Keywords: dam safety, feedforward neural network, adaptive neuro-fuzzy inference system, dam behavior, radial displacement, modeling.

The safety control of a dam is supported by monitoring activities and is based on models. Deterministic and statistical methods have been used to develop models to predict dam behaviour. A hybrid method has also been applied to forecast behaviour. The deterministic modelling requires solving differential equations for which closed form solutions may be difficult or impossible to obtain. The advantages of the statistical method, such as multiple linear regression, consists in the simplicity of formulation and the speed of execution.

The radial displacements of points of the dam are important behaviour indicators and they are nonlinear function of hydrostatic pressure, temperature and other unexpected unknown causes. In dam engineering, multilayer perceptron neural network models [1], models based on wavelet networks [2,3] and neuro-fuzzy models [4] have been developed for the prediction of dam displacements.

The major objective of the study presented in this paper is to construct a high-quality feedforward neural network (FNN) and adaptive neuro-fuzzy inference systems (ANFIS) to predict the radial displacement of arch dams and to demonstrate their application to identifying complex nonlinear relationships between input and output variables. The FNN and ANFIS models were developed and tested using experimental data which are collected over eleven years.

The two-layer network with a log-sigmoid transfer function at the hidden layer and a linear transfer function at the output layer were used. The optimal number of hidden neurons is sixteen. Fuzzy partitioning of the input variables of the ANFIS was realized by selection of the two primary fuzzy sets. The Gaussian membership function was adopted. The ANFIS models in the considered example have thirty-two rules. The FNN model has a slightly higher coefficient of correlation values for the training set, but a slightly lower coefficient of correlation values for the test set. Comparing the modelled values using FNN and ANFIS with the experimental data indicates that soft computing models provide accurate results. These models can be applied to the prediction of displacements.

1

J. Mata, "Interpretation of concrete dam behaviour with artificial neural network and multiple linear regression models", Engineering Structures, 33(3), 903-910, 2011. doi:10.1016/j.engstruct.2010.12.011

2

H. Su, Z. Wu, Z. Wen, "Identification Model for Dam Behavior Based on Wavelet Network", Computer-Aided Civil and Infrastructure Engineering, 22(6), 438-448, 2007. doi:10.1111/j.1467-8667.2007.00499.x

3

M. Cao, P. Qiao, Q. Ren, "Improved hybrid wavelet neural network methodology for time-varying behavior prediction of engineering structures", Neural Computing & Applications, 18(7), 821-832, 2009. doi:10.1007/s00521-009-0240-8

4

S. Demirkaya, "Deformation Analysis of an Arch Dam Using ANFIS", in A. Reiterer, U. Egly, M. Heinert, B. Riedel, (Editors), "Proceedings of the Second International Workshop on Application of Artificial Intelligence and Innovations in Engineering Geodesy", Braunschweig, Germany, 21-31, 2010.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £70 +P&P)