See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/255042167

In-place HEPA filter penetration test

Article

CITATION READS

1 37

5 authors, including:

Werner Bergman
Aerosol Science ,LLC
50 PUBLICATIONS   182 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Electric Filters View project

All content following this page was uploaded by Werner Bergman on 15 July 2019.

The user has requested enhancement of the downloaded file.

 IN-PLACE HEPA FILTER PENETRATION TEST*

W. Bergman, K. Wilson, J. Elliott, B. Bettencourt and J. W. Slawskil

Lawrence Livermore National Laboratory
Livermore, CA 94550

Abstract

We have demonstrated the feasibility of conducting penetration tests on high

efficiency particulate air (HEPA) filters as installed in nuclear ventilation systems.
The in-place penetration test, which is designed to yield equivalent penetration
measurements as the standard DOP efficiency test, is based on measuring the
aerosol penetration of the filter installation as a function of particle size using a
portable laser particle counter.“’ This in-place penetration test is compared to the
current in-place leak test using light scattering photometers for single HEPA filter
installations and for HEPA filter plenums using the shroud method. Test results
show the in-place penetration test is more sensitive than the in-place leak test, has
a similar operating procedure, but takes longer to conduct. Additional tests are
required to confirm that the in-place penetration test yields identical results as the
standard dioctyl phthalate (DOP) penetration test for HEPA filters with controlled
leaks in the filter and gasket and duct by-pass leaks. Further development of the
procedure is also required to reduce the test time before the in-place penetration
test is practical.

I. Introduction

Before a HEPA filtration system can be used in a DOE nuclear facility., the
ventilation system and the HEPA filters must pass acceptance tests described in
ASME N5 10 or AGl, and the HEPA filter must pass the MIL-STD-282 penetration test.
(I-3) The acceptance tests consist of leak tests of ducts and housings, airflow
capacity and distribution tests, and air-aerosol mixing uniformity tests. The
airflow distribution test is designed to insure that HEPA filters see a uniform air
flow, while the air-aerosol mixing test is performed to insure that the
concentration of aerosols challenging the filter is uniform. This will insure that
representative samples can be obtained before and after the filter for computing
the filter penetration.

-------_____________-------------------------------------------------
1 U.S. Department of Energy, Defense Programs(DP-45), Germantown, MD 20874

*This work was performed under the auspices of the U.S. Department of Energy by
Lawrence Livermore National Laboratory under contract no. W-7405ENG.45. The work
was supported by DOE’s Defense Program Office of Technical and Environmental Support,
DP-45.

649
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

The HEPA filter penetration test is given in MIL-STD-282.“’ This test requires
HEPA filters to have less than 0.03% penetration for 0.3 pm DOP aerosols as
measured by a light scattering photometer. The 0.3 pm aerosols were originally
selected because they were believed to be the most penetrating aerosols and would
yield the most conservative penetration values for the HEPA filters. These aerosols
were generated in a very large machine by a controlled condensation of DOP vapor
and were thought to be monodisperse.

After the HEPA filter is installed in a certified ductwork, and once a year
thereafter, the filter installation must be tested for leaks.(2-4) This in-place leak
test is performed to insure that the HEPA filter is properly installed and has not
been damaged, that there are no leaks in the mounting frame or between the
mounting frame and the housing, and that the system contains no bypassing that
would reduce the system penetration. The in-place leak test is not a filter
penetration test and can not be used in determining the penetration of HEPA
filters. The difference between the two tests is the particle size and the type of
aerosol generator used to challenge the filter: the DOP penetration test uses near
monodisperse 0.3 urn particles generated by a very large vapor condensation
generator, while the in-place test uses heterodisperse 0.7 pm particles generated by
small portable air or thermal generators.(2‘4’ ERDA 76-21 recommends an
acceptance criterion of 0.03% maximum penetration for the in-place DOP test.(4)

The HEPA filter leak test was implemented in 1960 in the U.S. to verify that the
installed filtration systems did not have leaks.(5) This test represented a second-
best choice at that time since it was not possible to conduct in-place penetration
tests using the available test equipment. The problem was that the particle
measuring instruments at that time could not distinguish between particle sizes,
and monodisperse 0.3 pm aerosol generators were not portable. The available light
scattering photometers were portable but could not distinguish between different
particle sizes. To measure HEPA filter penetration at 0.3 grn diameter, it was
necessary to have a monodisperse 0.3 urn diameter generator, which were not
portable. The only portable aerosol generators at that time produced
heterodisperse aerosols.

