien, W.A., (2006b) “A fuzzy-neural hybrid system of simulating typhoon waves,” Coastal Engineering, Vol. 53, pp. 737-748.

9. Deo, M.C., and Sridhar Naidu, C., (1999) “Real time wave forecasting using neural networks,” Ocean Engineering, Vol. 26, pp. 191-203.

10. Deo, M.C., and Kiran Kumar, N., (2000) “Interpolation of wave heights,” Ocean Engineering, Vol. 27, pp. 907-919.