Author(s): Saad Shauket Sammen, Thamer Ahmeed Mohamed, Abd Alhalim Ghazali, Lariyah Mohd Sidek, Azlan Abdul Aziz
Linked Author(s): Saad Shauket
Keywords: Dam safety, dam failure, failure time, generalized regression neural network.
Abstract:
Several approaches have been used for estimation of failure time when dam failure occurs. Generalized Regression Neural Network (GRNN) model has been proposed in this present study as a novel method to estimate the dam failure time. It has also made a comparison of the results of the GRNN models with the results obtained from the existing approaches. Various reservoir and dam characteristics have been used in the development of the GRNN models in order to estimate the dam failure time. To obtain the optimum results from the GRNN models, all models are optimized to smooth factor values. The values are found to range from 0.001 to 0.03. Furthermore, training the network takes up 85% of the total data and testing the network takes up 15% of the total data. Three statistical indices are used to assess the results of GRNN models. They are the Root Mean Square Error (RMSE), Mean Relative Error (MRE), and Coefficient of Correlation (R2 ). The results show that the value of the MRE could be decreased more than 50% by using the GRNN models as compared to the values of the existing empirical techniques. (2647, 68, 360)
Year: 2017