ENV 6438: Physical & Chemical Processes for Drinking Water Treatment

Department of Civil & Environmental Engineering

University of South Florida
Cunningham Spring 2020

Homework #5 Sedimentation Due Wed., March 4

Assignment for 2020: Problem 1 or 2,

then any three of the remaining problems

1. (55 pts) This problem is adapted from one written by Paul Roberts of Stanford University.
A spreadsheet or computer program is strongly recommended for this problem! Do
not try it by hand unless you have a lot of extra time and you have been wondering
how to fill it.
Consider the following particle-size distribution for a suspension of particles. You can
assume that the particles are spherical and have a density of 2.60 g/cm3 =2600 kg/m3.

particle geometric number

size range mean concentration
diameter ∆N
(µm) (µm) (particles/m3)
------------ ------------ -----------------
0.5 – 1 0.71 1.5 × 1012
1–2 1.4 3.75 × 1011
2–4 2.8 9.4 × 1010
4–8 5.7 2.3 × 1010
8 – 16 11. 5.9 × 109
16 – 32 23. 2.9 × 109
32 – 64 45. 7.3 × 108
64 – 128 91. 9.2 × 107

a. For each size class, calculate the particle mass concentration (g/m3), the particle surface
area concentration (m2/m3), and the particle volume fraction (m3/m3). Sum these to
calculate the overall mass concentration, surface area concentration, and volume
fraction.
b. Which particle size class dominates the number concentration? the mass concentration?
the surface area concentration? For each of these, report the percentage contributed by
the dominant size class.

problem 1 continues 

p 1/7
 ENV 6438: Physical & Chemical Processes for Drinking Water Treatment
Department of Civil & Environmental Engineering
University of South Florida
Cunningham Spring 2020

1. continued
c. Suppose you treated this water by gravity sedimentation in a rectangular sedimentation
basin with an overflow rate of 1.5 m/hr. The water temperature is 20 °C. Calculate the
percentage reduction in the (overall) number concentration, mass concentration, and
surface area concentration.
d. Explain the results you found in part (c). Why was one of the concentrations reduced
much more efficiently than another? How is this related to the way in which gravity
sedimentation works?
e. Repeat the calculations in part (c) if the density of the particles is 1.2 g/cm3 (as might be
expected for “fluffy” flocs leaving a flocculator for the sedimentation basin). How does
the density change affect the results? Explain briefly.
f. Repeat the calculations in part (c) if the water temperature is 5 °C instead of 20 °C. Use
the original particle density of 2.6 g/cm3. How does the change in water temperature
affect the results? Explain briefly.
g. Repeat the calculations in part (c) for the following overflow rates: 0.15 m/hr, 0.50 m/hr,
5.0 m/hr, 15 m/hr. Then make three graphs showing how the reduction in concentration
depends upon the overflow rate. Each graph should have five data points on it
corresponding to the five values of overflow rate that you considered. Briefly explain
how the reduction in concentration depends upon overflow rate, and why.
h. Suppose you needed to treat this water at a rate of 10 million gallons per day (pretty
typical for centralized water treatment -- not too big, not too small). You have to design
the sedimenation basins for this treatment. You want at least two sedimentation basins
in parallel -- more than two is OK. How long and how wide would you make the
sedimentation basins, and how many would you make? Briefly explain your reasoning.
Hint: make use of your results from part (g) along with practical design considerations.

p 2/7
 ENV 6438: Physical & Chemical Processes for Drinking Water Treatment
Department of Civil & Environmental Engineering
University of South Florida
Cunningham Spring 2020

2. (55 pts) Consider water at 25 °C that has just exited the flocculation process and is entering
a conventional sedimentation basin. There is a suspension of flocs in the water. Let’s
assume that all the flocs in the water are spheres. Also, the particle size distribution
function is
𝑑𝑑𝑑𝑑
= 𝐴𝐴𝑒𝑒 −𝑏𝑏𝑏𝑏
𝑑𝑑𝑑𝑑

where x indicates the diameter of the sphere, N indicates number concentration, and A and b
are parameters of the distribution. You might remember (from homework #3) the meaning
of this function: the number concentration ∆N of flocs between diameter x1 and diameter x2
is given by the following.
𝑥𝑥2
𝑑𝑑𝑑𝑑
∆𝑁𝑁 = � � � 𝑑𝑑𝑑𝑑
𝑥𝑥1 𝑑𝑑𝑑𝑑

