scholarly journals An Assessment of a Proposed Hybrid Neural Network for Daily Flow Prediction in Arid Climate

2014 ◽  
Vol 2014 ◽  
pp. 1-10 ◽  
Author(s):  
Milad Jajarmizadeh ◽  
Sobri Harun ◽  
Mohsen Salarpour
Keyword(s):  
Neural Network ◽  
Transfer Functions ◽  
Heuristic Method ◽  
Daily Rainfall ◽  
Hybrid Network ◽  
Coefficient Of Determination ◽  
Training Algorithms ◽  
Rainfall Runoff ◽  
Runoff Simulation ◽  
Hidden Layer

Rainfall-runoff simulation in hydrology using artificial intelligence presents the nonlinear relationships using neural networks. In this study, a hybrid network presented as a feedforward modular neural network (FF-MNN) has been developed to predict the daily rainfall-runoff of the Roodan watershed at the southern part of Iran. This FF-MNN has three layers—input, hidden, and output. The hidden layer has two types of neural expert or module. Hydrometeorological data of the catchment were collected for 21 years. Heuristic method was used to develop the MNN for exploring daily flow generalization. Two training algorithms, namely, backpropagation with momentum and Levenberg-Marquardt, were used. Sigmoid and linear transfer functions were employed to explore the network’s optimum behavior. Cross-validation and predictive uncertainty assessments were carried out to protect overtiring and overparameterization, respectively. Results showed that the FF-MNN could satisfactorily predict stream flow during testing period. The Nash-Sutcliff coefficient, coefficient of determination, and root mean square error obtained using MNN during training and test periods were 0.85, 0.85, and 39.4 and 0.57, 0.58, and 32.2, respectively. The predictive uncertainties for both periods were 0.39 and 0.44, respectively. Generally, the study showed that the FF-MNN can give promising prediction for rainfall-runoff relations.

scholarly journals Comparing the Selected Transfer Functions and Local Optimization Methods for Neural Network Flood Runoff Forecast

2014 ◽  
Vol 2014 ◽  
pp. 1-10 ◽  
Author(s):  
Petr Maca ◽  
Pavel Pech ◽  
Jiri Pavlasek
Keyword(s):  
Neural Network ◽  
Transfer Functions ◽  
Optimization Methods ◽  
Local Optimization ◽  
Multilayer Perceptrons ◽  
Training Algorithms ◽  
Activation Functions ◽  
Marquardt Algorithm ◽  
Runoff Forecast ◽  
Hidden Layer

The presented paper aims to analyze the influence of the selection of transfer function and training algorithms on neural network flood runoff forecast. Nine of the most significant flood events, caused by the extreme rainfall, were selected from 10 years of measurement on small headwater catchment in the Czech Republic, and flood runoff forecast was investigated using the extensive set of multilayer perceptrons with one hidden layer of neurons. The analyzed artificial neural network models with 11 different activation functions in hidden layer were trained using 7 local optimization algorithms. The results show that the Levenberg-Marquardt algorithm was superior compared to the remaining tested local optimization methods. When comparing the 11 nonlinear transfer functions, used in hidden layer neurons, the RootSig function was superior compared to the rest of analyzed activation functions.

scholarly journals Rainfall-Runoff Modeling Using the HEC-HMS Model for the Al-Adhaim River Catchment, Northern Iraq

Hydrology ◽  
2021 ◽  
Vol 8 (2) ◽  
pp. 58
Author(s):  
Ahmed Naseh Ahmed Hamdan ◽  
Suhad Almuktar ◽  
Miklas Scholz
Keyword(s):  
Geographical Information Systems ◽  
Hydrological Model ◽  
Daily Rainfall ◽  
Curve Number ◽  
Geographical Information ◽  
Coefficient Of Determination ◽  
Rainfall Runoff ◽  
Time Step ◽  
River Catchment ◽  
Northern Iraq

