Model for dam breach - erosion,deposition,entrainment #853
Conversation
|
@ShuangLi-1 It seems that there particles go inside the soil? |
I think that there is a soaking process, by which the yield of the solid is decreased dramatically. Such soaking process can be modeled by diffusion process and is enhanced by turbulent shear at soil surface. |
| FluidType &fluid_; | ||
| RiemannSolverType riemann_solver_; | ||
| }; | ||
| template <class RiemannSolverType, class KernelCorrectionType, typename... Parameters> |
There was a problem hiding this comment.
The problem is a typical FSI problem, we do not need add extra method in fluid mechanics.
| StdVec<DiscreteVariable<Vecd> *> dv_wall_vel_ave_, dv_wall_acc_ave_, dv_wall_n_; | ||
| StdVec<DiscreteVariable<Real> *> dv_wall_Vol_; | ||
| }; | ||
|
|
There was a problem hiding this comment.
We need to be refrained to add new class in the library. Please discuss with me first on such things.
Dam Breach Model Update for Coarse Particle DamsThis is the updated dam breach model, specifically designed for the breaching of dams composed of coarse particles. Algorithm Overview1. Shear Stress Computation
2. Strength Reduction
3. Viscous Modeling
4. Return Mapping Consistency
🔧Other Code Changes
Current ResultsWithout Explicit Shear AccelerationIf the fluid shear stress is not explicitly applied as a shear acceleration to soil particles, the results are relatively stable, but the model neglects actual surface shear:
With Direct Shear AccelerationIf the fluid shear stress is directly applied to soil particles, the non-uniformity of the shear force leads to velocity instabilities in surface particles and eventual simulation failure. Formula: 
Example Result: 
❓QuestionHow can we better handle surface shear stress in terms of its effect on the stress state and acceleration of soil particles? |
|
@ShuangLi-1 Do you applies a morris formulation of shear force? |
I did not calculate the shear force using the velocity gradient. Instead, I defined the shear force based on the friction velocity 
|
|
@ShuangLi-1 so your case is very similar to what Feng has faced. We need a stabilizer for it. |
|
In this case, we need to local shear direction from the soil normal. And Riemann solver along the normal direction. |
Thanks a lot. I will dicuss it with feng and you later. |
Improvements to the Erosion ModelModificationsThe following changes have been implemented to improve the erosion model:
Shear force direction (tangent selection): Vecd n = n_[index_i];
Vecd t1(-n[1], n[0]);
Vecd t2(n[1], -n[0]);
Vecd t_hat = (u_shear_i.dot(t1) > u_shear_i.dot(t2)) ? t1 : t2;Particle Relaxation Introduced particle relaxation to improve the initial particle distribution, especially for irregular boundaries. Numerical TreatmentDuring testing, we found that local numerical instabilities were related to anomalies in the computed shear force, which were in turn linked to unusually high observed velocities. When a limiter was introduced，the simulation can be stable: Real u_star = plastic_kernel_.getFrictionVelocity(SMIN(u_shear_i.norm(), 5.0), 0.5 * particle_spacing_);Maybe I need to revise the calculation of shear vleocity later to avoid the abnormally large, non-physical values。 ResultsWith the corrected shear force formulation and Riemann-based normal direction correction, numerical stability has improved significantly. Viscous forces are currently not included, because after yielding, the soil mass is still largely unaffected by fluid shear—shear values are too low to cause significant motion. ObservationTo verify this hypothesis, a test was conducted with soil strength parameters (c, φ) set to zero from the beginning. The soil behaved like a fluid in appearance but did not get carried away by the water. If the two phases were fully mixed, motion would be governed by the velocity gradient and pressure gradient. QuestionWhen the soil behaves like a fluid, is the turbulent boundary model based on friction velocity still valid? |
3 participants
💡 Fluid Shear Stress Coupling in Soil Plasticity Model
This pull request introduces a novel coupling mechanism between fluid-induced shear stress and the Drucker–Prager (DP) soil plasticity model within our SPH simulations. The primary objective is to enhance the model's capability by incorporating hydrodynamic erosion effects into the stress evolution of surface soil particles, aiming for a more physically accurate representation of soil-fluid interaction.
🧮 Physical Formulations
The bed shear stress exerted by the fluid on the soil surface is formulated as:
Where:
The total soil stress tensor is then updated by incorporating this fluid-induced shear:
Subsequently, a standard return mapping algorithm is applied to enforce the plasticity criterion:
🧪 Current Observations & Open Questions
While the fluid shear stress has been successfully integrated into the soil stress evolution, initial simulations reveal a negligible influence on the observed deformation or yield behavior under realistic physical values.
📊 Visual Comparison of Fluid Shear Stress Effects
The following visualizations illustrate the model's behavior under different fluid shear stress magnitudes:
Original fluid shear stress magnitude:
Amplified fluid shear stress (100×):
