UV
Disinfection
Guidance
Manual
C­
1
June
2003
Appendix
C.
Validation
of
UV
Reactors
To
receive
credit
for
Cryptosporidium,
Giardia,
or
virus
inactivation
using
UV
light,
the
Long
Term
2
Enhanced
Surface
Water
Treatment
Rule
(
LT2ESWTR)
requires
systems
to
demonstrate
that
the
UV
reactor
can
deliver
the
required
dose
through
validation
testing
(
40
CFR
141.729(
d)).
Furthermore,
validation
testing
must
determine
a
set
of
operating
conditions
that
can
be
monitored
by
the
control
system
to
ensure
that
the
UV
dose
required
for
a
given
pathogen
inactivation
credit
is
delivered
during
operation.
At
a
minimum,
these
operating
conditions
must
include
flowrate,
UV
intensity
measured
by
a
UV
intensity
sensor,
and
lamp
status.
The
validated
operating
conditions
must
account
for
the
following
factors
(
40
CFR
141,
Subpart
W,
Appendix
D):

 
Lamp
aging
 
Lamp
sleeve
fouling
 
UV
transmittance
of
the
water
 
Inlet
and
outlet
piping
or
channel
configurations
of
the
UV
reactor
 
Dose
distributions
arising
from
the
velocity
profiles
through
the
reactor
 
Failure
of
UV
lamps
or
other
critical
system
components
 
Measurement
uncertainty
of
on­
line
sensors
Unless
the
State
approves
an
alternative
approach,
validation
testing
must
involve
the
following
components:

 
Full­
scale
testing
of
a
UV
reactor,
which
conforms
uniformly
to
the
reactors
used
by
the
system
 
Inactivation
of
a
test
microorganism
whose
dose­
response
characteristics
have
been
quantified
with
a
low­
pressure
(
LP)
mercury
vapor
lamp
This
appendix
presents
one
approach
for
validating
UV
reactors.
Other
approaches
or
modifications
to
this
approach
may
be
used
at
the
discretion
of
the
State.
This
appendix
begins
with
an
overview
of
the
approach
for
conducting
validation
testing
and
interpreting
validation
results.
This
is
followed
by
a
description
of
the
materials,
equipment,
and
personnel
used
to
conduct
validation
testing
and
a
description
of
the
steps
involved
in
validating
UV
reactors.
The
appendix
ends
with
descriptive
examples
showing
how
validation
test
results
can
be
related
to
inactivation
credit.

Appendix
F
provides
more
detailed
background
information
on
validation
testing
and
includes
several
examples.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
2
June
2003
C.
1
Overview
UV
reactor
validation
should
provide
confidence
that
the
UV
reactor
is
appropriately
sized
for
a
given
disinfection
application
and
should
allow
a
water
treatment
plant
(
WTP)
to
receive
inactivation
credit
based
on
on­
line
measurements
of
flow,
UV
intensity,
lamp
status,
and,
in
some
cases,
UV
transmittance
(
UVT)
of
the
water
at
254
nm.
To
ensure
a
UV
reactor
is
appropriately­
sized
for
a
given
WTP,
validation
testing
should
provide
data
on
dose
delivery
and
monitoring
under
design
conditions
of
flow,
UVT,
and
lamp
output.
This
should
be
done
either
by
validating
UV
reactor
performance
under
those
conditions
or
by
validating
UV
reactor
performance
over
a
range
of
conditions
that
can
be
interpolated
to
obtain
performance
under
design
conditions.
To
allow
a
WTP
to
obtain
inactivation
credit
with
UV
disinfection,
validation
testing
should
provide
data
relating
on­
line
measurements
of
flow,
UV
intensity,
lamp
status,
and
UVT
to
UV
dose
levels
required
to
achieve
target
pathogen
inactivation
credit.
This
should
be
done
over
the
range
of
those
on­
line
measurements
expected
with
operation
of
the
UV
reactor
at
the
WTP.

UV
manufacturers
typically
produce
UV
reactors
as
part
of
a
product
line
where
each
reactor
is
manufactured
to
the
same
specifications.
If
a
representative
UV
reactor
from
that
product
line
undergoes
validation
testing,
the
test
results
can
be
applied
to
all
other
UV
reactors
within
that
product
line
if
those
reactors
are
manufactured
to
the
same
specifications
as
the
validated
reactor.
If
the
design
specifications
of
the
product
line
that
impact
dose
delivery
and
monitoring
change,
this
new
UV
reactor
design
must
be
re­
validated.

C.
1.1
Test
Protocol
The
validation
protocol
in
this
guidance
document
builds
on
well­
established
protocols
used
in
Europe
and
North
America.
A
UV
manufacturer
typically
delivers
a
UV
reactor
to
a
test
facility.
Test
personnel
inspect
the
UV
reactor
and
document
features
of
the
design
that
impact
dose
delivery
and
monitoring
(
e.
g.,
reactor
dimensions
and
sensor
properties).
The
UV
reactor
is
installed
within
a
biodosimetry
test
stand
with
inlet
and
outlet
piping
that
should
result
in
equal
or
worse
dose
delivery
than
with
the
reactor
installed
at
the
WTP.
The
UV
reactor
is
operated
under
various
test
conditions
of
flow,
UVT,
and
lamp
power.
The
test
condition
of
UVT
is
typically
obtained
using
a
UV­
absorbing
compound
injected
into
the
flow
upstream
of
the
UV
reactor.
A
challenge
microorganism
is
injected
into
the
flow
upstream
of
the
UV
reactor.
The
concentration
of
viable
challenge
microorganisms
is
measured
in
samples
collected
at
the
reactor's
inlet
and
outlet.
The
results
are
used
to
calculate
the
log
inactivation
of
the
challenge
microorganism
achieved
by
the
UV
reactor.
The
UV
dose­
response
of
the
challenge
microorganism
present
in
the
inlet
sample
is
measured
using
a
bench­
scale
device
termed
a
collimated
beam
apparatus.
The
UV
dose­
response
curve
is
used
to
relate
the
log
inactivation
observed
through
the
reactor
to
a
UV
dose
value
termed
the
Reduction
Equivalent
Dose
(
RED).
A
safety
factor
is
applied
to
the
results
to
account
for
any
bias
and
random
uncertainty
associated
with
the
validation
of
the
UV
reactor
and
the
on­
line
monitoring
approach
used
to
indicate
dose
delivery
both
during
validation
and
during
operation
at
the
WTP.
Last,
a
validation
report
is
prepared
that
describes
the
UV
reactor
tested,
the
test
protocol,
the
test
results,
and
the
inactivation
credits
that
can
be
assigned
to
the
UV
reactor
under
given
conditions
of
flow,
UVT,
and
lamp
output.
Figure
C.
1
presents
the
organization
of
this
validation
protocol
and
the
sections
within
this
appendix
that
address
each
of
these
issues.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
3
June
2003
Figure
C.
1
Elements
of
UV
Reactor
Validation
UV
System
Section
C.
2
Test
Facility
Section
C.
3
Provisions
for
Testing
Section
C.
2.1
UV
System
Documentation
Section
C.
2.2
Biodosimetry
Test
Stand
Section
C.
3.1
Collimated
Beam
Appendix
E
Microbiology
Methods
Appendix
D
UV
Intensity
Sensor
Test
Stand
Section
C.
3.2
Third
Party
Oversight
Section
C.
3.3
Testing
Section
C.
4
Evaluate
Influent
and
Effluent
Mixing
Section
C.
4.5
 
C.
4.6
UV
System
Inspection
and
Installation
Sections
C.
4.2
 
C.
4.4
Evaluate
UV
Sensors
Intensity
Section
C.
4.7
Evaluated
Impact
of
Lamp
Aging
Section
C.
4.8
Biodosimetry
Section
C.
4.9
Determining
Inactivation
Credit
Section
C.
4.10
Reporting
Section
C.
4.11
Develop
Approved
Test
Plan
Sections
C.
4.1
Examples
Section
C.
5
C.
1.2
Relating
RED
to
Target
Pathogen
Inactivation
Credit
Chapter
1
(
Table
1.4)
presents
the
UV
dose
needed
to
achieve
various
inactivation
credits
for
Cryptosporidium,
Giardia,
and
viruses.
The
dose
values
provided
in
Chapter
1
were
obtained
by
analyzing
UV
dose­
response
data
measured
using
a
bench­
scale
collimated
beam
device.
To
account
for
variability
in
the
dose­
response
of
the
pathogen,
an
80
percent
predictive
credible
interval
was
used
to
determine
dose
values
needed
to
achieve
a
given
log
inactivation
of
the
pathogen.
The
derivation
of
the
UV
dose
requirements
is
presented
in
Appendix
B.
This
assessment,
however,
does
not
account
for
the
measurement
uncertainty
associated
with
UV
reactor
validation
and
on­
line
dose
monitoring.
To
account
for
this
uncertainty,
the
RED
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
4
June
2003
measured
during
reactor
validation
should
be
equal
to
or
greater
than
a
target
RED
defined
using
the
following
equation:

(
)
P
Poly
RED
T
D
e
1
B
B
RED
×
+
×
×
=
Equation
C.
1
where
REDT
=
Target
RED
that
should
be
demonstrated
during
validation
BRED
=
RED
bias
BPoly
=
Polychromatic
bias
e
=
Expanded
uncertainty
expressed
as
a
fraction
DP
=
UV
dose
in
Chapter
1
(
Table
1.4)
required
for
a
given
level
of
target
pathogen
inactivation
credit.

The
RED
bias
term
accounts
for
the
difference
between
the
dose
delivered
to
the
target
pathogen
and
the
dose
measured
using
a
challenge
microorganism.
If
the
challenge
microorganism
is
more
resistant
to
UV
light
than
the
target
pathogen,
the
RED
measured
during
validation
will
be
greater
than
the
dose
delivered
to
the
pathogen.
The
magnitude
of
the
difference
will
depend
on
the
dose
distribution
of
the
UV
reactor
and
the
inactivation
kinetics
of
the
challenge
microorganism
and
the
target
pathogen.
If
the
challenge
microorganism
is
as
sensitive
or
more
sensitive
to
UV
light
than
the
target
pathogen,
the
RED
bias
has
a
value
of
1.00.
A
recommended
approach
for
obtaining
the
value
of
the
RED
bias
is
given
in
section
C.
4.10.2.

The
polychromatic
bias
term
accounts
for
spectral
differences
in
the
lamp
output,
lamp
sleeve
UV
transmittance,
UVT,
and
action
spectrum
of
the
challenge
microorganism
between
validation
and
operation
of
a
UV
reactor
equipped
with
medium­
pressure
(
MP)
lamps.
These
differences
can
cause
the
dose
delivered
at
the
WTP
to
differ
from
the
dose
measured
during
validation.
Depending
on
the
spectral
response
and
positioning
of
the
UV
intensity
sensor
and
the
dose
distribution
of
the
UV
reactor,
the
dose
delivered
at
the
WTP
can
be
less
than
dose
measured
during
validation
and
indicated
by
the
monitoring
system.
The
polychromatic
bias
term
accounts
for
this
issue.
The
polychromatic
bias
only
applies
to
UV
reactors
that
use
polychromatic
UV
lamps.
With
UV
reactors
using
LP
or
low­
pressure
high­
output
(
LPHO)
lamps,
the
polychromatic
bias
equals
1.00.
A
recommended
approach
for
obtaining
the
value
of
the
polychromatic
bias
is
given
in
section
C.
4.10.2.

The
expanded
uncertainty,
e,
accounts
for
the
uncertainty
in
the
measurements
taken
during
validation
and
used
with
dose
delivery
monitoring.
In
this
protocol,
the
numeric
value
of
the
expanded
uncertainty
is
estimated
using
an
80
percent
confidence
level
by
summing
the
individual
measurement
uncertainties
associated
with
on­
line
sensors
used
in
the
field
and
during
validation,
influent
and
effluent
challenge
microorganism
concentrations,
challenge
microorganism
UV
dose­
response,
and
quantification
of
the
UV
output
from
the
lamps.
This
approach
is
described
in
section
C.
4.10.2.

Two
approaches,
termed
Tier
1
and
Tier
2,
are
presented
in
section
C.
4.10
for
applying
the
RED
bias,
polychromatic
bias,
and
the
expanded
uncertainty
to
define
the
target
RED
values.

The
Tier
1
approach,
described
in
section
C.
4.10.1,
is
a
standardized
approach
that
uses
prescribed
values
for
the
RED
bias,
the
polychromatic
bias,
and
the
expanded
uncertainty
to
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
5
June
2003
define
RED
targets
to
be
demonstrated
during
validation.
To
use
the
Tier
1
approach,
the
dose
monitoring
and
validation
should
meet
defined
criteria
on
reactor
design,
challenge
microorganism
UV
dose­
response,
UV
absorber
used
during
validation,
sensor
properties,
monitoring
approach,
and
microbiology.

The
Tier
2
approach,
described
in
section
C.
4.10.2,
allows
the
user
to
calculate
the
values
of
the
RED
bias,
the
polychromatic
bias,
and
the
expanded
uncertainty,
and
to
use
those
values
to
define
the
RED
target
to
be
demonstrated
during
validation.
The
approach
does
not
prescribe
criteria
for
reactor
design,
challenge
microorganism
dose­
response,
the
UV
absorber
used
during
validation,
sensor
properties,
monitoring
approach,
or
microbiology.

C.
1.3
Other
Validation
Protocols
Validation
of
UV
reactors
used
in
drinking
water
applications
has
been
practiced
in
North
America
and
Europe
using
well­
established
protocols
that
include
the
following,
shown
in
chronological
order
of
development:

 
National
Sanitation
Foundation/
American
National
Standards
Institute
(
NSF/
ANSI)
Standard
55
 
Austrian
Standards
Institute
(
ÖNORM
;
Österreichisches
Normungsinstitut)
M
5873­
1
 
German
Association
for
Gas
and
Water
(
DVGW;
Deutsche
Vereinigung
des
Gas­
und
Wasserfaches)
W294
 
National
Water
Research
Institute/
American
Water
Works
Association
Research
Foundation
(
NWRI/
AwwaRF)
UV
Guidelines
UV
validation
conducted
as
per
DVGW
and
ÖNORM
demonstrates
that
a
UV
reactor
will
deliver
a
RED
of
40
mJ/
cm2
measured
using
Bacillus
subtilis
spores.
Validation
as
per
these
protocols
should
meet
criteria
for
the
UV
reactor
and
its
validation.
UV
validation
conducted
as
per
NWRI/
AwwaRF
Guidelines
and
NSF
Standard
55
both
use
MS2
bacteriophage
(
MS2)
as
a
challenge
microorganism.
NSF
standard
55
specifies
a
target
RED
of
40
mJ/
cm2
while
NWRI/
AwwaRF
Guidelines
does
not
specify
a
target
RED.
Validation
testing
as
per
NWRI/
AwwaRF
Guidelines
and
NSF
Standard
55
should
be
assessed
for
consistency
with
the
guidance
for
test
conditions
provided
in
section
C.
4.9.
Results
should
be
interpreted
as
per
the
guidance
provided
in
sections
C.
4.9
and
C.
4.10.

C.
1.4
Planning
UV
Validation
In
general,
validation
testing
will
be
conducted
either
for
a
UV
manufacturer
who
wishes
to
validate
a
given
UV
reactor
for
the
drinking
water
market
or
for
a
utility
that
wishes
to
validate
a
UV
reactor
for
a
specific
application.
Regardless
of
the
end
user,
parties
conducting
validation
testing
should
develop
a
test
plan
that
addresses
the
following
questions:

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
6
June
2003
 
Where
will
validation
take
place?

 
What
test
conditions
of
flow,
UVT,
and
lamp
output
should
be
tested?

 
What
UV
absorbers
and
challenge
microorganisms
should
be
used?

 
What
inlet
and
outlet
conditions
will
be
used
during
validation?

 
Who
will
provide
the
challenge
stock
solutions
and
assay
water
samples?

 
How
will
UV
intensity
sensor
properties
during
validation
be
verified?

 
Who
will
conduct
collimated
beam
testing?

 
What
is
the
expected
safety
factor
and
is
it
acceptable?

 
Who
will
provide
third
party
oversight?

 
What
State
review
and
approval
is
needed
for
the
test
protocol?

When
planning
how
validation
testing
will
be
done,
utilities
and
manufacturers
should
determine
if
they
want
to
evaluate
validation
results
under
Tier
1
or
Tier
2.
They
should
assess
if
the
planned
validation
will
meet
the
Tier
1
criteria
and
develop
preliminary
estimates
of
the
safety
factor
that
would
be
applied
under
Tier
2.
They
should
explore
opportunities
to
optimize
validation
testing
by
identifying
approaches
that
minimize
the
values
of
the
RED
bias,
polychromatic
bias,
and
expanded
uncertainty
terms
used
to
determine
the
safety
factor.
To
provide
flexibility
in
using
Tier
1
and
2,
one
approach
would
be
to
ensure
validation
meets
Tier
1
criteria
and
then
to
optimize
for
Tier
2.

C.
1.4.1
UV
Validation
for
Manufacturers
UV
manufacturers
will
conduct
validation
either
for
a
specific
WTP
or
to
allow
broad
application
of
their
UV
reactor
to
many
WTPs.
If
validation
is
being
done
to
allow
broad
application
of
the
UV
reactor,
the
test
conditions
of
flowrate,
UVT,
and
lamp
output
will
likely
span
a
larger
range
than
the
test
conditions
that
would
be
used
when
validating
for
a
specific
WTP.
The
UV
manufacturer
may
also
validate
the
UV
reactor
for
a
range
of
dose
targets
that
allow
the
UV
reactor
to
achieve
credit
for
a
range
of
pathogen
log
inactivation
values.
The
number
of
test
conditions
and
dose
targets
chosen
should
be
sufficient
to
allow
interpolation
of
validation
data
to
conditions
of
flowrate,
UVT,
and
lamp
output
specific
for
a
given
WTP
application.

For
broad
application
of
validation
results,
inlet
and
outlet
conditions
should
be
chosen
to
provide
a
conservative
yet
practical
representation
of
inlet/
outlet
piping
used
at
WTPs.
For
example,
if
the
UV
reactor
is
typically
applied
in
a
filter
gallery,
it
may
make
sense
to
test
with
a
90
degree
bend
immediately
upstream
of
the
reactor
to
represent
a
"
worst
case"
scenario.
On
the
other
hand,
if
a
UV
reactor
is
typically
installed
with
5
or
10
pipe
diameters
of
straight
pipe
upstream
of
the
reactor
inlet,
it
may
make
sense
to
test
with
a
90­
degree
bend
immediately
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
7
June
2003
upstream
of
a
5
pipe
diameters
of
straight
pipe.
UV
manufacturers
can
use
computational
fluid
dynamics
(
CFD)
as
a
tool
to
understand
the
impact
of
inlet
and
outlet
conditions
on
the
dose
delivery
of
their
UV
reactors
in
order
to
best
identify
the
inlet
and
outlet
conditions
most
representative
of
a
wide
range
of
applications.
In
order
to
facilitate
regulatory
approval
in
the
States,
validation
testing
should
be
conducted
using
recognized
and
accepted
protocols.
Alternatively,
the
UV
manufacturer
should
solicit
feedback
and
approval
for
the
validation
test
plan
from
the
State(
s)
before
testing.

C.
1.4.2
UV
Validation
for
Utilities
Utilities
have
the
option
of
validating
UV
reactors
either
at
a
UV
test
facility
or
on­
site
at
their
WTP.
Utilities
considering
on­
site
validation
should
address
recommendations
on
water
quality,
disposal,
and
test
train
requirements
provided
in
section
C.
3.1.
Potential
issues
include
obtaining
water
with
a
sufficiently
high
UVT
that
allows
validation
over
the
entire
UVT
range
expected
at
the
WTP,
providing
sufficient
mixing
of
additives
prior
to
entering
the
UV
reactor
and
mixing
of
the
challenge
microorganisms
after
the
reactor,
and
obtaining
permits
for
the
disposal
of
the
water
used
for
validation.
Utilities
considering
off­
site
validation
at
a
test
facility
should
ensure
that
the
inlet
and
outlet
conditions
used
during
validation
are
representative
of
those
conditions
used
at
the
WTP.
Recommendations
for
inlet
and
outlet
conditions
to
be
used
during
UV
validation
are
provided
in
section
C.
3.1.5.

If
on­
site
validation
is
considered,
the
utility
should
identify
who
will
provide
microbiological
support
for
validation
testing.
The
utility
could
use
either
their
own
microbiological
lab
or
a
third
party
lab.
Regardless
of
the
approach,
the
microbiology
lab
should
have
demonstrated
experience
working
with
the
challenge
microorganism
and
be
able
to
provide
timely
analysis
of
water
samples
collected
during
validation
testing.
Appendix
D
provides
detail
on
the
microbiological
lab
qualifications
and
includes
growth
and
assay
methods
for
two
commonly
used
challenge
microorganisms.

With
on­
site
validation,
the
utility
should
also
identify
how
it
will
verify
the
performance
of
UV
intensity
sensors
used
during
validation.
Because
utility
staff
typically
do
not
have
experience
in
optoelectronic
instrumentation,
they
should
use
a
third
party
laboratory
to
benchmark
sensor
performance.
Sections
C.
3.2
and
C.
4.7
describe
the
laboratory
needs
and
the
measurements
used
to
benchmark
sensor
performance.

C.
2
UV
Reactor
This
section
describes
the
hardware
and
documentation
that
the
UV
manufacturer
should
provide
to
the
validation
facility.

C.
2.1
Provisions
for
Testing
The
UV
manufacturer
should
provide
for
validation
a
UV
reactor
with
the
following
characteristics:

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
8
June
2003
 
A
UV
reactor
that
matches
the
technical
description
in
the
documentation
provided
as
per
section
C.
2.2.

 
UV
lamps
that
have
undergone
appropriate
burn­
in.
The
recommended
burn­
in
period
is
100
hours.

 
Lamps
aged
to
give
end­
of­
lamp­
life
conditions
if
the
reactor
is
to
be
tested
with
aged
lamps.

 
Provisions
to
reduce
lamp
output
as
per
section
C.
4.9.4.

 
Provisions
to
measure
the
UV
output
of
each
lamp
and
the
electrical
power
delivered
to
the
lamps
as
per
section
C.
4.9.2.

 
On­
line
and
reference
UV
intensity
sensors
that
meet
the
technical
description
provided
in
the
documentation.

 
A
safety
cut­
off
switch
to
prevent
overheating
if
LPHO
or
MP
lamps
are
used.

C.
2.2
UV
Reactor
Documentation
Prior
to
validation
testing,
the
UV
manufacturer
should
provide
to
the
party
conducting
the
tests
documentation
identifying
and
describing
the
UV
reactor.
Documentation
should
include
all
reactor
and
component
information
that
impacts
dose
delivery
and
monitoring
including
the
following:

 
Technical
descriptions
of
the
reactor
and
all
internal
components,
including
lamps,
sleeves,
UV
intensity
sensors,
baffles,
and
cleaning
mechanisms.
The
technical
description
should
include
dimensions
and
placement
of
all
wetted
components.

 
Technical
descriptions
of
the
inlet
and
outlet
piping
to
the
reactor
undergoing
validation,
including
the
length
and
cross­
sectional
dimensions
of
any
pipes,
channels,
and
bends,
and
dimensions
of
any
hydraulic
structures
affecting
flow.
If
reactors
are
validated
in
series,
technical
descriptions
of
the
piping
between
reactors
should
be
provided.

 
Lamp
specification
stating
the
lamp
manufacturer
and
product
number,
electrical
power
rating,
length
from
electrode
to
electrode,
spectral
output
of
new
and
aged
lamps,
mercury
content,
and
envelope
diameter.
The
spectral
output
should
be
specified
for
5
nm
intervals
or
less
over
a
wavelength
range
that
includes
the
response
range
of
the
UV
intensity
sensors
and
the
germicidal
range.

