Without global variable

[Xiangyu-Hu]

1. revise the transport velocity formulation so that it correct all particles for multiphase flow.
2. revised the droplet and wetting case so that the simulation can be stable to long physical time.

[Shuaihao-Zhang]

The result seems good for the case test_2d_square_droplet.

For the test_2d_wetting_effects case, the particle distribution is not uniform. Maybe other reasons related to the boundary condition.

I found two other problems, which may be not related to the current modification.
(1) The first one is that, for the test_2d_wetting_effects case, the air phase and the water phase share the same layer of particle on the interface, as shown below:

I found this problem when increasing the resolution. For very low resolutions, we don't have this problem. I guess the bug is from the particle generation process.

(2) The second one is that the advection time step is always the same as the acoustic time step for the two cases.

[Xiangyu-Hu]

@Shuaihao-Zhang probably the transport velocity correction is too much.