...
The simplification of a single baffle turn eliminates eliminated important behavior within the flocculator. Multiple baffles enable separation regions to interact with each other. The velocity magnitude plot below in Figure 1 displays how the flow before separation at the second turn is not uniform and moves fastest at the edge of the baffle. The enabled us to see the interaction between different baffles.
...
Figure 1: Velocity Magnitude for three turns case
...
Figure 1 shows the flocculation tank with three baffles turning. As can be seen, the fast moving fluid after first turning propagated to the second turning.
Intuition will tell us that the velocity at the second turning will further increase because of the faster incoming flow at second bend. Surprisingly enough, at second turn, the outer edge velocity is lower compared to the first turn.
...
Figure 2: Turbulence Dissipate Rate for three turns
...
One explanation of such phenomena is that the faster moving fluid can't turn fast enough and impinges on the edge of the third baffle. The contours wall. Upon investigating contour of turbulence dissipation rate are , illustrated in Figure 2.
Figure 1: Velocity Magnitude for three turns case
figure 2, we see that there is a very high region of energy dissipation rate right after the second turn. The high energy dissipation rate region explains why we have a smaller velocity.
...
Flocculation Tank with Five Baffles Turning
The five baffle turning case results below shown in Figure 3 reveals a similar velocity profile after each turn. In this case, there is not a high velocity region followed by a high dissipation region, and the flow seems more uniform. 
The flow obviously relates to the geometry of the problem. Altering the geometry of the problem will result in different flows with a different distribution of energy dissipation.
Automation of Mesh Creation Process
In order to save time, a script was written to automate the mesh generation process in Gambit. The script file can be accessed here.
Using the script, the user just have to change the parameter of baffle spacings, clearance height and flocculation tank height. Upon changing those parameter, the user can execute Gambit and run the journal files for mesh generation. The script can be accessed here.
Similarly, a script in FLUENT can process a completed mesh, and automatically set up the solver, initial conditions, fluid properties, convergence criteria, and save the convergence solution data file to be analyzed later. The incomplete script can be accessed here.
Nondimensional Analysis
Dimensional analysis relates ε to K,V, π and baffle width, as follows:
ε = KV^3 /(2*π-cell*b)
Thus, 2ε*b/KV^3 represents a dimensionless quantity where K is the minor loss coefficient (the drop in the pressure coefficient per baffle), b is the baffle spacing, and V is the average velocity flowing through the channel when the space is b.
turns enables high dissipation regions to interact. The geometry to investigate is presented below in the automization of the mesh in GAMBIT. Below is the case where baffle spacing is fixed at .1, and the flocculator height is fixed at .1. This results in a regions of high velocity magnitude, "undeveloped" energy dissipation regions, and a uniform energy dissipation value.
.
...
Figure 3: Velocity and Turbulence Dissipate Rate for five turns
...
This problem is rich in geometric variables, and our investigation found that a larger height baffles is more optimal. Figure 4 shows the velocity and energy dissipation plots for flocculation height of .3:
...
Figure 4: Velocity and Turbulence Dissipate Rate for fh = 0.3
...
Note that the overall epsilon value range (max at .01) is lower than the plot above which has a max of (.05), but has comparatively more uniform distribution for baffles turns 3 and 4For the optimized geometry below, the baffle spacing is fixed at .1, and baffle height enables a fully developed dissipation region: 
The nondimensionalized epsilon plot is shown below:
The plot is equivalent to 50*ε (since (2*.1)/(.1^3*4.2)=48). Similarly for the 1-turn baffle case the nondimensional parameters can be plotted:
The nondimensional values are one order of magnitude lower corresponding to how multiple baffles induce a much higher level of mixing and turbulent energy dissipation independent of the geometry and velocity.