United States
                    Environmental Protection
                    Agency
Water Engineering
Research Laboratory
Cincinnati, OH 45268
                    Research and Development
EPA/600/S2-86/005 Mar. 1986
&EPA          Project Summary
                   Ultraviolet  Disinfection  of
                   Wastewaters from  Secondary
                   Effluent  and  Combined  Sewer
                   Overflows

                   0. Karl Scheible, Maureen C. Casey, and Angelika Forndran
                      A 2-year,  pilot-scale investigation
                   was conducted at New York City's Port
                   Richmond Water Pollution Control
                   Plant to demonstrate the application of
                   ultraviolet (UV) disinfection to second-
                   ary wastewater effluent and to deter-
                   mine the feasibility of applying  UV
                   radiation to the disinfection of a waste-
                   water similar to combined sewer over-
                   flow (CSO). Three  systems were  op-
                   erated:  Two were  submerged quartz
                   units that differed only in the spacing of
                   the lamps; the third  used Teflon* tubes
                   to carry the liquid, with the UV lamps
                   surrounding the tubes.
                      The  UV process  was very effective
                   in the disinfection of secondary efflu-
                   ent. The performance of the process
                   could be described  empirically by the
                   initial coliform density,  the suspended
                   solids, the UV absorbance coefficient,
                   and the system loading rate, as defined
                   by Q/W, the ratio  of the flow to  the
                   actual UV output of the system. Over-
                   all, the study demonstrated  that  log
                   survival ratios of -3 to -4 could be
                   achieved consistently at practical sys-
                   tem loadings.  Similarly, the study
                   showed that a log survival ratio of as
                   low as  -3 could  be achieved with
                   primary effluent.
                      A mathematical expression was
                   developed and was found to respond
                   correctly to the variables associated
                   with the UV process. The study demon-
                   strated  that coliforms,  which are  oc-
                   cluded by suspended particles, are not

                   •Mention  of trade names  or commercial products
                   does not constitute endorsement or recommendation
                   for use
affected by UV light and, in effect, set
the limiting final density. The inactiva-
tion rate of the coliform was related to
the calculated  intensity  of the  UV
reactor.
   Suggestions are made with regard to
the maintenance and monitoring of the
system to enhance the efficiency and
cost  effectiveness of the process. A
cost analysis of the system shows it to
be cost effective and competitive with
chlori nation.

   This Project Summary was devel-
oped by EPA's Water Engineering Re-
search Laboratory, Cincinnati. OH. to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).

Introduction
   The use of ultraviolet (UV) radiation
for the disinfection of wastewaters is
accepted as an effective and economi-
cally attractive alternative to the use of
chlorine or ozone. As with any newly
emerging  technology, however, direct
field experience is limited, and system
designs have generally relied on empiri-
cal information. The primary objectives of
this study related to design considera-
tions for UV systems. The study was to
establish and demonstrate a design ap-
proach that would account for the major
process variables and that  would be
generically applicable  to alternative
equipment configurations  and waste-
water  applications. The program  also

-------
addressed the water quality parameters
appropriate to UV disinfection. Operation
and maintenance (O&M) requirements
were  assessed, with particular regard
to those tasks needed to maintain effec-
tive long-term performance. Finally, an
analysis of capital and O&M costs was
conducted.

