نوع مقاله : مقاله عمران
نویسندگان
1 دانشکده مهندسی عمران، دانشگاه تبریز،
2 دانشکده مهندسی عمران، دانشگاه تبریز، تبریز، ایران
3 دانشکده مهندسی عمران، دانشگاه تبریز
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
The significant wave height is a highly important parameter in the design, analysis, and operation of marine structures. Therefore, predicting this parameter significantly contributes to improving the design and analysis of marine structures. Numerous models and methods, including empirical, semi-empirical, and numerical approaches, have been developed to predict significant wave height. In the present study, the SWAN model has been developed for shallow and semi-deep coastal waters. Wind speed data were used as input for the models. To calculate and predict significant wave height, three models including SWAN, Support Vector Machine (SVM), and Gene Expression Programming (GEP) were applied in the Anzali region. For this purpose, field data on wind and waves recorded by the Anzali buoy over a specific time period were utilized. After running the models and plotting the relevant graphs, the error rates of each model were calculated. Based on statistical evaluation criteria, it was concluded that the overall trend of significant wave height predicted by all three models shows acceptable agreement with the actual values recorded by the buoy. Additionally, both the Support Vector Machine and Gene Expression Programming models were able to predict significant wave height with high accuracy. A comparison of the results from the models revealed that the Support Vector Machine model, with statistical indices (Bias = 0.13, RMSE = 0.198, CC = 0.981, and CV = 0.147), estimated the target parameter more accurately than the Gene Expression Programming model. However, both models showed lower accuracy in simulating peak values.
کلیدواژهها [English]