 
Sleeve
specifications
indicating
sleeve
dimensions,
material,
and
UV
transmittance
from
200
to
400
nm.

 
Technical
description
of
the
placement
of
the
lamp
within
the
sleeve.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
9
June
2003
 
Specifications
for
the
reference
and
on­
line
UV
intensity
sensors
indicating
manufacturer
and
product
number,
external
dimensions,
and
measurement
properties.
Measurement
properties
include
spectral
and
angular
response,
working
range
and
linearity,
calibration
factor,
temperature
stability,
long­
term
stability,
and
measurement
uncertainty.
Data
and
calculations
should
be
provided
showing
how
the
total
measurement
uncertainty
of
the
sensor
is
derived
from
the
individual
sensor
properties.
Table
C.
1
gives
an
example
of
the
calculation
of
sensor
measurement
uncertainty
from
the
uncertainty
that
arises
due
to
each
sensor
property.

Table
C.
1
Example
of
a
UV
Intensity
Sensor
Uncertainty
Datasheet
Property
Uncertainty
(%)
Calibration
8
Linearity
5
Temperature
response
3
Angular
response
5
Spectral
response
1
Long
term
drift
10
Total
Uncertainty1
15
1
Total
uncertainty
is
calculated
as
the
square
root
of
the
sum
of
the
squared
individual
uncertainties.
In
this
example,
total
uncertainty
is
(
82+
52+
32+
52+
12+
102)
1/
2
=
15%.

 
Specifications
for
the
UV
intensity
sensor
port
indicating
all
dimensions
and
tolerances
that
impact
the
positioning
of
the
sensor
relative
to
the
lamps.

 
If
the
sensor
port
contains
a
monitoring
window
separate
from
the
sensor,
specifications
giving
the
window
material,
thickness,
and
UV
transmittance
from
200
to
400
nm
should
be
provided.

 
Technical
description
of
the
algorithm
used
by
the
reactor
to
monitor
dose
delivery,
including
the
use
of
sensors,
signal
processing,
and
calculations.

Documentation
should
also
be
provided
on
the
proper
installation
and
operation
of
the
reactor
to
ensure
proper
and
safe
validation
testing,
including:

 
Flowrate,
headloss,
and
pressure
rating
of
the
reactor
 
Assembly
and
installation
instructions
 
Electrical
requirements
including
required
line
frequency,
voltage,
amps,
and
power
 
Operation
and
maintenance
manuals
that
include
cleaning
procedures,
required
spare
parts,
and
safety
requirements.
Safety
requirements
should
include
information
on
electrical
lockouts,
eye
and
skin
protection
from
UV
light,
safe
handling
of
lamps,
and
mercury
cleanup
recommendations
in
the
event
of
a
lamp
breakage
Lastly,
the
UV
manufacturer
should
consider
providing
the
following
information
relevant
to
the
test
procedure:

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
10
June
2003
 
Specifications
for
the
challenge
microorganism
to
be
used
during
validation
that
includes
protocols
required
for
growth
and
enumeration,
expected
UV
dose­
response,
and
suitability
for
use
in
validation
testing
as
discussed
in
section
F.
1.4.

 
Specifications
for
the
UV
absorber
to
be
used
during
validation.

 
A
description
of
the
test
conditions
of
flowrate,
UVT,
and
lamp
output
used
to
validate
the
reactor,
and
the
expected
measurements
of
UV
intensity
and
challenge
microorganism
RED.

C.
3
Test
Equipment,
Facilities,
and
Personnel
This
section
describes
the
test
equipment,
laboratory
facilities,
and
personnel
that
are
typically
used
during
validation
testing,
including
the
following
components:

 
Biodosimetry
test
stand
for
measuring
challenge
microorganism
inactivation
by
the
UV
reactor
 
UV
intensity
sensor
test
stand
for
measuring
sensor
properties
 
Third
party
oversight
Appendix
D
provides
information
on
the
microbiological
laboratory
with
specific
information
on
the
growth
and
assay
of
MS2
bacteriophage
and
B.
subtilis
spores.
Appendix
E
provides
information
on
collimated
beam
apparatus
used
to
measure
the
UV
dose­
response
of
the
challenge
microorganism.

C.
3.1
Biodosimetry
Test
Stand
The
biodosimetry
test
stand
is
used
to
measure
the
inactivation
of
a
challenge
microorganism
by
the
UV
reactor
operating
under
controlled
conditions
of
flowrate,
UVT,
and
lamp
output.

Figure
C.
2
presents
a
block
diagram
of
such
a
test
stand
with
the
following
features:

 
Water
supply
with
rate­
of­
flow
control
and
backflow
prevention
 
Dosing
pumps
and
ports
for
injecting
the
challenge
microorganism,
the
UV­
absorbing
compound,
and,
if
required,
a
disinfectant
residual­
quenching
agent
 
Influent­
mixing
device
(
static
mixer
or
length
of
pipe)
upstream
of
the
reactor
to
ensure
the
challenge
microorganism
and
UV­
absorbing
compound
are
well­
mixed
prior
to
entering
the
reactor
 
Influent
sampling
port
after
the
influent­
mixing
device
and
before
the
reactor
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
11
June
2003
 
Inlet
and
outlet
piping
to
the
reactor
that
results
in
a
dose
delivery
equal
to
or
less
than
the
dose
delivery
expected
with
the
installation
of
the
reactor
at
a
WTP
 
UV
reactor
under
test
 
Ports
to
allow
head­
loss
measurements
across
the
UV
reactor
 
Effluent­
mixing
device
(
static
mixer
or
length
of
pipe)
downstream
of
the
reactor
to
ensure
that
the
challenge
microorganisms
that
survive
inactivation
by
the
reactor
are
well­
mixed
prior
to
sampling
 
Effluent
sampling
port
after
the
effluent­
mixing
device
 
Water
disposal
facilities
Figure
C.
2
Block
Diagram
of
the
Biodosimetry
Test
Stand
Challenge
Microbe
Backflow
Prevention
Static
Mixer
Static
Mixer
Flow
meter
UV
Reactor
Valve
To
Waste
Effluent
Sample
Port
Influent
Sample
Port
Valve
Water
Supply
Inlet
Piping
UV
Absorber
Outlet
Piping
Pressure
Gage
Pressure
Gage
Influent
Quenching
Agent
C.
3.1.1
Water
Supply
Validation
testing
should
prove
that
the
monitoring
of
dose
delivery
by
the
UV
reactor
is
valid
over
the
full
range
of
UVT
values
expected
with
the
application
of
the
UV
reactor
at
the
WTP.
Typically,
the
UVT
of
the
water
supply
used
for
validation
is
high
and
UV
absorbing
chemicals
are
added
upstream
of
the
reactor
to
simulate
different,
lower
UVTs
over
the
test
range.
For
validation
results
to
be
generally
applied
to
all
WTPs,
the
water
supply
should
have
a
UVT
at
254
nm
greater
than
97
percent
(
UV
absorption
coefficient
less
than
0.013
cm­
1
with
a
10
nm
path
length).

Whether
coagulants
are
naturally
present
(
e.
g.,
reduced
iron
in
ground
water)
or
added
as
part
of
water
treatment,
they
can
affect
the
challenge
microorganism
concentration,
the
turbidity,
and
the
UVT
of
water
samples
collected
during
reactor
validation
(
Petri
et
al.
2000).
Coagulation
of
the
challenge
microorganism
can
lead
to
reduced
counts
and
poor
sample­
tosample
repeatability.
To
avoid
these
effects,
the
water
supply
should
not
contain
coagulants
that
interfere
with
the
validation
results.
Alternatively,
chelating
agents
or
coffee
can
be
used
as
an
additive
to
counter
these
effects
(
Petri
et
al.
2000).

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
12
June
2003
The
water
passing
through
the
reactor
should
not
contain
disinfectant
residuals
that
inactivate
the
challenge
microorganism
during
testing.
If
the
water
does
contain
a
disinfectant
residual,
a
quenching
agent
should
be
injected
into
the
water
upstream
of
the
microorganism
injection
port.
The
quenching
agent
should
have
a
minimal
impact
on
the
UVT.

The
water
supply
(
volume
and
flowrate)
should
be
sufficient
to
allow
testing
over
the
rated
flow
range
of
the
UV
reactor.
A
flow­
control
device
(
e.
g.,
variable
speed
pump
or
valve)
can
be
used
to
vary
the
flow
over
that
range.
A
flowmeter
with
a
known
measurement
uncertainty
should
monitor
the
flowrate
through
the
UV
reactor.

Backflow
prevention
should
be
used
with
a
potable
water
supply.
Backflow
prevention
can
be
obtained
using
reduced
pressure
zone
(
RPZ)
backflow
preventers,
air
gaps,
or
check
valves.

C.
3.1.2
Dosing
of
Additives
Challenge
microorganisms,
UV­
absorbing
compounds,
and
possibly
disinfectant
quenching
agents
may
be
injected
into
the
flow
upstream
of
the
UV
reactor
during
validation.
If
pumps
are
used
to
inject
the
additives,
they
should
provide
a
pulseless
flowrate
or
have
a
cycle
time
an
order
of
magnitude
less
than
the
residence
time
of
the
reactor.
The
flowrate
generated
by
the
pump
should
be
stable
over
the
time
required
to
take
samples
as
per
section
C.
4.9.5.
An
injection
port
using
standardized
injector
technologies
can
be
used
to
disperse
the
additives
into
the
flow.

C.
3.1.3
Mixing
of
Reactor
Influent
and
Effluent
Additives
passed
through
the
reactor
should
be
well­
mixed
through
the
cross­
section
of
the
pipe
prior
to
the
reactor
influent
sampling
port.
The
challenge
microorganisms
surviving
UV
disinfection
should
be
well­
mixed
through
the
pipe
cross­
section
prior
to
the
reactor
effluent
sampling
port.
Mixing
can
be
achieved
either
using
static
mixers
or
by
relying
on
the
turbulent
mixing
present
in
the
lengths
of
pipe
upstream
of
the
sampling
ports.
If
the
water
passed
through
the
UV
reactor
is
obtained
from
a
large
tank,
the
additives
can
be
premixed
in
the
tank
to
obtain
a
uniform
concentration
for
testing.

C.
3.1.4
Sample
Taps
The
sample
taps
should
be
located
to
provide
representative
samples
of
undisinfected
water
entering
the
reactor
and
the
disinfected
water
leaving
the
UV
reactor.
If
the
influent
sample
tap
is
located
too
close
to
the
reactor
influent,
the
samples
collected
may
be
exposed
to
UV
light,
resulting
in
underestimation
of
the
influent
concentration
of
the
challenge
microorganism.
If
the
effluent
sample
tap
is
located
too
close
to
the
reactor
effluent,
the
effluent
samples
will
be
collected
before
full
exposure
to
UV
light
and
the
effluent
concentration
of
the
challenge
microorganism
will
be
overestimated.
The
UVT
of
the
water
can
be
used
to
calculate
how
far
UV
light
from
the
reactor
penetrates
the
water
upstream
and
downstream
of
the
reactor.
The
sampling
points
should
be
located
far
enough
from
the
UV
reactor
that
the
germicidal
UV
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
13
June
2003
intensity
at
the
point
of
sampling
is
less
than
0.1
percent
of
the
germicidal
intensity
within
the
UV
reactor.

Sample
taps
may
sample
from
a
single
point
within
the
flow
or
from
multiple
points
at
the
same
time.
Samples
taken
from
multiple
points
within
the
flow
should
have
the
same
concentration
of
additives
and
microorganisms
within
the
measurement
error.

Sampling
taps
should
remain
open
over
the
duration
of
the
test.
Sample
collection
should
meet
standards
of
good
practice
as
defined
by
Standard
Methods
Section
9060
(
APHA
et
al.
1995).
Samples
should
be
collected
in
bottles
that
have
been
cleaned
and
sterilized.
Samples
collected
should
be
immediately
stored
on
ice
within
a
cooler
in
the
dark
until
needed
for
analysis.

C.
3.1.5
UV
Reactor
Inlet
and
Outlet
Conditions
As
stated
previously,
the
inlet
and
outlet
structures
to
the
UV
reactor
during
validation
should
result
in
equal
or
worse
dose
delivery
than
with
the
reactor
installed
at
the
WTP.
EPA
recommends
using
any
one
or
combination
of
the
following
approaches:

 
Inlet
and
outlet
conditions
used
at
the
WTP
match
those
used
during
validation
for
at
least
10
pipe
diameters
upstream
and
5
pipe
diameters
downstream
of
the
reactor.

 
UV
reactor
is
validated
either
with
a
90­
degree
bend
immediately
upstream
of
the
reactor
inlet
or
a
with
90­
degree
bend
followed
by
a
length
of
straight
pipe
immediately
upstream
of
the
reactor
inlet.
The
reactor
is
installed
at
the
WTP
with
a
length
of
straight
pipe
immediately
upstream
of
the
reactor
equal
to
5
pipe
diameters
plus
any
length
used
after
the
90­
degree
bend
during
validation.
To
avoid
jetting
effects,
piping
upstream
of
the
straight
pipe
length
should
not
have
expansions
for
at
least
10
pipe
diameters
and
any
valves
located
in
that
length
of
pipe
should
always
be
fully
open
during
operation
of
the
reactor.
With
this
approach,
it
is
assumed
that
the
90­
degree
bend
immediately
upstream
of
the
reactor
inlet
provides
worse
hydraulics
than
the
installation.
This
approach
assumes
that
the
reactor
design
has
not
been
optimized
for
the
90­
degree
bend
inlet.

 
Velocity
of
the
water
measured
at
evenly­
spaced
points
through
a
given
cross
section
of
the
flow
upstream
and
downstream
of
the
reactor
is
within
20
percent
of
the
theoretical
velocity
with
both
the
validation
test
stand
and
the
installation.
The
theoretical
velocity
is
defined
as
the
flowrate
divided
by
the
cross­
sectional
area.

CFD­
based
dose
modeling
can
be
used,
in
tandem
with
one
of
the
above­
mentioned
approaches,
to
show
that
dose
delivery
with
the
installation
is
better
than
dose
delivery
during
validation
for
given
conditions
of
flowrate,
UVT,
and
lamp
output.
To
account
for
uncertainty
in
CFD
predictions
of
dose
delivery
(
Petri
and
Olson
2001,
Wright
and
Hargreaves
2002),
CFD
predictions
of
dose
delivery
during
validation
should
be
at
least
20
percent
greater
than
predictions
of
dose
delivery
at
the
WTP.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
14
June
2003
C.
3.1.6
Quality
Assurance
and
Quality
Control
Flowmeters,
injection
pumps,
pressure
gauges,
and
other
measuring
devices
used
should
bear
evidence
of
being
in
calibration.
Accuracy
of
instrumentation
should
be
checked
by
comparison
with
standard
measurements.
The
documentation
describing
the
test
facility
should
be
provided
and
verified
including
the
following
items:

 
A
description
of
the
validation
test
stand,
including
all
piping,
valves,
flowmeters,
mixers,
pumps,
sampling
locations,
and
measurement
instrumentation
 
The
measurement
uncertainty
and
the
last
calibration
date
of
all
measurement
instrumentation
Comparisons
of
on­
line
instrumentation
with
standard
measurements
 

C.
3.2
UV
Intensity
Sensor
Test
Stand
The
properties
of
the
on­
line
and
reference
UV
intensity
sensors
should
be
measured
by
an
independent
laboratory
that
is
equipped
to
confirm
sensor
calibration
and
measure
the
sensor's
angular
and
spectral
response,
linearity
over
the
working
range,
and
temperature
response.
Measurements
should
be
National
Institute
of
Standards
and
Technology
(
NIST)
traceable
or
equivalent
with
quantified
measurement
uncertainties.
Personnel
who
test
UV
intensity
sensors
should
be
qualified
to
undertake
optical
testing,
understand
the
test
protocols
for
the
sensors
as
provided
by
the
manufacturer,
and
be
aware
of
all
safety
requirements
associated
with
UV­
irradiation
devices.

C.
3.3
Third­
Party
Oversight
Validation
of
UV
reactors
and
their
components
should
be
conducted
at
facilities
and
by
personnel
that
are
acceptable
to
the
State.
At
a
minimum,
personnel
independent
of
the
manufacturer
of
the
UV
reactor
should
oversee
validation
testing.
A
registered
professional
engineer
with
knowledge
and
experience
in
testing
and
evaluating
UV
reactors
should
witness
the
validation
testing
to
verify
that
the
documented
validation
protocol
was
followed
and
the
reported
data
and
results
are
accurate.
The
engineer
should
be
responsible
for
supervising
the
preparation
of
the
engineering
report
on
validation
testing
and
should
review
and
approve
that
report
prior
to
its
release.
The
engineer
should
not
have
a
personal
stake
in
the
outcome
of
the
validation
testing
or
any
conflict
of
interest
with
respect
to
the
ultimate
use
of
the
UV
reactor
being
tested.
Where
necessary,
the
engineer
should
use
other
third
parties
to
provide
expert
opinion
on
various
aspects
of
UV
validation
testing.

C.
4
Testing
This
section
describes
the
recommended
steps
for
validating
the
UV
reactor
provided
by
the
UV
manufacturer.
At
the
discretion
of
the
State,
variations
or
alternatives
to
the
procedures
or
steps
may
be
accepted.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
15
June
2003
C.
4.1
Develop
Approved
Test
Plan
The
first
step
in
validating
a
UV
reactor
should
be
the
development
and
review
of
a
test
plan.
The
test
plan
should
be
developed
with
input
and
approval
from
the
utility,
manufacturer,
third
party
oversight,
and
the
State.
The
test
plan
should
resolve
the
questions
identified
in
section
C.
1.4.

C.
4.2
UV
Reactor
Inspection
Prior
to
installing
the
UV
reactor
in
the
biodosimetry
test
stand,
the
UV
reactor
should
be
inspected
to
confirm
that
it
matches
the
descriptions
and
dimensions
provided
in
the
manufacturer's
documentation
as
described
in
section
C.
2.2.

C.
4.3
UV
Reactor
Installation
The
UV
reactor
and
its
inlet
and
outlet
piping
should
be
installed
at
the
test
facility
in
accordance
with
the
manufacturer's
installation
and
assembly
instructions.
If
reactors
are
installed
in
series,
the
piping
between
the
reactors
should
conform
to
specifications
provided
by
the
UV
reactor
manufacturer.
The
piping
should
be
inspected
to
ensure
compliance
with
the
manufacturer's
documentation.

C.
4.4
Headloss
and
Integrity
Evaluation
The
physical
integrity
of
the
UV
reactor
and
the
test
train
should
be
checked
before
conducting
further
testing.
Personnel
who
operate
the
UV
reactor
during
all
tests
should
be
familiar
with
its
operation
and
maintenance
manual
and
with
any
safety
requirements.

Procedure
1.
Pass
water
through
the
reactor
at
the
minimum
and
maximum
flowrates.

2.
Measure
and
record
the
headloss
across
the
reactor
at
each
flowrate.

3.
On
completion
of
the
test,
visually
inspect
the
sleeves,
UV
intensity
sensors,
and/
or
monitoring
windows
for
mechanical
integrity.

4.
If
the
headloss
across
the
reactor
exceeds
specifications
provided
by
the
manufacturer,
or
if
component
integrity
has
been
compromised,
investigate
the
cause
and
resolve
the
issue
before
further
testing.

C.
4.5
Evaluation
of
the
Mixing
of
Additives
The
mixing
of
the
UV­
absorbing
chemical
and
the
challenge
microorganism
prior
to
entering
the
UV
reactor
should
be
confirmed.
Mixing
can
be
confirmed
by
comparing
the
UV
absorbance
of
the
water
at
254
nm
(
A254)
of
samples
collected
at
the
influent
and
effluent
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
16
June
2003
sampling
ports
using
the
following
procedure.
This
test
should
not
be
necessary
if
a
static
mixer
is
used
between
the
injection
port
and
the
reactor
entrance
and
the
flowrate
through
the
static
mixer
meets
manufacturer
specifications.

Procedure
1.
Prepare
a
stock
solution
of
the
UV­
absorbing
compound.

2.
Pass
water
through
the
reactor
at
the
minimum
flowrate.

3.
Inject
sufficient
UV­
absorbing
compound
into
the
flow
of
water
passing
through
the
reactor
to
give
a
UVT
less
than
the
minimum
that
will
be
used
during
challenge
testing.

4.
Collect
water
samples
from
the
influent
and
effluent
sampling
ports
at
1­
minute
intervals
and
measure
the
UVT.
The
sample
volume
should
be
less
than
5
mL
and
collected
over
a
time
not
exceeding
2
seconds.

5.
Calculate
the
A254
from
the
measured
UVT.
Mixing
of
the
injected
compounds
should
be
sufficient
if
the
average
A254
of
the
influent
samples
and
the
average
A254
of
the
effluent
samples
agree
within
2
percent
and
the
standard
deviation
of
each
is
less
than
5
percent.
If
these
conditions
are
not
met,
the
mixing
between
the
injection
port
and
the
influent
sampling
port
should
be
increased
and
retested.

C.
4.6
Evaluation
of
the
Mixing
of
Surviving
Microorganisms
Mixing
of
the
surviving
challenge
microorganisms
leaving
the
UV
reactor
should
be
confirmed.
Mixing
can
be
confirmed
by
comparing
the
challenge
microorganism
concentration
of
samples
collected
at
the
effluent
sampling
port
and
a
sampling
port
downstream
of
the
effluent
sampling
port
using
the
following
procedure.
This
test
should
not
be
necessary
if
a
static
mixer
is
located
between
the
reactor
exit
and
the
effluent
sampling
port
and
the
flowrate
through
the
static
mixer
meets
manufacturer
specifications.

Procedure
1.
Prepare
a
stock
solution
of
the
challenge
microorganism
and
a
stock
solution
of
the
UV­
absorbing
compound.

2.
Pass
water
through
the
reactor
at
the
minimum
flowrate
that
will
be
used
during
challenge
testing.

3.
Operate
the
UV
reactor
with
the
lamps
power
set
at
100
percent.

4.
Inject
the
challenge
microorganism
into
the
water
flowing
through
the
reactor.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
17
June
2003
5.
Collect
at
least
three
UV­
disinfected
samples
spaced
1
minute
apart
from
the
effluent
sampling
point
and
from
a
location
at
least
5
pipe
diameters
downstream
of
the
effluent
sampling
point.

6.
Measure
the
concentration
of
the
challenge
microorganism
in
each
sample
in
triplicate.

7.
If
the
concentration
in
the
effluent
samples
is
below
the
detection
limit,
repeat
steps
2
to
6
with
the
UV
absorber
injected
into
the
flow
to
reduce
the
dose
delivery
by
the
reactor.

8.
Repeat
steps
3
to
7,
passing
the
water
through
the
reactor
at
the
minimum
flowrate
that
will
be
used
during
the
challenge
test.

9.
The
mixing
should
be
sufficient
if
there
is
no
statistical
difference
at
a
95
percent
confidence
level
between
the
geometric
means
of
the
samples
collected
from
the
two
effluent
sample
points.
If
statistical
differences
are
observed,
the
mixing
between
the
reactor
and
the
effluent
sampling
port
should
be
increased
and
the
test
repeated.

C.
4.7
UV
Intensity
Sensor
Evaluation
The
measurement
uncertainty
of
the
UV
intensity
sensors
used
on
the
UV
reactors
should
be
confirmed.
This
may
be
achieved
either
by
comparing
the
UV
intensity
sensor
measurements
made
on
the
reactor
to
a
reference
measurement,
or
by
measuring
the
properties
of
the
sensors
using
a
UV
intensity
sensor
test
stand.
The
following
sections
discuss
each
of
these
approaches.