It has become necessary to estimate the quantities of runoff by knowing the amount of rainfall to calculate the required quantities of water storage in reservoirs and to determine the likelihood of flooding. The present study deals with the development of a hydrological model named Hydrologic Engineering Center (HEC-HMS), which uses Digital Elevation Models (DEM). This hydrological model was used by means of the Geospatial Hydrologic Modeling Extension (HEC-GeoHMS) and Geographical Information Systems (GIS) to identify the discharge of the Al-Adhaim River catchment and embankment dam in Iraq by simulated rainfall-runoff processes. The meteorological models were developed within the HEC-HMS from the recorded daily rainfall data for the hydrological years 2015 to 2018. The control specifications were defined for the specified period and one day time step. The Soil Conservation Service-Curve number (SCS-CN), SCS Unit Hydrograph and Muskingum methods were used for loss, transformation and routing calculations, respectively. The model was simulated for two years for calibration and one year for verification of the daily rainfall values. The results showed that both observed and simulated hydrographs were highly correlated. The model’s performance was evaluated by using a coefficient of determination of 90% for calibration and verification. The dam’s discharge for the considered period was successfully simulated but slightly overestimated. The results indicated that the model is suitable for hydrological simulations in the Al-Adhaim river catchment.

scholarly journals Neural network emulation of a rainfall-runoff model

2007 ◽  
Vol 4 (1) ◽  
pp. 287-326 ◽  
Author(s):  
R. J. Abrahart ◽  
L. M. See
Keyword(s):  
Neural Network ◽  
Transfer Functions ◽  
Network Models ◽  
Exact Equation ◽  
Rainfall Runoff ◽  
Neural Network Models ◽  
Linear Relationships ◽  
Non Linear ◽  
Rainfall Runoff Model ◽  
Runoff Model

Abstract. The potential of an artificial neural network to perform simple non-linear hydrological transformations is examined. Four neural network models were developed to emulate different facets of a recognised non-linear hydrological transformation equation that possessed a small number of variables and contained no temporal component. The modeling process was based on a set of uniform random distributions. The cloning operation facilitated a direct comparison with the exact equation-based relationship. It also provided broader information about the power of a neural network to emulate existing equations and model non-linear relationships. Several comparisons with least squares multiple linear regression were performed. The first experiment involved a direct emulation of the Xinanjiang Rainfall-Runoff Model. The next two experiments were designed to assess the competencies of two neural solutions that were developed on a reduced number of inputs. This involved the omission and conflation of previous inputs. The final experiment used derived variables to model intrinsic but otherwise concealed internal relationships that are of hydrological interest. Two recent studies have suggested that neural solutions offer no worthwhile improvements in comparison to traditional weighted linear transfer functions for capturing the non-linear nature of hydrological relationships. Yet such fundamental properties are intrinsic aspects of catchment processes that cannot be excluded or ignored. The results from the four experiments that are reported in this paper are used to challenge the interpretations from these two earlier studies and thus further the debate with regards to the appropriateness of neural networks for hydrological modelling.

MODELING OF HIGH-SPEED FINISH MILLING PROCESS USING SOFT COMPUTING

2010 ◽  
Vol 01 (03) ◽  
pp. 405-420 ◽  
Author(s):  
B. SUREKHA ◽  
PANDU R. VUNDAVILLI ◽  
M. B. PARAPPAGOUDAR ◽  
K. SHYAM PRASAD
Keyword(s):  
Neural Network ◽  
Soft Computing ◽  
High Speed ◽  
Transfer Functions ◽  
Forward Modeling ◽  
Milling Process ◽  
Rule Base ◽  
Machining Parameters ◽  
Layer Transfer ◽  
Hidden Layer

In the present study, forward modeling of high-speed finish milling process has been solved using soft computing. Two different approaches, namely neural network (NN) and fuzzy logic (FL), have been developed to solve the said problem. The performance of NN and FL systems depends on the structure (i.e. number of neurons in the hidden layer, transfer functions, connection weights, etc.) and knowledge base (i.e. rule base and data base), respectively. Here, an approach is proposed to optimize the above-mentioned parameters of NN and FL systems. A binary coded genetic algorithm (GA) has been used for the said purpose. Once optimized, the NN and FL-based models will be able to provide optimal machining parameters online. The developed approaches are found to solve the above problem effectively, and the performances of the developed approaches have been compared among themselves and with that of the results of existing literature.

