Problem faced when running for longer time and using mdbc

Hi there,

I faced a problem when I simulate an experimental flume with a rock structure using DualSPhysics. The flume is about 20m x 1.5m x 1.2 m (length x width x height), and the wave generator is piston which will generate irregular wave with JONSWAP theory. I need to run 1200 seconds (20 minutes) to collect the overtopping data (in XML file) across the crest of the structure. I run in 2D case and dp=0.015 m, with time step = 10s.

Figure below shows that the initial water level in paraview:

Figure below shows the water level after simulation:

The problem is that:

  1. When I run for mdbc, there is a warning come out, saying that: " There are 1289 of 1289 fixed or moving particles boundary without normal data". But if I run with dbc, there is no warning. I hope I could this problem to reduce the voids between the boundary and particles.
  2. The structure crest height is 0.39m. And it is 17m away from the piston. When I use the physical experimental data in the report which tells me that water level = 0.318m, wave height=0.1 m can cause overtopping. But in my simulation case, there is no water overtop. So I change the data to new case: (as shown in the picture) water level=0.35m and wave height =0.15m. * and I found that the new case give me the overtopping at first but not long after that I observed the weird scenario that occur in Paraview: the water particles is gradually reduced (but I didn't get any warning after completion of the simulation). I think that is the main problem why I still cannot get the water overtop behind the structure in my first case. Can anyone tell me why?? how to solve the problems??


  • In regards to the first warning, this means that you have no normal vectors for the boundary particles i.e. mDBC is not being used.

    When a particle does not have a normal, it defaults to DBC. Therefore you have been running simulations using DBC in reality.

    Kind regards

  • Thank you for your responds. Yes, I also realised the problem.... but why when I run longer time (for example 1200s), 10% of particles out at the end of the simulation? What should I do?

Sign In or Register to comment.