United States
Environmental Protection
Agency
Municipal Environmental Research ;
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-81-063 June 1981
Project Summary
Field Evaluation of a
Swirl Degritter at
Tamworth, New South Wales,
Australia
G. J. Shelley, P. B. Stone, and A. J. Cullen
The overall objective of this field
evaluation program was to provide
information on the behavior of a full-
scale swirl degritter designed and con-
structed in accordance with the shapes
and proportions developed during
model studies.
The swirl degritter was designed to
pretreat river water before it enters
into the rising main in order to reduce
wear and tear on the raw water pumps
and also to reduce the solids loading of
the rising main and that of the balance
tank of the water treatment works.
Results of the solids removal had
been evaluated in terms of three
parameters: solids larger than 0.2 mm
(the classical size aimed at in grit
chambers); solids larger than 0.088
mm; and total settleable solids. In
general, the tests proved the validity
of the laboratory results, and at design
flowrates, 98% removal efficiencies
were achieved.
Tests at flowrates higher than the
design showed slightly better efficien-
cies than predicted.
The field evaluation tests carried out
at Tamworth, New South Wales,
Australia, prove the validity of the
system in terms of its hydraulic effi-
ciency. When compared with a con-
ventional, constant-velocity, longitu-
dinal flow grit chamber, the construction
cost is halved and operation and main-
tenance costs are considerably lower.
This Project Summary was devel-
oped by EPA's Municipal Environ-
mental Research Laboratory, Cincinnati,
OH, to announce key findings of the
research project that is fully docu-
mented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
A considerable number of Australian
towns rely on rivers for the source of
their water supply. These rivers are
generally intermittent flowing streams
with highly variable flows, and their
degree of pollution depends on the
stage of the river and on the prior history
of floods.
Many water intake installations are
constructed in such a manner that
gravitational flow feeds a raw water
pumping station, which may be at a
considerable distance from town. The
water is then pumped to a treatment
works closer to the town; from there, it
is distributed to the service reservoirs
Difficulties have been experienced with
solids settling in the gravity lines and
also with the excessive wear and tear on
the raw water pumps due to the sand
content of the water.
Pretreatment of the raw water—
removing sand particles larger than 0.2
mm diameter entering the system—had
been adopted to alleviate these problems.
-------
Conventional grit removal systems,
whether of the longitudinal flow, con-
stant velocity, or aerated grit chamber
type, are generally costly installations
requiring a high degree of maintenance
of the mechanical grit removal devices.
The swirl degritter developed by the
American Public Works Association
(APWA) and the U.S. Environmental
Protection Agency (USEPA) was selected
because its capital cost is about 50% of
the alternative systems and because its
operational and maintenance costs are
considerably lower as moving compo-
nents may not be needed. Figure 1
illustrates an isometric view of the swirl
degritter.
This swirl degritter was constructed
between 1979 and 1980, and being the
largest of its kind, an extensive field
evaluation program had been carried
out to compare its performance with the
design manual published by USEPA1
and also with the results of a previous
field evaluation program of a prototype
at Denver, Colorado.2
Originally the swirl degritter was to be
tested with natural water expecting that
the river stages would vary considerably
during the test period. Owing to an
unusually dry period, this did not happen,
and the suspended solids content of the
river water remained practically constant.
To subject the degritter to most of the
expected service conditions, sand was
added to the influent, and the whole
range of service flows was reproduced
using the raw water pumps. Removal
efficiency rates were determined for
three classes of material: 0.2 mm and
larger (the classical size aimed at to
remove with a grit chamber); particles
larger than 0.088 mm; and total settle-
able solids. A summary of the removal
efficiency evaluations and comparisons
is depicted by the curves on Figure 2.
The test results essentially proved the
predictions of the design manual, which
were based on a n extensive model study
carried out by the LaSalle Hydraulics
Laboratories and compared well with
the Denver Study.
The swirl degritter constructed at
Tamworth, New South Wales, Australia
(Figure 2), is about 45 ft (14 m) under-
ground, and the bottom of the grit col-
lector is a further 10 ft (3 m) lower. The
collected grit is removed from the
hopper with a hydraulically operated jet
pump and discharged back into the
river. This arrangement coupled with a
grit level detector allows for automatic
operation through telemetry governed
by the central control board of the water
supply system.
The field evaluation program consisted
of sampling the influent, sampling the
effluent, laboratory analyses of the
particle size distribution of the samples,
and evaluating the results.
Sampling the Influent
The influent was sampled in the inlet
conduit just before it entered the de-
gritter, about 45 ft (14 m) underground
with three inlet ports: one near the
invert, the second in the center, and the
third close to the obvert to the square
conduit. The samples taken at the
lowest intake port had not been included
in this evaluation because of their in-
determinate solids content—they may
have contained either suspended load
or bedload. The diameter of the sampling
ports was 1.5 in. (39 mm); both the
sampling suction and delivery lines
were 1.125-in.-diameter (29 mm) un-
plasticized PVC conduits with a flexible
hose at the discharge end. The sampling
pump was attached to the intermediate
landing of the access ladder, and it
delivered the withdrawn liquid into 54
gal (205 L) steel drums sitting on the lid
of the structure on the surface.