C.
4.7.1
Assessing
Uncertainty
Using
Reference
Sensors
If
the
measurement
uncertainty
of
the
reference
intensity
sensor
is
known,
the
following
procedure
can
be
used
to
check
the
uncertainty
of
the
UV
intensity
sensors
used
during
validation.

Procedure
1.
Pass
water
through
the
reactor
without
the
addition
of
UV­
absorbing
chemicals.

2.
Using
at
least
three
recently
calibrated
reference
sensors,
install
each
sensor
on
the
UV
reactor
at
each
port
and
record
the
measured
UV
intensity.
Repeat
using
each
duty
sensor.
If
the
sensors
can
be
rotated,
then
measure
the
minimum
and
maximum
sensor
readings
with
rotation.

3.
Record
the
water
temperature
as
an
indicator
of
the
operating
temperature
of
the
sensors.

4.
Repeat
the
test
with
the
UVT
decreased
to
the
minimum
value
expected
during
testing.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
18
June
2003
5.
For
a
given
lamp
output
and
UVT,
the
difference
between
the
reference
sensor
measurements
should
follow
Equation
C.
2:

(
)
2
1
2
2
Ref
2
1
Ref
2
Ref
1
Ref
100
*
1
I
I
 
+
 
 
 
Equation
C.
2
where
I
=
Intensity
measured
by
a
reference
sensor
designated
by
the
subscript
 
=
Measurement
uncertainty
of
reference
sensor
designated
by
the
subscript
(%)

6.
For
a
given
lamp
output
and
UVT,
the
difference
between
the
reference
and
duty
sensor
measurements
should
follow
Equation
C.
3:

(
)
2
1
2
Duty
2
Ref
Ref
uty
D
100
*
1
I
I
 
+
 
 
 
Equation
C.
3
where
IRef
=
Intensity
measured
by
the
reference
sensor
IDuty
=
Intensity
measured
by
the
duty
sensor
 Ref
=
Measurement
uncertainty
of
the
reference
sensor
(%)
 Duty
=
Measurement
uncertainty
of
the
duty
sensor
(%)

7.
UV
intensity
sensors
that
do
not
meet
these
criteria
should
be
replaced.
Alternatively,
the
UV
manufacturer
can
re­
evaluate
their
stated
measurement
uncertainty
and
use
a
higher
value.

C.
4.7.2
Assessing
Uncertainty
Using
a
Sensor
Test
Stand
The
measurement
uncertainty
of
the
UV
intensity
sensors
can
be
assessed
by
a
laboratory
capable
of
confirming
sensor
calibration
and
properties
with
a
known
measurement
uncertainty.
The
laboratory
should
measure
linearity,
spectral
and
angular
response,
and
temperature
response.
Results
should
be
used
to
calculate
the
measurement
uncertainty.
Sensors
that
do
not
meet
manufacturer
specifications
should
be
replaced.
Alternatively,
the
UV
manufacturer
can
re­
evaluate
their
stated
measurement
uncertainty
and
use
a
higher
value.

C.
4.8
Evaluation
of
Lamp
and
Sleeve
Aging
on
Dose
Monitoring
With
operation
over
time,
UV
lamps
and
sleeves
can
experience
non­
uniform
aging
along
their
length
and
around
their
circumference.
Lamps
can
also
experience
spectral
shifts
in
output
and
sleeves
can
experience
spectral
shifts
in
UV
transmittance.
If
these
effects
have
a
significant
impact
on
how
the
dose
delivery
indicated
by
the
monitoring
system
compares
to
the
delivered
dose,
validation
should
be
conducted
using
both
new
and
aged
lamps
and
sleeves.
The
following
procedure
compares
dose
delivery
monitoring
with
new
and
aged
lamps
to
identify
if
validation
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
19
June
2003
should
be
conducted
with
both
new
and
aged
lamps
and
sleeves.
Alternatively,
data
on
the
UV
output
of
new
and
aged
lamps
and
the
UV
transmittance
of
new
and
aged
sleeves
can
be
compared
and
used
to
demonstrate
if
validation
should
be
conducted
with
new
and
aged
lamps
and
sleeves.
In
both
approaches,
an
aged
lamp
or
sleeve
is
one
that
has
reached
the
end
of
its
useful
service
life.

Procedure
1.
Prepare
a
stock
solution
of
the
challenge
microorganism.

2.
Fit
the
UV
reactor
with
aged
lamps
and
sleeves.

3.
Pass
water
through
the
reactor
at
a
constant
UVT
and
at
the
maximum
flowrate
that
will
be
used
during
challenge
testing.

4.
Operate
the
UV
reactor
at
peak
lamp
power.

5.
Inject
the
challenge
microorganism
into
the
flow
passing
through
the
reactor.

6.
Collect
at
least
three
microbiological
samples
spaced
one
minute
apart
from
the
influent
and
effluent
sampling
ports.

7.
Record
the
UV
intensity
sensor
measurements.

8.
Fit
the
UV
reactor
with
new
lamps
that
have
undergone
100­
hour
burn­
in
and
new
sleeves.

9.
Lower
the
lamp
power
to
give
a
UV
intensity
sensor
reading
equivalent
to
the
reading
obtained
in
step
7.

10.
Repeat
steps
5
and
6.

11.
If
the
mean
log
inactivation
achieved
with
new
lamps
differs
from
the
mean
log
inactivation
achieved
with
aged
lamps,
lamp
aging
impacts
the
relationship
between
dose
delivery
and
UV
intensity
sensor
reading,
and
validation
with
aged
lamps
and
sleeves
should
be
considered.

C.
4.9
Dose
Delivery
Validation
Dose
delivery
validation
via
biodosimetry
provides
an
assessment
of
dose
delivery
and
monitoring
by
the
UV
reactor
under
specific
conditions
of
flowrate,
UVT,
and
lamp
output.

C.
4.9.1
Preparation
of
Challenge
Microorganism
Stock
Solution
The
challenge
microorganism
is
used
to
measure
the
dose
delivery
of
the
UV
reactor
during
validation.
Because
MS2
and
B.
subtilis
spores
are
typically
used,
methods
for
their
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
20
June
2003
preparation
and
assay
are
provided
in
this
manual
in
Appendix
D.
Other
peer­
reviewed
methods
may
be
used.
A
rationale
for
selecting
challenge
microorganisms
other
than
MS2
and
B.
subtilis
spores
is
provided
in
section
F.
1.

The
challenge
microorganism
stock
solution
should
be
prepared
in
accordance
with
peerreviewed
methods.
The
source
of
the
challenge
microorganism,
the
source
of
the
host
(
if
used),
a
description
of
all
media
used,
the
steps
involved
in
propagating
the
challenge
microorganism,
and
the
steps
involved
in
purifying
the
challenge
microorganism
to
create
a
mono­
disperse
stock
solution
should
be
documented.
The
volume
of
stock
solution
needed
should
be
estimated
prior
to
testing
based
on
the
test
plan
and
the
expected
stock
concentration.

C.
4.9.2
Reactor
Preparation
If
the
number
of
sensors
is
less
than
the
number
of
lamps,
the
UV
intensity
sensors
should
be
directly
monitoring
the
lamps
with
the
highest
output
and
those
lamps
should
be
the
closest
lamps
to
the
sensor.
The
lamps
with
the
highest
output
can
be
identified
by
taking
measurements
using
either
a
dedicated
test
stand
or
the
UV
reactor.
One
approach
for
using
the
UV
reactor
is
described
below.
This
preparation
should
not
be
necessary
if
the
UV
reactor
has
one
UV
intensity
sensor
per
lamp.

Procedure
1.
Install
a
lamp
within
a
lamp
sleeve
located
near
one
of
the
reactor's
UV
intensity
sensors.

2.
Pass
water
through
the
reactor
at
a
constant
flowrate
and
UVT.

3.
With
only
the
lamp
under
evaluation
on,
record
the
measured
UV
intensity.

4.
Repeat
the
test
for
each
lamp
and
rank
the
results.

5.
Install
the
lamps
in
the
UV
reactor
so
that
the
lamps
with
the
highest
output
are
closest
to
the
UV
intensity
sensors
monitoring
those
lamps.

C.
4.9.3
Flowrates
At
a
minimum,
the
reactor
should
be
validated
at
the
minimum
and
maximum
flowrates
as
defined
by
the
UV
manufacturer.
Other
flowrates
within
that
range
can
be
tested.
For
interpolation
of
validation
results
as
a
function
of
flowrate,
a
recommended
approach
for
selecting
intermediate
flowrates
is
to
approximate
a
geometric
series
using
Equation
C.
4:

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
21
June
2003
n
Max
n
Q
Q
 
=
1
 
Equation
C.
4
where
Qn
=
nth
flowrate
to
be
tested
Qmax
=
Maximum
flowrate
to
be
tested
 
=
Rate
term
with
a
recommended
value
between
1.5
and
2
n
=
Number
of
flowrates
to
be
tested
The
value
of
 
should
not
exceed
2
and
should
be
sufficient
to
obtain
at
least
three
measured
data
points
for
interpolation.

Example.
Interpolation
will
be
used
to
predict
RED
as
a
function
of
flowrate
for
a
UV
reactor
rated
over
a
flow
range
of
2
to
20
mgd.
If
a
rate
term
of
2
was
used
with
Equation
C.
4,
the
UV
reactor
would
be
validated
at
flowrates
of
20,
10,
5,
2.5,
and
2
mgd.
If
a
rate
term
of
1.5
was
used
with
Equation
C.
4,
the
UV
reactor
would
be
validated
at
flowrates
of
20,
13,
8.9,
5.9,
4.0,
2.6,
and
2
mgd.

C.
4.9.4
Lamp
Power
and
UV
Transmittance
At
a
given
flowrate,
the
UV
reactor
should
be
validated
under
conditions
of
UVT
and
lamp
output
that
demonstrate
the
UV
reactor
is
sized
to
deliver
a
given
dose
and
the
UV
reactor's
dose
monitoring
system
provides
a
valid
measure
of
that
dose.
Typically,
the
UVT
of
the
source
water
used
during
validation
is
high
and
UV
absorbing
chemicals
are
added
to
that
water
to
achieve
the
UVT
used
during
validation
testing.
Different
levels
of
lamp
output
can
be
obtained
using
one
or
more
of
the
following
approaches:

 
Using
new
and
aged
lamps
 
Using
different
lamp
types
with
the
same
spectral
output
(
e.
g.,
using
LP
and
LPHO
lamps)

 
Changing
the
ballasts'
power
settings
 
Using
specially
modified
ballasts
capable
of
operating
at
different
power
levels
 
Changing
the
supply
voltage
to
the
lamp
ballasts
If
lamp
aging
affects
the
relationship
between
the
inactivation
achieved
by
the
UV
reactor
and
the
measurements
made
by
the
on­
line
UV
intensity
sensor,
aged
lamps
should
be
used
when
validation
testing
involves
reduced
lamp
output.

The
conditions
of
lamp
power
and
UVT
used
during
validation
should
depend
on
the
monitoring
approach
of
the
UV
reactor.
The
next
three
sections
describe
recommended
approaches
for
defining
these
test
conditions
for
UV
reactors
that
use
the
following
monitoring
approaches:

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
22
June
2003
 
UV
intensity
setpoint
approach
 
UV
intensity
and
UVT
setpoint
approach
 
Calculated
dose
approach
Section
F.
2
provides
background
on
the
development
of
these
approaches.

UV
Intensity
Setpoint
Approach
With
the
UV
intensity
setpoint
approach,
measurements
of
UV
intensity
and
flowrate
are
used
directly
to
indicate
dose
delivery.
Dose
delivery
at
or
above
a
given
level
is
indicated
when
the
measured
intensity
reads
above
an
alarm
setpoint
value
defined
as
a
function
of
flowrate.

With
the
UV
intensity
setpoint
approach,
the
UV
intensity
sensor
is
positioned
within
the
UV
reactor
to
respond
to
the
impacts
of
both
lamp
output
and
UVT.
As
such,
dose
delivery
can
be
monitored
without
the
need
to
measure
the
UVT.

Strategies
for
implementing
this
approach
include:

1.
Using
a
single
UV
intensity
setpoint
value
from
minimum
to
maximum
flow
to
verify
dose
delivery
at
some
minimum
level.

Example.
A
UV
intensity
setpoint
of
10
mW/
cm2
is
used
to
verify
a
minimum
MS2
RED
of
40
mJ/
cm2
from
1
to
5
mgd.

2.
Several
UV
intensity
setpoint
values
are
used,
each
one
applying
over
a
specific
flow
range.

Example.
UV
intensity
setpoints
of
10
and
20
mW/
cm2
are
used
to
verify
a
minimum
MS2
RED
of
40
mJ/
cm2
from
1
to
2.5
mgd
and
from
2.5
to
5
mgd,
respectively.

3.
UV
intensity
setpoint
values
are
interpolated
as
a
function
of
flowrate.

Example.
UV
intensity
setpoints
defined
by
the
following
equation
are
used
to
indicate
an
MS2
dose
of
39
mJ/
cm2
from
1
to
2.4
mgd:

Intensity
setpoint
(
mW/
cm2)
=
15.6
x
flow
rate
(
mgd)
+
3.9
4.
UV
intensity
setpoints
are
defined
as
a
function
of
flowrate
for
multiple
levels
of
dose
delivery.

Example.
A
UV
intensity
setpoint
of
10
mJ/
cm2
is
used
to
verify
a
minimum
RED
of
40
mJ/
cm2
from
1
to
5
mgd.
A
UV
intensity
setpoint
value
of
7
mW/
cm2
is
used
to
verify
a
minimum
RED
of
25
mW/
cm2
from
1
to
5
mgd.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
23
June
2003
With
UV
reactors
using
this
monitoring
approach,
validation
testing
provides
data
on
the
relationship
between
dose
delivery
and
measured
intensity
at
a
given
flowrate.
Dose
delivery
at
a
given
flowrate
and
UV
intensity
is
measured
under
two
conditions
of
lamp
power
and
UVT,
described
as
follows:

1.
Lamps
at
peak
power
and
the
UVT
decreased
to
give
a
UV
intensity
sensor
reading
at
a
setpoint
value.

2.
High
UVT
and
the
lamp
power
lowered
to
give
a
UV
intensity
sensor
reading
at
a
setpoint
value.

The
RED
assigned
to
the
reactor
is
the
lower
value
observed
between
the
two
test
conditions.

If
the
lamp
power
cannot
be
sufficiently
lowered
to
obtain
a
UV
intensity
sensor
reading
at
the
setpoint
value,
an
alternative
to
the
second
test
condition
is
to
test
with
the
lowest
possible
lamp
power
setting
and
the
UVT
decreased
until
an
intensity
reading
at
the
setpoint
is
obtained.
This
alternative
second
test
condition
is
acceptable
if
the
following
conditions
are
met:

 
The
adjusted
lamp
power
results
in
a
lamp
output
equal
to
or
lower
than
the
lamp
output
used
for
sizing
the
UV
reactor
for
a
WTP.
The
lamp
output
used
for
sizing
the
UV
reactor
is
the
product
of
the
lamp­
aging
factor
and
the
fouling
factor.

 
The
RED
measured
with
the
second
condition
is
equal
to
or
greater
than
the
RED
measured
with
the
first
test
condition
or
the
UVT
with
the
second
test
condition
is
less
than
the
UVT
expected
at
the
WTP.

There
are
several
approaches
for
defining
the
UV
intensity
setpoint
values
evaluated
during
validation
testing:

1.
If
a
UV
reactor
is
being
validated
for
an
application
with
specific
design
conditions
of
flowrate,
lamp
output,
and
UVT,
the
intensity
setpoint
at
design
flow
is
equal
to
or
greater
than
the
intensity
reading
obtained
with
the
reactor
operating
under
these
design
conditions.

2.
A
UV
reactor
manufacturer
can
usually
provide
model
estimates
of
dose
delivery
as
a
function
of
flowrate
and
UV
intensity.
The
model
estimates
would
be
used
to
define
the
intensity
setpoint
values
associated
with
a
target
dose
delivery.
Since
model
estimates
may
not
be
accurate,
trial
and
error
testing
may
be
used
to
establish
the
optimal
intensity
setpoint
necessary
for
a
target
level
of
dose
delivery.
Alternatively,
testing
can
be
used
to
define
the
relation
between
dose
delivery
and
measured
intensity,
and
interpolation
can
be
used
to
define
the
optimal
setpoint
associated
with
a
target
dose
delivery.

During
the
validation
of
a
UV
reactor
using
the
intensity
setpoint
monitoring
approach,
the
UVT
used
will
likely
be
less
than
the
design
UVT
and
the
lamp
output
will
be
less
than
the
design
lamp
output.
While
it
may
appear
that
these
test
conditions
are
more
stringent
than
the
design
conditions,
it
should
be
recognized
that
design
conditions
do
not
represent
the
worst­
case
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
24
June
2003
conditions
that
can
occur
at
a
WTP.
For
example,
lamps
can
age
below
their
expected
end­
oflife
output,
lamp
sleeves
can
foul
internally,
wiper
mechanisms
can
fail,
and
dose­
pacing
strategies
can
reduce
lamp
output.
These
factors
in
combination
can
result
in
a
UV
output
well
below
the
design
output.
If
the
design
UVT
is
selected
at
a
95
percent
confidence
level,
then
a
UVT
below
the
design
value
is
expected
5
percent
of
the
time.
Because
intensity
setpoints
should
provide
a
valid
measure
of
dose
delivery,
regardless
of
the
combination
of
lamp
output
and
UVT
values,
a
UV
reactor
using
intensity
setpoint
monitoring
should
be
validated
over
the
full
range
of
conditions
giving
rise
to
the
setpoint,
even
if
they
exceed
design
conditions.

Example.
A
UV
reactor
that
uses
the
intensity
setpoint
approach
for
monitoring
is
sized
using
a
design
UVT
of
90
percent,
a
lamp
aging/
fouling
factor
of
70
percent,
and
a
flow
of
5
mgd.
With
lamp
power
and
UVT
adjusted
to
70
and
90
percent,
respectively,
the
UV
intensity
sensor
reads
14
mW/
cm2.
The
UV
reactor
is
tested
at
a
flow
of
5
mgd
using
the
following
conditions
of
lamp
output
and
UVT
that
give
rise
to
a
UV
intensity
of
14
mW/
cm2:

 
100
percent
lamp
power,
87
percent
UVT
 
27
percent
lamp
power,
98
percent
UVT
By
testing
the
reactor
using
these
conditions,
dose
delivery
associated
with
a
setpoint
of
14
mW/
cm2
is
validated.

UV
Intensity
and
UVT
Setpoint
Approach
With
the
UV
intensity/
UVT
setpoint
approach,
measurements
of
UV
intensity,
UVT,
and
flowrate
are
used
to
indicate
dose
delivery.
Dose
delivery
at
or
above
a
given
level
is
indicated
when
both
the
measured
UV
intensity
and
UVT
read
above
their
respective
alarm
setpoint
values.
Strategies
for
implementing
this
approach
include:

1.
Using
a
single
UV
intensity
setpoint
value
and
UVT
setpoint
value
from
minimum
to
maximum
flowrate
to
indicate
dose
delivery
at
some
level.

Example.
A
minimum
MS2
RED
of
40
mJ/
cm2
from
1
to
5
mgd
is
verified
when
the
measured
UV
intensity
is
equal
to
or
greater
than
10
mW/
cm2
and
the
measured
UVT
is
equal
to
or
greater
than
85
percent.

2.
Several
sets
of
UV
intensity
and
UVT
setpoint
values
are
used,
each
set
applying
over
a
specific
flow
range.

Example.
For
an
MS2
RED
of
40
mJ/
cm2,
a
UV
intensity
setpoint
value
of
10
mW/
cm2
and
a
UVT
setpoint
value
of
80
percent
are
used
from
1
to
2.5
mgd.
A
UV
intensity
setpoint
value
of
20
mW/
cm2
and
a
UVT
setpoint
value
of
85
percent
are
used
from
2.5
to
5
mgd.

3.
Sets
of
UV
intensity
and
UVT
setpoint
values
are
interpolated
as
a
function
of
flowrate.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
25
June
2003
4.
Sets
of
UV
intensity
and
UVT
setpoint
values
are
defined
as
a
function
of
flowrate
for
multiple
levels
of
dose
delivery.

With
UV
reactors
using
this
monitoring
approach,
validation
testing
provides
data
on
the
dose
delivery
with
the
reactor
operating
at
the
setpoint
values
and
proof
that
the
sensor
is
appropriately
positioned
for
this
monitoring
approach.
As
such,
each
set
of
UV
intensity
and
UVT
setpoints
should
be
tested
using
two
conditions
as
follows:

1.
UVT
decreased
to
give
a
reading
at
the
UVT
setpoint
followed
by
a
decrease
in
lamp
power
to
give
a
UV
intensity
sensor
reading
at
the
UV
intensity
setpoint.

2.
Lamp
power
at
100
percent
and
UVT
decreased
to
give
a
UV
intensity
sensor
reading
at
the
intensity
setpoint.

The
first
condition
provides
data
on
dose
delivery
with
the
reactor
operating
with
UV
intensity
and
UVT
at
the
setpoint
values.
The
second
condition
provides
data
on
the
positioning
of
the
UV
intensity
sensor.
If
the
RED
measured
with
the
second
test
condition
is
greater
than
the
RED
measured
with
the
first,
the
UV
intensity
sensor
is
not
appropriately
positioned
for
this
monitoring
strategy
and
this
monitoring
strategy
cannot
be
used
(
see
section
F.
2
for
a
rational
for
this
criteria).

There
are
several
approaches
for
defining
the
UV
intensity
and
UVT
setpoints
used
during
validation
testing.

1.
At
design
flow,
the
UVT
setpoint
is
the
design
UVT.
The
intensity
setpoint
is
the
UV
intensity
measured
with
the
lamp
output
and
UVT
adjusted
to
their
design
values.

2.
At
other
flowrates,
model
estimates
of
dose
as
a
function
of
UVT
and
lamp
output
can
be
used
to
identify
the
setpoint
values
that
will
be
assessed
during
validation
testing.
Trial
and
error
testing
or
interpolation
of
test
results
can
be
used
to
refine
and
optimize
those
values
for
a
given
target
dose
delivery.

Example.
A
UV
reactor
that
uses
the
UV
intensity
and
UVT
setpoint
approach
for
monitoring
is
sized
for
a
WTP
using
a
design
UVT
of
90
percent,
a
design
lamp
fouling/
aging
factor
of
70
percent,
and
a
design
flowrate
of
5
mgd.
Operating
under
those
conditions,
the
intensity
sensor
measures
14
mW/
cm2.
The
UV
reactor
is
validated
under
two
test
conditions
at
a
flowrate
of
5
mgd:

 
70
percent
lamp
power
and
90
percent
UVT
resulting
in
a
UV
intensity
reading
of
14
mW/
cm2
 
100
percent
lamp
power
and
75
percent
UVT
resulting
in
a
UV
intensity
reading
of
14
mW/
cm2
The
first
condition
provides
data
on
the
dose
delivery
of
the
reactor
operating
at
the
setpoint.
The
second
condition
provides
data
to
assess
the
positioning
of
the
UV
intensity
sensor.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
26
June
2003
Calculated
Dose
Approach
With
the
calculated
dose
approach,
dose
delivery
is
calculated
from
measurements
of
UV
intensity,
UVT,
and
flowrate
using
an
algorithm
developed
by
the
UV
reactor
manufacturer.
For
UV
reactors
that
use
this
approach,
the
UV
reactor
should
be
tested
over
a
range
of
combinations
of
flowrate,
UVT,
and
lamp
power
that
result
in
a
given
calculated
dose.
At
a
given
flowrate,
that
range
should
include
the
following
combinations:

 
Maximum
power
and
decreased
UVT
 
Maximum
UVT
and
decreased
lamp
power
 
One
or
two
intermediate
combinations
of
UVT
and
lamp
power
If
the
algorithm
for
calculating
dose
accounts
for
lamps
operating
at
different
power
levels
or
specific
lamps
operating
either
on
or
off,
test
conditions
should
include
combinations
of
these
conditions.

