Seite - 868 - in Book of Full Papers - Symposium Hydro Engineering

Fig. 4 Probabilistic bearing capacity of rough foundation In this paper, Monte Carlo’s simulation method as one of the reliability analysis methods is used to investigate the influence of the variation of friction angle (φ) on the reliability of dam foundations. In this regard, the performance function is built with Michalowski’s equation. In the presented research, based on the results, the following conclusions are drawn: 1. For foundations on granular soils, the probability of failure of foundation is more sensitive to the variation of friction angle of soils, since the COVs of main parameters of bearing capacity like the friction angle (φ) is considered. 2. The proposed probabilistic approach accounts for parameter uncertainty of soil friction angle, and enables bearing capacity to be quantified in the form of a cumulative probability distribution function that provides bearing capacity predictions corresponding to certain reliability levels. This indicates the importance of adopting probabilistic analysis in favor of the factor of safety. It was also shown that the developed probabilistic method can be used to predict bearing capacity of dam foundation for reliability levels of 90% and 95%. 3. The results obtained from Monte Carlo’s simulation method and deterministic method indicate that the deterministic method is not reliable enough in comparison to the probabilistic results. 4. A numerical example was given to illustrate the use of Monte Carlo’s simulation. The results indicate that the suggested factor of safety of 3 that usually used by available deterministic models is conservative. However, it seems not to be economically beneficial. 5. A substantial influence of the dilatancy angle on 𝑁 𝛾 is found and this is a reason for concern. Since the non-associativity of plastic deformation of soils is usually not accounted and especially in high levels of friction angle, is usually an essential phenomenon in soils under applied loads. Taking into account the reliability analysis with non-associativity, one can highlight the aim of the current study. 868