Improving event-based rainfall–runoff modeling using a combined artificial neural network–kinematic wave approach

2010 ◽  
Vol 390 (1-2) ◽  
pp. 92-107 ◽  
Author(s):  
Lloyd H.C. Chua ◽  
Tommy S.W. Wong
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Kinematic Wave ◽  
Rainfall Runoff ◽  
Wave Approach ◽  
Runoff Modeling ◽  
Artificial Neural ◽  
Event Based

scholarly journals OPTIMIZATION RAINFALL-RUNOFF MODELING FOR CIUJUNG RIVER USING BACK PROPAGATION METHOD

SINERGI ◽  
2018 ◽  
Vol 22 (3) ◽  
pp. 193
Author(s):  
Ika Sari Damayanthi Sebayang ◽  
Agus Suroso ◽  
Alnis Gustin Laoli
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Urban Areas ◽  
Back Propagation ◽  
Water Volume ◽  
Rainfall Runoff ◽  
Back Propagation Algorithm ◽  
Runoff Modeling ◽  
Artificial Neural ◽  
The Relationship

The rainfall-runoff model is required to ascertain the relationship between rainfall and runoff. Hydrologists are often confronted with problems of prediction and estimation of runoff using the rainfall date. In actual fact the relationship of rainfall-runoff is known to be highly non-linear and complex. The spatial and temporal precipitation patterns and the variability of watershed characteristics create a more complex hydrologic phenomenon. Runoff is part of the rain water that enters and flows and enters the river body. Rainfall-runoff modeling in this study using Artificial Neural Network, back propagation method and sigmoid binary activation function. This model is used to simulate single or long-term continuous events, water volume, making it very appropriate for urban areas. Back propagation is an inherited learning algorithm and is commonly used by perceptron with multiple layers to change the weights associated with neurons in the hidden layer. Back propagation algorithm uses output error to change the values of its weight in the backward direction. The location of the review is the Ciujung River Basin (DAS), the data used are rainfall and debit data of Ciujung River from 2011-2017. Based on training and simulation results, obtained R2 value: 2012 = 0,85102; 2013 = 0,78661; 2014 = 0,81188; 2015 = 0,77902; 2016 = 0,7279. on model 2 = 0,8724. On model 3 R2:  January = 0,96937; February = 0,92984; March = 0,90666; April = 0,92566; May = 0,9128; June = 0,87975; July= 0,85292; August = 0,95943; September = 0,88229; October = 0,90537; November = 0,93522; December = 0,9111. with MSE (Mean Squared Error) of 0,0018479. The closer value of MSE to 0 and the value of R2 close to 1 then the better designed artificial neural network. If the data used for training more, the artificial neural network will produce a larger R2 value.

scholarly journals ÁRVORE MODELO FRENTE A UMA REDE NEURAL ARTIFICIAL PARA A MODELAGEM CHUVA-VAZÃO

Nativa ◽  
2019 ◽  
Vol 7 (5) ◽  
pp. 527
Author(s):  
Aline Bernarda Debastiani ◽  
Sílvio Luís Rafaeli Neto ◽  
Ricardo Dalagnol da Silva
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Regression Tree ◽  
Superior Performance ◽  
Multi Layer Perceptron ◽  
Rainfall Runoff ◽  
Model Tree ◽  
Runoff Modeling ◽  
Artificial Neural ◽  
Artificial Neural Network Ann

O objetivo deste estudo é investigar o desempenho da árvore modelo (M5P) e sua sensibilidade à poda e comparação com o desempenho de uma Rede Neural Artificial (RNA) para a simulação da vazão média diária mensal. A motivação para esta análise está na maior simplicidade e velocidade de processamento da M5P comparado às RNAs e a carência de estudos aplicando este método na modelagem hidrológica. O estudo foi desenvolvido na bacia hidrográfica do Alto Canoas, tendo um delineamento experimental composto por um período de treinamento, um de validação cruzada e dois períodos de testes. A RNA utilizada foi a Multi Layer Perceptron (MLP), implementada no software MATLAB, e a M5P (com e sem poda), disponível do software WEKA. O algoritmo M5P se mostrou sensível à poda em somente metade dos tratamentos. A M5P apresentou bom ajuste na modelagem, porém a RNA apresentou desempenho superior em todos os tratamentos.Palavras-chave: rede neural artificial; árvore de regressão; Bacia do Alto Canoas. MODEL TREE IN COMPARISON TO ARTIFICIAL NEURAL NETWORK FOR RAINFALL-RUNOFF MODELING ABSTRACT: The aim of this study is to investigate the performance of the model tree (M5P) and its sensitivity to pruning and comparison to the performance of an Artificial Neural network (ANN) for the simulation of daily average discharge of the month. The motivation for this analysis is on simplicity and speed of processing M5P compared the RNAs. The study was developed in the Alto Canoas watershed, having an experiment consisting of a training period, a cross-validation and two testing periods. The ANN used was the Multi Layer Perceptron (MLP), implemented in MATLAB software, and M5P (with and without pruning), available from the WEKA software. M5P algorithm proved sensitive to pruning in half of the treatments. The M5P showed good fit in the modeling, but the RNA presented superior performance in all treatments.Keywords: artificial neural network; regression tree; Basin Alto Canoas.

scholarly journals Conjunction of emotional ANN (EANN) and wavelet transform for rainfall-runoff modeling

2018 ◽  
Vol 21 (1) ◽  
pp. 136-152 ◽  
Author(s):  
Elnaz Sharghi ◽  
Vahid Nourani ◽  
Amir Molajou ◽  
Hessam Najafi
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Wavelet Transform ◽  
Rainfall Runoff ◽  
Runoff Modeling ◽  
Dry Conditions ◽  
Model Efficiency ◽  
Artificial Neural ◽  
Unique Framework ◽  
Emotional System

Abstract The current research introduces a combined wavelet-emotional artificial neural network (WEANN) approach for one-time-ahead rainfall-runoff modeling of two watersheds with different geomorphological and land cover conditions at daily and monthly time scales, to utilize within a unique framework the ability of both wavelet transform (to mitigate the effects of non-stationary) and emotional artificial neural network (EANN, to identify and individualize wet and dry conditions by hormonal components of the artificial emotional system). To assess the efficiency of the proposed hybrid model, the model efficiency was also compared with so-called EANN models (as a new generation of ANN-based models) and wavelet-ANN (WANN) models (as a multi-resolution forecasting tool). The obtained results indicated that for daily scale modeling, WEANN outperforms the other models (EANN and WANN). Also, the obtained results for monthly modeling showed that WEANN could outperform the WANN and EANN models up to 17% and 35% in terms of validation and training efficiency criteria, respectively. Also, the obtained results highlighted the capability of the proposed WEANN approach to better learning of extraordinary and extreme conditions of the process in the training phase.

Sign in / Sign up

Export Citation Format

Share Document