Earth dams play a crucial role in water supply and power generation; therefore, ensuring their stability, safety and identifying various potential failure mechanisms are of great importance.
Among different failure modes, internal erosion (piping) is one of the primary factors contributing to the instability and eventual failure of earth dams. In this study, a probabilistic framework is proposed to assess the risk of internal erosion in earth dams considering uncertainty in the permeability coefficient of materials. The Lar Dam was selected as a case study. Seepage analysis was performed using the finite element method in SEEP/W software under transient conditions based on a 22-year time series of reservoir water levels. The uncertainty in permeability coefficients of dam materials was modeled using normal probability distributions. To quantify the risk of internal erosion, two point-estimate methods, namely Rosenblueth and Harr methods, were applied. The results showed that the probability of internal erosion is 0.22 and 0.21 using Rosenblueth and Harr methods, respectively, with a standard deviation of approximately 0.10. These results indicate a relatively high risk of internal erosion in Lar Dam. Furthermore, the close agreement between the two methods demonstrates that the Harr method can be used as a computationally efficient alternative for probabilistic risk assessment in cases with symmetric distributions. The proposed approach provides a practical framework for incorporating uncertainty into internal erosion risk analysis of earth dam.