ارزيابي عملكرد روش هاي داده گرا در تخمين بار كل رسوبي رودخانه هاي شني

Evaluating the Performance of Data-Driven Methods for Prediction of Total Sediment Load in Gravel-Bed Rivers

انجام ﻣﻄﺎﻟﻌﺎت ﻓﺮاوان در راﺑﻄﻪ ﺑﺎ اﻧﺘﻘﺎل رﺳﻮب و ﺑﻪوﯾﮋه ﭘﯿﺶﺑﯿﻨﯽ اﯾﻦ ﭘﺪﯾﺪه ﻧﺸﺎﻧﮕﺮ اﻫﻤﯿﺖ ﺑﺴﯿﺎر ﺑﺎﻻي آن در ﻋﻠﻮم ﻣﺮﺗﺒﻂ ﺑﺎ ﻣﻬﻨﺪﺳﯽ و ﻣﺪﯾﺮﯾﺖ ﻣﻨﺎﺑﻊ آب ﻣﯽﺑﺎﺷﺪ. در اﯾﻦ ﺑﯿﻦ روشﻫﺎي ﻫﻮﺷﻤﻨﺪ در ﺳﺎلﻫﺎي اﺧﯿﺮ ﺑﻪ ﻃﻮر ﻣﻮﻓﻘﯿﺖآﻣﯿﺰي در ﭘﯿﺶﺑﯿﻨﯽ ﺑﺎر ﺑﺴﺘﺮ، ﺑﺎر ﻣﻌﻠﻖ و ﻫﻤﭽﻨﯿﻦ ﺑﺎر ﮐﻞ رﺳﻮب ﺑﻪ ﮐﺎر ﮔﺮﻓﺘﻪ ﺷﺪه اﺳﺖ. ﺑﺎ اﯾﻦ ﺣﺎل ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ ﮐﻤﺒﻮد دادهﻫﺎي ﻣﺮﺗﺒﻂ ﺑﻪ ﺑﺎر ﮐﻞ ﺑﺮاي رودﺧﺎﻧﻪﻫﺎي ﺑﺎ ﺑﺴﺘﺮ ﺷﻨﯽ، ﻣﻄﺎﻟﻌﺎت اﻧﺠﺎم ﮔﺮﻓﺘﻪ در اﯾﻦ راﺳﺘﺎ ﻣﺤﺪود ﻣﯽﺑﺎﺷﺪ. ﻫﺪف از ﺗﺤﻘﯿﻖ ﺣﺎﺿﺮ اﺳﺘﻔﺎده از روشﻫﺎي ﻗﺪرﺗﻤﻨﺪ ﻣﺎﺷﯿﻦ ﺑﺮدار ﭘﺸﺘﯿﺒﺎن، ﺷﺒﮑﻪ ﻋﺼﺒﯽ ﻣﺼﻨﻮﻋﯽ و رﮔﺮﺳﯿﻮن ﻓﺮآﯾﻨﺪ ﮔﺎوﺳﯽ ﺑﻪ ﻣﻨﻈﻮر ﭘﯿﺶﺑﯿﻨﯽ ﺑﺎر ﮐﻞ رﺳﻮب در 19 رودﺧﺎﻧﻪ ﺷﻨﯽ واﻗﻊ در اﯾﺎﻻتﻣﺘﺤﺪه آﻣﺮﯾﮑﺎ و ﻣﻘﺎﯾﺴﻪ ﻧﺘﺎﯾﺞ ﺣﺎﺻﻞ ﺑﺎ روشﻫﺎي ﮐﻼﺳﯿﮏ ﻣﺮﺳﻮم ﻣﯽﺑﺎﺷﺪ. ﺑﺪﯾﻦ ﻣﻨﻈﻮر ﭘﺎراﻣﺘﺮﻫﺎي ﺑﺪون ﺑﻌﺪ ﻣﺨﺘﻠﻔﯽ ﻣﺒﺘﻨﯽ ﺑﺮ ﻫﯿﺪروﻟﯿﮏ ﺟﺮﯾﺎن و ﻣﺸﺨﺼﺎت رﺳﻮب ﺗﻌﺮﯾﻒ و ﻋﻤﻠﮑﺮد روشﻫﺎي ﻣﺬﮐﻮر ﻣﻮرد ارزﯾﺎﺑﯽ ﻗﺮار ﮔﺮﻓﺖ. ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ ﻧﺘﺎﯾﺞ ﺑﻪ دﺳﺖ آﻣﺪه ﺷﺒﮑﻪ ﻋﺼﺒﯽ ﻣﺼﻨﻮﻋﯽ ﺑﺎ دارا ﺑﻮدن ﺿﺮﯾﺐ ﻫﻤﺒﺴﺘﮕﯽ و ﻣﻌﯿﺎر ﻧﺎش- ﺳﺎﺗﮑﯿﻒ ﺑﻪ ﺗﺮﺗﯿﺐ ﺑﺮاﺑﺮ ﺑﺎ 0/952 =R و 0/903 =NSE ﺑﺮاي دادهﻫﺎي ﺻﺤﺖﺳﻨﺠﯽ از ﻋﻤﻠﮑﺮد ﺑﻬﺘﺮي ﻧﺴﺒﺖ ﺑﻪ دو روش دﯾﮕﺮ ﺑﺮﺧﻮردار ﻣﯽﺑﺎﺷﺪ. در ﻧﻬﺎﯾﺖ ﺑﺎ اﻧﺠﺎم ﺗﺤﻠﯿﻞ ﺣﺴﺎﺳﯿﺖ، ﭘﺎراﻣﺘﺮ ﻧﺴﺒﺖ ﺳﺮﻋﺖ ﻣﺘﻮﺳﻂ ﺑﻪ ﺳﺮﻋﺖ ﺑﺮﺷﯽ ﺟﺮﯾﺎن ﺑﻪ ﻋﻨﻮان ﺗﺄﺛﯿﺮﮔﺬارﺗﺮﯾﻦ ﭘﺎراﻣﺘﺮ در ﭘﯿﺶﺑﯿﻨﯽ ﺑﺎر ﮐﻞ رﺳﻮب ﻣﻌﺮﻓﯽ ﺷﺪ.

Numerous studies on sediment transport, especially prediction of this phenomenon, indicate its high importance in the sciences related to engineering and water resources management. In recent years, intelligent methods have been applied successfully to predict bed, suspended and total sediment load. However, due to the lack of measured data, limited researches have been done to deal with prediction of total load in gravel-bed rivers. The aim of this study is to apply Support Vector Machine (SVM), Artificial Neural Network (ANN) and Gaussian Process Regression (GPR) to predict total sediment load for 19 gravel-bed rivers and to compare the obtained results with well- known classic methods. For this purpose, different non-dimensional parameters based on hydraulic condition and sediment characteristics were defined and the performance of these methods was evaluated. According to the obtained results, the ANN model with correlation coefficient of R =0.952 and Nash–Sutcliffe efficiency (NSE=0.903) showed a better performance as compared to the other methods. Finally, by performing sensitivity analysis, the ratio of mean flow to shear velocity was introduced as the most effective parameter in predicting total sediment load.