Discharge Coefficient Estimation of Sharp-Crested Curved-Plan Spillways using Soft Computing Approaches

Document Type : Original Article

Authors

1 Assistant Professor, Civil Engineering Department, Quchan University of Technology, Quchan, Iran

2 Head of Applied Research Department, Khorasan Razavi Regional Water Company, Iran

Abstract

The use of labyrinth and side weirs and the design of curved-plan form weirs are inevitable in some cases due to design and implementation limitations. The present study proposes dimensionless equations based on the vertex angle and the head ratio over the spillway to the spillway height to estimate the discharge coefficient of such weirs using the gene expression programming (GEP). Moreover, several other soft computing techniques including adaptive neuro-fuzzy inference system (ANFIS), artificial neural network (ANN), and support vector machine (SVM) approaches have been adopted. Three performance evaluation criteria including Nash-Sutcliffe index (NS), Root Mean Square Error (RMSE), and Mean Normalized Error (MNE) have been used for comparison purposes. To calibrate and validate the proposed models, available experimental data were collected and reanalyzed. A comparison of the calculated and experimental discharge coefficients of curved weirs revealed that the proposed equation had satisfactory accuracy. For the training dataset, the ANFIS, SVM ANN, and GEP have the NS values 0.971, 0.965, 0.962, 0.956, respectively. Moreover, the results of these models were compared with available equation of the prediction of the discharge coefficients of sharp-crested curved-plan form weirs, and it was find that the ANFIS and GEP respectively give 609% and 500% improved results than the existing model in terms of RMSE.

Keywords

Main Subjects

Files

History

  • Receive Date: 05 November 2023
  • Revise Date: 05 December 2023
  • Accept Date: 23 December 2023

Share

How to cite

Statistics