Document Type : پژوهشی
Authors
1 2PhD in Water Sciences and Engineering, Department of Soil Conservation and Watershed Management, Lorestan Province Agriculture and Natural Resources Research and Education Center, Agricultural Research, Education and Extension Organization,
2 Research Assistant Professor, Department of Soil Protection and Watershed Management, Lorestan Province Agriculture and Natural Resources Research and Education Center, Agricultural Research, Education and Extension Organization, Khorramabad, Iran
3 PhD in Watershed Science and Engineering, Soil Conservation and Watershed Management Research Department, Agricultural and Natural Resources Research and Education Center of Lorestan Province, Agricultural Research, Education and Extension Organizati
Abstract
Measuring river discharge has always been a fundamental challenge in river management. The use of precise instruments for its calculation is essential. Numerical, analytical, artificial intelligence, and empirical methods are the most common approaches for measuring daily river flow. In this research, a hybrid intelligent model based on the Support Vector Regression (SVR) model approach has been developed for river discharge simulation. To this end, three optimization algorithms, including Wavelet, innovative gunner (AIG), and Black Widow Optimization (BWO), were used to simulate river discharge. For modeling, statistical data and information from hydrometric stations of the Karkheh River basin, including Cham Anjir, Kashkan, Pol-e Zal, and Jelogir, were used as a case study over four combined scenarios of input parameters from 1993 to 2023. To evaluate the performance of the models, the evaluation criteria of the correlation coefficient, root mean square error, mean absolute error, and Nash-Sutcliffe efficiency coefficient were used. Also, scatter, relative error, and Taylor diagrams were used to analyze the results of the models. The results showed that combined scenarios in the studied models improve the performance of the model. Also, the results from the evaluation criteria showed that the Support Vector Regression-Wavelet model has a correlation coefficient of 0.941-0.974, root mean square error (m3/s) of 0.022-0.054, mean absolute error (m3/s) of 0.025-0.011, and Nash-Sutcliffe efficiency coefficient of 0.962-0.986 in the validation stage. Overall, the results showed that the use of intelligent models based on the Support Vector Regression approach can be an effective approach in river engineering.
Keywords
Main Subjects