Now, a variety of instruments and aerosol generators are commercially

available that can be used for measuring in-place filter penetration. Portable
particle spectrometers are available that can measure specific particle sizes in
heterodisperse aerosols. Portable aerosol generators are also available that can
generate monodisperse aerosols. Thus it is now possible to measure in-place HEPA
filter penetration at 0.3 pm using portable equipment consisting of either a
particle size spectrometer and a heterodisperse aerosol generator or an integrated
particle analyzer (e.g. photometer, condenstation nuclei counter) and a
monodisperse aerosol generator. We will only address the in-place penetration
method using laser spectrometers and heterodisperse aerosols in this paper.

650
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

II. Difference Between Penetration. In-Place Leak. and In-Place Penetration Tests

The difference between the results of the penetration and the in-place leak tests
can be illustrated with a typical HEPA filter penetration curve shown as a function
of particle size in Figure 1. The penetration is a maximum at O.l5um, decreases
rapidly with increasing particle size and is negligible at 0.7 pm for HEPA filters with
no leaks. Although the penetration measurement at 0.3 pm is significantly less
than the maximum, it still provides a sensitive measurement of the filter
penetration. In contrast to the in-tact IIEPA filter installation in Figure 1, particle
penetration through leaks is independent of particle size. Thus any penetration
that is measured at 0.7 pm diameter during the in-place leak test can be
attributed to leaks.

0.1 1
Diameter, urn
Figure 1. Plot of HEPA filter penetration measurements as a function of
particle size for dioctyl sebacate (DOS) aerosols with two different laser
spectrometers. Nuclear grade, 1,000 cfm HEPA filter.

651
 Two different laser particle counters (Particle Measurement Systems, Bolder, CO)
were used to generate the curve in Figure 1: the LAS-HS laser counter, which
measures particles from 0.067 to 0.95 urn diameter and the LASAIR laser counter,
which measures particles from 0.14 to 2.4 urn diameter. The diameter
measurements are based on the logarithm midpoint of each of the counter
channels. A 100: 1 diluter (TSI, Minneapolis, MN) was used to dilute the upstream
measurements to avoid coincidence counting. The dioctyl sebacate aerosols were
generated with a Laskin nozzle aerosol generator (Virtis, Gardiner, NY). Details of
the test procedure are described in previous reports.@‘) The agreement between
the two instruments is good.

It is possible to conduct filter penetration tests as described in ASME N-510 an d

ASME AG-1 using a laser particle counter during in-place filter tests.(2’3) If the laser
counter is used for measuring the total number of particles without regard to
particle size, then the filter test becomes another leak test. However, if the laser
counter is used to discriminate between different particle sizes, such as 0.3 urn,
then the laser test becomes an in-place penetration test. Using the laser particle
counter also allows the maximum filter penetration, as shown in Figure 1, to be
determined with the in-place penetration test. A description of the filter efficiency
test using the laser particle counter is given by Bergman and Biermann and by
Scripsick et a1.‘6-8’

The in-place penetration test using the laser particle counter is a measurement
of the penetration of the total filtration system. This test incorporates the aerosol
penetration from both the HEPA filter and leaks in the filter housing or gaskets. In
separate filter penetration and leak tests, the total penetration of the filtration
system is determined from the sum of the filter penetration and the leak
penetration. In separate penetration and leak tests, once the filter is installed, it is
only possible to determine system leaks with the light scattering photometer and
assume the filter penetration remains the same. The in-place leak test using the
light scattering photometer can only detect a major deterioration in filter
penetration.

The increased sensitivity of the laser particle counter allows filter penetration
measurements of two stages of HEPA filters for both the leak test and the
penetration test. This capability, which is not possible for the standard
photometer based leak test, is advantageous because of the reduced testing time
and the difficulty in measuring the penetration of individual stages in systems
having minimal space between stages. Schuster and Osetck were the first to use a
laser particle counter to measure the filter penetration of one-stage and two-stage,
size 1 HEPA filters.“’ They found typical DOP penetrations of 0.003% for single
stage and 0.000005% for two stage HEPA filters. However measurements of
penetration versus particle size were only reported for the single stage HEPA
filters.“’