Facilities and Experimental
Program

Facilities
   Port Richmond, located on the north
shore of Staten  Island, is 1 of 12 waste-
water treatment plants owned and op-
erated by the New York City Department
of Environmental Protection. This step-
aeration, activated-sludge facility has a
design capacity for the secondary system
of 227 million L/day (60 mgd). Excess
primary effluent can be bypassed around
the secondary portion of the plant. The
pilot facility was located so that there
was convenient access to both the sec-
ondary effluent and the  bypassed  pri-
mary effluent. A layout of the pilot facility
is presented in Figure 1. Either secondary
or primary effluent would be pumped to a
constant-head  tank, with the overflow
returned  to  the  plant bypass channel.
Flow would then be directed by gravity
to the UV  units.
   Units 1  and 2 were submerged quartz
systems. Each had 100 lamps in a sym-
metrical (10 by 10) array perpendicular to
flow. The  lamps were Voltarc 40 Watt
(nominal)  G36T6VH. They were 0.9 m
long and had an arc length of 0.75m. The
rated output for each lamp at 253.7 nm is
14W nominal. Each lamp was sheathed
in a quartz  sleeve with  a 2.3-cm di-
ameter. The only difference between the
two systems was that  the spacing be-
tween the quartz surfaces was 5.0 cm for
Unit 1 (7.3-cm  centerline spacing) and
1.25 cm for Unit 2 (3.55-cm centerline
spacing). Each of the quartz units was
equipped  with  mechanical wipers to
maintain the outer quartz surfaces. The
units were tested at flow rates between
800 and  2500 L/M; residence times
were generally between 1 and 20 sec.
   Unit 3 used Teflon tubes to carry the
liquid. The test compartment of Unit 3
contained eight Teflon tubes, each 3 m
long and 8.9 cm in diameter. The tubes
were  on  15-cm centerlines. The lamps
were placed on the outside of the Teflon
tubes. Two lamps were required to ex-
tend the length of the Teflon. The lamps,
each with an effective a re length of 1.5 m,
were  also placed on 15-cm centerlines
parallel to the Teflon tube.
            Unit 3
Unit
 1
                                            Meters
                                                 Power Panel
                             5. cm i
                                     Units
                                           —: 1.25 cm
                                    Eff. Tanks--'
                               g - Butterfly Valves -

                                    Inf. Tanks

                                      U.V.
Unit
 2
                        Power Supply
                    Lighting Supply
                                                         From Feed Pumps

                                                         —»~ To
                                                              Bypass
                                                                   To Bypass

                                                               -. Palmer-Bowk
                                                                Flumes
                                         Units 1 and 2
                                         Effluent
Figure 1.    Schematic layout of UV pilot facility at the Port Richmond Water Pollution Contr
           Plant, Staten Island. New York.
Experimental Program
   The Field program was  initiated in
December 1981 with the startup of the
two quartz systems. The Teflon unit was
in place by August 1982. The operational
period  for testing the  units extended
through September  1983. The  experi-
mental program was designed to monitor
system performance over a wide range of
loading conditions. The operational vari-
ables imposed on the units were the rate
of flow and the  number of lamps in
operation.
   Influent and effluent samples were
enumerated for total and fecal coliform
densities by direct membrane filter (MF)
techniques.  Influent  samples were
analyzed for suspended solids, turbidity,
pH, temperature, and  UV  absorbance
(total and filtered) and chemical  oxygen
demand. The  UV absorbance measure-
ment was conducted by standard spec-
trophotometric  procedures  and by  a
procedure that corrected  for the scat-
tering of light. Special studies that were
incorporated into the experimental pro-
gram included detailed  hydraulic tracer
                analyses, photoreactivation by the stati
                light  and  dark  bottle procedure, an
                specific procedures to monitor direct!
                the transmissibility of the quartz an
                Teflon enclosures and the UV output c
                the lamps.

                Results and Discussion
                  The total and fecal coliforms average
                (geometric mean) 1.02 x 106  and 3.61
                10B colonies/100  mL  for the 2-ye<
                period,  respectively. Relatively littl
                seasonal variation occurred for the col
                form   densities;  significant  variatior
                were seen, however, in the routine wati
                quality parameters such as the CO
                (average  = 44.5  mg/L) and  suspend*
                solids (average = 14.3 mg/L). These vari
                tions were generally seasonal in natur
                the higher-quality effluent occurre
                during  the warmer  summer  month
                when nitrification was typically  ai
                complished at the plant.
                  The UV absorbance coefficients (ba:
                e) were determined  by  two measun
                ment  techniques for both  total ar
                filtered samples.  The first  method w<