The velocity in the 1.125-in. (23-mm)
sampling line was 2.8 fps (0.85 m/s)
during runs 1 through 34 and 4.5 fps
(1.37 m/s) during runs 36 through 55.
Each sample was collected in a steel
drum lined with removable semirigid
polyethylene liner. The collection times
were about 4 min during runs 1 through
34 and about 2.5 min during runs 36
through 55. Thus the samples repre-
sented continuous composite samples.
The velocity of the main stream in the
conduit varied between 0.75 fps (0.23
m/s) and 6.10 fps (1.86 m/s).
Sampling the Effluent
The effluent was sampled immedi-
ately downstream of the raw water
pumps with a slotted 0.375-in.-inside-
diameter (10 mm) copper tube, inserted
into the main 12-in.-diameter (300 mm)
cement-lined cast iron pipe. A flexible
hose was attached to the discharge end
of the copper tube. The sampling velocity
depended on the pressures produced in
the main stream by the raw water pump.
During runs 1 through 34, filling time of
one 54-gal (205-L) drum varied between
12 and 25 min; during runs 36 through
55, two 54-gal (205-L) drums were filled
in between 17 and 28 min, depending
on the pressure available. This procedure
resulted in continuous composite
samples. The velocity in the main stream
varied between 6.4 fps (1.94 m/s) and
15.4 fps (4.7 m/s), and the sampling
velocity at the intake slot varied between
0.4 fps (0.13 m/s) and 1.3 fps (0.4 m/s).
Thus the ratio of the intake to main
stream velocity was very low.
Laboratory Analysis of Particle
Size Distribution
All the samples collected were left for
at least 1 hr in the drums, allowing for
quiescent settlement of all particles ol
about 25 /urn and larger. Floaters were
then placed on the surfaces, and the
supernatant water was syphoned off
always from about 0.5 in. (12 mm) frorr
the surface. This decanting procedure,
which took about 20 to 25 min, allowec
additional fine solids to settle because
the body of the water remained always
still. The liquid-solid mixture left in the
bottom 2 in. (50 mm) of water (4 to 7 gal;
15 to 25 L) then were lifted out in the
polyethylene liners, sealed, and trans-
ported by road to the laboratory in
Sydney. In the laboratory, the transpar-
ent polyethylene liners were hung up
for at least 2 days so that all the solids
dropped into one corner. They were
then pierced, and the supernatant watei
was allowed to escape gradually without
resuspending the solids.
The solids were dried in an oven anc
weighed; the material larger than 1.68
mm was removed by sieving; the sample
was again weighed; the material smallei
than 0.088 mm was removed by wash
ing; and after drying in the oven, the
sample was again weighed. The fre-
quency of the particle size distribution ol
the fraction between 0.088 mm anc
1.68 mm wasdetermined with a settling
column and recorded on a Hewleti
Packard 21 MX and Data Logger. The
results were both tabulated and plottec
by a computer.
In previous works,3"5 which publishec
the efficiency of sampling (the ratio ol
concentration of solids in the sample
and concentration in the main stream
in a graphical form, it had been estab
lished that samplers pick up the concen-
tration of solids at the point of sampling
only if the sampling and main streanr
velocities are equal. Should the sampling
velocity be less than that of the mair
stream, the sample would contain i
higher concentration of solids than the
sampled liquid; conversely, sampling
velocities higher than the main streanr
velocities result in sample concentra
-------
.Finished ground leveL
R.L 391-OOOm
0 1 2 3 4 5m
(1227-54 ft)
A Inlet
B Deflector
C Weir and Weirtray
D Spoiler
E Floor
F Conical Hopper
Plan
Figure 1. General arrangement of the Tamworth swirl degritter.
Isometric view of the swirl degritter.
-------
700
5?
o
.§
ki
1
o
6
90
70
Total Settleable Solids
+42
„ +44
3®3$$$L+D2
Legend
+34 Run No. 34 Tamworth
Run No. 2 Denver
+D2
200 300 400 500 600 700 800 900 1000 L/s
5.0
10.0
100
15.0
Flowrate
20.0
25.0 mil gal/day
9U
x
u
C
•2
5
£
80
70
+40
Legend
+54 Run No. 34 Tarn worth
+D2 Run No. 2 Denver
150i200 300 400 500 600 700 800 900 1000
5.0
10.0
15.0
Flowrate
20.0
100r+12t
90
u
2
o
I
80
70
+40
Legend
+34 Run No. 34 Tamworth
+D2 Run No. 2 Denver
+D4
+D3
+D2
200 300 400 500 600 700 800^00 1000 1100 L/s
Figure 2.
5.0 10.0 15.0
Flowrate
Removal efficiencies.
20.0
25.0 mil gal/day
tions lower than those in the sampled
liquid. It had also been established that
the efficiency of sampling depends on
the construction and orientation of the
point sampler.