Suppose that for this particular distribution of flocs, A = 2.3×109 particles/m4 and b = 2100
m–1.
a. (5 pts) Graph dN/dx vs x. Let x go from 0.1 µm (10–7 m) up to 1.0 cm (10–2 m). Present
the graph three ways: once with both axes drawn on a linear scale, once with the
abscissa (y-axis) drawn on a logarithmic scale, and once on a log-log scale.
b. (5 pts) Estimate/calculate the total number concentration of flocs in the water, NT. This
means you have to integrate over all possible values of x, i.e., from x = 0 to x = ∞.
Report your answer in units of particles/m3. (The integral is not very difficult.)
c. (5 pts) The volume of a sphere of diameter x is 𝜋𝜋𝑥𝑥 3 ⁄6. Therefore the volume
concentration of spheres in the water (i.e., volume of flocs per volume of water) is
given by
∞
𝜋𝜋𝑥𝑥 3 𝑑𝑑𝑑𝑑
� � � � � 𝑑𝑑𝑑𝑑
0 6 𝑑𝑑𝑑𝑑
Estimate the volume concentration (volume fraction) Ω for the floc suspension in
question. Hint: you will end up with an integral that you probably want to look up in an
integral table – you can evaluate it yourself using integration by parts multiple times,
but that is a pain – it is probably easier to just look it up.
d. (5 pts) If the flocs have a density of 1500 kg/m3 = 1500 g/L, estimate/calculate the mass
concentration of flocs in the water, in units of mg/L.
problem 2 continues 

p 3/7
 ENV 6438: Physical & Chemical Processes for Drinking Water Treatment
Department of Civil & Environmental Engineering
University of South Florida
Cunningham Spring 2020

2. continued
e. (5 pts) Suppose the overflow rate in the sedimentation basin is 1.5 m/hr.
Estimate/calculate the diameter of the smallest flocs that will be 100% removed by
sedimentation. In other words, flocs with diameter greater than or equal to a diameter
of xcrit will be 100% removed; flocs with diameter smaller than xcrit will be partially
removed; find xcrit. If you use Stokes’ law, double-check that the Reynolds number is
in the proper range.
f. (5 pts) Now let’s think about the flocs exiting the sedimentation basin. For a floc of
size x, argue or show that the fraction of flocs remaining in the water after
sedimentation is
fremaining = 0 if x ≥ xcrit,
fremaining = 1 – c x2 if x < xcrit.
Give the formula for the parameter c in terms of the other parameters of the problem.
g. (5 pts) Plot fremaining vs x, for x going from 0 to 100 µm. Use the formula from part (f).
h. (5 pts) The particle size distribution for flocs exiting the sedimentation basin is given by
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
� � = �𝑓𝑓remaining � � � = �𝑓𝑓remaining �(𝐴𝐴𝑒𝑒 −𝑏𝑏𝑏𝑏 )
𝑑𝑑𝑑𝑑 exiting 𝑑𝑑𝑑𝑑 entering
Plot this function vs x, for x going from 0 to 100 µm. Make both axes linear (not
logarithmic). On the same graph, plot dN/dx for the entering flocs, i.e., the same
function as in part (a), but only going from 0 to 100 µm. Based on the graph, does
sedimentation remove a lot of particles?
i. (10 pts) Now let’s put it all together. Find the total number concentration of flocs
exiting the sedimentation basin. That is, repeat the calculation of part (b), but using the
particle size distribution for the effluent rather than the influent. You can do it either
analytically or numerically. I am not sure yet which way is easier, analytically or
numerically – I plan to do it analytically. NOTE: I thought about asking for the effluent
mass concentration too, but I decided that would be too much math for now. Maybe
next year I will add that on.
j. (5 pts) How well did the sedimentation basin do? Did you see a significant removal of
particles? Do you think the water is now particle-free, or will filtration still be
necessary?

p 4/7
 ENV 6438: Physical & Chemical Processes for Drinking Water Treatment
Department of Civil & Environmental Engineering
University of South Florida
Cunningham Spring 2020

3. (15 pts)
a. Suppose you were interested in deriving an equation for the terminal settling velocity of
particles during “type 1” settling, for cases where the Reynolds number is between 1 and
100. One estimate for the relationship between Re and drag coefficient (CD) in this
regime is CD = 18.5 Re–0.6. Using this relationship, derive a closed-form expression for
the settling velocity of the particle. Show your work. How does the settling velocity
depend upon the particle diameter in this regime?
b. Repeat your derivation for the range 5 < Re < 1000, using the estimated relationship
CD = 13 Re–0.5. How does settling velocity depend on particle diameter according to this
relationship?

4. (15 pts)
a. Answer questions 10-1A and 10-1C in the text book.
b. Which particle settles faster? What is the dominant reason why that one settles faster?

5. (15 pts) We know that for sufficiently low values of Reynolds number (Re), the drag
coefficient CD = 24/Re. However, for Re > 1 (or so), this relationship does not hold. A
number of empirical relationships have been used to estimate how CD depends on Re in the
“transition region” where 1 < Re < 1000. In class, I gave you four such relationships.
Graph CD as a function of Re for the range 1 < Re < 1000, using all four relationships.
Make both axes logarithmic. Make sure all the lines are distinguishable from each other.
Do all four relationships agree reasonably well? If so, then it does not matter which one we
choose, but if not, then we might need to explore further to determine if one of these
relationships is better or worse than the others. Do any of the four curves deviate strongly
from the other three? – if so, perhaps we can eliminate that curve from consideration.