-------
simply a spectrophotometric  measure-
ment of the absorbance of a direct beam
of light (at a wavelength of 253.7 nm)
that is passed through a quartz cell with
a 1-cm pathlength. By this method, light
that does not pass directly  through the
cell and reach the detector \s considered
to be absorbed. The second method in-
corporates  an integrating  sphere  and
accounts for  light that  is scattered and
still available for germicidal  activity. The
spherical method is, in effect, a more
accurate measure of the true absorbance
than is the direct method.
   The direct absorbance measurements
were always higher than the spherical
absorbance  measurements,  indicating
that the more routinely practiced direct
method overestimates the loss of energy
in the  liquid. The direct UV absorbance
coefficient averaged 0.466 cm"1  and
0.404"1 for the total and filtered samples,
respectively. Average total  and filtered
UV coefficients of 0.372 cm"1 and 0.358
cm"1, respectively, were measured by
the spherical  method. The direct method
measures an absorbance for the total
sample that is approximately 17 percent
higher than that measured by the spheri-
cal method. A reasonable approximation
can be made by using the direct method
on a filtered sample; this approach could
be used in cases where the instrumenta-
tion  is not available to correct for scat-
tering effects.
   Tests were conducted over 2 weeks
using primary effluent as the feed to the
UV systems. The total and fecal coliforms
averaged  3.17  x 107 and  1.25 x 107
colonies/100 mL, respectively.  The
average  suspended  solids  were 80.9
mg/L; the UV absorbance  coefficients
averaged 0.865 and 0.747 cm"1 for the
direct  method, total and  filtered,  re-
spectively, and 0.593 and 0.533 cm"1 for
the spherical method, total  and filtered,
respectively.

Hydraulic  Characterization
   A procedure was demonstrated  for
experimentally constructing  a residence
time distribution (RTD)fortheopen chan-
nel quartz systems. This procedure in-
volved the steady-state injection of a
tracer upstream of the unit; the injection
would be discontinued, and the die-away
to a second steady-state  level  (back-
ground) would be measured downstream
of the  unit. The derivative of this curve
would  then be determined  and plotted
with time to yield the RTD curve.
   A series of RTD curves was developed
and analyzed to estimate the dispersion
coefficient, E, of the system. This coeffi-
cient measures the spread of the RTD
curve and indicates the unit's hydraulic
behavior with regard to plug flow (E ap-
proaches zero) and complete mix (E ap-
proaches infinity) conditions. The values
of E for the two quart systems were esti-
mated to be 1.5 cm2/sec and 15 cmVsec
for Units 1  and 2, respectively. By calcu-
lation, the Teflon  unit  was shown  to
approximate a plug flow reactor.  Data
were also  presented  to  demonstrate
turbulence for all three reactors.

Estimation of Available
UV Energy
   An important element in evaluating
a  UV system's performance  or in the
design of a system is the actual energy
available in the germicidal range. The key
is to understand  how efficiently the
253.7-nm energy of the low pressure
mercury  arc  lamp is  being used  and,
conversely, how it is being  lost. A con-
siderable effort was expended during the
study to directly monitor the output of the
lamps  and to quantify  the amount  of
energy available for disinfection. A sum-
mary of these data for the quartz systems
is presented in Figure 2, which estimates
the average UV output of the lamps with
     I
    IS
    
