Multivariate Adaptive Regression Spline Based Reliability Analysis of Stability of Durgawati Earthen Dam

In the present paper, multivariate adaptive regression splines (MARS) method is used to carry out reliability analysis of Durgawati earthen dam. The steady and transient state seepage is considered while calculating reliability index \(\left( \beta \right)\). The earthen dam is an important water retaining structure and serves the catchment area of Durgawati River in Bihar, India. The factor of safety (FOS) of the earthen dam against slope failure under steady and transient state is performed using GeoStudio-07 software. In order to account for the variability of material parameters over the body of the dam, a data set of material parameters [i.e., cohesion (\(c\)), angle of internal friction \(\left( \varphi \right)\) and soil unit weight (\(\gamma\))] are prepared, and MARS is used to calculate reliability index \(\left( \beta \right)\) for different realizations over the dam geometry. The reliability index \(\left( \beta \right)\) thus estimated shows that the dam is stable and safe from slope failure point of view under different seepage loadings.