Identification of Nonlinear Dynamic Processes with Unknown and Variable Dead Time Using an Internal Recurrent Neural Network

Cheng, Yi; Karjala, Thomas W.; Himmelblau, D. M.

doi:10.1021/ie00044a025

Cited by 25 publications

(12 citation statements)

References 5 publications

(5 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A similar network can be obtained from the already trained model eliminating all the weights from the hidden layer to the lost output nodes. Nevertheless, the purpose here is to verify if embedding of the output time series into the model outputs increases network performances, as suggested by Cheng et al (1995). To compare single-output and multiple-output models, a new neural network model was trained with the same number of hidden nodes as before (four epochs of 1000 iterations each, with a learning rate decreasing from 0.2 to 0.05).…”

Section: Resultsmentioning

confidence: 99%

Artificial neural network approach to flood forecasting in the River Arno

Campolo

Soldati²,

Andreussi

2003

Hydrological Sciences Journal

186

View full text Add to dashboard Cite

The basin of the River Arno is a flood-prone area where flooding events have caused damage valued at more than 100 billion euro in the last 40 years. At present, the occurrence of an event similar to the 1966 flood of Firenze (Florence) would result in damage costing over 15.5 billion euro. Therefore, the use of flood forecasting and early warning systems is mandatory to reduce the economic losses and the risk for people. In this work, a flood forecasting model is presented that exploits the real-time information available for the basin (rainfall data, hydrometric data and information on dam operation) to predict the water-level evolution. The model is based on artificial neural networks, which were successfully used in previous works to predict floods in an unregulated basin and to predict water-level evolution in the Arno basin under low flow conditions. Accurate predictions are obtained using a two-year data set and a special treatment of input data; which allows a balance to be found between the spatial and temporal resolution of rainfall information and the model complexity. The prediction of water-level evolution remains accurate within a forecast time ahead of 6 h, which is the minimum time lag for the river to respond to dam releases under saturated conditions of the basin. The predicted flow rate percentage error ranges from 7 to 15% from the 1-h ahead to 6-h ahead predictions, and the accuracy of prediction increases for each time ahead of prediction, as the flow rate increases, suggesting that the model is particularly suited for flood forecasting purposes.Key words flood forecasting; artificial neural network; system response identification; nonlinear modelling; rainfall-runoff; River Arno, Italy Une approche à base de réseau de neurones artificiels pour la prévision des crues du fleuve Arno Résumé Le bassin du fleuve Arno est une zone sujette au phénomène des inondations, où le coût des dégâts dus aux inondations durant les 40 dernières années se chiffre à plus de 100 milliards d'euros. De nos jours, un événement aussi grave que l'inondation de 1966 à Florence produirait des dommages pour plus de 15.5 milliards d'euros. C'est pourquoi l'utilisation de systèmes de prévision et d'annonce précoce de crues s'avère nécessaire en vue d'une réduction des pertes économiques et du risque pour les personnes. Nous présentons dans ce travail un modèle de prévision de crues exploitant les informations en temps réel disponibles pour le bassin (données de précipitations, enregistrements hydrométriques et opérations de barrage), dans le but de prévoir l'évolution du niveau de l'eau. Le modèle est basé sur des réseaux de neurones artificiels qui ont été employés avec succès dans des travaux développés précédemment pour prévoir les crues dans un bassin non-régulé et pour prévoir Marina Campolo et al. 382l'évolution du niveau de l'eau dans le bassin de l'Arno en conditions d'écoulement réduit. Des prévisions précises sont obtenues à partir d'un jeu de données de deux ans et d'un traitement spécial des données ...

show abstract

Section: Resultsmentioning

confidence: 99%

Artificial neural network approach to flood forecasting in the River Arno

Campolo

Soldati²,

Andreussi

2003

Hydrological Sciences Journal

186

View full text Add to dashboard Cite

show abstract

“…ANNs were successfully used in many applications as nonlinear input-output maps for process data, in the identification and modeling of linear and nonlinear systems [6][7][8], and in various process control [9][10][11], and pattern recognition [12,13] applications. Besides the above-mentioned contributions, which mainly focus on the applications of neural networks, there has been a recent increase in the number of studies concerning the control-relevant properties of neural networks [14], as well as the improvement of network training by using different network structures, transfer functions and learning algorithms [6,[15][16][17].…”

Section: Introductionmentioning

confidence: 99%

Model based control of a yeast fermentation bioreactor using optimally designed artificial neural networks

Nagy

2007

Chemical Engineering Journal

176

105

View full text Add to dashboard Cite

“…In the six output nodes the water levels for the 6 upcoming days are given. Embedding the input and output time series segments into the input and output arrays of the model was found useful to achieve good model performance at low cost [see Cheng et al, 1995].…”

Section: Water Level Forecasting Modelmentioning

confidence: 99%

“…Recently, a number of complex processes have been modeled with the aid of neural networks [Bishop, 1994;Maier and Dandy, 1996;Cheng et al, 1995], which may be properly clas- The flow rate of the Arno is highly variable during the year, with minima less than 1 m3/s and maxima about 4000 m3/s. During high-flow periods the aim of water management policy is to control and, if possible, to prevent floods.…”

Section: Introductionmentioning

confidence: 99%

Forecasting river flow rate during low‐flow periods using neural networks

Campolo¹,

Soldati²,

Andreussi³

1999

Water Resources Research

View full text Add to dashboard Cite

Abstract. The pollution in the river Arno downstream of the city of Florence is a severe environmental problem during low-flow periods when the river flow rate is insufficient to support the natural waste assimilation mechanisms which include degradation, transport, and mixing. Forecasting the river flow rate during these low-flow periods is crucial for water quality management. In this paper a neural network model is presented for forecasting river flow for up to 6 days. The model uses basin-averaged rainfall measurements, water level, and hydropower production data. It is necessary to use hydropower production data since during low-flow periods the water discharged into the river from reservoirs can be a major fraction of total flow rate. Model predictions were found to be accurate with root-mean-square error on the predicted river flow rate less then 8% over the entire time horizon of prediction. This model will be useful for managing the water quality in the river when employed with river quality models.

show abstract

Identification of Nonlinear Dynamic Processes with Unknown and Variable Dead Time Using an Internal Recurrent Neural Network

Cited by 25 publications

References 5 publications

Artificial neural network approach to flood forecasting in the River Arno

Artificial neural network approach to flood forecasting in the River Arno

Model based control of a yeast fermentation bioreactor using optimally designed artificial neural networks

Forecasting river flow rate during low‐flow periods using neural networks

Contact Info

Product

Resources

About