-------
   The energy  loss resulting from the
fouling of surfaces through which the UV
radiation must pass was measured under
varying conditions  and  after  specific
cleaning exercises.  With regard to the
quartz systems, this fouling occurred on
both the inner and outer surfaces of the
quartz. Cleaning was generally accom-
plished by a combined acidic/detergent
solution. The Teflon surfaces also be-
came dirty, and estimates were made of
the tube transmittance  with time. The
lower  line  in  Figure  2  is therefore an
estimate of the actual output being used,
which can be as low as 30 to 40 percent
of the nominal output. This information is
used to estimate the total  wattage  at
253.7 nm being transferred to the liquid
at the time of a given sampling.
Empirical Analysis of Process
Performance
   The  performances of the UV units
were empirically evaluated by a series of
multiple  linear regression analyses.
These correlated performance with the
operation of the units and with thequality
of the  Port  Richmond  secondary  and
primary effluents.  The  least  squares
method was used; the regression is cal-
culated stepwise, ordering the indepen-
dent  variants by decreasing degree of
significance. The dependent variable in
all cases was the log of the survival ratio
(Log  L/Lo) for either total or fecal coli-
forms.  The independent variables that
were tested to reflect wastewater quality
were the suspended solids, turbidity, and
either the direct or the spherical UV ab-
sorbance coefficients. The  variant se-
lected to  represent the operating condi-
tion of the particular unit was the ratio of
the flow  rate to the estimated total UV
output at 253.7 nm at the time of sam-
pling. This ratio, Q/W, is measured in
L/min  per watt.  The UV output, W, is
estimated from the number of lamps in
operation at the time of sampling and the
estimated average lamp output at the
time  (Figure 2).
   Performance was best predicted by
the ratio Q/W, the suspended solids con-
centration,  and the  spherical  UV ab-
sorbance  coefficient.  An   example of
solutions to the regression equations for
Units 1 and 2 is presented  in Figure 3.
This  figure shows the  loading to the
system (as defined by the ratio Q/W), as
a function of the UV absorbance  coef-
ficient  and  the  suspended solids to
achieve log fecal coliform survival ratios
of -3 and -4.
Application of the Proposed
Disinfection Model
   A mathematical model was developed
as part of the Port Richmond project to
describe the process performance of a UV
disinfection  system.  The expression is
written:
L =  Lo exp [ ux{
             2E
  + Lparticulate
                       +  4KE)1/2}]
                          U2
                                  (1 )
where:
         L  = the bacterial density re-
              maining  after exposure
              to  UV  (coliforms/100
              ml)

         Lo  = the initial bacterial den-
              sity  measured immedi-
              ately before entry into
              the  UV  reactor  (coli-
              forms/100 mL)
 x = the distance traveled b\
      an  element of watei
      while under direct ex
      posure to UV light (cm)
 u = the velocity of the waste
      water as it travels through
      the UV reactor (cm/sec)
      This quantity is calculate*
      as:
      u = x/(Vv/Q)
      where Q  is the flow rat(
      (L/sec) and V» is the re
      actor void volume (L)
 E = the  dispersion  coeffici
      ent (cmVsec)
 K = the  rate  coefficient fo
      the  inactivation of coli
      forms (sec"1).
iate = the bacterial density as
      sociated  with the  par
      ticulates  in  the waste
      water  (coliforms/10<
      mL)
                                                  r-SS{mg/l)
                                                 '
                                                                 Unit 2
                                                                 Fecal Coliform

                                                                     L  -3.0
            0.2
                                                                      0.8
   8.0
   6.0
 x

 Q.
  . 4.O
 a
   2.0
              SS (mg/l)
                          Unit 1
                          Fecal Coliform

                          Log  	= -4.0
                              1-0
                   0.4
                         0.6
                                         8.0
                                          6.0
                                          4.0
                                          2.0
                                           0
                                      1.0    O
              Unit 2
              Fecal Coliform

              Log — = -4.0
                  f-o
     SS Img/l)
                                                                 I
                                                                       I
 0.2
                                                                      0.8
                    a, (cm'1)
       Unit 1, Fecal Coliform — Secondary Effluent

       Log
           -L = 0.3/5 (Q/W) + 0.032 (SS)
           L°    + 2.44 (crj - 5.59
            0.4    0.6

              Us (cm''1)

Unit 2, Fecal Conform—Secondary Eff.

 Log — = 0.321 (Q/W) + 0.029 (SS)
      L°     + 2.52 (ctj - 5.56
 Figure 3.    Sizing requirements for the quartz units to achieve fecal coliform logL/L0of-3 and-
            in a secondryeffluent as a function of suspended solids and absorbance coefficier

-------
   Thus aside from knowing (or estab-
lishing) the physical dimensions of the
system (x, Vv)  and the  system loading
conditions (Q and L0), the model requires
knowledge of the hydraulics of the sys-
tem (E), the sensitivity of the coliforms to
UV (K), and the characteristics of the
wastewater (coliform occlusion in the
participates, and, as will  be discussed
shortly, the UV absorption properties of
the fluid). The  dispersion  coefficient of
the two units  (quartz)  has  been dis-
cussed. The inactivation rate is estimated
as a function of the intensity within the
UV reactor. Thus it is important that the
intensity be quantified for a given system
configuration.