Example.
A
UV
reactor
that
uses
a
calculated
dose
for
compliance
will
be
used
at
a
WTP
with
a
design
UVT
of
90
percent,
a
design
lamp
fouling/
aging
factor
of
70
percent,
and
a
design
flowrate
of
5
mgd.
The
target
RED
is
40
mJ/
cm2.
At
5
mgd,
test
conditions
that
result
in
a
calculated
dose
of
40
mJ/
cm2
by
the
monitoring
system
are
as
follows:

 
100
percent
lamp
power,
80
percent
UVT
 
58
percent
lamp
power,
90
percent
UVT
 
34
percent
lamp
power,
98
percent
UVT
C.
4.9.5
Measuring
Challenge
Microorganism
Inactivation
by
the
UV
Reactor
The
reactor
should
be
operated
at
each
of
the
test
conditions
of
flowrate,
UVT,
and
lamp
power
in
accordance
with
sections
C.
4.9.3
and
C.
4.9.4.
Prior
to
sampling,
steady­
state
conditions
should
be
confirmed
by
monitoring
the
UV
intensity
sensor
measurements
and
the
UVT.
The
challenge
microorganism
should
be
injected
into
the
flow
upstream
of
the
reactor
and
well­
mixed
prior
to
its
entering
the
UV
reactor.
At
least
three
influent
and
effluent
samples
should
be
collected
for
each
test
condition.
The
time
interval
between
sample
collections
should
be
greater
than
or
equal
to
the
residence
time
between
the
inlet
and
outlet
sampling
ports.
Water
samples
should
be
collected
by
personnel
who
are
familiar
with
good
sampling
practices
as
specified
in
Standard
Methods
(
APHA
et
al.
1995)
and
the
guidance
for
collecting
UV­
irradiated
samples.
Sample
volumes
should
be
sufficient
for
assessing
the
challenge
microorganism
concentrations
in
the
influent
and
effluent.

Before
and
after
the
samples
are
collected,
the
flowrate
through
the
reactor,
all
UV
intensity
sensor
measurements,
on­
line
UVT
measurements,
and
any
calculated
dose
values
should
be
measured
and
recorded.
With
the
validation
of
LP
or
LPHO
UV
reactors,
the
UVT
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
27
June
2003
should
be
measured
and
recorded
with
each
influent
sample.
With
MP
reactors,
the
UVT
from
200
to
400
nm
should
be
measured
and
recorded.
The
electrical
power
delivered
to
the
lamps
by
each
ballast
should
also
be
measured
and
recorded.
The
challenge
test
should
be
repeated
if
the
flowrate,
UV
intensity,
lamp
power,
or
UVT
changes
by
more
than
the
error
of
the
measurement
over
the
course
of
sampling.

The
challenge
microorganism
concentration
in
the
samples
should
be
measured
within
24
hours
of
collection
using
a
peer­
reviewed
method.
Suggested
methods
for
measuring
MS2
and
B.
subtilis
spore
concentrations
in
water
samples
are
provided
in
Appendix
D.
Reported
challenge
microorganism
concentrations
should
include
dilutions,
volumes
used,
and
the
number
of
plaques
or
colonies
counted
on
each
plate.

C.
4.9.6
Quality
Assurance
and
Quality
Control
Samples
During
testing
of
the
UV
reactor,
samples
should
be
collected
to
ensure
quality
assurance
and
control
(
QA/
QC)
including:

 
Trip
controls
­
sample
bottles
of
challenge
microorganism
stock
solution
of
known
concentration
that
travel
with
the
stock
solution
from
the
microbiological
laboratory
to
the
location
of
reactor
testing
and
back
to
the
laboratory.
The
concentration
of
the
challenge
microorganism
in
the
trip
controls
measured
at
the
beginning
and
end
should
be
the
same
at
a
90
percent
confidence
level.

 
Reactor
blanks
­
influent
water
samples
taken
without
any
addition
of
challenge
microorganism
to
the
flow
passing
through
the
reactor.
The
concentration
of
the
challenge
microorganism
measured
with
the
blank
should
not
interfere
with
the
determination
of
RED
delivered
by
the
reactor.

 
Reactor
controls
­
influent
and
effluent
water
samples
taken
with
the
UV
lamps
(
in
the
reactor)
turned
off.
The
challenge
microorganism
concentrations
in
both
samples
should
be
the
same
at
a
90
percent
confidence
level.

 
Method
blanks
­
sample
bottle
of
sterilized
reagent
grade
water
that
undergoes
the
challenge
microorganism
assay
procedure.
The
concentration
of
challenge
microorganism
with
the
method
blank
should
be
non­
detectable.

C.
4.9.7
Challenge
Microorganism
Dose­
Response
The
UV
dose­
response
of
the
challenge
microorganism
within
samples
collected
from
the
reactor
influent
should
be
measured
with
the
collimated
beam
apparatus
as
described
in
Appendix
E.
At
least
two
dose­
response
curves
should
be
generated.
One
sample
should
have
UVT
unadjusted
by
UV­
absorbing
additives
and
one
sample
should
have
UVT
adjusted
to
give
the
minimum
UVT
used
in
section
C.
4.9.4.
A
one­
liter
influent
sample
should
be
sufficient
for
measuring
the
challenge
microorganism
UV
dose­
response.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
28
June
2003
The
collimated
beam
tests
should
be
conducted
within
24
hours
of
sample
collection.
Based
on
the
expected
dose­
response
of
the
challenge
microorganism,
UV
doses
should
be
applied
to
achieve
log
reductions
of
approximately
0.5,
1.0,
2.0,
3.0,
4.0,
and
5.0.
For
each
log
reduction,
at
least
three
aliquots
of
the
influent
sample
should
be
irradiated.
Three
aliquots
should
also
be
collected
as
zero
dose
samples.
Aliquots
should
be
packed
on
ice
and
stored
in
the
dark
until
they
are
assayed.
Aliquots
should
be
assayed
within
24
hours
of
irradiation.

The
log
inactivation
for
each
applied
dose
delivered
by
the
collimated
beam
should
be
calculated
using
Equation
C.
5:

 
 

 
 

 
 

 
 
=
N
N
on
Inactivati
Log
0
log
Equation
C.
5
where
N0
=
Average
concentration
of
the
challenge
microorganism
in
the
zero
dose
aliquots
N
=
Challenge
microorganism
concentration
in
an
aliquot
of
sample
Fitting
Dose­
Response
Data
The
dose­
response
of
the
challenge
microorganism
should
be
plotted
as
UV
dose
versus
log
inactivation.
An
equation
that
best
expresses
the
UV
dose
as
a
function
of
log
(
N0/
N)
should
be
obtained
using
regression
analysis.
A
linear
equation
should
best­
fit
first­
order
kinetics.
A
quadratic
equation
should
provide
a
better
fit
with
tailing,
and
other
equations
should
be
used
if
inactivation
kinetics
involves
shoulders
(
DVGW
1997,
ONORM
2001).
Equation
coefficients
obtained
from
the
regression
analysis
should
be
significant
at
a
95
percent
confidence
level.
The
differences
between
the
values
measured
and
predicted
by
the
equation
should
be
randomly
distributed
around
zero
and
not
show
a
dependence
on
dose.
Confidence
intervals
for
the
fit
should
be
determined
at
an
80
percent
confidence
level.
The
equation
should
be
used
for
interpolating
dose­
response
data
but
should
not
be
used
for
extrapolation
outside
of
the
measured
UV
dose
range.

Example.
The
dose­
response
of
MS2,
presented
in
the
following
table,
was
measured
using
a
collimated
beam
apparatus.

UV
Dose
(
mJ/
cm2)
Log
Inactivation
Log
Inactivation
0
0.016
­
0.119
10
0.805
1.06
30
1.87
2.16
60
3.40
3.62
100
4.71
4.83
Regression
analysis
was
used
to
fit
the
equations
to
the
MS2
dose­
response
data:

B
on
Inactivati
Log
A
Dose
+
×
=

and
(
)
2
on
Inactivati
Log
D
on
Inactivati
Log
C
Dose
×
+
×
=

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
29
June
2003
The
following
table
lists
the
coefficients
derived
from
the
regression
analysis,
the
pstatistics
for
those
coefficients,
and
the
R­
squared
value
for
the
fit.

Equation
Coefficient
Value
p­
statistic
A
20.5
3.13
×
10­
7
B
on
Inactivati
Log
A
Dose
+
×
=
R­
squared
=
0.967
B
­
6.01
0.15
C
8.90
1.22
×
10­
4
(
)
2
on
Inactivati
Log
D
on
Inactivati
Log
C
Dose
×
+
×
=
R­
squared
=
0.995
D
2.47
4.39
×
10­
5
In
evaluating
the
two
equations,
a
first
check
was
done
to
determine
if
the
equation
coefficients
were
significant
at
a
95
percent
confidence
level.
While
the
R­
squared
value
for
the
first
equation
was
high,
the
p­
statistic
for
coefficient
B
was
greater
than
0.05,
indicating
that
it
was
not
significant
at
a
95
percent
confidence
level.
Thus,
the
first
equation
was
not
a
good
fit
to
the
dose­
response
data.
On
the
other
hand,
the
p­
statistics
for
coefficients
C
and
D
with
the
second
equation
were
both
less
than
0.05,
indicating
that
they
were
significant
at
a
95
percent
confidence
level.
Thus,
Equation
2
was
a
valid
fit
to
the
dose­
response
data.

A
second
check
of
the
two
equations
was
to
determine
if
the
difference
between
the
measured
and
predicted
dose
was
randomly
distributed
as
a
function
of
the
log
inactivation.
Figure
C.
3
presents
the
dose­
response
data
and
the
fits
to
the
data
with
confidence
levels.
As
shown,
the
first
equation
under­
predicts
UV
dose
at
low
and
high
levels
of
inactivation
and
overpredicts
dose
at
mid
levels
of
dose.
On
the
other
hand,
the
second
equation
does
not
show
a
bias
in
the
prediction
of
dose
as
a
function
of
log
inactivation.
This
second
check
further
demonstrates
that
the
second
equation
was
a
valid
fit
to
the
dose­
response
data
while
the
first
equation
was
not
valid.

To
illustrate
the
importance
of
using
an
appropriate
equation
to
fit
the
dose­
response
data,
the
following
table
compares
the
dose
predicted
using
the
two
equations
for
2­
log
inactivation.

UV
Dose
for
2
log
inactivation
(
mJ/
cm2)
Equation
Mean
Lower
Bound
B
on
Inactivati
Log
A
Dose
+
×
=
35
30
(
)
2
on
Inactivati
Log
D
on
Inactivati
Log
C
Dose
×
+
×
=
28
26
As
shown,
the
first
equation
over­
predicts
the
mean
dose
needed
for
2­
log
inactivation
by
27
percent,
as
compared
to
the
second
equation.
Large
errors
can
occur
predicting
the
UV
dose
associated
with
a
given
log
inactivation
if
the
equation
used
to
fit
the
data
is
not
appropriate.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
30
June
2003
Figure
C.
3
UV
Dose
Plotted
as
a
Function
of
MS2
Log
Inactivation
and
Fitted
Using
Two
Equations
Combining
Dose­
Response
Data
During
validation,
the
UV
dose­
response
of
the
challenge
microorganism
is
used
to
relate
the
inactivation
measured
through
the
reactor
under
each
test
condition
to
an
RED
value.
Typically,
it
is
assumed
that
the
dose­
response
measured
with
a
subset
of
the
test
conditions
assessed
during
validation
can
be
used
to
calculate
the
RED
for
all
test
conditions.
This
assumption
is
valid
if
the
dose­
response
of
the
challenge
microorganism
does
not
vary
from
test
condition
to
test
condition.
To
prove
this
assumption,
the
regression
coefficients
generated
for
each
set
of
dose­
response
data
should
be
equal
at
a
95
percent
confidence
level
(
Draper
and
Smith
1981).
If
the
coefficients
are
the
same,
the
equation
fitting
the
combined
dataset
should
be
used
for
determining
the
RED.
If
the
coefficients
are
different,
the
cause
of
the
difference
should
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
31
June
2003
be
determined.
Difference
in
UV
dose­
response
could
occur
if
the
dose­
response
was
determined
with
different
batches
of
the
challenge
microorganism
or
if
water
quality
interferences
are
impacting
the
dose­
response
(
e.
g.,
MS2
coagulation).
The
following
example
presents
an
approach
that
can
be
used
to
determine
if
two
sets
of
UV
dose­
response
data
can
be
combined.

Example.
The
following
table
gives
the
dose­
response
data
for
MS2
measured
on
two
influent
samples
during
validation
testing.

Log
Inactivation
(
N0/
N)
UV
Dose
(
mJ/
cm2)
Influent
Sample
1
Influent
Sample
2
0
0.02
0.09
0.24
­
0.10
10
0.33
0.709
0.54
0.40
20
1.1
1.4
1.0
1.4
40
1.8
2.4
2.3
2.3
60
2.7
3.2
3.2
3.3
80
3.5
4.4
3.4
3.9
100
3.9
4.4
4.3
4.8
Each
dataset
can
be
described
using
the
following
equation:

2
0
0
N
N
Log
B
N
N
Log
A
Dose
 

  

 

 

  
 

  
 

  
 
×
+

  
 

  
 
×
=

To
determine
if
the
two
datasets
could
be
combined,
a
general
equation
is
defined
for
both
datasets
as:

2
0
0
2
0
0
N
N
Log
d
D
N
N
Log
d
C
N
N
Log
B
N
N
Log
A
Dose
 

  

 

 

  
 

  
 

  
 
×
×
+

  
 

  
 
×
×
+

 

  
 

 

  
 

  
 

  
 
×
+

  
 

  
 
×
=

The
term
d
is
set
to
zero
with
the
first
dataset
and
set
to
one
with
the
second
dataset.
Multiple
regression
analysis
using
the
full
dataset
is
used
to
determine
the
values
of
coefficients
A,
B,
C
and
D
with
the
following
results:

Coefficient
Value
p­
statistic
A
17.5
0.000
B
1.03
0.202
C
­
2.43
0.553
D
0.435
0.689
As
shown
by
the
p­
statistic,
the
term
A
was
statistically
significant
at
the
95
percent
confidence
level
(
p
 
0.05)
and
the
terms
B,
C,
and
D
were
not
(
p
 
0.05).
The
regression
analysis
was
repeated
in
a
step­
wise
fashion,
removing
the
term
with
the
highest
p­
value
from
the
equation.
With
the
second
regression,
the
terms
A
and
B
were
statistically
significant
and
the
term
C
was
not.
With
the
third
regression,
the
terms
A
and
B
were
both
statistically
significant.
Because
neither
terms
C
nor
D
were
significant,
it
can
be
concluded
that
the
regression
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
32
June
2003
coefficients
generated
by
the
fits
to
each
dose­
response
are
equal
at
a
95
percent
confidence
level.
Thus,
the
two
datasets
can
be
combined.

C.
4.9.8
Reactor
Log
Inactivation
and
RED
For
each
condition
of
flowrate,
UVT,
and
lamp
output
as
defined
in
sections
C.
4.9.3
and
C.
4.9.4,
the
arithmetic
mean
and
standard
deviation
of
the
log
of
the
influent
and
effluent
challenge
microorganism
concentrations
should
be
calculated.
For
each
test
condition,
the
log
inactivation
should
be
calculated
using
equation
C.
6:

(
)
(

E
I
N
log
N
log
on
Inactivati
Log
 
=
)
Equation
C.
6
where
log(
NI)
=
Mean
challenge
microorganism
log
concentration
of
the
influent
samples
log(
NE)
=
Mean
challenge
microorganism
log
concentration
of
the
effluent
samples
The
uncertainty
of
the
log
inactivation
should
be
calculated
using
Equation
C.
7:

(
)
(
)

%
on
Inactivati
Log
n
t
n
t
E
E
E
I
I
I
100
2
1
2
2
×
 
 

 
 

 
 

 
 
+

=
 
 
in
U
Equation
C.
7
where
Uin
=
Percent
uncertainty
of
the
log
inactivation
through
the
UV
reactor
tI
=
t­
statistic
of
the
influent
samples
at
an
80
percent
confidence
level
 I
=
Standard
deviation
of
the
challenge
microorganism
log
concentration
of
the
influent
samples
nI
=
Number
of
influent
samples
tE
=
t­
statistic
of
the
effluent
samples
at
an
80
percent
confidence
level
 E
=
Standard
deviation
of
the
challenge
microorganism
log
concentration
of
the
effluent
samples
nE
=
Number
of
effluent
samples
The
RED
should
be
calculated
from
the
log
inactivation
using
the
equation
describing
the
UV
dose­
response
curve
of
the
challenge
microorganism.
The
percent
measurement
uncertainty
of
the
RED
can
be
calculated
using
Equation
C.
8:

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
33
June
2003
(
)
2
1
2
D
2
DR
2
in
RED
U
U
U
U
+
+
=
Equation
C.
8
where
URED
=
Percent
uncertainty
of
the
measured
RED
UDR
=
Percent
uncertainty
of
the
regression
equation
fitting
the
challenge
microorganism's
UV
dose­
response
data
at
an
80
percent
confidence
level
(
see
section
C.
4.8.7,
Figure
C.
3)
UD
=
Percent
uncertainty
of
the
collimated
beam
dose
calculation
that
is
not
captured
in
the
variability
of
the
measured
dose­
response
data
(
see
Appendix
E).
This
typically
includes
the
uncertainty
of
the
radiometer
and
the
Petri
factor
Example.
A
UV
reactor
was
validated
using
MS2.
The
UV
dose­
response
measured
using
a
collimated
beam
apparatus
is
given
in
Figure
C.
3.
The
dose­
response
was
fitted
using
the
following
equation:

(
)
2
on
Inactivati
Log
47
.
2
on
Inactivati
Log
90
.
8
Dose
×
+
×
=

The
uncertainty
of
the
radiometer
used
with
the
collimated
beam
apparatus
was
8
percent.
The
Petri
factor
was
measured
with
an
uncertainty
of
2
percent.
Thus
the
uncertainty
of
the
collimated
beam
dose
calculation,
UD
is
calculated
as
follows:

(
)
%
2
.
8
2
8
U
2
1
2
2
D
=
+
=

The
following
table
presents
the
microbiology
results
obtained
with
the
influent
and
effluent
samples
collected
with
one
of
the
test
conditions
assessed
during
validation.

The
mean
and
standard
deviation
of
the
influent
and
effluent
log
concentrations
of
the
MS2
are
6.32
±
0.075
and
4.26
±
0.13,
respectively.
The
log
inactivation
through
the
reactor
is
calculated
as
follows:

06
2
26
4
32
6
.
.
.
on
Inactivati
Log
=
 
=

Table
C.
2
Estimated
Log
Inactivation
and
Corresponding
RED
Values
Using
Bioassay
Results
Plate
Counts
­
Dilution
=
104
Plate
Counts
­
Dilution
=
105
Concentration
Before
UV
1
2
3
1
2
3
PFU/
mL
log
Sample
1
148
180
TNTC
15
18
20
1.77
×
106
6.24
Sample
2
173
TNTC
TNTC
11
32
22
2.17
×
106
6.33
Sample
3
TNTC
192
150
37
15
22
2.47
×
106
6.39
Plate
Counts
­
Dilution
=
102
Plate
Counts
­
Dilution
=
103
Concentration
After
UV
1
2
3
1
2
3
PFU/
mL
Log
Sample
1
166
181
TNTC
17
18
42
2.57
×
104
4.40
Sample
2
133
TNTC
101
13
28
10
1.70
×
104
4.23
Sample
3
165
141
123
17
14
12
1.43
×
104
4.15
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
34
June
2003
The
t­
statistic
for
3
samples
and
an
80
percent
confidence
level
is
1.88.
The
percent
uncertainty
of
the
log
inactivation
is
calculated
as
follows:

(
)
(
)

%
.
.
.
.
.
.

U
in
91
7
100
06
2
3
13
0
88
1
3
075
0
88
1
2
1
2
2
=
×
 

  
 

 

  
 
×
+
×
=

The
RED
associated
with
a
log
inactivation
of
2.06
is
calculated
as
follows:

(
)
2
2
mJ/
cm
8
.
28
06
.
2
47
.
2
06
.
2
90
.
8
Dose
=
×
+
×
=

The
percent
uncertainty
of
the
regression
equation,
UDR,
at
a
log
inactivation
of
2.06
is
6
percent.
The
percent
uncertainty
of
the
RED
is
calculated
as
follows:

(
)
%
9
.
12
2
.
8
6
9
.
7
U
2
1
2
2
2
RED
=
+
+
=

C.
4.9.9
Interpretation
of
Results
Interpretation
of
the
results
should
depend
on
the
monitoring
approach
used
to
guarantee
dose
delivery:

 
With
the
UV
intensity
setpoint
approach,
the
UV
reactor
should
be
rated
at
the
lowest
inactivation
observed
for
each
setpoint
condition
evaluated.

 
With
the
UV
intensity
and
UVT
setpoint
approach,
the
UV
reactor
should
be
rated
at
the
inactivation
observed
with
UV
reactor
operation
under
setpoint
conditions.

 
With
the
calculated
dose
approach,
the
UV
reactor
should
be
rated
at
the
lowest
inactivation
observed
for
each
calculated
dose
setpoint
evaluated.

C.
4.9.10
Interpolation
of
Results
The
RED
measured
by
validation
testing
can
be
interpolated
as
a
function
of
flowrate,
UVT,
and
UV
intensity
by
fitting
an
equation
to
the
data
being
interpolated.
If
the
RED
is
interpolated
as
a
function
of
the
measured
intensity
or
the
inverse
flowrate,
the
equation
used
should
pass
through
the
origin
(
0,0).
The
equation
coefficients
should
be
significant
at
a
95
percent
confidence
level.
The
differences
between
the
values
measured
and
predicted
by
the
equation
should
be
randomly
distributed
around
zero.
The
equation
should
be
used
for
interpolating
between
measured
data
but
should
not
be
used
for
extrapolation.

The
uncertainty
of
the
equation
used
to
interpolate
the
RED
should
be
assessed
by
determining
the
80
percent
confidence
level.
If
significant,
the
uncertainty
should
be
included
as
an
uncertainty
term
in
the
determination
of
the
expanded
uncertainty,
as
described
in
section
C.
4.10.2.3.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
35
June
2003
C.
4.10
Determining
Inactivation
Credit
This
guidance
presents
two
approaches,
termed
Tier
1
and
Tier
2,
which
can
be
used
to
relate
the
RED
demonstrated
during
reactor
validation
to
target
pathogen
inactivation.
Other
approaches
or
modifications
to
this
approach
may
be
used
at
the
discretion
of
the
State.

With
both
approaches,
the
RED
demonstrated
during
validation
should
be
equal
to
or
greater
than
a
target
RED
that
is
related
to
the
dose
tables
in
Chapter
1
using
a
safety
factor.
With
Tier
1,
fixed
safety
factors
have
been
defined
and
applied
to
the
dose
tables
in
Chapter
1
to
define
target
RED
values.
The
Tier
1
safety
factors
are
based
on
specific
Tier
1
criteria
for
the
UV
reactor
and
its
validation
protocol.
The
Tier
1
approach
can
be
used
with
a
given
UV
reactor
provided
it
meets
all
the
Tier
1
criteria.
With
Tier
2,
the
safety
factors
are
calculated
based
on
the
validation
results
for,
and
certain
properties
of,
the
UV
reactor
that
are
calculated
from
the
validation
results
and
certain
properties
of
the
UV
reactor
undergoing
validation.

