کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
768200 1462713 2014 9 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Predicting optimum parameters of a protective spur dike using soft computing methodologies – A comparative study
موضوعات مرتبط
مهندسی و علوم پایه سایر رشته های مهندسی مکانیک محاسباتی
پیش نمایش صفحه اول مقاله
Predicting optimum parameters of a protective spur dike using soft computing methodologies – A comparative study
چکیده انگلیسی


• A protective spur dike is used to reduce scour depth around main spur dikes.
• SVR, ANFIS and ANN were compared in predicting the best parameters of a protective spur dike.
• Polynomial and radial basis function are applied as the kernel function of SVR.
• SVR_rbf (radial basis function) produced better results than other developed models.
• Optimized parameters of the protective spur dike are presented.

This study proposes a new approach for determining the optimum parameters of a protective spur dike to control scour around existing spur dikes. Several parameters of a protective spur dike were studied to determine their optimum values, including the angle of the protective spur dike relative to the flume wall, its length, and its distance from the main spur dikes, flow intensity, and the diameters of the sediment. To build an effective prediction model, the polynomial and radial basis function are applied as the kernel function of support vector regression (SVR) for prediction of protective spur dike parameters for scour controlling around spur dikes and their performance were compared to Adaptive Neuro Fuzzy System (ANFIS), and Adaptive Neural Network (ANN). Instead of minimizing the observed training error, Polynomial-based SVR (SVR_poly) and radial basis function SVR (SVR_rbf) attempt to minimize the generalization error bound so as to achieve generalized performance. The performance of proposed optimizer was confirmed by experimental results. The results showed that an improvement in predictive accuracy and capability of generalization based SVR can serve as a promising alternative for existing prediction models.

ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Computers & Fluids - Volume 97, 25 June 2014, Pages 168–176
نویسندگان
, , , , , ,