UV Intensity
   Complex, multi-lamp  systems do not
allow for the direct measurement of the
actual UV intensity at any point within a
reactor. A microorganism moving through
a complex lamp system will'be exposed to
radiation  from  all  directions;  current
detector  systems  are  not  capable  of
adequately  accounting  for  all energy
under such conditions. However, a com-
putational method has been developed to
approximate the light intensity within a
system on the basis of the physical prop-
erties of the UV lamps, the configuration
of the multi-lamp reactors, and the prop-
erties  of  the  aqueous medium.  This
method is based on the  point source
summation technique, which presumes
the lamp  to be a finite series of point
sources that radiate energy radially in all
directions. The  intensity  at a given point
in a reactor is, then, the sum of intensi-
ties from each of these point sources. The
full report presents  the computational
framework for the method.
   The final  product of the  intensity
computation is  the average nominal in-
tensity in a reactor as a function of the UV
absorbance coefficient of the liquid. This
product is presented in Figure 4 for each
of the units operated at  Port Richmond.
An important note applies to these solu-
tions: The intensity is calculated at the
nominal output  of the lamp and assumes
that the quartz and Teflon enclosures will
transmit 100 percent of the energy. Thus
to estimate the actual intensity  under a
given set of conditions, it is necessary to
adjust the nominal  average intensity:
   Average intensity = (nominal average
   intensity)
   x (output fraction relative to nominal
   lamp output)

   x (transmittance relative to nominal
   enclosure transmittance)
     5
  X
 CM
  5
 •5!  4
  <2
 I
 S  3
  8
  I   i
  I
     Low Pressure Mercury Arc
      Rated Nominal Output ~
         18.2 Watts/m arc
          (at 253.7 nm)
                              Unit 2
Unit 3
  5:    0   0.2   0.4   0.6   0.8   1.0
  •^
          UV Absorbance Coefficient at
              253.7nm, asfcm'^i

Figure 4.   Solutions of intensity model for
           units 1, 2, and 3 as a function of
           the UV absorbance coefficient.

Estimate of Coliform Density
Associated with Particulates
   Coliforms that are heavily aggregated
or are retained in the suspended matter
typical of primary or secondary effluents
will not be affected by the UV radiation. A
select set of Port Richmond data were
analyzed to estimate the  densities as-
sociated with the particulates, as defined
by the suspended solids concentration.
The  samplings were those conducted
under very high-dose conditions  (Q/W
very low). The rationale was  that any
coliforms measured after such exposure
would  be attributable to the coliforms
retained in  the particulates.  A  linear
regression of the log of the effluent fecal
coliforms as a function of the log  of the
suspended solids,  when  transformed,
yields an expression in the form,
   Lparticulate — C Oo
where  SS  is the  suspended solids
(mg/L); in this case, c and d were deter-
mined to be 0.26 and 1.96, respectively.

Estimate of the Inactivation
Rate as a Function of the
Intensity
   Insimilarfashion, a subset of data was
selected to estimate the inactivation rate.
In this case,  the samplings are those in
which the dose was low enough  that a
significant  coliform  density would re-
main in the  exposed effluent. The rate
was estimated for  each  sampling  by
manipulating the model equation and
solving for K. The correlation of the Log K
as a function of the Log (average was found
to be linear; the data for fecal coliforms
are presented  in Figure 5. The trans-
formed expression has the form,
   K = a lb average
where (average has the units /uW/cm2. The
coefficients were determined to be
0.0000145 and  1.3 for a and b, respec-
tively.