C.
4.10.1
Tier
1
Approach
For
a
UV
reactor
using
LP
or
LPHO
lamps,
Table
C.
3
presents
the
Tier
1
RED
values
that
should
be
demonstrated
during
validation
to
achieve
the
specified
log­
inactivation
credits
for
Cryptosporidium,
Giardia,
and
virus.
Table
C.
4
presents
the
Tier
1
RED
values
for
MP
reactors.
The
Tier
1
RED
values
are
applicable
with
all
UV
reactors
that
meet
the
Tier
1
criteria
provided
in
this
section.

Example.
To
receive
2.5
log
Cryptosporidium
inactivation
credit,
a
LP
reactor
under
Tier
1
should
demonstrate
an
RED
of
28
mJ/
cm2.

Table
C.
3
Tier
1
RED
Targets
for
UV
Reactors
with
LP
or
LPHO
Lamps
RED
Target
(
mJ/
cm2)
Log
Inactivation
Credit
Cryptosporidium
Giardia
Virus
0.5
6.8
6.6
55
1.0
11
9.7
81
1.5
15
13
110
2.0
21
20
139
2.5
28
26
169
3.0
36
34
199
3.5
­
­
227
4.0
­
­
259
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
36
June
2003
Table
C.
4
Tier
1
RED
Targets
for
UV
Reactors
with
MP
Lamps
RED
Target
(
mJ/
cm2)
Log
Inactivation
Credit
Cryptosporidium
Giardia
Virus
0.5
7.7
7.5
63
1.0
12
11
94
1.5
17
15
128
2.0
24
23
161
2.5
32
30
195
3.0
42
40
231
3.5
­
­
263
4.0
­
­
300
Tier
1
criteria
for
the
UV
reactor
are
as
follows:

 
UV
reactors
equipped
with
MP
lamps
should
be
equipped
with
one
sensor
per
lamp.
UV
reactors
equipped
with
LP
or
LPHO
lamps
should
be
equipped
with
at
least
one
sensor
per
bank
of
lamps.

 
The
standard
deviation
of
the
UV
output
of
LP
or
LPHO
lamps
should
be
15
percent
or
less
of
the
mean
output.
The
standard
deviation
should
be
determined
using
either
life
test
or
field
data
on
aged
lamps.

 
UV
intensity
sensors
should
view
a
point
along
the
length
of
the
lamp
that
is
within
25
percent
of
the
arc
length
away
from
the
electrode.

 
UV
intensity
sensors
should
have
a
spectral
response
that
peaks
between
250
and
280
nm.
When
mounted
on
the
UV
reactor
and
viewing
the
lamps
through
water,
the
measurement
of
UV
light
greater
than
300
nm
made
by
the
sensor
should
be
less
than
10
percent
of
the
total
measurement
made
by
the
sensor.
Conformance
to
these
criteria
can
be
demonstrated
using
UV
intensity
field
modeling.
Figure
C.
4
presents
an
example
of
how
two
sensors
would
conform
to
this
criterion.

 
The
UV
intensity
sensors
used
during
validation
and
the
duty
and
reference
sensors
used
during
operation
of
the
UV
reactor
at
the
WTP
should
provide
NIST
traceable
measurements
with
a
measurement
uncertainty
of
±
15
percent
or
less
at
an
80
percent
confidence
level.

 
During
operation
of
the
UV
reactor
at
the
WTP,
measurements
made
by
the
duty
UV
intensity
sensor
should
be
checked
using
a
reference
UV
intensity
sensor.
The
difference
between
the
measurement
made
by
the
duty
and
reference
sensors
should
meet
the
following
criteria:

(
)
2
1
2
Duty
2
Ref
Ref
uty
D
100
1
I
I
 
+
 
 
×
 

  
 

 

  

 
 
Equation
C.
9
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
37
June
2003
Figure
C.
4
Comparison
of
the
Spectral
Response
of
Two
UV
Intensity
Sensors
Estimated
Using
UV
Intensity
Field
Modeling
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
200
250
300
350
400
Wavelength
(
nm)
Spectral
Response
Relative
to
254
nm
Unfiltered
Sensor
Filtered
Sensor
250
280
Filtered
Sensor.
Detected
UV
light
with
a
0
cm
sensor­
to­
lamp
water
layer.
Detected
UV
>
300
nm
is
0.7%
of
total
UV
light
detected.
Unfiltered
Sensor.
Detected
UV
light
with
a
0
cm
sensor­
to­
lamp
water
layer.
Detected
UV
>
300
nm
is
41%
of
total
UV
light
detected.
0
5
10
15
20
25
200
250
300
350
400
Wavelength
(
nm)
Detected
Irradiance
0
10
20
30
40
50
200
250
300
350
400
Wavelength
(
nm)
Detected
Irradiance
0.0000
0.0005
0.0010
0.0015
0.0020
0.0025
200
250
300
350
400
Wavelength
(
nm)
Detected
Irradiance
0.0000
0.0100
0.0200
0.0300
0.0400
200
250
300
350
400
Wavelength
(
nm)
Detected
Irradiance
Unfiltered
Sensor.
Detected
UV
light
with
a
20
cm
sensor­
to­
lamp
water
layer.
Detected
UV
>
300
nm
is
85%
of
total
UV
light
detected.
Filtered
Sensor.
Detected
UV
light
with
a
20
cm
sensor­
to­
lamp
water
layer.
Detected
UV
>
300
nm
is
5%
of
total
UV
light
detected.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
38
June
2003
 
If
the
dose
monitoring
strategy
uses
an
on­
line
UVT
monitor,
the
A254
calculated
from
the
measured
UVT
should
have
a
measurement
uncertainty
of
±
10
percent
or
less
at
an
80
percent
confidence
level.

Tier
1
criteria
for
the
flow
measurements
are
as
follows:

 
The
flow
measurements
during
validation
and
during
operation
of
the
UV
reactor
at
the
WTP
should
have
a
measurement
uncertainty
of
±
5
percent
or
less
at
an
80
percent
confidence
level.

Tier
1
criteria
for
the
collimated
beam
apparatus
are
as
follows:

 
The
calculated
dose
delivered
by
the
collimated
beam
apparatus
should
have
a
measurement
uncertainty
of
±
15
percent
or
less
at
an
80
percent
confidence
level.

Tier
1
criteria
for
the
challenge
microorganism
dose­
response
are
as
follows:

 
Over
the
range
of
doses
within
one
log
of
the
log
reduction
demonstrated
during
validation,
the
UV
sensitivity
of
the
challenge
microorganism
should
be
less
than
or
equal
to
25
mJ/
cm2
per
log
inactivation
(
the
dose­
response
of
a
resistant
strain
of
MS2).
For
example,
if
you
measure
log
inactivation
values
between
1.5
and
3.5
log,
the
test
organism
you
use
should
have
a
dose­
response
less
than
or
equal
to
25
mJ/
cm2
per
log
inactivation
between
0.5
and
4.5
log
inactivation.

 
If
the
dose­
response
of
the
challenge
microorganism
has
a
shoulder,
that
shoulder
should
not
occur
over
a
dose
range
greater
than
50
percent
of
the
RED
demonstrated
during
validation.
The
shoulder
is
defined
by
extrapolating
the
exponential
reduction
region
of
the
dose­
response
curve
to
the
dose­
axis.

 
If
the
dose­
response
demonstrates
tailing,
the
tailing
should
not
occur
until
one
log
reduction
greater
than
the
highest
log
reduction
demonstrated
during
validation.

Tier
1
criteria
for
the
UVT
used
for
validating
UV
reactors
using
medium­
pressure
lamps
are
as
follows:

 
The
UVT
at
254
nm
of
the
water
during
validation
should
be
greater
than
the
values
specified
in
Figure
C.
5
for
a
given
sensor­
to­
lamp
water
layer
and
UV­
absorbing
chemical
(
the
polychromatic
bias
should
be
1.0).
The
sensor­
to­
lamp
water
layer
is
defined
as
the
distance
traveled
through
water
by
UV
light
passing
from
the
lamp
to
the
sensor.
The
values
in
Figure
C.
5
were
taken
from
Figure
C.
7
for
a
polychromatic
bias
of
1.2.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
39
June
2003
Figure
C.
5
Criteria
for
the
Minimum
UVT
of
MP
UV
Reactors
under
Tier
1
Tier
1
criteria
for
the
challenge
microorganism
dose­
response
data
are
as
follows:

 
A
plot
of
dose
versus
log
inactivation
should
have
an
80
percent
confidence
level
of
10
percent
or
less
at
the
log
inactivation
demonstrated
by
the
UV
reactor.

Tier
1
criteria
for
the
challenge
microorganism
measurements
through
the
reactor
are
as
follows:

 
Five
influent
and
five
effluent
samples
should
be
collected
per
test
condition
evaluated
as
per
section
C.
4.9.5.

 
The
standard
deviation
of
the
challenge
microorganism
concentration
measured
with
the
influent
and
the
effluent
samples
should
be
less
than
or
equal
to
0.20
log.

Tier
1
criteria
for
the
interpolation
of
challenge
microbe
results
are
as
follows:

 
The
uncertainty
of
the
interpolation
should
be
10
percent
or
less
at
an
80
percent
confidence
level.

C.
4.10.2
Tier
2
Approach
The
safety
factor
used
to
relate
the
RED
demonstrated
during
validation
to
the
dose
required
to
inactivate
the
target
pathogen
should
be
defined
using
Equation
C.
10:

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
40
June
2003
(

e
B
B
SF
Poly
RED
+
×
×
=
1
)
Equation
C.
10
where
SF
=
Safety
Factor
BRED
=
RED
bias
BPoly
=
Polychromatic
bias
e
=
Expanded
uncertainty
as
a
fraction
The
following
sections
describe
an
approach
for
defining
each
of
these
terms.

Determining
the
RED
Bias
If
a
single
challenge
microorganism
is
used
to
demonstrate
dose
delivery
during
validation,
the
RED
bias
should
be
determined
using
Figure
C.
6
and
the
following
procedure.
(
Section
F.
1
provides
the
background
on
the
development
of
Figure
C.
6
and
the
procedure
for
determining
the
RED
bias.)

Procedure
1.
Calculate
the
UV
sensitivity
of
the
target
pathogen
as
the
dose
requirement
specified
in
Chapter
1
divided
by
the
corresponding
log
inactivation
credit.

2.
Calculate
the
UV
sensitivity
of
the
challenge
microorganism
as
the
calculated
RED
divided
by
the
log
inactivation.

3.
If
the
target
pathogen
is
more
resistant
to
UV
light
than
the
challenge
microorganism,
the
RED
bias
equals
1.0.
Otherwise,
calculate
the
RED
bias
using
Equation
C.
11:

P
C
RED
RED
Bias
RED
=
Equation
C.
11
where
REDC
=
RED
of
the
challenge
microorganism
obtained
from
Figure
C.
6
REDP
=
RED
of
the
target
pathogen
obtained
from
Figure
C.
6
Example.
An
MS2
inactivation
of
2
log
corresponding
to
an
RED
of
36
mJ/
cm2
is
measured
during
validation.
A
2­
log
Cryptosporidium
credit
is
required.
The
UV
dose
required
to
achieve
that
level
of
inactivation
from
Chapter
1
is
5.8
mJ/
cm2.
Thus,
the
UV
sensitivity
of
MS2
and
Cryptosporidium
is
defined
as
36/
2.0
=
18
and
5.8/
2.0
=
2.9
mJ/
cm2
per
log
inactivation,
respectively.
Because
MS2
is
more
resistant
than
Cryptosporidium,
the
RED
bias
is
greater
than
one.
In
Figure
C.
6,
REDs
of
19
and
8.2
correspond
to
UV
sensitivities
of
18
and
2.9
mJ/
cm2
per
log,
respectively.
Thus,
using
Equation
C.
11,
the
RED
bias
is
19/
8.2
=
2.3.

Example.
An
MS2
inactivation
of
4
log
and
a
corresponding
RED
of
80
mJ/
cm2
is
measured
during
validation.
A
2.0­
log
adenovirus
credit
requiring
a
dose
of
100
mJ/
cm2
is
required.
Thus,
the
UV
sensitivity
of
the
challenge
microorganism
and
pathogen
are
20
and
50
mJ/
cm2
per
log
inactivation,
respectively.
Because
the
UV
sensitivity
of
adenovirus
is
greater
than
that
of
the
challenge
microorganism,
the
RED
bias
equals
1.0.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
41
June
2003
Figure
C.
6
RED
versus
Microorganism
UV
Sensitivity
for
Use
in
Determining
the
RED
Bias
0
5
10
15
20
25
30
0
10
20
30
40
5
UV
Sensitivity
(
mJ/
cm2
per
log
inactivation)
Reduction
Equivalent
Dose
(
mJ/
cm2)

0
If
two
challenge
microorganisms
with
different
UV
sensitivities
are
used
during
validation
to
demonstrate
dose
delivery,
the
RED
delivered
to
the
target
pathogen
can
be
determined
by
interpolation
using
the
following
procedure.
(
Section
F.
1.3
provides
the
background
on
the
use
of
two
challenge
microorganisms
to
demonstrate
RED
delivered
to
a
target
pathogen.)

Procedure
1.
For
a
given
test
condition
of
flowrate,
UVT,
and
lamp
output,
calculate
the
UV
sensitivity
of
the
challenge
microorganisms
as
their
respective
measured
REDs
divided
by
their
corresponding
log
inactivations.

2.
Determine
the
UV
sensitivity
of
the
target
pathogen
as
the
dose
listed
in
Chapter
1
divided
by
the
log
inactivation.

3.
Calculate
the
RED
delivered
to
the
target
pathogen
using
the
following
equation:

(
)(
)
(
)
1
C
2
C
1
C
P
1
C
2
C
1
C
P
10
D
10
D
10
D
10
D
RED
RED
RED
RED
 
 
 
+
=
Equation
C.
12
where
REDP
=
Estimate
of
the
target
pathogen's
RED
REDC1
=
The
RED
measured
with
the
first
challenge
microorganism
REDC2
=
The
RED
measured
with
the
second
challenge
microorganism
D10P
=
UV
sensitivity
of
the
target
pathogen
(
mJ/
cm2
per
log
inactivation)
D10C1
=
UV
sensitivity
of
the
first
challenge
microorganism
(
mJ/
cm2
per
log
inactivation)
D10C2
=
UV
sensitivity
of
the
second
challenge
microorganism
(
mJ/
cm2
per
log
inactivation)

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
42
June
2003
4.
Calculate
the
percent
uncertainty
of
the
estimated
RED
of
the
target
pathogen
using
Equation
C.
13:

(
)
(
)
(
{
})

P
2
1
2
1
C
,
RED
2
C
2
1
C
,
RED
1
C
2
1
C
2
C
1
C
p
2
1
C
,
RED
1
C
REDp
RED
U
RED
U
RED
10
D
10
D
10
D
10
D
U
RED
U
 
 

 
 

 
 

 
 
+

 

  
 

 

  
 
 
 
+

=

Equation
C.
13
where
REDp
U
=
Percent
uncertainty
of
the
RED
estimated
for
the
pathogen
1
C
,
RED
U
=
Percent
uncertainty
of
the
RED
measured
with
the
first
challenge
microorganism
(
see
Equation
C.
8)

As
an
alternative
two­
microorganism
approach,
the
log
inactivation
measured
with
the
challenge
microorganisms
can
be
interpolated
as
a
function
of
the
microorganisms'
first­
order
inactivation
coefficients.

Example.
A
UV
reactor
was
validated
using
MS2
and
 X174
at
1
and
2
mgd.
The
UV
sensitivities
of
the
MS2
and
 X174
were
18
and
2
mJ/
cm2
per
log
inactivation,
respectively.
The
following
table
gives
the
RED
and
percent
uncertainties
measured
with
MS2
and
 X174.
At
the
lower
flowrate
of
1
mgd,
the
 X174
was
inactivated
to
below
the
detection
limit
and
the
measured
RED
was
estimated
as
greater
than
10
mJ/
cm2.
The
table
also
gives
the
RED
delivered
to
Cryptosporidium
estimated
using
Equation
C.
12
and
the
percent
uncertainty
of
that
RED
estimated
using
Equation
C.
13.
These
estimations
assumed
a
UV
sensitivity
of
Cryptosporidium
of
4.0
mJ/
cm2
per
log
inactivation
based
on
the
dose
in
Chapter
1
for
a
3.0­
log
inactivation
credit.

MS2
 X174
Cryptosporidium
Flow
(
mgd)
RED
(
mJ/
cm2)
Uncertainty
(%)
RED
(
mJ/
cm2)
Uncertainty
(%)
RED
(
mJ/
cm2)
Uncertainty
(%)
1
40
6
>
10
0
14
3.2
2
20
11
9
4
10
2.9
Determining
the
Polychromatic
Bias
For
a
UV
reactor
using
a
germicidal
UV
intensity
sensor
(
the
spectral
response
meets
Tier
1
criteria),
the
polychromatic
bias
can
be
assigned
a
value
of
one
if
the
UV
intensity
sensor
is
located
where
dose
delivery
is
proportional
to
measured
UV
intensity
or
closer
to
the
lamps
than
that
location.
This
can
be
shown
experimentally
by
demonstrating
under
fixed
conditions
of
flow
and
measured
UV
intensity
that
the
RED
obtained
with
peak
UVT
and
lowered
lamp
power
is
greater
than
or
equal
to
the
RED
measured
with
peak
lamp
power
and
lowered
RED.

If
data
are
not
available
showing
the
UV
intensity
sensor
location
meets
the
above
criteria,
the
polychromatic
bias
should
be
determined
by
calculating,
at
a
given
flowrate,
UVT,

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
43
June
2003
and
measured
UV
intensity,
the
ratio
of
RED
during
validation
to
the
RED
at
the
WTP.
This
calculation
should
be
done
conservatively
by
assuming
ideal
dose
delivery
where
dose
is
the
product
of
the
average
intensity
within
the
reactor
and
the
theoretical
mean
residence
time.
The
calculation
should
include
the
following
factors:

 
The
spectral
UV
transmittance
of
the
water
during
validation
and
at
the
WTP.

 
The
spectral
lamp
output
during
validation
and
expected
at
the
WTP
with
aged
lamps.

 
The
spectral
sleeve
UV
transmittance
during
validation
and
expected
at
the
WTP
with
aged
and
fouled
sleeves.

 
The
spectral
response
of
the
sensor
used
during
validation
and
at
the
WTP.

 
The
action
spectra
of
the
challenge
microorganism
used
during
validation
and
the
action
spectra
of
the
target
pathogen
taken
from
the
literature.

If
the
above
ratio
is
less
than
one,
the
polychromatic
bias
should
be
assigned
a
value
of
one.

Figures
C.
7
to
C.
9
present
the
polychromatic
bias
for
reactors
with
UV
intensity
sensor
spectral
response
curves
shown
in
Figure
C.
10.
Each
figure
presents,
for
a
given
sensor
spectral
response,
the
polychromatic
bias
as
a
function
of
the
UVT,
the
sensor
to
sleeve
water
layer,
and
the
UV
absorbing
chemical
used
during
validation
(
coffee,
lignin
sulphonate,
and
natural
organic
matter
(
NOM)).
The
spectral
UV
absorption
coefficient
of
the
UV
absorbers
and
the
WTP
water
used
to
define
the
polychromatic
bias
values
is
provided
in
Figures
C.
11
and
C.
12.
Figures
C.
7
to
C.
9
can
be
used
to
determine
the
polychromatic
bias
if
the
spectral
response
of
the
UV
intensity
sensor
used
in
the
figure
is
representative
of
the
spectral
response
the
UV
reactor's
intensity
sensor.
Alternatively,
the
polychromatic
bias
can
be
calculated
using
a
model
that
meets
the
above­
mentioned
criteria.

The
polychromatic
bias
shown
in
Figures
C.
7
to
C.
9
was
determined
for
an
annular
reactor
with
a
reactor
radius
of
18.8
cm
and
a
sleeve
radius
of
3.81
cm.
The
UV
intensity
field
was
calculated
using
a
radial
intensity
model.
Section
F.
4.2
presents
details
on
the
models
used
to
develop
Figures
C.
7
to
C.
9.

The
polychromatic
bias
values
in
Figures
C.
7
to
C.
9
only
account
for
differences
between
the
spectral
UV
absorbance
during
validation
and
the
spectral
UV
absorbance
at
the
WTP.
They
do
not
account
for
the
impact
of
spectral
shifts
in
the
optical
properties
of
the
UV
reactor
(
e.
g.,
lamp
output,
sleeve
UVT).
If
spectral
shifts
in
UV
reactor
properties
occur
with
operation
of
the
UV
reactor
at
the
WTP,
the
polychromatic
bias
should
be
multiplied
by
terms
that
account
for
those
shifts.
Section
F.
4.3
describes
spectral
shifts
and
provides
estimates
of
the
polychromatic
biases
that
can
occur
with
those
shifts.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
44
June
2003
Figure
C.
7
Polychromatic
Bias
as
a
Function
of
Water
UVT
and
Sensor­
to­
Lamp
Water
Layer
for
UV
Reactors
using
Sensors
with
Germicidal
Response
(
response
A
in
Figure
C.
10)
Validated
using
Coffee,
Lignin
Sulphonate,
or
NOM
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
45
June
2003
Figure
C.
8
Polychromatic
Bias
as
a
Function
of
Water
UVT
and
Sensor­
To­
Lamp
Water
Layer
for
UV
Reactors
Using
Sensors
with
SiC
Response
(
Response
B
In
Figure
C.
10)
Validated
Using
Coffee,
Lignin
Sulphonate,
or
NOM
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
46
June
2003
Figure
C.
9
Polychromatic
Bias
as
a
Function
of
Water
UVT
and
Sensor­
To­
Lamp
Water
Layer
for
UV
Reactors
Using
Sensors
with
Germicidal
Response
(
Response
C
In
Figure
C.
10)
Validated
Using
Coffee,
Lignin
Sulphonate,
or
NOM
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
47
June
2003
Figure
C.
10
Spectral
Response
of
Sensors
Used
in
Defining
Figures
C.
7
to
C.
9
Figure
C.
11
UV
Absorption
Coefficient
of
Coffee,
Lignin
Sulphonate,
and
the
Target
Water
used
to
Define
Figures
C.
7
to
C.
9
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
48
June
2003
Figure
C.
12
Spectral
UV
Absorbance
of
NOM
and
the
Target
Water
Used
to
Define
Figures
C.
7
to
C.
9
0.00
0.05
0.10
0.15
0.20
200
250
300
350
400
Wavelength
(
nm)
UVa
(
cm­
1)

Target
Water
NOM
Example.
A
UV
reactor
equipped
with
"
Sensor
B"
located
10
cm
from
the
lamp
sleeve
(
10
cm
water
layer)
is
validated
using
coffee
as
a
UV
absorbing
chemical.
The
UV
reactor
is
validated
at
three
intensity
setpoints,
each
tested
at
lowered
UVT
values
of
95
percent,
90
percent,
and
85
percent.
The
polychromatic
bias
values
taken
from
Figure
C.
8
are
1.11,
1.29,
and
1.56,
for
UVT
values
of
95
percent,
90
percent
and
85
percent,
respectively.

Example.
A
UV
reactor
equipped
with
"
Sensor
A"
located
15
cm
from
the
sleeve
(
20
cm
water
layer)
is
being
considered
at
a
WTP
with
a
design
UVT
of
80
percent.
From
Figure
C.
7,
the
polychromatic
bias
with
coffee,
lignin
sulphonate,
and
NOM
are
1.7,
1.3,
and
1.2,
respectively.
Comparing
these
values,
a
strong
incentive
exists
to
select
the
UV
absorber
that
minimizes
the
polychromatic
bias.

