Estimation of Dam Failure Peak Outflow using Neural Network Approach

Content Provider	Semantic Scholar
Author	Tahershamsi, Ahmad Hooshyaripor, Farhad Sheikholeslami, R.
Copyright Year	2016
Abstract	This article presents an artificial neural network (ANN) for prediction of peak outflow from breached embankment dams based on considering height and volume of water behind the dam at the time of failure. Two different algorithms are used for training the ANN. They are Imperialist Competitive Algorithm (ICA) as a new evolutionary algorithm and LevenbergMarquardt (LM) algortihm. The comparison of results between the proposed method and those conventional approaches which are based on regression analysis shows a better performance of the ANN based models. To evaluate the uncertainty of the two training algorithms, a Monte-Carlo simulation is used to sample 1000 sets from the database of historical dam failures for different sets of training and test in the ANN model. Three statistical measures ( i.e. 95PPU, d-factor, and DDR) are used to compare the uncertainty analysis. The obtained results indicate a better performance of ICA compared to LM algorithm.
File Format	PDF HTM / HTML
Alternate Webpage(s)	https://repository.uwl.ac.uk/id/eprint/2829/1/Paper-Edit%20KB.pdf
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article