Modeling of Gas Holdup and Pressure Drop Using ANN for Gas-Non-Newtonian Liquid Flow in Vertical Pipe

2014 ◽  
Vol 917 ◽  
pp. 244-256 ◽  
Author(s):  
Nirjhar Bar ◽  
Sudip Kumar Das
Keyword(s):  
Pressure Drop ◽  
Liquid Flow ◽  
Multilayer Perceptron ◽  
Transfer Functions ◽  
Gas Holdup ◽  
Newtonian Liquid ◽  
Training Algorithms ◽  
Vertical Pipe ◽  
Frictional Pressure Drop ◽  
Hidden Layer

This paper is an attempt to compare the the performance of the three different Multilayer Perceptron training algorithms namely Backpropagation, Scaled Conjugate Gradient and Levenberg-Marquardt for the prediction of the gas hold up and frictional pressure drop across the vertical pipe for gas non-Newtonian liquid flow from our earlier experimental data. The Multilayer Perceptron consists of a single hidden layer. Four different transfer functions were used in the hidden layer. All three algorithms were useful to predict the gas holdup and frictional pressure drop across the vertical pipe. Statistical analysis using Chi-square test (χ2) confirms that the Backpropagation training algorithm gives the best predictability for both cases.

scholarly journals A modified topographic index that incorporates the hydraulic and physical properties of soil

2016 ◽  
Vol 48 (2) ◽  
pp. 370-383 ◽  
Author(s):  
Lu Yi ◽  
Wan-Chang Zhang ◽  
Chang-An Yan
Keyword(s):  
Daily Rainfall ◽  
Hydrological Models ◽  
Rainfall Runoff ◽  
Runoff Simulation ◽  
Stream Network ◽  
Topographic Index ◽  
Climate Conditions ◽  
Performance Improvements ◽  
And Performance ◽  
Properties Of Soil

Aiming at quantifying the impacts of soil properties on rainfall–runoff processes, the soil saturated hydraulic conductivity (Ks) and the soil erodibility factor (K) were selected and incorporated into the classical topographic index ln(α/tanβ) (TI) to construct a modified topographic index (TI′). Stream network extractions and performance evaluations of topography-based hydrological models based on TI and TI′ were carried out in three watersheds with different climate conditions. The investigations indicated that: (1) the changes of stream networks caused by the incorporation of Ks·K could correctly present the phenomenon that the points would show greater potential to be saturated to become contributing areas if their underlying soils possess higher hydraulic conductivities and stronger erodibility; and (2) the performances of the topography-based hydrological models TOPMODEL and TOPX were improved when simulating the daily rainfall–runoff processes with the input of ln(α/(tanβ·Ks·K) (TI3). TI3 was suitable for rainfall–runoff simulation in arid and semi-arid, humid and semi-humid, and humid regions. The performance improvements increased as the spatial heterogeneity of Ks·K enlarged. Based on these investigations, TI3 was recommended for the modified form of TI′.

scholarly journals A comparison between lumped and distributed hydrological models for daily rainfall-runoff simulation

2021 ◽  
Vol 958 (1) ◽  
pp. 012016
Author(s):  
F Vilaseca ◽  
S Narbondo ◽  
C Chreties ◽  
A Castro ◽  
A Gorgoglione
Keyword(s):  
Computational Cost ◽  
Model Performance ◽  
Daily Rainfall ◽  
Complex Model ◽  
The Other ◽  
Hydrological Models ◽  
Rainfall Runoff ◽  
Runoff Simulation ◽  
Distributed Hydrological Models ◽  
Social Importance

Abstract In Uruguay, the Santa Lucía Chico watershed has been studied in several hydrologic/hydraulic works due to its economic and social importance. However, few studies have been focused on water balance computation in this watershed. In this work, two daily rainfall-runoff models, a distributed (SWAT) and a lumped one (GR4J), were implemented at two subbasins of the Santa Lucía Chico watershed, with the aim of providing a thorough comparison for simulating daily hydrographs and identify possible scenarios in which each approach is more suitable than the other. Results showed that a distributed and complex model like SWAT performs better in watersheds characterized by anthropic interventions such as dams, which can be explicitly represented. On the other hand, for watersheds with no significant reservoirs, the use of a complex model may not be justified due to the higher effort required in modeling design, implementation, and computational cost, which is not reflected in a significant improvement of model performance.

Sign in / Sign up

Export Citation Format

Share Document