Determining
the
Random
Uncertainty
The
random
uncertainty
associated
with
monitoring
and
validation
should
be
calculated
at
an
80
percent
confidence
level
using
the
uncertainty
of
the
terms
listed
in
Table
C.
5.
The
expanded
uncertainty
should
be
calculated
as
the
square
root
of
the
sum
of
squares
of
uncertainties
of
each
term.

If
one
challenge
microorganism
is
used
during
validation,
the
uncertainty
of
the
RED
is
calculated
using
Equation
C.
8.
If
two
challenge
microorganisms
are
used,
the
uncertainty
of
the
RED
is
calculated
using
Equation
C.
13.
The
uncertainty
of
the
interpolation
is
obtained
from
the
confidence
bands
of
the
equation
used
for
the
interpolation
(
see
section
C.
4.9.10).
The
uncertainty
of
the
UV
intensity
sensors
used
during
validation
and
used
at
the
WTP
should
be
obtained
from
manufacturer
data
with
supporting
documentation
as
per
Table
C.
1
in
section
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
49
June
2003
Table
C.
5
Factors
Impacting
Expanded
Uncertainty
of
Dose
Delivery
Monitoring
and
Validation
Uncertainty
Measured
RED
Any
interpolation
of
RED
as
a
function
of
flowrate,
UVT,
or
UV
intensity
Sensors
used
during
validation
(
UV
intensity,
UVT)
On­
line
and
reference
sensors
used
at
the
WTP
(
UV
intensity,
UVT)
Lamp
output
quantification
C.
2.2.
If
the
dose
monitoring
approach
uses
a
UVT
monitor,
include
the
measurement
uncertainty
of
the
UVT
monitor
obtained
from
data
provided
by
the
manufacturer.

The
uncertainty
of
lamp
output
quantification
is
zero
if
each
lamp
is
monitored
by
an
individual
UV
intensity
sensor.
Otherwise,
the
uncertainty
can
be
calculated
using
Equation
C.
14:

2
1
28
.
1
n
n
y
Uncertaint
 
=
Equation
C.
14
where
 
=
Standard
deviation
of
lamp­
to­
lamp
output
expressed
as
a
percentage
of
the
mean
n1
=
Number
of
banks
of
lamps
in
series
in
the
reactor
n2
=
Number
of
sensors
monitoring
each
bank
The
variability
of
UV
output
from
lamp­
to­
lamp
can
be
obtained
from
either
life
test
or
field
data
on
aged
lamps.

Example.
A
UV
reactor
consists
of
two
banks
of
four
lamps.
Each
bank
is
equipped
with
two
UV
intensity
sensors.
Dose
delivery
is
monitored
using
the
UV
intensity
setpoint
approach.
The
manufacturer
provides
data
showing
the
standard
deviation
of
lamp­
to­
lamp
output
is
12
percent
of
the
mean
output
at
the
end
of
lamp
life.
Thus,
the
lamp
output
quantification
uncertainty
is
1.28
×
12/(
20.5
×
20.5)
=
7.7
percent.
When
operating
at
a
WTP,
the
online
UV
intensity
sensors
have
a
measurement
uncertainty
of
20
percent.
The
on­
line
sensors
will
be
checked
using
a
reference
sensor
with
an
uncertainty
of
5
percent.
During
validation,
the
flowmeter
and
UV
intensity
sensors
had
an
uncertainty
of
0.5
percent
and
5
percent.
The
collimated
beam
dose
calculation
has
an
uncertainty
of
8
percent.
The
regression
fit
to
the
doseresponse
of
the
phage
has
an
uncertainty
of
10
percent.
The
UV
reactor
is
tested
at
peak
flowrate
with
the
results
shown
in
Table
C.
6.
The
uncertainty
of
the
measured
log
inactivation
is
determined
as
4.4
percent.
As
summarized
in
Table
C.
7,
a
total
uncertainty
of
26
percent
is
calculated
as
the
square
root
of
the
sum
of
the
squares
of
the
individual
uncertainties.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
50
June
2003
Table
C.
6
Sample
Calculation
of
the
Log
Inactivation
Uncertainty
Influent
Effluent
N
Log
N
N
Log
N
3.60
×
105
5.56
154
2.19
4.90
×
105
5.69
206
2.31
4.10
×
105
5.61
263
2.42
Mean
5.62
Mean
2.31
St
Dev.
0.067
St
Dev.
0.116
T­
statistic
2.92
T­
statistic
2.92
Uncertainty
0.0729
Uncertainty
0.126
Inactivation
Log
Inactivation
3.31
Uncertainty
0.145
Uncertainty
(%)
4.40
Table
C.
7
Sample
Calculation
of
the
Expanded
Uncertainty
Uncertainty
Uncertainty
(%)
Uncertainty
Squared
Log
Inactivation
by
reactor
4.4
19
Collimated
beam
dose
calculation
8
64
Regression
fit
to
UV
Dose­
Response
Data
10
100
Validation
UV
intensity
sensor
5
25
WTP
on­
line
UV
intensity
sensor
20
400
WTP
reference
UV
intensity
sensor
5
25
Quantification
of
lamp­
to­
lamp
variability
7.7
59
Expanded
Uncertainty
26
692
Determining
the
Safety
Factor
The
safety
factor
relating
the
RED
measured
during
validation
to
the
pathogen
inactivation
requirements
should
be
calculated
as
the
product
of
the
RED
bias,
the
polychromatic
bias,
and
the
expanded
uncertainty
as
per
Equation
C.
10.

Example.
MS2
inactivation
of
2.0
log
corresponding
to
an
RED
of
40
mJ/
cm2
is
measured
during
validation
with
a
LP
reactor.
The
expanded
uncertainty
of
35
percent
is
calculated.
Because
LP
lamps
are
used,
the
polychromatic
bias
is
1.00.
An
RED
bias
of
2.0
is
determined
using
the
observed
UV
sensitivity
of
MS2
and
the
UV
sensitivity
associated
with
a
3.0­
log
Cryptosporidium
inactivation
credit.
A
safety
factor
of
(
1+
0.35)
×
2.0
×
1.0
=
2.7
is
calculated.
Hence,
the
Cryptosporidium
RED
demonstrated
by
validation
is
40
/
2.7
=
15
mJ/
cm2.
Because
the
demonstrated
Cryptosporidium
RED
is
greater
than
the
3.0­
log
requirement
of
12
mJ/
cm2,
the
UV
reactor
is
validated
for
a
3.0­
log
Cryptosporidium
inactivation
credit.

Example.
The
UV
reactor
in
the
above
example
is
instead
equipped
with
MP
lamps
monitored
with
UV
intensity
sensors
matching
the
spectral
response
of
Sensor
A.
The
UV
intensity
sensors
view
the
lamp
through
a
15
cm
water
layer.
The
UV
reactor
is
validated
using
lignin
sulphonate
at
a
maximum
UVT
of
80
percent.
Using
Figure
C.
7,
the
polychromatic
bias
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
51
June
2003
of
1.3
is
determined.
Thus,
the
safety
factor
is
(
1+
0.35)
×
2.0
×
1.3
=
3.5
and
the
Cryptosporidium
RED
demonstrated
by
validation
is
40/
3.5
=
11.4
mJ/
cm2.
In
this
case,
the
demonstrated
RED
is
less
than
the
required
RED
of
12
mJ/
cm2
for
3.0­
log
Cryptosporidium
inactivation
credit,
and
the
UV
reactor
should
not
be
considered
validated
for
3.0­
log
inactivation
of
Cryptosporidium.
However,
for
a
2.5­
log
Cryptosporidium
inactivation
requiring
a
dose
of
8.5
mJ/
cm2,
the
RED
bias
is
2.1,
resulting
in
a
safety
factor
of
3.7
and
a
demonstrated
Cryptosporidium
RED
of
40
/
3.7
=
10.8
mJ/
cm2.
Because
the
demonstrated
RED
of
10.8
mJ/
cm2
is
greater
than
the
target
RED
of
8.5
mJ/
cm2,
the
UV
reactor
can
be
considered
validated
for
a
2.5
log
Cryptosporidium
inactivation
credit.

C.
4.11
Validation
Test
Report
The
engineer
responsible
for
third­
party
oversight
should
collect
all
documentation
and
test
results
and
prepare
summary
and
detailed
reports.

C.
4.11.1
Summary
Report
The
summary
report
should
describe
the
UV
reactor
validated
under
this
protocol
in
general
terms
including
the
following
components:

 
Inlet
and
outlet
conditions
 
Number
of
UV
lamps
and
their
location
within
the
reactor
 
Lamp
characteristics
including
type,
electrical
power
consumption,
and
spectral
output
 
Monitoring
and
controls
approach
used
for
dose
compliance
 
Number
of
UV
intensity
sensors
and
their
locations
 
UVT
monitor,
if
used
 
Safety
features
used
to
ensure
water
disinfection
The
summary
report
should
provide
the
challenge
microorganism
UV
dose­
response,
including
the
regression
fit
and
the
confidence
intervals.
The
report
should
tabulate
each
reactor
test
condition
evaluated,
including
the
flowrate,
UV
intensity
setpoint,
UVT
setpoint
(
if
used),
calculated
dose
(
if
used),
log
inactivation
achieved,
and
calculated
RED.
The
number
of
samples
evaluated,
the
standard
deviation
of
the
influent
and
effluent
samples,
and
the
uncertainty
of
the
inactivation
through
the
reactor
should
also
be
tabulated.

If
interpolation
of
bioassay
results
is
part
of
dose
monitoring,
tables
or
charts
should
present
the
results
of
the
interpolation.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
52
June
2003
If
the
reactor
is
evaluated
under
Tier
1,
documentation
should
be
provided
supporting
that
the
validation
met
Tier
1
criteria.
The
report
should
state
the
pathogen
credits
that
the
UV
reactor
can
achieve
based
on
the
Tier
1
designation.

If
the
UV
reactor
is
evaluated
under
Tier
2,
documentation
should
be
provided
describing
the
Tier
2
analysis
including
the
determination
of
the
RED
bias,
polychromatic
bias,
and
expanded
uncertainty.
For
the
expanded
uncertainty,
each
term
used
in
the
calculation
should
be
provided.
The
report
should
state
the
pathogen
credits
that
the
UV
reactor
can
achieve
based
on
the
Tier
2
results.

Based
on
the
values
used
to
determine
the
safety
factor
applied
to
the
validation
data
(
Tier
1
or
2),
the
summary
report
should
specify
all
criteria
for
the
measurement
uncertainty
of
the
UV
intensity
sensors,
and
UVT
monitors
used
at
the
WTP.

C.
4.11.2
Detailed
Report
The
detailed
report
should
provide
a
comprehensive
description
of
the
test
methodology
that
includes
the
following
components:

 
Identity
and
qualifications
of
personnel
involved
in
the
validation
test
 
UV
reactor
specifications
 
UV
intensity
sensor
specifications
and
calibration
documentation
 
Physical
test
set­
up
 
Summary
of
QA/
QC
procedures
 
Materials
and
methods
employed
during
the
test
 
Complete
test
results,
including
raw
data
and
analyses
performed
C.
5
UV
Reactor
Validation
Examples
This
section
provides
examples
of
UV
reactor
validation
for
the
following
reactors
and
monitoring
approach
combinations:

 
LP
reactor
using
a
single
intensity
setpoint
(
section
C.
5.1)

 
LP
reactor
using
multiple
setpoints
as
a
function
of
flowrate
(
section
C.
5.2)

 
LP
reactor
using
multiple
setpoints
as
a
function
of
flowrate
and
dose
(
section
C.
5.3)

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
53
June
2003
 
MP
reactor
using
a
single
intensity
setpoint
and
UVT
setpoint
(
section
C.
5.4)

 
MP
reactor
using
the
calculated
dose
method
for
monitoring
(
section
C.
5.5)

C.
5.1
LP
Reactor
Using
a
Single
Intensity
Setpoint
A
UV
reactor
consists
of
two
banks
in
series
of
nine
LPHO
lamps
oriented
perpendicular
to
flow.
Dose
delivery
is
monitored
using
the
UV
intensity
setpoint
approach.
Each
bank
is
equipped
with
one
UV
intensity
sensor.

The
UV
reactor
is
considered
for
use
at
a
WTP.
The
application
requires
a
2.5
log
inactivation
credit
of
Cryptosporidium.
The
design
flowrate
and
UVT
at
the
WTP
are
500
gpm
and
90
percent,
respectively.
The
UV
manufacturer
states
the
lamp
fouling/
aging
factor
for
the
reactor
is
70
percent.
During
operation
at
a
WTP,
the
on­
line
and
reference
UV
intensity
sensors
are
expected
to
have
a
measurement
uncertainty
of
15
and
5
percent,
respectively.
The
reactor
will
operate
at
the
WTP
using
a
single
intensity
setpoint
to
indicate
dose
delivery
over
a
flow
range
of
100
to
500
gpm.

The
reactor
is
validated
using
coffee
as
the
UV
absorber
and
MS2
as
the
challenge
microorganism.
Figure
C.
13
gives
the
dose­
response
of
the
MS2
measured
during
validation
with
a
collimated
beam.
The
dose­
response
is
fitted
using
the
following
equation:

3
.
4
N
N
log
3
.
20
Dose
0
+

  

 

  
 
×
=

Figure
C.
13
also
provides
80
percent
confidence
levels
for
the
fit
and
the
percent
uncertainty
of
the
UV
dose
calculated
from
those
confidence
levels.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
54
June
2003
Figure
C.
13
Dose­
Response
of
the
MS2
Challenge
Microorganism
Used
in
Example
C.
5.1
A
reference
sensor
is
used
to
monitor
UV
intensity
during
validation
testing.
The
intensity
setpoint
to
be
validated
is
determined
by
operating
the
reactor
under
design
conditions
of
70
percent
lamp
output
and
90
percent
water
UVT.
Under
these
conditions,
the
UV
intensity
sensor
reads
5.0
mW/
cm2.

Table
C.
8
gives
the
validation
test
conditions
and
results.
The
reactor
is
tested
at
flowrates
of
100
and
500
gpm
with
the
intensity
sensor
reading
5.0
mW/
cm2.
At
each
flowrate,
the
reactor
is
tested
under
conditions
of
low
UVT
­
high
lamp
output
and
high
UVT
­
low
lamp
output.
Each
test
condition
is
evaluated
using
five
influent
and
five
effluent
samples.

Table
C.
8
Validation
Test
Conditions
and
Results
for
Example
C.
5.1
Test
Conditions
Test
Results
Flow
(
gpm)
UVT
(%)
Lamp
(%)
UV
Intensity
(
mW/
cm2)
Influent
(
log)
Effluent
(
log)
Inactivation
(
log)
RED
(
mJ/
cm2)

100
98
44
4.98
4.97
±
0.08
<
0
>
4.97
>
105
100
84
100
4.90
5.02
±
0.10
<
0
>
5.02
>
106
500
98
44
4.98
5.03
±
0.06
4.02
±
0.08
1.00
24.6
500
84
100
4.92
5.02
±
0.10
3.52
±
0.18
1.49
34.6
Note.
Influent
and
effluent
data
presented
as
mean
±
standard
deviation.

Based
on
the
results,
the
reactor
is
rated
at
an
MS2
RED
of
24.6
mJ/
cm2
for
a
flow
range
of
100
to
500
gpm
and
a
sensor
setpoint
of
5
mW/
cm2.

A
Tier
2
analysis
was
used
to
assess
if
the
reactor
achieved
2.5
log
Cryptosporidium.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
55
June
2003
RED
Bias.
Since
a
2.5
log
inactivation
of
Cryptosporidium
requires
a
dose
of
8.5
mJ/
cm2,
the
UV
sensitivity
of
Cryptosporidium
is
defined
as
8.5/
2.5
=
3.4
mJ/
cm2
per
log
inactivation.
Since
1.00­
log
MS2
inactivation
occurred
with
a
dose
of
24.6
mJ/
cm2,
the
UV
sensitivity
of
MS2
is
defined
as
24.6
mJ/
cm2
per
log
inactivation.
In
Figure
C.
6,
an
RED
of
9.2
and
21
mJ/
cm2
occurs
with
a
UV
sensitivity
of
3.4
and
24.6
mJ/
cm2
per
log
inactivation,
respectively.
Accordingly,
the
RED
bias
is
21/
9.2
=
2.28.

Polychromatic
Bias.
The
polychromatic
bias
equals
1.0
because
the
UV
reactor
uses
LPHO
lamps.

Expanded
uncertainty.
A
t­
statistic
of
1.53
is
associated
with
5
samples
and
an
80
percent
confidence
level.
Using
the
standard
deviations
for
the
influent
and
effluent
counts
in
Table
C.
8,
the
uncertainty
of
the
log
inactivation
through
the
reactor
is
calculated
as
follows
using
Equation
C.
7:

(
)
(
)

%
8
.
6
%
100
*
00
.
1
5
53
.
1
08
.
0
5
53
.
1
06
.
0
Error
2
1
2
2
=
 

  
 

 

  
 
×
+
×
=

The
uncertainty
of
the
collimated
beam
dose
calculation
was
determined
to
be
8.9
percent.
At
a
UV
dose
of
24.6
mJ/
cm2,
the
uncertainty
in
the
dose
calculation
based
on
the
confidence
bands
in
Figure
C.
13
is
9.6
percent.

The
uncertainties
of
the
sensors
used
during
validation
and
at
the
WTP
are
as
follows:

 
Validation
UV
intensity
sensor
5
percent
 
WTP
on­
line
UV
intensity
sensor
15
percent
 
WTP
reference
UV
intensity
sensor
5
percent
The
total
uncertainty
of
the
sensors
is
calculated
according
to
the
following
equation:

(
)
%
6
.
16
5
15
5
Error
2
1
2
2
2
=
+
+
=

The
UV
vendor
states
the
standard
deviation
of
the
UV
output
from
lamp
to
lamp
is
25
percent.
Given
two
banks
of
lamps
and
one
UV
intensity
sensor
per
bank,
the
uncertainty
of
the
lamp
output
is
calculated
as
follows
using
Equation
C.
14:

Error
%
.
.
6
22
2
1
25
28
1
=
×
=

Including
each
of
these
random
uncertainty
terms,
the
expanded
uncertainty
is
calculated
as
follows:

Error(
)
%
.
.
.
.
.
.
7
31
6
22
6
16
6
9
9
8
8
6
2
1
2
2
2
2
2
=
+
+
+
+
=

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
56
June
2003
Safety
factor.
Using
Equation
C.
10,
the
safety
factor
is
calculated
as
follows:

(
)
00
3
00
1
28
2
317
0
1
.
.
.
.
SF
=
×
×
+
=

Based
on
this
safety
factor
value
and
the
Cryptosporidium
dose
target
for
2.5­
log
inactivation
credit,
the
MS2
RED
demonstrated
during
validation
should
be
as
follows:

2
mJ/
cm
5
.
25
5
.
8
00
.
3
RED
2
MS
=
×
=

Because
the
demonstrated
RED
of
24.6
mJ/
cm2
is
less
than
this
value,
the
reactor
cannot
get
2.5­
log
Cryptosporidium
inactivation
credit
operating
at
a
sensor
setpoint
of
15
mW/
cm2.
However,
with
a
2.0­
log
Cryptosporidium
credit
target,
the
RED
bias
would
be
2.6,
resulting
in
a
safety
factor
of
3.42
and
an
MS2
RED
target
of
19.8
mJ/
cm2.
Because
the
demonstrated
MS2
RED
is
greater
than
this
value,
the
reactor
can
get
2.0
log
Cryptosporidium
credit
operating
at
a
setpoint
of
15
mW/
cm2
over
a
flow
range
of
100
to
500
gpm.

The
reactor
does
not
meet
Tier
1
criteria
because
the
standard
deviation
of
the
UV
output
from
lamp­
to­
lamp
is
greater
than
15
percent.
If
the
reactor
did
meet
all
Tier
1
criteria,
the
reactor
would
receive
credit
for
2.0­
log
Cryptosporidium
based
on
a
comparison
of
the
demonstrated
MS2
RED
of
24.6
mJ/
cm2
with
the
dose
criteria
in
Table
C.
3.

C.
5.2
LP
Reactor
with
a
Intensity
Setpoint
Interpolation
as
a
Function
of
Flow
A
UV
reactor
consists
of
four
banks
of
six
LPHO
lamps
oriented
perpendicular
to
the
flow.
Dose
delivery
is
monitored
using
the
UV
intensity
setpoint
approach.
Each
bank
is
equipped
with
two
UV
intensity
sensors.

The
UV
reactor
is
rated
by
the
manufacturer
for
flows
ranging
from
0.9
to
2.4
mgd.
The
manufacturer
states
that
sensor
setpoints
of
6.0,
7.5,
10,
and
14
mW/
cm2
should
indicate
a
3.0­
log
Cryptosporidium
inactivation
credit
at
flows
of
0.9,
1.2,
1.7
and
2.4
mgd,
respectively.

During
operation
at
a
WTP,
the
on­
line
and
reference
UV
intensity
sensors
will
have
a
measurement
uncertainty
of
15
and
5
percent
respectively.

The
reactor
is
validated
using
lignin
sulphonate
as
the
UV
absorber
and
MS2
as
the
challenge
microorganism.
Figure
C.
14
gives
the
dose­
response
of
the
MS2
measured
during
validation
with
a
collimated
beam.
The
dose­
response
is
fitted
using
the
following
equation:

144
.
0
N
N
log
6
.
15
Dose
0
 

  

 

  
 
×
=

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
57
June
2003
Figure
C.
14
Dose­
Response
of
the
MS2
Challenge
Microorganism
Used
in
Example
C.
5.2
Confidence
intervals
are
fitted
to
the
data
at
an
80
percent
level.

Table
C.
9
gives
the
validation
test
conditions
and
results.
The
reactor
is
tested
at
four
flowrates,
0.9,
1.2,
1.7
and
2.4
mgd,
with
the
lamp
power
and
UVT
adjusted
to
give
a
UV
intensity
sensor
reading
at
the
setpoint
values.
At
each
flowrate,
the
reactor
is
tested
under
conditions
of
reduced
UVT
­
maximum
lamp
output
and
maximum
UVT
­
reduced
lamp
output.
A
reference
sensor
with
an
uncertainty
of
5
percent
is
used
during
validation
to
measure
UV
intensity.
Each
test
condition
is
evaluated
using
five
influent
and
five
effluent
samples.

