Related parameter problems for the verification of the settlement of a single ball in a fluid

I want to verify a classic example of the settlement of a single ball in a fluid, but there is only the same trend, and there is a big gap in data comparison(DualSPHysics v5.0). I have tried almost every relevant parameter, but the data has not been compared. On, may I ask where is the problem? Attached is my example file(One is a modeling file containing relevant parameters, and the other is a batch file). The picture is a comparison between my calculation results and the calculation results in the literature( Thanks!




  • I have not looked at your simulation files, but it seems like your objects are a bit too bouyant, since they do not get the same depth as experiments. If you are using DBC, try to scale down your ball model a few percent and see if it helps.

    Kind regards

  • I have been playing with different options such as using mDBC and results do not change too much.

    To be honest, I think some information is missing in that paper.

    I would suggest you to check the original experimental work described in:

    I do not understand how they determine the initial velocity of 2.17 m/s when sphere is at free surface... it seems that sphere falls down from upper positions in the experimental work



  • edited January 8

    @Alex Thanks for your suggestion very much. The original experiment is indeed that the sphere falls from a certain position, and the speed of the sphere entering the water comes from the following: the speed of the sphere detected by the experimental instrument when it touches the water surface. I have a question, this is a 3D experiment, my file is a 2D simulation, how should the quality of the sphere in the 3D experiment be given in the 2D simulation? I'm assigning the density of the sphere relative to the density of water, is this the correct way? Do I have to use 3D simulation? For this question, I also made some attempts

    In addition, I found that other scholars have simulated this problem: (DOI: 10.1007/s10064-019-01558-3). The size of the particle spacing has a great influence on the simulation results. He also uses this software, but combined with another open source code. I use the same particle spacing as him but Can't get the same simulation results, is this because of his code changes? When my particle number reaches close to 30 million, it can be similar to his simulation result.But I'm not sure if it's the right thing to do by increasing the number of particles too much (reducing the spacing between particles).

    Thanks very much!


Sign In or Register to comment.