[Robb13]

I'm modelling massless particles in 2D axisymmetric pipe flow. The turbulence model used is RSM with quadratic pressure-strain and the particles are modeled using unsteady particle tracking and the discrete random walk model to incorporate the turbulence effects. Thereby, the particles act as fluid tracers.

The particles are injected when the flow is developed, and they are injected such that all the particles are placed in the same weighted crosssectional area in a corresponding 3D case. This results in a linear distribution of particles along the radius; a small amount of particles in the center and more near the wall (for the 2D geometry).

However, although all the fluid properties (velocities and reynolds stresses) agree well with experimental and DNS data, the particles end up evenly distributed along the radius in the 2D case. So there are the same amount of particles per delta radius in near the center as near the wall.

This means that there are more particles than one would expect in the middle and less near the wall. This is confirmed by the average speed of the particles, since the particles move faster on average than the fluid bulk velocity.

Is there someone who knows what could be wrong?

Thanks in advance