: Depth averaged velocity prediction in Skewed Compound Channel with Inclined Floodplains using soft computing techniques based on group method of data handling model - دانشکده فنی و مهندسی
: Depth averaged velocity prediction in Skewed Compound Channel with Inclined Floodplains using soft computing techniques based on group method of data handling model
نوع: Type: Thesis
مقطع: Segment: masters
عنوان: Title: : Depth averaged velocity prediction in Skewed Compound Channel with Inclined Floodplains using soft computing techniques based on group method of data handling model
ارائه دهنده: Provider: Ali nikpoor Nikche
اساتید راهنما: Supervisors: Dr. Bahram Rezaei
اساتید مشاور: Advisory Professors: Dr. Abbas Parsaei
اساتید ممتحن یا داور: Examining professors or referees: Dr. Jalal Sadeghian, Dr. Jalal Akbari
زمان و تاریخ ارائه: Time and date of presentation: 2025
مکان ارائه: Place of presentation: 33
چکیده: Abstract: Reliable prediction and modeling of depth averaged velocity is essential for boundary shear stress and flow discharge, which can be calculated using the Darcy-Weisbach equation and numerical integration. The importance of accurate modeling of flow hydraulic variables in compound channels is crucial in river engineering designs such as determining the bed boundary, bank protection, and flood routing, for which researchers have used analytical methods and soft computing models. In this study, the depth averaged velocity in skewed compound channels with inclined floodplains was modeled and predicted using soft computing models including the Artificial Neural Network (ANN) model and the Group Method of Data Handling (GMDH) model. The dimensionless variables used in this study are relative depth, ratio of discharge to gravity acceleration and channel depth, , floodplain slope, relative distance, bed slope and angle of deviation, relative width, distance from the center of the compound channel cross-section in the main channel and distance from the center of the compound channel cross-section in the floodplain . The results showed that the models used showed high accuracy in predicting the depth averaged velocity, and the artificial neural network model provided more accurate results in estimating the depth averaged velocity than the group method of data handling model, considering the ratios (R2 = 0.979, RMSE = 0.024, MAPE = 5.674%) for the model training stage and (R2 = 0.976, RMSE = 0.024, MAPE = 4.146%) for the testing stage. Also, by examining the results of the sensitivity analysis, it was determined that the most influential variables in estimating the depth averaged velocity in skewed compound channels with inclined floodplains for the ANN model are the relative distance variables\left(\mathbit{S}_{\mathbit{fp}}\right), distance from the center of the compound channel cross-section in the main channel {(\mathbit{Y}}_{\mathbit{n}\mathbf{1}}), distance from the center of the compound channel cross-section in the floodplain (\mathbit{Y}_{\mathbit{n}\mathbf{2}}),the deviation angle \left(\mathbit{\theta}\right) and the bed slope \left(\mathbit{S}_\mathbf{0}\ \right), respectively, and for the GMDH model, the deviation angle variables \left(\mathbit{\theta}\right) and distance from the center of the compound channel cross-section in the main channel {(\mathbit{Y}}_{\mathbit{n}\mathbf{1}}).