Calibrated Disinfect/on Model
   The foregoing discussions presented
the analyses required to determine the
appropriate  model  coefficients. These
were  a and b to describe Kas a function of
the intensity,  c and d to describe the
Lparticuiate as a function of the suspended
solids, and the dispersion coefficient, E.
These coefficients can now be used in the
calibrated model to predict performance.
The predicted values were compared to
the observed  values as a test of the
validity of the model expression. These
analyses, which are presented in detail in
the report, showed the model expression
to respond  correctly to the variables
associated with UV design.
   Model solutions  were developed as
part  of the  study to demonstrate the
utility of the  model for design and for the
evaluation of existing systems. An exam-
ple is provided in Figure 6, which  pre-
sents the predicted performance of each
of the three  system configurations  as a
function of the system loading  and the
calculated intensity.  Note that these do
not account for the effect of the coliform
density associated with solids; the latter
would be additive to the levels predicted
by Figure 6. Several design examples are
presented to demonstrate the use of the
model for estimating design sizing for
varying wastewater and operating  con-
ditions.
                                 Cost Analysis
                                   A detailed  cost  analysis was con-
                                 ducted to determine current capital and
                                 O&M costs for the UV process.  The
                                 capital costs are presented on the basis of
                                 both equipment and installed costs; also
                                 included is a discussion of the facilities
                                 required to support the process.  The
                                 O&M is broken down to several  cost
                                 elements,  including labor, energy,  ma-
                                 terials, and system replacement parts.
                                 The details of the cost evaluation cannot
                                 be presented within the context of this
                                 summary,  but it  is critical that these
                                 details be  understood before the reader
                                 can make effective use of the cost curves
                                 presented  in the report. For this reason,
                                 we  refer to the full report only for a
                                 discussion  of the costs. The following
                                 figures can be used  as preliminary

-------
        20.0
    u
    01
    I
    I
        10.0

         8.0

         6.0


         4.0
2.0
         1.0

         0.8

         0.6


         0.4
         0.2
                          Fecal Coliform
                (Effluent Density Corrected for SS)
                     Unit 1 O
                     Unit 2 •
   ~  LogK-1.3Loglmr4.84
            [K = 0.0000145 /»,a13]
                 r = 0.914
                   \
                    \
                      \
\   \
J	\
600   WOO
                  2000     4000  6000
                     Average Intensity, l^
                                                  10,000   20,000
                                 60.000
Figure 5.   Estimation of the inactivation rate for fecal conforms as a function of the calculated
           average intensity.
screening factors in estimating the costs
associated with UV disinfection.
   The equipment  costs (reactor, ancil-
lary equipment, and replacement parts)
will range from $4800/kW for the larger
systems  (greater  than 400 lamps)  to
$7900/kW for the smaller systems (less
than 20 lamps). Note that these are 1984
costs; kW is total wattage (generally use
80 watts/lamp). The installation cost is
1.5 to 2.0 times  the equipment cost
(housing, piping, electrical, engineering,
etc.). A ballpark figure of $15,000/kW
can be used as an estimate of installed
cost.
   Exclusive of capital  amortization,
annual O&M costs are between $1200
and $1700/kW for the smaller systems
(less than lOkW) and between $600and
$800/kW for the larger systems (greater
than 300 kW). Setting the trend from
smaller to larger systems, the materials
cost accounts for 20 to 40 percent of the
annual O&M costs (75 percent of which
is the replacement of  lamps), power ac-
counts for 10 to 30 percent, and labor 70
to 30 percent.
                                   6
                               Conclusions
                                 The mathematical model expression
                               developed as part of this project correctly
                               responds to the major UV process and
                               equipment variables. When calibrated to
                               a specific wastewater application, the
                               model  can  be used to develop design
                               curves for specific equipment configura-
                               tions, and it can be used to describe the
                               operations  of  a system. The  critical
                               wastewater parameters required are the
                               design flow, suspended solids,  UV ab-
                               sorption coefficient, and the  initial
                               density.
                                 The ideal hydraulic design of  a UV
                               reactor is one with radially turbulent plug
                               flow. These conditions can be achieved:
                               (1) by designing for effective approach
                               and exit conditions to yield an even dis-
                               tribution  of flow across the entire lamp
                               reactor, (2)  by designing at higher veloci-
                               ties to encourage turbulence, and  (3)  by
                               having high  aspect ratios  (length  to
                               hydraulic radius).
                                 The average intensity of UV radiation
                               in a reactor can be calculated by the point
                               source summation method and described
as a function of the absorbance coef-
ficients of the wastewater. This estimate
must then take into account the actual
output of the UV source and the trans-
missibility of the enclosures separating
the source from the liquid.
   An effective parameter in describing
the sizing of a system is the radio of the
flow rate to the system's UV output,
Q/W. This ratio can be used empirically
and as an output parameter for the dis-
infection model. The wattage  must ac-
count for the losses associated with the
aging of the lamps and the degradation ol
the enclosure surfaces. Performance ol
the Port Richmond units was best de-
scribed by this ratio coupled  with the
suspended solids and the UV absorbance
coefficient.
   Log survival ratios greater than -5 car
be achieved for secondary effluents; the
effect of the solids, however, will be add!
live. Thus if the solids contribute densi
ties equal to  1 to 2 logs, then the actua
performances that can be achieved are
between -3 and -4; this level is  generally
sufficient for secondary effluents. U\
radiation is also effective in thedisinfec
tion of primary effluents, achieving lo(
survival  ratios as low as  -3  at  initia
densities greater than 107 colonies/10(
ml. Much of the residual density will b<
associated with the suspended solids ir
the effluent, and  the hydraulic loading;
are relatively low because of  the  higt
absorbance characteristics of the waste
water.
   The quartz systems  were  most ef
ficient than the Teflon system,  based or
the level of energy required to achieve
equivalent levels of performance. Nc
significant performance  difference;
were  found  between  the two quart;
systems,  which  differed only in  th<
spacing between the quartz surfaces.
   Essential to the proper design ant
operation of the system  is a very clea
understanding of the UV output of thi
reactor  and the transmissibility of thi
quartz  and  Teflon enclosures. Carefu
control of the average lamp output am
the transmittance can affect the costs fo
O&M of the system.
   With regard to the UV equipment, thi
process  lends itself to  simplicity an<
flexibility, attributes that should be main
tained  in the fabrication of the equip
ment. The reactor should be accessibl
for easy maintenance  and/or replace
ment  of the lamps  and quartz/Teflo
enclosures. Proper mating of the lamp
and  ballasts is  critical,  and  adequat
ventilation of the power panel should b
provided to protect the ballasts from ovei
heating.

