Radial Basis Neural Network and Particle Swarm Optimization-based equations for predicting the discharge capacity of triangular labyrinth weirs |
| |
Affiliation: | 1. Laboratory of Hydraulics, Hydrology and Glaciology (VAW), Swiss Federal Institute of Technology (ETH) Zurich, CH-8093 Zürich, Switzerland;2. Faculty of Engineering and Science, Universitetet i Adger, Kristiansand, Norway;3. CEris/Cehidro and DECivil, Instituto Superior Técnico, Universidade de Lisboa, Lisboa, Portugal;1. IHC MTI, Smitweg 6, 2961 AW Kinderdijk, The Netherlands;2. Delft University of Technology, Faculty of Mechanical, Maritime and Materials Engineering, Section of Dredging Engineering, Mekelweg 2, 2628 CD Delft, The Netherlands;3. Specialist Advisor Instrumentation, Deltares, PO Box 177, 2600 MH Delft, The Netherlands;1. Department of Chemical and Process Engineering, Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia, 46300 UKM Bangi, Selangor, Malaysia;2. Department of Mechanical Engineering, Graduate School of Science and Engineering, Tokyo Metropolitan University, 1-1 Minamioosawa-Osawa, Hachiouji 192–0397, Japan;3. Graduate School of Mechanical Engineering, Division of Artificial System Science, Chiba University, 1–33 Yayoi, Inage, Chiba 263–8522, Japan |
| |
Abstract: | Conventional weirs are utilized for controlling, measuring and adjusting the flow depth in hydraulic structures, such as those found in irrigation and drainage networks. Various weirs with modified shapes are utilized to increase the discharge capacity. The main goal of this study is to investigate the discharge coefficient (Cd) of triangular labyrinth weirs using soft computing methods. The performance of the Radial Basis Neural Network (RBNN) is compared with that of Multiple Nonlinear and Multiple Linear Particle Swarm Optimization (MNLPSO and MLPSO). Models developments are conducted using published experimental data from the literature. Comparing the RBNN, MLPSO and MNLPSO results obtained through these soft computing techniques with experimental data shows that all models perform well in predicting the discharge coefficient of a triangular labyrinth weir. Performance of the proposed approaches which demonstrated explicit equation given by MNLPSO model provided the discharge capacity with lower error (RMSE=0.0223) is compared with the MLPSO (RMSE=0.0346) and RBNN (RMSE=0.045) approaches. |
| |
Keywords: | Radial Basis Neural Network Particle Swarm Optimization Evolutionary algorithm Discharge coefficient Triangular labyrinth weir |
本文献已被 ScienceDirect 等数据库收录! |
|