Accuracy and Uncertainty Analysis of Intelligent Techniques for Predicting the Longitudinal Dispersion Coefficient in Rivers

آب و فاضلاب

Abbas Akbarzadeh¹  Amir Khakpour²  Roohollah Noori³  Ashkan Farokhnia⁴  Mohammad Salman Sabahi⁵ 

[1] Assist. Prof., Water Research Institute, Ministry of Energy, Tehran;Management Director of CELCO., Ph.D. Student of Environmental Eng., Dept. of Environmental, Tehran University;Member of Water Research Institute, Ministry of Energy, Ph.D. Student of Environmental Eng., Dept. of Environmental Eng., Tehran University;Member of Water Research Institute, Ministry of Energy, Ph.D. Student of Hydraulic Structures Eng., Tarbiat Modarres University, Tehran.;Ph.D. Student of Environmental Eng., Dept. of Environmental, Tehran University;

关键词: Monte-Carlo;    Longitudinal dispersion coefficient;    neural network;    adaptive neuro-fuzzy inference system;    Bound Wide Factor;   

DOI  :  

来源: DOAJ

【 摘 要 】

Accurate prediction of longitudinal dispersion coefficient (LDC) can be useful for the determination of pollutants concentration distribution in natural rivers. However, the uncertainty associated with the results obtained from forecasting models has a negative effect on pollutant management in water resources. In this research, appropriate models are first developed using ANN and ANFIS techniques to predict the LDC in natural streams. Then, an uncertainty analysis is performed for ANN and ANFIS models based on Monte-Carlo simulation. The input parameters of the models are related to hydraulic variables and stream geometry. Results indicate that ANN is a suitable model for predicting the LDC, but it is also associated with a high level of uncertainty. However, results of uncertainty analysis show that ANFIS model has less uncertainty; i.e. it is the best model for forecasting satisfactorily the LDC in natural streams.

【 授权许可】

Unknown