Table
C.
9
Validation
Test
Conditions
and
Results
for
Example
C.
5.2
Test
Conditions
Test
Results
Inactivation
Flow
(
mgd)
UVT
(%)
Lamp
(%)
UV
Intensity
(
mW/
cm2)
Influent
(
Logs)
Effluent
(
Logs)
Log
Uncertainty
(%)
RED
(
mJ/
cm2)

0.90
98
37
6.15
5.99
±
0.096
2.95
±
0.080
3.04
3.8
47.5
0.90
70
100
6.06
5.94
±
0.127
3.21
±
0.087
2.73
5.4
42.5
1.2
98
45
7.48
6.09
±
0.100
2.98
±
0.108
3.11
4.5
48.5
1.2
75
100
7.46
6.04
±
0.070
3.34
±
0.088
2.70
4.0
42.1
1.7
98
61
10.1
6.03
±
0.150
3.27
±
0.112
2.76
6.5
43.1
1.7
83
100
10.1
5.98
±
0.116
3.45
±
0.120
2.52
6.2
39.4
2.4
98
83
13.8
6.03
±
0.102
3.37
±
0.090
2.67
4.9
41.6
2.4
92
100
13.7
6.02
±
0.136
3.37
±
0.062
2.66
5.4
41.4
Table
C.
10
presents
the
MS2
RED
and
reactor
setpoint
assigned
to
each
flowrate
based
on
the
validation
results.
A
Tier
2
analysis
was
used
to
determine
the
Cryptosporidium
inactivation
credit
that
can
be
assigned
to
the
reactor
given
the
validation
test
results.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
58
June
2003
Table
C.
10
Summary
of
Validation
Results
for
Example
C.
5.2
Flow
(
mgd)
UV
Intensity
Setpoint
(
mW/
cm2)
MS2
RED
(
mJ/
cm2)

0.90
6.06
42.5
1.2
7.46
42.1
1.7
10.1
39.4
2.4
13.7
41.4
RED
Bias.
Since
a
3.0­
log
inactivation
credit
for
Cryptosporidium
requires
a
dose
of
12
mJ/
cm2,
the
UV
sensitivity
of
Cryptosporidium
is
defined
as
12/
3.0
=
4.0
mJ/
cm2
per
log
inactivation.
The
UV
sensitivity
of
MS2
is
16
mJ/
cm2
per
log
inactivation
42.5/
2.73
=
16
mJ/
cm2.
In
Figure
C.
6,
an
RED
of
10
and
18
mJ/
cm2
is
associated
with
a
UV
sensitivity
of
4.0
and
16
mJ/
cm2
per
log
inactivation.
Accordingly,
the
RED
bias
is
18/
9.8
=
1.84.

Polychromatic
Bias.
The
polychromatic
bias
equals
1.0
because
the
UV
reactor
uses
LPHO
lamps.

Expanded
uncertainty.
The
uncertainty
of
the
log
inactivation
through
the
reactor,
calculated
using
Equation
C.
7,
is
tabulated
in
Table
C.
9.
A
mean
value
of
5.1
percent
is
used
as
the
uncertainty
of
the
log
inactivation
in
this
analysis.
The
uncertainty
of
the
collimated
beam
dose
calculation
was
determined
as
8.9
percent.
For
an
RED
near
40
mJ/
cm2,
the
uncertainty
in
the
RED
arising
from
the
scatter
in
the
dose­
response
in
Figure
C.
14
is
4
percent.

The
uncertainties
of
the
sensors
used
during
validation
and
at
the
WTP
are
as
follows:

 
Validation
UV
intensity
sensor
5
percent
 
WTP
On­
line
UV
intensity
sensor
10
percent
 
WTP
Reference
UV
intensity
sensor
5
percent
The
total
uncertainty
of
the
sensors
is
calculated
as
follows:

(
)
%
2
.
12
5
10
5
Error
2
1
2
2
2
=
+
+
=

The
UV
vendor
states
the
standard
deviation
of
the
UV
output
from
lamp
to
lamp
is
15
percent.
Given
four
banks
of
lamps
and
two
sensors
per
bank,
the
uncertainty
associated
with
the
number
of
sensors
is
calculated
as
follows
using
Equation
C.
14:

%
8
.
6
4
2
15
28
.
1
Error
=
×
=

Including
each
of
these
random
uncertainty
terms,
the
expanded
uncertainty
is
calculated
as
follows:

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
59
June
2003
(
)
%
8
.
17
8
.
6
2
.
12
4
9
.
8
1
.
5
Error
2
1
2
2
2
2
2
=
+
+
+
+
=

Safety
Factor.
Using
Equation
C.
10,
the
safety
factor
is
calculated
as
follows:

(
)
17
.
2
00
.
1
84
.
1
178
.
0
1
SF
=
×
×
+
=

Cryptosporidium
Credit.
Using
this
safety
factor,
the
target
RED
that
should
be
demonstrated
during
validation
is
12
×
2.17
=
26
mJ/
cm2.
Because
the
demonstrated
RED
of
39.4
mJ/
cm2
is
greater
than
this
number,
the
UV
reactor
operating
at
the
validated
intensity
setpoints
can
get
credit
for
3.0­
log
Cryptosporidium
inactivation.

The
validation
results
can
be
used
to
define
three
strategies
for
operating
the
UV
reactor
at
a
WTP:

1.
The
UV
reactor
can
operate
using
one
intensity
setpoint
over
the
full
range
of
flowrates.
In
this
case,
a
setpoint
of
13.7
mW/
cm2
can
be
used
to
indicate
a
3.0­
log
Cryptosporidium
inactivation
at
all
flows
of
2.4
mgd
or
less.

2.
The
UV
reactor
can
operate
using
multiple
intensity
setpoints
where
each
setpoint
functions
over
a
given
range
of
flows.
In
this
case,
a
setpoint
of
13.7
mW/
cm2
would
be
used
at
all
flows
from
1.7
to
2.4
mgd,
a
setpoint
of
10.1
mW/
cm2
would
be
used
at
all
flows
from1.2
to
1.7
mgd,
and
a
setpoint
of
7.46
mW/
cm2
would
be
used
at
all
flows
from
0.90
to
1.2
mgd.

3.
The
UV
reactor
can
be
operated
using
intensity
setpoints
interpolated
as
a
function
of
flowrate
using
the
validation
data.
In
this
case,
using
the
plot
of
sensor
setpoint
versus
flowrate
in
Figure
C.
15,
a
setpoint
value
of
11.7
mW/
cm2
can
be
used
at
a
flow
of
2
mgd
to
indicate
an
MS2
RED
of
39.4
mJ/
cm2
and
hence
a
3.0­
log
Cryptosporidium
inactivation
credit.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
60
June
2003
Figure
C.
15
Interpolation
of
Intensity
Setpoint
Values
Indicating
an
MS2
Dose
of
39.3
mJ/
cm2
0
20
40
60
80
100
60
70
80
90
100
Water
UV
Transmittance
(%)
Lamnp
Output
(%)

6.1
mW/
cm2
7.5
mW/
cm2
10.1
mW/
cm2
13.8
mW/
cm2
The
UV
reactor
and
validation
test
conditions
met
all
prerequisites
to
be
considered
under
Tier
1.
The
Tier
1
requirement
for
3.0­
log
inactivation
of
Cryptosporidium
by
a
LPHO
reactor
is
36
mJ/
cm2.
Since
the
RED
demonstrated
during
validation
is
greater
than
this
amount,
the
reactor
can
receive
3.0­
log
Cryptosporidium
inactivation
credit
under
Tier
1.

Validation
data
obtained
in
this
example
can
be
related
to
design
criteria
by
plotting
combinations
of
lamp
output
and
water
UVT
that
result
in
a
given
measured
UV
intensity
setpoint
value.
For
example,
Figure
C.
16
plots
combinations
of
lamp
output
and
water
UVT
that
result
in
the
intensity
setpoint
values
validated
in
Table
C.
9.
Any
combination
of
UVT
and
lamp
output
along
that
curve
can
be
used
as
design
criteria
for
each
setpoint
value
shown.
For
example,
a
setpoint
of
10.1
mW/
cm2
indicates
3.0­
log
Cryptosporidium
inactivation
at
a
flow
of
1.7
mgd.
A
setpoint
of
10.1
mW/
cm2
occurs
with
a
combination
of
70
percent
lamp
output
and
93
percent
UVT.
Thus,
the
reactor
could
be
used
in
a
design
application
where
the
design
flow,
UVT,
and
lamp
fouling/
aging
factor
are
1.7
mgd,
93
percent,
and
70
percent,
respectively.
The
setpoint
of
10.1
mW/
cm2
is
also
obtained
with
a
combination
of
80
percent
lamp
output
and
89
percent
UVT.
Thus,
the
reactor
could
also
be
used
in
a
design
application
where
the
design
flow,
UVT,
and
lamp
fouling/
aging
factor
is
1.7
mgd,
89
percent,
and
80
percent,
respectively.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
61
June
2003
Figure
C.
16
Combinations
of
Lamp
Output
and
Water
UV
Transmittance
that
Result
in
Given
Sensor
Setpoint
Values
C.
5.3
LP
Reactor
with
Intensity
Setpoint
Interpolation
as
a
Function
of
Flow
and
Target
Inactivation
A
UV
reactor
consists
of
twelve
rows
of
twelve
LPHO
lamps
oriented
perpendicular
to
flow.
Dose
delivery
is
monitored
using
the
UV
intensity
setpoint
approach.
Each
row
is
equipped
with
one
UV
intensity
sensor.

The
UV
reactor
is
rated
by
the
UV
vendor
for
flows
ranging
from
5
to
20
mgd.
During
operation
at
a
WTP,
the
on­
line
and
reference
UV
intensity
sensors
will
have
a
measurement
uncertainty
of
15
and
5
percent,
respectively.

The
UV
manufacturer
wants
to
validate
the
UV
reactor
using
test
conditions
that
allow
interpolation
of
intensity
setpoints
as
a
function
of
flowrate
and
measured
RED.
Table
C.
11
gives
the
validation
test
conditions
and
results.
To
allow
interpolation
of
sensor
setpoints
as
a
function
of
flowrate,
the
reactor
is
tested
at
a
three
flowrates
of
5,
10,
and
20
mgd.
To
allow
interpolation
of
sensor
setpoints
as
a
function
of
dose
delivery,
the
reactor
is
tested
at
each
flowrate
at
setpoint
values
that
the
manufacturer
states
will
result
in
MS2
RED
values
of
10,
20,
and
30
mJ/
cm2.
At
each
setpoint
evaluated,
the
reactor
is
tested
under
conditions
of
reduced
UVT
­
maximum
lamp
output
and
maximum
UVT
­
reduced
lamp
output.
Each
test
condition
is
evaluated
using
five
influent
and
five
effluent
samples.
A
reference
sensor
with
an
uncertainty
of
5
percent
is
used
during
validation
to
measure
the
UV
intensity.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
62
June
2003
Table
C.
11
Validation
Test
Conditions
and
Results
for
Example
C.
5.3
Test
Conditions
Test
Results
Inactivation
Flow
(
mgd)
UVT
(%)
Lamp
(%)
UV
Intensity
(
mW/
cm2)
Influent
(
logs)
Effluent
(
logs)
log
Uncertainty
(%)
RED
(
mJ/
cm2)

5
98
31
5.20
6.00
±
0.074
1.86
±
0.098
4.14
2.8
42.8
5
66
100
5.10
5.98
±
0.136
2.68
±
0.090
3.30
4.7
34.4
5
97.5
20
3.28
6.02
±
0.088
3.23
±
0.039
2.80
3.3
29.4
5
57
100
3.27
6.02
±
0.129
3.87
±
0.060
2.15
6.3
23.0
5
80
20
1.81
5.97
±
0.075
4.55
±
0.176
1.42
12.8
15.8
5
47
100
1.84
5.96
±
0.118
4.84
±
0.110
1.12
13.7
12.8
10
98
55
9.20
6.09
±
0.141
2.31
±
0.114
3.79
4.6
39.3
10
80
100
9.10
5.96
±
0.076
2.66
±
0.121
3.30
4.1
34.4
10
98
33
5.50
6.00
±
0.068
3.74
±
0.070
2.26
4.1
24.1
10
68
100
5.60
6.00
±
0.130
4.04
±
0.086
1.97
7.6
21.2
10
90.5
20
2.62
5.96
±
0.066
4.91
±
0.072
1.05
8.8
12.1
10
53
100
2.63
5.99
±
0.135
4.92
±
0.076
1.07
13.7
12.3
20
98
91
15.1
5.97
±
0.080
2.93
±
0.104
3.04
4.1
31.8
20
95
100
15.2
5.97
±
0.117
2.97
±
0.125
3.00
5.4
31.5
20
98
67
11.2
6.03
±
0.117
3.62
±
0.121
2.41
6.7
25.6
20
86
100
11.2
5.91
±
0.079
3.82
±
0.038
2.09
4.0
22.4
20
98
33
5.50
6.00
±
0.167
4.91
±
0.104
1.09
17.2
12.5
20
68
100
5.60
5.97
±
0.032
4.89
±
0.110
1.08
10.2
12.4
During
validation,
lignin
sulphonate
and
MS2
are
used
as
the
UV
absorber
and
challenge
microorganism,
respectively.
Figure
C.
17
gives
the
dose­
response
of
the
MS2
measured
during
validation
with
a
collimated
beam
apparatus.
The
dose­
response
is
fitted
using
the
following
equation:

70
.
1
N
N
log
91
.
9
Dose
0
+

  

 

  
 
×
=

Confidence
intervals
are
fitted
to
the
data
at
an
80
percent
level.

Table
C.
12
presents
the
MS2
RED
assigned
to
each
reactor
setpoint
based
on
the
validation
results.
For
a
given
flowrate,
Figure
C.
18
presents
the
measured
RED
interpolated
as
a
function
of
measured
UV
intensity.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
63
June
2003
Figure
C.
17
Dose­
Response
of
the
MS2
Challenge
Microorganism
Used
in
Example
C.
5.3
Table
C.
12
Summary
of
Validation
Results
for
Example
C.
5.3
Flow
(
mgd)
UV
Intensity
(
mW/
cm2)
MS2
RED
(
mJ/
cm2)
5
5.10
34.4
5
3.27
23.0
5
1.84
12.8
10
9.10
34.4
10
5.60
21.2
10
2.63
12.3
20
15.2
31.5
20
11.2
22.4
20
5.60
12.4
A
Tier
2
analysis
was
used
to
determine
the
Cryptosporidium
inactivation
credit
that
can
be
assigned
to
the
UV
reactor
given
the
validation
test
results.
Because
the
validation
results
will
be
interpolated
as
a
function
of
dose
delivery,
the
Tier
2
safety
factors
are
determined
as
a
function
of
measured
RED.
For
1.5,
2.0,
2.5,
and
3.0
log
Cryptosporidium
inactivation,
Table
C.
13
presents
the
RED
bias
as
a
function
of
the
measured
RED.
RED
bias
values
were
determined
using
the
approach
cited
in
section
C.
4.10.2.
The
polychromatic
bias
equals
1.0
because
the
UV
reactor
uses
LPHO
lamps.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
64
June
2003
Table
C.
13
RED
Bias
as
a
Function
of
the
Target
Pathogen
Target
Inactivation
and
the
Demonstrated
RED
in
Example
C.
5.3
RED
Bias
for
a
Cryptosporidium
log
Inactivation
of
Demonstrated
RED
(
mJ/
cm2)
Challenge
Microorganism
UV
Sensitivity
(
mJ/
cm2
per
log)
1.5
log
2.0
log
2.5
log
3.0
log
12
11.6
1.98
1.89
1.76
1.65
16
11.3
1.96
1.87
1.74
1.63
20
11.1
1.95
1.86
1.73
1.62
24
11.0
1.94
1.85
1.72
1.61
28
10.8
1.93
1.84
1.71
1.61
32
10.8
1.93
1.84
1.71
1.60
36
10.7
1.92
1.83
1.70
1.60
Figure
C.
18
Measured
RED
as
a
Function
of
Sensor
Setpoint
Values
for
Given
Flowrates
y
=
6.64
+
0.89
y
=
3.45
+
2.83
y
=
1.97
+
1.09
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
0.0
5.0
10.0
15.0
20.0
Setpoint
(
mW/
cm2)
RED
(
mJ/
cm2)

20
mgd
10
mgd
5
mgd
Table
C.
14
presents
the
random
uncertainty
terms
and
the
expanded
uncertainty
as
a
function
of
the
demonstrated
RED.
Using
data
from
Table
C.
11,
Figure
C.
19
presents
the
uncertainty
of
the
log
inactivation
as
a
function
of
the
demonstrated
RED.
An
empirical
fit
to
this
data
was
used
to
obtain
the
uncertainty
of
the
log
inactivation
as
a
function
of
demonstrated
RED
in
Table
C.
14.
The
uncertainty
of
the
RED
due
to
the
dose­
response
data
was
obtained
from
Figure
C.
17.
The
uncertainty
of
the
collimated
beam
dose
calculation
was
8.9
percent.
The
uncertainties
of
the
sensors
used
during
validation
and
at
the
WTP
are
as
follows:

 
Validation
UV
intensity
sensor
5
percent
 
WTP
On­
line
UV
intensity
sensor
15
percent
 
WTP
Reference
UV
intensity
sensor
5
percent
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
CThe
total
uncertainty
of
the
sensors
is
calculated
as
follows:

(
)
%
6
.
16
5
15
5
Error
2
1
2
2
2
=
+
+
=

The
UV
vendor
states
the
standard
deviation
of
the
UV
output
from
lamp
to
lamp
is
25
percent.
Given
four
rows
of
lamps
and
two
sensors
per
row,
the
uncertainty
associated
with
the
number
of
sensors
is
calculated
as
follows:

%
2
.
9
12
1
25
28
.
1
Error
=
×
=

Table
C.
14
Random
Uncertainty
Terms
as
a
Function
of
the
Demonstrated
RED
for
Example
C.
5.3
Uncertainty
(%)
Demonstrated
RED
(
mJ/
cm2)
Challenge
Microorganism
Log
Inactivation
Challenge
Microorganism
Dose­
response
Collimated
Beam
Dose
Calculation
Intensity
and
Flow
Sensors
Number
of
Sensors
Total
Expanded
Uncertainty
12
12.4
11.7
8.9
16.6
9.2
27.0
14
10.5
9.7
8.9
16.6
9.2
25.3
16
9.1
8.2
8.9
16.6
9.2
24.3
18
8.0
7.1
8.9
16.6
9.2
23.5
20
7.1
6.3
8.9
16.6
9.2
23.0
22
6.4
5.6
8.9
16.6
9.2
22.6
24
5.9
5.1
8.9
16.6
9.2
22.3
Figure
C.
19
Uncertainty
of
the
Measured
Log
Inactivation
as
a
Function
of
Demonstrated
RED
y
=
198.01x­
1.10
0
5
10
15
20
0.0
10.0
20.0
30.0
40.0
50.0
Demonstrated
RED
(
mJ/
cm2)
Log
Inactivation
Uncertainty
65
June
2003
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
66
June
2003
Safety
factors
calculated
from
the
RED
bias,
polychromatic
bias,
and
expanded
uncertainty
are
tabulated
in
Table
C.
15.
The
safety
factors
were
multiplied
by
the
dose
requirements
for
Cryptosporidium
to
obtain
target
RED
values
which
are
tabulated
in
Table
C.
16.
For
a
given
demonstrated
RED,
the
UV
reactor
can
achieve
a
given
level
of
Cryptosporidium
credit
if
the
demonstrated
RED
is
greater
than
the
target
RED.
Interpolation
of
these
data
can
be
used
to
identify
the
RED
required
to
obtain
a
given
level
of
Cryptosporidium
inactivation.
Using
this
approach,
an
RED
of
14.0,
19.1,
and
25
mJ/
cm2
is
required
to
show
2.0,
2.5,
and
3.0­
log
Cryptosporidium
inactivation,
respectively.

For
each
flowrate
validated,
interpolation
of
the
data
in
Figure
C.
18
will
provide
the
UV
intensity
setpoints
that
will
indicate
an
RED
of
14.0,
19.1,
and
25
mJ/
cm2.
For
a
given
RED,
Figure
C.
20
presents
those
intensity
setpoints
as
a
function
of
flowrate.
Interpolation
of
the
data
in
Figure
C.
20
can
be
used
to
identify
the
intensity
setpoint
required
at
a
given
flowrate.
For
example,
at
a
flowrate
of
15
mgd,
intensity
setpoints
of
4.7,
6.7,
and
9.0
mW/
cm2
can
be
used
to
indicate
Cryptosporidium
log
inactivation
of
2.0,
2.5,
and
3.0.

Intensity
setpoints
obtained
from
Figure
C.
20
for
a
given
design
flow
can
be
related
to
design
values
of
water
UVT
and
lamp
output
using
an
approach
similar
to
that
used
in
section
C.
5.2
(
see
Figure
C.
16).

Table
C.
15
Safety
Factors
Applicable
to
Validation
Results
Safety
Factors
Needed
Given
a
Cryptosporidium
Inactivation
of
Demonstrated
RED
(
mJ/
cm2)
1.5
log
2.0
log
2.5
log
3.0
log
12
2.54
2.42
2.25
2.11
14
2.48
2.37
2.20
2.06
16
2.44
2.33
2.17
2.03
18
2.42
2.30
2.14
2.01
20
2.40
2.28
2.13
1.99
22
2.38
2.27
2.11
1.98
24
2.37
2.26
2.10
1.97
Table
C.
16
Comparison
of
Demonstrated
RED
and
RED
Required
for
Various
Log
Inactivation
of
Cryptosporidium
for
Example
C.
5.3
RED
Needed
to
Achieve
a
Cryptosporidium
Inactivation
of
Demonstrated
RED
(
mJ/
cm2)
1.5
log
2.0
log
2.5
log
3.0
log
12
9.9
14.0
19.1
25
14
9.7
13.7
18.7
25
16
9.5
13.5
18.4
24
18
9.4
13.3
18.2
24
20
9.4
13.2
18.1
24
22
9.3
13.2
17.9
24
24
9.2
13.1
17.9
24
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
67
June
2003
Figure
C.
20
Intensity
Setpoint
Values
Indicating
Various
Log
Inactivation
of
Cryptosporidium
for
example
C.
5.2
y
=
0.001x2
+
0.517x
+
0.872
y
=
0.005x2
+
0.176x
+
0.968
y
=
0.003x2
+
0.346x
+
0.920
0
2
4
6
8
10
12
14
0
5
10
15
20
25
Flowrate
(
gpm)
Intensity
setpoint
(
mW/
cm2)

2.0
log
Crypto
2.5
log
Crypto
3.0
log
Crypto
C.
5.4
MP
Reactor
Using
a
Single
UV
Intensity
Setpoint
and
UV
Transmittance
Setpoint
A
UV
reactor
consists
of
two
MP
lamps
oriented
parallel
to
the
flow.
Each
lamp
is
monitored
by
a
UV
intensity
sensor.
Lamps
are
spaced
40
cm
apart
and
20
cm
from
the
wall.
The
lamp
sleeve
radius
is
5
cm.
The
sensor
is
located
on
the
wall,
20
cm
away
from
the
lamp.
The
sensor's
spectral
response
matches
that
of
"
Sensor
A"
in
Figure
C.
10.
Dose
delivery
is
indicated
using
the
UV
intensity
and
UVT
setpoint
approach.

The
reactor
is
rated
for
a
flow
from
0.1
to
0.5
mgd.
The
reactor
will
be
used
with
a
design
UVT
of
85
percent.
The
manufacturer
states
that
the
lamp
output
at
the
end­
of­
lamp
life
will
be
78
percent
of
the
100
hr
burn­
in
value.
The
fouling
factor
for
the
reactor
is
90
percent.
Accordingly,
the
lamp
output
factor
for
the
reactor
is
0.78
×
0.90
=
0.70.
A
UV
intensity
setpoint
of
2.8
mW/
cm2
is
obtained
by
measuring
the
UV
intensity
with
the
water
UVT
set
to
85
percent
and
the
lamp
output
lowered
to
70
percent.

During
operation
at
a
WTP,
the
on­
line
and
reference
UV
intensity
sensors
will
have
a
measurement
uncertainty
of
10
and
5
percent,
respectively.