-------
                                                                W/cm2)
                                                                 10,000
                                                                 Quartz (Unit 1)
                                                              Spacing 5.0 cm
                                                              5000
                                                             10.000
                                                             15.000
                                                             20.000
                                                                 Quartz (Unit 2)
                                                               Spacing 1.25 cm 	
                                                     70,000 Assumes 1.1 meters arc
                                                                per meter tube
                                                                18.2 Wn/m arc
              0.2   0.4  0.6  0.8  1.0   1.2   1.4  1.6  1.8   2.0  2.2  2.4

                                          1 pm/U. V.
               (1)  L = L0exp
                                                       ,
                                                       '
Figure  6.
                  L  2E               °     J

       where a - 0.0000145 and b = 1.3

   (2)  L,a would be additive.

   (3)  Assumes velocity is greater minimum velocity.


Design solutions for three Port Richmond system configurations showing per-
formance as a function of loading and intensity.
   The full report, prepared by  Hydro-
Qual,  Inc., was  submitted by the  New
York City Department of Environmental
Protection in fulfillment of Cooperative
Agreement No.  CR 807556 under the
partial sponsorship of the U.S. Environ-
mental Protection Agency.
                                                                                   "A" U. S. GOVERNMENT PRINTING OFFICE:! 986/646-116/20785

-------
                                         0. Karl Scheible and Maureen C. Casey are with HydroQual. Inc., Mahwah. NJ
                                           07430; and Angelika Forndran is with New York City Department of Environ-
                                           mental Protection. Wards Island, NY 10035.
                                         Albert D. Venosa is the EPA Project Officer (see below).
                                         The complete report, entitled "Ultraviolet Disinfection of Wastewaters from
                                           Secondary Effluent and Combined Sewer Overflows," (Order No. PB 86-145
                                           182/AS; Cost: $34.95, 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 Officer can be contacted at:
                                                Water Engineering Research Laboratory
                                                U.S. Environmental Protection Agency
                                                Cincinnati, OH 45268
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300

EPA/600/S2-86/005
              OC00329   PS
              U  S  EHVIR PROTECTION AGENCY
              RfSIOH  5  tIBJMRT   ^^
              230  S DEARBORN  STREET
              CHicaeo               IL

-------