The samplers at Tamworth satisfied
all the conditions for fully representative
sampling except for the parity in veloci-
ties. A further problem was that, in a
large conduit with comparatively low
laminar flows, the distribution of solids
in the main stream is not uniform in the
vertical plane. A wide variety of infor-
mation is available in the literature
regarding the solids distribution in free
surface flows, but there appears to have
been no comprehensive work published
for flows under pressure in compara-
tively large conduits.
To calculate the removal efficiencies,
we had to make sure that the samples
were representative of the actual
concentrations.
In the case of sampling the effluent,
the ratio of the sampling velocities to the
main stream velocities were always
considerably below the unit. For the
purposes of this report, we assumed
that the sampled concentrations were
the equivalent of the main stream
concentrations, knowing that any errors
in this assumption would result in a
higher effluent concentration and,
therefore, a lower removal efficiency of
the swirl degritter.
In the case of the influent, however,
the same assumptions could not be
made; the amount of the concentration
of solids in the influent had to be
adopted either from a previously known
value or from the calibrated efficiency of
the samplers.
It had been established that the solids
content of the natural water was so low
(3.4 mg/L) and so fine (d (particle
effective diameter) 0.088 mm) that it
could safely be neglected, and so the
solids content of tested samples de-
pended solely on the artificially intro-
duced sand.
Dewatering and physical inspections
after runs 18 and 34 revealed that
significant sand deposits and some
scouring of the deposited sand had
occurred in the inlet conduit between
the point of adding sand and that of
sampling; thus, the concentration of
added sand was not necessarily equiva-
lent to the average concentration of
solids reaching the sampling points or
reaching the swirl degritter.
-------
Evaluating the Results
The efficiency of sampling could be
calculated by synthesizing the diagrams
of the FIASP report3 into one single
equation:
E = fi(d)x(10V)f2(d>- 100
where E = sampling error (%)
d = particle effective diameter
(mm)
V= ratio of sampling and main
stream velocities
The knowledge of the local concentra-
tion of solids, however, was insufficient
for calculating the removal efficiency of
the swirl degritter because the distribu-
tion of solids in the vertical plane at
various main stream velocities was still
unknown.
Another set of tests was then con-
ducted, whereby frequent dewatering
and inspection ascertained that during
the test runs no significant deposits of
solids occurred between the point of
adding sand and the influent sampling
(in other words, all the sand added
reached the swirl degritter).
Thus runs 36 through 55 achieved a
dual purpose. By comparing the known
concentration of the artificially added
sand with the sampled concentration of
the effluent, the removal efficiency of
the swirl degritter could be calculated.
The ratio of the local concentrations to
the average concentration of solids for
all the three classes of materials (total
solids, solids larger than 0.088 mm
diameter, and solids larger than 0.2 mm
diameter) at the various main stream
velocities could also be established.
After multiple regression analyses, the
equations took the general shape of:
f3(V) = an + bn:
I
C top (n)
.(n)
.±E,
(n)
Caveraae (n)
where
V = the main stream veloc-
ity (m/s)
an, bn, cn = regression coefficients
Ccemer(n) = concentration of solids
of the class of materials
considered in the sam-
ple withdrawn from the
center of the conduit
(local concentration)
Ctop (n) = concentration of solids
of the class of materials
considered in the sam-
ple withdrawn from the
top of the conduit (lo-
cal concentration)
E(n)
i (n) = concentration of solids
of the class of materials
considered as intro-
duced by adding sand
(average concentration)
= standard error of the
regression analysis in
the f3(V) value
It was then possible to calculate:
1 . The average concentrations in the
conduit from the two local concen-
trations:
(n) =
b(n) Ccen,er (n)
.op
f3(V) + a ± E ,„,
2. The removal efficiencies of the swirl
degritter:
^effluent (n)
b(n)Ccenter(n) + c(n)Ctop(n)
.x(F(V) + a)±
x100
b(n)CCente,
-------
G. J. Shelley is a Consulting Engineer in Cammeray, New South Wales,
Australia 2062; P. B. Stone and A. J. Cullen are with the Public Works Depart-
ment of New South Wales. Australia.
Richard Field and Hugh Masters are the EPA Project Officers (see below).
The complete report, entitled "Field Evaluation ofaSwirlDegritterat Tamworth,
New South Wales. Australia." (Order No. PB 81-187 247; Cost: $11.00.
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officers can be contacted at:
Storm and Combined Sewers Section
Municipal Environmental Research Laboratory—Cincinnati
U.S. Environmental Protection Agency
Edison, NJ 08837
> US QOVERNMENT PRINTINQ OFFICE 1M1 -757-012/7130
-------
, r- , r, ^ Postage and
United States Center for Environmental Research Fees Paid
Environmental Protection Information Environmental
Agency Cincinnati OH 45268 Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
RETURN POSTAGE GUARANTEED Third-Class
Bulk Rate
LOJ W TILLEY
RfcGIUi^ v tPA
230 S DtAHBURiM i>
CHICAGO IL t>060<4
------- |