The
UV
reactor
is
validated
using
lignin
sulphonate
as
a
UV
absorber
and
MS2
as
a
challenge
microorganism.
The
measured
dose­
response
of
the
challenge
microorganism
is
provided
in
Figure
C.
14.
Table
C.
17
gives
the
validation
test
conditions
and
results.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
68
June
2003
Table
C.
17
Validation
Test
Conditions
and
Results
for
Example
C.
5.4
with
the
Sensor
Located
20
cm
from
the
Lamp
Test
Conditions
Test
results
Inactivation
Flow
(
mgd)
UVT
(%)
Lamp
(%)
UV
Intensity
(
mW/
cm2)
Influent
(
Logs)
Effluent
(
logs)
Log
Uncertainty
(%)
RED
(
mJ/
cm2)

0.1
85
70
2.9
6.00
±
0.07
0.00
>
6.0
0
­
>
82.6
0.5
85
70
2.9
6.06
±
0.08
3.04
±
0.16
3.03
5.6
47.1
0.5
82.7
100
2.8
5.99
±
0.11
2.13
±
0.14
3.85
4.3
60.0
0.5
93.5
20
2.9
6.12
±
0.11
4.52
±
0.07
1.60
7.6
24.8
The
first
two
test
conditions
evaluate
dose
delivery
at
minimum
and
maximum
flow
with
the
reactor
operating
at
the
intensity
and
UVT
setpoint
values.
Based
on
these
results,
the
UV
reactor
is
rated
at
an
MS2
RED
of
47.1
mJ/
cm2
when
operating
at
the
setpoint
conditions.

The
last
two
test
conditions
evaluate
the
sensor
position
and
the
validity
of
using
the
UV
intensity
and
UVT
setpoint
approach
for
indicating
dose
delivery.
As
indicated,
the
UV
reactor
delivers
an
MS2
RED
of
60.0
mJ/
cm2
when
operating
with
peak
lamp
output
and
the
UVT
lowered
to
82.7
percent
to
give
a
measured
intensity
at
the
setpoint.
The
UV
reactor
delivers
a
dose
of
24.8
mJ/
cm2
when
operating
at
high
UVT
and
lowered
lamp
output
to
give
a
measured
intensity
at
the
setpoint
value.
In
other
words,
an
intensity
setpoint
of
2.9
mW/
cm2
and
a
UVT
setpoint
of
85
percent
does
not
ensure
the
reactor
delivers
an
MS2
RED
of
47.1
mJ/
cm2.

The
manufacturer
has
three
options
for
resolving
this
problem:

 
Relocate
the
sensor
closer
to
the
lamp.

 
Switch
from
the
dose
monitoring
method
to
the
calculated
dose
approach.

 
Switch
from
the
dose
monitoring
method
to
the
intensity
setpoint
approach
either
with
the
sensor
in
its
current
location
or
with
the
sensor
in
a
more
optimized
location.

In
this
example,
the
manufacturer
chooses
to
relocate
the
sensor
to
8
cm
from
the
lamp
and
revalidates
the
UV
reactor.
Table
C.
18
gives
the
test
conditions
and
results.
In
this
case,
a
UV
intensity
of
41.0
mW/
cm2
is
measured
with
the
UV
reactor
operating
with
a
UVT
of
85
percent
and
a
lamp
output
of
70
percent.
This
value
is
greater
than
the
UV
intensity
measured
with
the
sensor
on
the
wall
because
the
sensor
is
located
closer
to
the
lamp.
Based
on
the
results,
the
UV
reactor
is
rated
at
an
MS2
RED
of
48.5
mJ/
cm2
when
operating
at
setpoint
conditions
of
85
percent
UVT
and
a
41.0
mW/
cm2
UV
intensity
value.
With
the
measured
intensity
at
the
intensity
setpoint
value,
the
UV
reactor
delivers
an
RED
greater
than
48.5
mJ/
cm2
when
operating
with
a
UVT
greater
than
the
UVT
setpoint
value
and
an
RED
less
than
48.5
mJ/
cm2
when
operating
with
a
UVT
less
than
the
setpoint
value.
Thus,
the
intensity
sensor
is
properly
located
for
using
the
intensity
and
UVT
setpoint
approach
for
indicating
dose
delivery
and
the
setpoints
will
ensure
the
dose
delivery
meets
an
RED
of
48.5
mJ/
cm2.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
69
June
2003
Table
C.
18
Validation
Test
Conditions
and
Results
for
Example
C.
5.4
with
the
Sensor
Located
12
cm
from
the
Lamp
Test
Conditions
Test
Results
Inactivation
Flow
(
mgd)
UVT
(%)
Lamp
(%)
UV
Intensity
(
mW/
cm2)
Influent
(
Logs)
Effluent
(
Logs)
Log
Uncertainty
(%)
RED
(
mJ/
cm2)

0.1
85
70
41.0
6.00
±
0.07
0.00
>
6.00
­
>
82.6
0.5
85
70
41.0
6.01
±
0.10
2.90
±
0.10
3.12
4.3
48.5
0.5
75
100
41.4
6.03
±
0.07
3.26
±
0.18
2.77
6.7
43.0
0.5
98
46
41.1
5.96
±
0.12
1.59
±
0.12
4.37
3.7
68.1
A
Tier
2
analysis
is
used
to
determine
the
Cryptosporidium
inactivation
credit
that
can
be
assigned
to
the
UV
reactor
given
the
validation
test
results.

RED
Bias.
Since
a
3.0
log
inactivation
of
Cryptosporidium
requires
a
dose
of
12
mJ/
cm2,
the
UV
sensitivity
of
Cryptosporidium
is
defined
as
12/
3.0
=
4
mJ/
cm2
per
log
inactivation.
The
UV
sensitivity
of
MS2
is
48.5/
3.12
=
15.5
mJ/
cm2.
In
Figure
C.
6,
a
RED
of
9.78
and
18.0
mJ/
cm2
is
associated
with
UV
sensitivities
of
4
and
15.5
mJ/
cm2
per
log
inactivation,
respectively.
Accordingly,
the
RED
bias
is
18.0/
9.78
=
1.84.

Polychromatic
Bias.
The
sensor­
to­
lamp
water
layer
is
3
cm.
For
a
sensor
with
the
response
of
"
Sensor
A"
in
Figure
C.
10,
a
polychromatic
bias
1.00
is
obtained
from
Figure
C.
7
for
a
UVT
of
85
percent.

Expanded
uncertainty.
The
uncertainty
of
the
log
inactivation
through
the
UV
reactor
calculated
using
Equation
C.
7
is
4.3
percent.
The
uncertainty
of
the
collimated
beam
dose
calculation
was
8.9
percent.
The
uncertainty
in
the
RED
arising
from
the
scatter
in
the
doseresponse
obtained
from
Figure
C.
14
is
3.9
percent
at
an
RED
of
48.5
mJ/
cm2.
The
uncertainties
of
the
sensors
used
during
validation
and
at
the
WTP
are
as
follows:

 
Validation
UV
intensity
sensor
5
percent
 
WTP
On­
line
UV
intensity
sensor
10
percent
 
WTP
Reference
UV
intensity
sensor
5
percent
The
total
uncertainty
of
the
sensors
is
calculated
as
follows:

(
)
%
2
.
12
5
10
5
Error
2
1
2
2
2
=
+
+
=

Because
each
lamp
is
monitored,
the
uncertainty
knowing
the
output
of
the
lamps
is
0
percent.
Including
each
of
these
random
uncertainty
terms,
the
expanded
uncertainty
is
calculated
as
follows:

(
)
%
2
.
16
0
2
.
12
9
.
3
9
.
8
3
.
4
Error
2
1
2
2
2
2
2
=
+
+
+
+
=

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
70
June
2003
Safety
factor.
Using
Equation
C.
10,
the
safety
factor
is
calculated
as
follows:

(
)
14
.
2
00
.
1
84
.
1
162
.
0
1
SF
=
×
×
+
=

Cryptosporidium
Credit.
Using
this
safety
factor,
the
target
RED
that
should
be
demonstrated
during
validation
is
12
×
2.14
=
26
mJ/
cm2.
Because
the
demonstrated
RED
is
greater
than
this
number,
the
UV
reactor
operating
at
or
above
the
validated
intensity
and
UVT
setpoints
can
get
credit
for
3.0­
log
Cryptosporidium
inactivation.

C.
5.5
MP
Reactor
Using
Calculated
Dose
Monitoring
A
UV
reactor
consists
of
twelve
MP
lamps
oriented
perpendicular
to
the
flow.
Each
lamp
is
monitored
by
a
UV
intensity
sensor
whose
spectral
response
matches
that
of
"
Sensor
A"
in
Figure
C.
10.
The
UV
intensity
sensors
view
the
UV
lamps
through
a
15
cm
water
layer.
During
operation
at
a
WTP,
the
on­
line
and
reference
UV
intensity
sensors
will
have
a
measurement
uncertainty
of
10
and
5
percent
respectively.
During
validation,
a
reference
sensor
is
used
to
measure
UV
intensity.

The
UV
reactor
is
validated
at
flows
ranging
from
10
to
40
mgd
using
lignin
sulphonate
as
a
UV
absorber
and
MS2
as
a
challenge
microorganism.
Figure
C.
3
gives
the
dose­
response
of
the
challenge
microorganism.
The
lamp's
power
supplies
vary
lamp
power
from
30
to
100
percent.
The
UV
reactor
is
validated
at
flowrate,
UVT,
and
lamp
power
combinations
that
give
a
calculated
UV
doses
of
30,
20
and
10
mJ/
cm2.
Table
C.
19
gives
the
validation
test
conditions
and
results.

Table
C.
19
Validation
Test
Conditions
and
Results
for
Example
C.
5.5
Test
Conditions
Test
Results
Inactivation
Flow
(
mgd)
UVT
(%)
Lamp
(%)
Calculated
RED
(
mJ/
cm2)
Intensity
(
mW/
m2)
Log
Uncertainty
(%)
Measured
RED
(
mJ/
cm2)

40
98
40.5
30.1
22.7
2.38
3.0
35.2
40
90
68
30.2
11.8
2.43
3.2
36.3
40
82.8
100
30.2
5.8
2.38
4.2
35.1
20
90
33.8
30.0
5.9
2.57
3.0
39.1
20
85
44.5
30.1
3.6
2.36
3.6
34.7
20
75
70
29.7
1.2
2.31
3.5
33.8
10
79
30
30.3
1.0
2.22
4.8
32.0
10
75
35.5
30.1
0.6
2.76
2.9
43.3
40
96.5
30
20.1
13.5
1.42
4.0
17.7
40
85
59.5
20.1
4.9
1.34
6.2
16.3
40
73.5
100
19.9
1.3
1.31
7.3
15.9
20
85
30
20.3
2.5
1.75
4.0
23.1
20
75
47
19.9
0.8
1.72
2.9
22.7
40
84.5
30
9.9
2.3
0.96
6.5
10.8
40
75
47
10.0
0.8
1.24
8.1
14.9
40
80
38
10.1
1.4
0.93
7.8
10.4
20
70
30
10.4
0.2
0.86
10.4
9.5
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
71
June
2003
Figure
C.
21
provides
a
plot
of
the
measured
RED
as
a
function
of
the
calculated
RED.
As
shown,
there
is
a
range
of
measured
RED
values
associated
with
a
given
calculated
RED.
The
UV
reactor
is
rated
at
the
lower
end
of
that
range
for
a
given
calculated
dose.
A
power
function
(
y=
AxB)
is
used
to
define
the
relationship
between
the
calculated
RED
and
the
lower
bound
of
the
measured
RED.

Figure
C.
21
Relationship
Between
Measured
and
Calculated
Dose
for
the
MP
Reactor
in
Example
C.
5.5
A
Tier
2
analysis
was
used
to
determine
the
calculated
RED
values
required
for
3.0­
log
credit
for
Cryptosporidium.
For
various
log
inactivation
credit
values
for
Cryptosporidium,
Table
C.
20
gives
the
RED
bias
as
a
function
of
the
MS2
RED
predicted
from
the
calculated
dose
using
the
power
function
in
Figure
21.

Table
C.
20
RED
Bias
as
a
Function
of
the
Target
Pathogen
Target
Inactivation
and
the
Calculated
Dose
in
Example
C.
5.5
RED
Bias
for
Cryptosporidium
log
inactivations
of
Calculated
Dose
(
mJ/
cm2)
MS2
RED
(
mJ/
cm2)
MS2
Log
Inactivation
MS2
Sensitivity
(
mJ/
cm2
per
log
inactivation)
3.0
log
2.5
log
2.0
log
1.5
log
1.0
log
10
4.8
0.43
11.2
1.61
1.73
1.86
1.95
1.99
12
6.6
0.58
11.3
1.62
1.74
1.87
1.97
2.00
14
8.6
0.75
11.5
1.63
1.76
1.89
1.98
2.01
16
10.8
0.92
11.7
1.64
1.77
1.90
1.99
2.03
18
13.3
1.11
11.9
1.65
1.78
1.92
2.01
2.04
20
15.9
1.30
12.2
1.67
1.80
1.93
2.03
2.06
22
18.7
1.50
12.5
1.68
1.82
1.95
2.05
2.08
24
21.8
1.69
12.9
1.70
1.84
1.97
2.07
2.11
26
25.0
1.88
13.3
1.72
1.86
2.00
2.10
2.13
28
28.4
2.07
13.7
1.74
1.88
2.02
2.12
2.16
30
32.0
2.25
14.2
1.77
1.91
2.05
2.15
2.19
Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
72
June
2003
Table
C.
21
gives
the
random
uncertainty
terms
and
the
resulting
total
expanded
uncertainty
as
a
function
of
the
MS2
RED
and
calculated
dose.
Using
data
from
Table
C.
19,
Figure
C.
22
presents
the
uncertainty
of
the
log
inactivation
as
a
function
of
measured
MS2.
An
empirical
fit
to
this
data
was
used
to
predict
the
uncertainty
of
the
log
inactivation
as
a
function
of
MS2
RED
in
Table
C.
21.
The
uncertainty
of
the
RED
due
to
the
dose­
response
data
was
obtained
from
Figure
C.
3.
The
uncertainty
of
the
collimated
beam
dose
calculation
was
8.0
percent.
The
uncertainties
of
the
sensors
used
during
validation
and
at
the
WTP
are
as
follows:

 
Validation
UV
intensity
sensor
5
percent
 
WTP
On­
line
UV
intensity
sensor
10
percent
 
WTP
Reference
UV
intensity
sensor
5
percent
The
total
uncertainty
of
the
sensors
is
calculated
as
follows:

(
)
%
2
.
12
5
10
5
Error
2
1
2
2
2
=
+
+
=

Because
each
lamp
is
monitored
by
a
UV
intensity
sensor,
the
uncertainty
associated
with
quantifying
lamp
output
is
zero.

Table
C.
21
Random
Uncertainty
Terms
as
a
Function
of
the
Calculated
Dose
for
Example
C.
5.5
Uncertainty
(%)
Calculated
Dose
(
mJ/
cm2)
MS2
RED
(
mJ/
cm2)
Challenge
Microorganism
Log
Inactivation
Challenge
Microorganism
Dose­
response
Collimated
Beam
Dose
Calc
Intensity
Sensors
Total
Expanded
Uncertainty
10
4.8
13.9
1.9
8.0
12.2
20.3
12
6.6
11.1
1.3
8.0
12.2
18.4
14
8.6
9.2
0.9
8.0
12.2
17.3
16
10.8
7.8
0.7
8.0
12.2
16.6
18
13.3
6.8
0.5
8.0
12.2
16.1
20
15.9
5.9
0.4
8.0
12.2
15.8
22
18.7
5.3
0.3
8.0
12.2
15.6
24
21.8
4.8
0.3
8.0
12.2
15.4
26
25.0
4.3
0.2
8.0
12.2
15.3
28
28.4
3.9
0.2
8.0
12.2
15.1
30
32.0
3.6
0.2
8.0
12.2
15.1
Table
C.
22
gives
the
polychromatic
bias
as
a
function
of
the
UVT.
The
polychromatic
bias
values
were
taken
from
Figure
C.
7
for
sensor
"
A"
located
with
a
15
cm
water
layer
and
lignin
sulphonate
as
the
UV­
absorbing
chemical.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
73
June
2003
Figure
C.
22
Uncertainty
of
the
Measured
Log
Inactivation
as
a
Function
of
Demonstrated
RED
Table
C.
22
Polychromatic
Bias
for
Example
C.
5.5
UVT
(%)
Polychromatic
Bias
98
1.03
95
1.06
90
1.10
85
1.20
80
1.30
75
1.55
Safety
factors
calculated
from
the
RED
bias,
polychromatic
bias,
and
expanded
uncertainty
are
tabulated
in
Table
C.
23.
The
safety
factors
were
multiplied
by
the
3.0­
log
dose
requirement
for
Cryptosporidium
of
12
mJ/
cm2
to
obtain
target
RED
values
which
are
tabulated
in
Table
C.
24.
For
a
given
calculated
dose,
the
UV
reactor
can
achieve
receive
3.0­
log
Cryptosporidium
credit
if
the
measured
MS2
RED
associated
with
that
calculated
dose
is
greater
than
the
target
RED.

Table
C.
25
presents
the
calculated
dose
needed
to
achieve
a
given
level
of
Cryptosporidium
inactivation
and
the
lower
limit
of
UVT
over
which
the
calculated
dose
applies.
The
values
in
Table
C.
22
for
3.0­
log
inactivation
credit
were
obtained
from
Table
C.
21.
The
values
for
other
log
inactivation
credit
levels
were
obtained
by
repeating
the
analysis
in
Tables
C.
23
and
C.
24.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
74
June
2003
Table
C.
23
Safety
Factors
for
3­
Log
Cryptosporidium
Inactivation
as
a
Function
of
the
Calculated
Dose
for
Example
C.
5.5
Safety
Factors
for
3.0­
log
Cryptosporidium
Inactivation
for
UVT
of
Calculated
Dose
(
mJ/
cm2)
MS2
RED
(
mJ/
cm2)
98%
95%
90%
85%
80%
75%
10
4.8
1.99
2.05
2.13
2.32
2.51
3.00
12
6.6
1.97
2.03
2.10
2.30
2.49
2.97
14
8.6
1.97
2.02
2.10
2.29
2.48
2.96
16
10.8
1.97
2.03
2.10
2.29
2.48
2.96
18
13.3
1.98
2.03
2.11
2.30
2.49
2.97
20
15.9
1.99
2.05
2.12
2.32
2.51
2.99
22
18.7
2.00
2.06
2.14
2.34
2.53
3.02
24
21.8
2.02
2.08
2.16
2.36
2.55
3.04
26
25.0
2.04
2.10
2.18
2.38
2.58
3.08
28
28.4
2.07
2.13
2.21
2.41
2.61
3.11
30
32.0
2.09
2.16
2.24
2.44
2.64
3.15
Table
C.
24
Comparison
of
the
Target
MS2
RED
Needed
to
Demonstrate
3.0­
Log
Cryptosporidium
Inactivation
Credit
to
the
Calculated
Dose
and
Measured
MS2
RED
Target
MS2
RED
(
mJ/
cm2)
Calculated
Dose
(
mJ/
cm2)
MS2
RED
(
mJ/
cm2)
98%
95%
90%
85%
80%
75%
10
4.8
23.9
24.6
25.5
27.8
30.1
35.9
12
6.6
23.7
24.3
25.3
27.6
29.9
35.6
14
8.6
23.6
24.3
25.2
27.5
29.8
35.5
16
10.8
23.6
24.3
25.2
27.5
29.8
35.5
18
13.3
23.7
24.4
25.3
27.6
29.9
35.7
20
15.9
23.9
24.6
25.5
27.8
30.1
35.9
22
18.7
24.1
24.8
25.7
28.0
30.4
36.2
24
21.8
24.3
25.0
25.9
28.3
30.6
36.5
26
25.0
24.5
25.2
26.2
28.6
31.0
36.9
28
28.4
24.8
25.5
26.5
28.9
31.3
37.3
30
32.0
25.1
25.9
26.8
29.3
31.7
37.8
Table
C.
25
Dose
and
UVT
Alarm
Setpoints
for
Various
Log
Inactivation
Credit
Levels
of
Cryptosporidium
3.0
log
2.5
log
2.0
log
1.5
log
1.0
log
Dose
(
mJ/
cm2)
UVT
(%)
Dose
(
mJ/
cm2)
UVT
(%)
Dose
(
mJ/
cm2)
UVT
(%)
Dose
(
mJ/
cm2)
UVT
(%)
Dose
(
mJ/
cm2)
UVT
(%)
25
98.0
22
94.7
19
88.8
15
93.0
12
89.6
26
95.6
23
88.5
20
83.2
16
84.7
13
80.9
27
90.4
24
83.9
21
79.4
17
79.5
14
76.3
28
85.9
25
80.5
22
76.8
18
76.3
15
77.5
29
82.2
26
78.1
23
75.1
­
­
­
­
30
79.2
27
76.3
­
­
­
­
­
­
­
­
28
75.0
­
­
­
­
­
­

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
75
June
2003
The
data
in
Table
C.
25
represents
calculated
dose
and
UVT
alarm
setpoints
that
can
be
used
to
ensure
the
reactor
delivers
a
given
log
inactivation
of
Cryptosporidium.
Alternatively,
as
shown
in
Table
C.
26,
a
single
calculated
dose
alarm
setpoint
can
be
defined
over
the
validated
range
of
UVT.

Table
C.
26
Dose
Setpoints
for
Various
Log
Inactivation
Credit
Levels
of
Cryptosporidium
Cryptosporidium
Log
Inactivation
Credit
Calculated
Dose
Setpoint
(
mJ/
cm2)
UVT
Range
(%)

1.0
14
75
­
98
1.5
18
75
­
98
2.0
23
75
­
98
2.5
28
75
­
98
3.0
30
79
­
98
C.
6
References
APHA/
AWWA/
WEF.
1995.
Standard
methods
for
the
Examination
of
Water
and
Wastewater,
19th
Edition.
American
Public
Health
Association,
American
Water
Works
Association,
and
Water
Environment
Federation,
Washington,
DC.

Chang,
J.
C.
H.,
S.
F.
Osoff,
D.
C.
Lobe,
M.
H.
Dorfman,
C.
M.
Dumais,
R.
G.
Qualls
and
J.
D.
Johnson.
1985.
UV
inactivation
of
pathogenic
and
indicator
microorganisms.
Applied
and
Environmental
Microbiology
49,
no.
6:
1361­
1365.

Draper,
N.
and
Smith,
H.
1981.
Applied
regression
analysis,
Second
Edition.
New
York:
Wiley.

DVGW.
1997.
UV
Disinfection
Devices
for
Drinking
Water
Supply
 
 
Requirements
and
Testing.
German
Gas
and
Water
Management
Union
(
DVGW),
Bonn,
Germany.

NWRI/
AwwaRF.
2000.
Ultraviolet
Disinfection
Guidelines
for
Drinking
Water
and
Water
Reuse.
National
Water
Research
Institute
and
the
AwwaRF
ÖNORM.
2001.
ÖNORM
M
5873­
1
Plants
for
the
Disinfection
of
water
Using
Ultraviolet
Radiation:
Requirements
and
Testing
Low
Pressure
Mercury
Lamp
Plants.
Österreichisches
Normungsinstitut,
Vienna,
Austria.

Petri,
B.
M.,
G.
Fang,
J.
P.
Malley,
D.
C.
Moran,
and
H.
Wright.
2000.
Ground
water
UV
Disinfection:
Challenges
and
Solutions.
Proceedings
of
the
American
Water
Works
Association
Water
Quality
Technology
Conference.
Salt
Lake
City,
UT,
November
5­
9,
2000.

Proposal
Draft
Appendix
C.
Validation
of
UV
Reactors
UV
Disinfection
Guidance
Manual
C­
76
June
2003
Petri,
B.
and
D.
Olson.
2001.
Bioassay­
validated
numerical
models
for
UV
reactor
design
and
scale­
up
IUVA
First
International
Congress
on
UV
Technologies,
Washington,
DC,
June
14­
16,
2001.

Wright,
N.
G.
and
D.
M.
Hargreaves.
2001.
The
use
of
CFD
in
the
evaluation
of
UV
treatment
systems
Journal
of
Hydroinformatics
3:
59­
70.

Proposal
Draft