6. (15 pts) All of our calculations for settling velocity are based on an assumption of spherical
particles. But we know that particles are not necessarily spherical; in homework 3, we had
“flaky” particles. Re-do the force balance (p 646 of the text, or notes from class #11) for the
settling of a “flaky” elliptical particle like we had in homework 3. Derive a formula for the
terminal settling velocity of such particles. If possible, express the settling velocity in terms
of the sphericity, ψ, which we saw previously. As particles become less spherical, how does
that affect the settling velocity? If we pretend that particles are spherical (when they are
not), are we being overly optimistic or overly pessimistic?

p 5/7
 ENV 6438: Physical & Chemical Processes for Drinking Water Treatment
Department of Civil & Environmental Engineering
University of South Florida
Cunningham Spring 2020

7. (15 pts) Suppose a sand particle is settling in a sedimentation basin at 20 °C. The particle is
roughly spherical in shape with a diameter of 0.73 mm and a density of 2.60 g/cm3. Assume
that the particle is in the “transition” regime, i.e., Re > 5 or so. Estimate the particle’s
settling velocity using each of the four different models for the relationship between drag
coefficient and Reynolds number (based on your notes from class). Report the settling
velocity and the Reynolds number that you estimate using each of the four models. How
much variation is there between the predictions of the four models? Do you think it matters
which one you choose, or do they all give you pretty much the same results? (Also verify
that the Reynolds number is in the transition regime, not the laminar/Stokes regime.)

8. (15 pts) In class, we derived a formula for the terminal settling velocity of spherical
particles. However, we did not determine how long it takes a particle to reach its terminal
settling velocity, assuming that the particle starts at rest. If the time to reach terminal
velocity is long compared to the residence time in a sedimentation basin, then a lot of our
calculations are off. Hopefully, we will see that terminal velocity is reached in a time much
less than the residence time of the basin – in that case, our methodologies are intact. So let’s
find out.
Force balance for a spherical particle: Fnet = (π/6)*dp3*(ρp–ρL)*g – 0.5*CD*ρL*Aproj*v2
Drag coefficient: CD = 24/Re assuming that we are in the Stokes’ flow regime
Reynolds number: Re = ρL * v * dp / µ
Newton’s law: Fnet = m * dv/dt, where m = (π/6)*dp3*ρp
Initial condition: v(t=0) = 0
Combine the first four equations to derive a differential equation dv/dt = f(v). Solve that
differential equation subject to the initial condition. You should get an answer that is of the
following general form:
v(t) = vterminal * [1 – exp(-kt)]
where k is an apparent rate coefficient that depends on physical parameters of the problem.
We can estimate that terminal velocity is reached within a time 5/k.
For a spherical particle that has a diameter of 0.1 µm and a density of 2500 kg/m3, how long
does it take to reach terminal velocity? What if the diameter is 100 µm instead? Is terminal
velocity reached rapidly enough that we can assume the velocity is constant?

p 6/7
 ENV 6438: Physical & Chemical Processes for Drinking Water Treatment
Department of Civil & Environmental Engineering
University of South Florida
Cunningham Spring 2020

9. (15 pts) Answer question 10-7D in the Crittenden text. (Only recommended if you are
interested in solids thickening, which is not “type 1” settling. We did not discuss this in
class. You will have to figure it out on your own.)

10. (15 pts) Suppose you are designing a circular sedimentation basin that uses center feed with
radial collection (see Figure 10-13 in your text). Base your design on a flow rate of 20
million gallons per day, a water temperature of 20 °C, and a particle (floc) removal
efficiency of 99%. Assume that the flocs are spherical with diameter 0.1 mm and density
1200 kg/m3. The depth of the clarifier is 4.5 m, and the radius of the inlet zone is 1.5 m.
What is the minimum outer diameter for the clarifier?

11. (15 pts) In class we derived expressions for settling velocity assuming a spherical shape for
the particles. If the particles are not spherical, then you can proceed as follows. Find the
diameter of an “equivalent” sphere that has the same surface-area-to-volume ratio as the
particle of interest. Then find the settling velocity of that “equivalent” sphere. That will be
a good estimate for the settling velocity of the particle of interest.
a. Use this method to estimate the settling velocity of the smallest size class of “flaky
particles” from homework #3. (Hint: if you’re clever enough, you can use the sphericity
to help save some time on this part.)
b. Repeat for the largest size class of “flakes” from homework #3. You should notice
something interesting. Explain briefly.
c. Estimate the fractional removal of the two size classes of flaky particles in a
sedimentation basin with an overflow rate of 1.5 m/hr. Is sedimentation an effective
removal mechanism for these flaky particles?

p 7/7