Geomorphology
Saeid Roustami; Babak Shahinejad; Hojatolah Younesi; Hassan Torabipoudeh; Reza Dehghani
Abstract
Flood is one of the natural phenomena that causes a lot of human and financial losses in the world every year and creates many problems for the economic and social development of countries. Therefore, in order to reduce the damage, control and guidance of this phenomenon, estimating flood discharge and ...
Read More
Flood is one of the natural phenomena that causes a lot of human and financial losses in the world every year and creates many problems for the economic and social development of countries. Therefore, in order to reduce the damage, control and guidance of this phenomenon, estimating flood discharge and identifying the factors affecting it is very important. In this study, in order to estimate the flood discharge of Kashkan catchment located in Lorestan province, new hybrid artificial intelligence models including artificial neural network - innovative gunner, artificial neural network - black widow spider and artificial neural network - chicken crowding during the period 1300-1400 were used. To evaluate the simulation performance, statistical indices of determination coefficient (R2), absolute mean error (MAE), Nash-Sutcliffe productivity coefficient (NSE), bias percentage (PBIAS) were used. The results showed that hybrid artificial intelligence models improve the performance of the single model. The results showed that the artificial neural network- innovative gunner model has more accuracy and less error than other models. Overall, the results showed that the use of hybrid artificial intelligence models is effective in estimating flood discharge and can be considered as a suitable and rapid solution in water resources management.
Ata Allah Nadiri; Saeed Yousefzadeh
Volume 4, Issue 10 , June 2017, , Pages 21-40
Abstract
An accurate estimation of the hydrogeological parameters such as hydraulic conductivity, which is essential for careful management and protection of groundwater resources, is an important part of hydrogeological studies. Various field and laboratory methods, generally done using hydrogeological data, ...
Read More
An accurate estimation of the hydrogeological parameters such as hydraulic conductivity, which is essential for careful management and protection of groundwater resources, is an important part of hydrogeological studies. Various field and laboratory methods, generally done using hydrogeological data, have already been proposed for estimating hydraulic conductivity. One of the best and the most complete methods is the field pumping test which is very time-consuming and expensive. In addition, hydrogeological parameters estimated by it have an inherent uncertainty. In this study, we tried to use artificial intelligence methods, widely considered in recent years, such as artificial neural network (ANN), mamdani fuzzy logic(MFL), sugeno fuzzy logic(SFL), and adoptive neuro-fuzzy inference system (ANFIS) for the estimation of the hydraulic conductivity. In this study, for the accurate estimation of the hydraulic conductivity in Maraghe-Bonab plain by these models, geophysical and hydrogeological data were used as models' inputs. Their results were compared with the evaluation criteria, and the best model based on the RMSE was selected. Accordingly, the ANFIS model, compared to other models, with an RMSE of 1.12 in the test phase has high power in the estimation of the hydraulic conductivity. Radius of clustering, number of fuzzy rules, and number of clusters are very important in fuzzy and neuro-fuzzy models. Radius of clustering in the ANFIS model, based on the minimum RMSE amount, was equal to 0.4 and the numbers of clusters, based on if-then fuzzy rules, was 9. The methods presented in this study, which demonstrated superior performance in estimating hydraulic conductivity of Maragheh-Bonab plain, can be used in estimating hydraulic conductivity of other plains with similar hydrogeological conditions.