A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Modeling dimensionless longitudinal dispersion coefficient in natural streams using artificial intelligence methods
Abstract Learning the Longitudinal Dispersion (LD) mechanism in natural channel is vitally important to be able to control water pollution and to prevent different stratification in flow that crucial for water resources conservation for both human and aquatic life. Many related studies can be found in the existing literature. However, almost all studies aim to investigate the mechanism of the LD in natural channels or to model dimensional LD coefficient. The main goal of this work is to develop three models based on different artificial neural network techniques to predict dimensionless (not dimensional) LD coefficients in natural channels. Another goal of this study is to present a large and critical assessment on the existing studies made on both dimensional and dimensionless LD. The data sets obtained from the literature concerning more than 30 rivers at different times in the United States of America, include the depth, the width, and the mean cross-sectional velocity of the flow, shear velocity, and dimensionless longitudinal dispersion coefficient. The results have been compared with the data at hand, a fuzzy based model, and seven conventional equations proposed in literature by using statistical magnitudes, error modes, and contour map method. It is observed that the feed forward neural network yields the best reliable results.
Modeling dimensionless longitudinal dispersion coefficient in natural streams using artificial intelligence methods
Abstract Learning the Longitudinal Dispersion (LD) mechanism in natural channel is vitally important to be able to control water pollution and to prevent different stratification in flow that crucial for water resources conservation for both human and aquatic life. Many related studies can be found in the existing literature. However, almost all studies aim to investigate the mechanism of the LD in natural channels or to model dimensional LD coefficient. The main goal of this work is to develop three models based on different artificial neural network techniques to predict dimensionless (not dimensional) LD coefficients in natural channels. Another goal of this study is to present a large and critical assessment on the existing studies made on both dimensional and dimensionless LD. The data sets obtained from the literature concerning more than 30 rivers at different times in the United States of America, include the depth, the width, and the mean cross-sectional velocity of the flow, shear velocity, and dimensionless longitudinal dispersion coefficient. The results have been compared with the data at hand, a fuzzy based model, and seven conventional equations proposed in literature by using statistical magnitudes, error modes, and contour map method. It is observed that the feed forward neural network yields the best reliable results.
Modeling dimensionless longitudinal dispersion coefficient in natural streams using artificial intelligence methods
Toprak, Z. Fuat (author) / Hamidi, Nizamettin (author) / Kisi, Ozgur (author) / Gerger, Reşit (author)
KSCE Journal of Civil Engineering ; 18 ; 718-730
2013-09-13
13 pages
Article (Journal)
Electronic Resource
English
Predicting Longitudinal Dispersion Coefficient in Natural Streams
| Online Contents | 1998
Predicting Longitudinal Dispersion Coefficient in Natural Streams
| British Library Online Contents | 1998
Predicting Longitudinal Dispersion Coefficient in Natural Streams by Artificial Neural Network
| Online Contents | 2005
Predicting Longitudinal Dispersion Coefficient in Natural Streams by Artificial Neural Network
| British Library Online Contents | 2005