For the velocity, we find for a velocity inlet of 1.8 m/s so Re=3000 this profil

We see that there is a recirculation zone near the transversal walls and the velocity is nulle in the bottom
The following section is the interesting one: we will put particles in our flow and we will try to follow them.
 

Particle tracking

After having the results of our simulation, StarCD permits us to make particle tracking quite easily . First of all, we choose the place where we will inject the particles:

 
                                                    - the initial positions of the particles -

Now, we are able to choose the different characteristics of our particles, and to let them go in the flow!
We will make different tries for different diameters and different densities.

D=0 , rho=0

If we set all characteristics of the particles to 0, what we get is the stream lines. Given that it is interesting, here they are.

For :
    -D=10-4m, rho=2000kg/m3
    -D=10-5m, rho=2000kg/m3
    -D=10-6m, rho=2000kg/m3

Apparently, we obtain the same results. It is the same result because we plot a continuous track but if we plot a segment track ( option in Particle tracks, plot tracks), this give us an animation of the particles and we see
that the particles have not the same velocity when we change their diameter.

Things began more interesting for the next settings

    -D=4*10-4m, rho=2000kg/m3

Some particles fall down , but what is strange is that the particle near the transition zone didn't fall down.
with a segment plot of the tracks, we see that this particle has the more less velocity .
Once the particles higher began to fall, the inertie force became more strong and the fall is unelectable

if we increase the diameter
    -D=6*10-4m, rho=2000kg/m3   , we obtain

The Precipitator is more efficient imn this case

In my first analysis, the velocity inlet seems to be a little high comparatively to the dimensions of my precipitator, so i decided to decrease it to m/s and to start a new analysis
Re=1700
the velocity profile is

for
   -D=0m, rho=0
    -D=10-6m, rho=2000kg/m3
    -D=10-5m, rho=2000kg/m3
we obtain, apparently, the same result

The same remarque is applicable here
For     D=10-5m, rho=2000kg/m3, we obtain this result

So, the precipitator became more efficient.
It will be more and more efficient  if we decrease the velocity but the problem is that our flow will become not turbulent so the principle of this precipitator will be not applied