The hydrological unstructured domain simulator (shud) is a multi process, multi-scale hydrological model, in which the main hydrological processes are fully coupled by the semi discrete finite volume method (FVM). Continuous applications of the shud model include hydrological analysis of slope at regional scale, water resources and stormwater management, as well as interdisciplinary research on lacustrine issues, agriculture, geochemistry, geomorphology, water quality, ecology, climate and land use change. The advantage of shud lies in its flexibility. It is a scientific and resource evaluation tool that needs modeling and simulation.

A large number of reservoir inflow with preserved data, spillway discharge and discharge of breach ditch (unstable flow with sediment transport capacity) formed along erosion are comprehensively applied in the model. It is assumed that the slope ratio at the bottom of the breach is basically the same as that at the downstream of the dam. The development of breach ditch depends on the material properties of dam body (D50 size, unit weight, friction angle, bond strength). The model takes into account the following possible complexities: 1) the material properties of the dam core and the outside of the dam are different; 2) before the real breach caused by overtopping flow is formed, the erosion ditch along the downstream slope surface of the dam is determined; 3) the downstream slope of the dam may be covered with grass or contain materials larger than the external components of the dam; 4) the shear strength and cohesive force are exceeded The results show that more than one sudden structural collapse mechanism is enlarged due to the hydrostatic pressure; 5) the width of the breach obtained from the slope stability theory is enlarged; 6) the free surface breach flow is developed by piping, which causes the collapse; and 7) the non cohesive material (gravel) or cohesive material (clay) may be carried away by erosion. The discharge water level hydrograph is obtained by time step iterative method, which can be completed in only a few seconds on a main computer. The model is not constrained by numerical stability and convergence stability.

The sand pile avalanche model simulates the occurrence of avalanches in sand and plots the frequency of these avalanches. With the increase of sand particles, the shape and size of sand pile are easy to model. But even though the shape of the pile is conical, a new grain of sand can also trigger avalanches. Some sand particles fall off the cone. And because these avalanches are unpredictable, it's impossible to predict whether the next grain of sand will lead to avalanches.