652
 Ortiz determined the filter leaks in a number of 20,000 cfm two-stage HEPA
filter systems. (lo) He did not discriminate between particle size, but rather used
the total particle count before and after the filters to determine the system leaks.
The test was therefore a leak test and not a penetration test. The leak
measurements for ten systems varied from 0.0067% to 0.00000009%. The
maximum allowable leakage for two stage HEPA filters is 0.000009%. This study
was significant not only because the test system was demonstrated under field
conditions, but also because it showed the laser particle counter detected filter
system failures that were not seen with the standard single stage method described
in ASME N5 10.“’

Ortiz et al also conducted a round robin test of two-stage HEPA filtration system
in which they measured filter penetration as a function of particle size using a
laser spectrometer.‘“’ In this configuration, the filter test was an in-place
penetration test. To avoid coincidence counting, the upstream concentration was
diluted. The test apparatus and procedure were incorporated into an ASTM test
method for evaluating HEPA filters.‘r2’

The Los Alamos National Laboratory (LANL) uses a laser spectrometer and
heterodisperse aerosols as developed by Ortiz and incorporated in the ASTM
standard for conducting in-place HEPA filter leak tests in all of their facilities.(10-13’
Since the particle measurements are made by adding all of the sizes into a single
count, the LANL in-place filter measurements can not be used for determining filter
penetration, but rather for leaks. Adding together the particle counts in the
different particle size bins destroys the ability to measure filter penetration with
heterodisperse aerosols. However, by keeping the particle counts in the different
size bins separate, the LANL test procedure for leaks can be converted to a test of
filter penetration test.

III. Correlation of In-Place Penetration Test With Standard Penetration Test

In order to claim that an in-place filter penetration test is equivalent to the

standard HEPA filter penetration test at 0.3 urn, it is necessary to establish a
correlation between the in-place penetration test with the standard penetration
test specified in MIL-STD-282.“’ Such a correlation would include penetration
measurements on HEPA filters with varying defects in the filter and the gasket as
well as by-pass leaks in the ventilation ducting. These correlation tests have not
yet been completed. However Scripsick et al conducted tests on 849 new HEPA
filters using laser measurements at 0.31 urn and the standard Q-107 measurements
at 0.3 um.‘8’ The correlation between the laser measurements at 0.31 urn and the
Q-107 measurements at 0.3 urn is good, as shown in Figure 2. (8) Note that the,
correlation becomes worse at smaller penetration values. This is not surprising
considering the photometer in the Q-107 measurements is increasingly noisy below
0.01% penetration. We plan to conduct similar correlations using filters with
controlled leaks in the media and gaskets and using controlled by-pass leaks in the
ducting.

653
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

SIZE
4A
5 0 :
5A v i

6 0 i

d
0.001 U.Ul lJ.1

Q107 Penetration at 0.30 pm, 96

Figure 2. Correlation of HEPA filter penetration between laser spectrometer an d

Q107 photometer.‘8)

For measurements of the maximum filter penetration, it is not necessary to

conduct correlation tests with the Q107 tester because it only measures the
penetration at 0.3 urn. The Q107 can not be used to determine the maximum
filter penetration at 0.15 urn, as seen in Figure 1. In fact, there are no standard
reference tests for the maximum filter penetration. The laser spectrometer can be
used in a primary test standard for the maximum filter penetration if the particle
size range is sufficient to clearly show a maximum as seen in Figure 1.

IV. Correlation of In-Place Penetration Test With Standard Leak Test

We have conducted a series of filter penetration tests on a HEPA filter with a n

increasing number of pin holes to establish a correlation between the in-place
penetration test and the standard leak test. A nuclear grade, 1,000 cfm HEPA filter
was used in these correlation tests. Two different laser spectrometers were used to

654
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

determine the in-place filter penetration as a function of size: the LAS-HS laser
counter, which measures particles from 0.067 to 0.95 nm diameter and the IASAIR
laser counter, which measures particles from 0.14 to 2.4 pm diameter. A 100: 1
diluter (TSI, Minneapolis, MN) was used to dilute the upstream measurements to
avoid coincidence counting. The dioctyl sebacate aerosols used in the in-place
penetration tests were generated with a Laskin nozzle aerosol generator (Virtis,
Gardiner, NY). Filter penetration was determined from the ratio of the
downstream concentration divided by the upstream concentration after correcting
for the upstream dilution and subtracting background aerosols. Figure 1 shows the
penetration of the new HEPA filter as a function of particle size,

The standard leak test was conducted using a TDA-2GN light scattering
photometer (ATI, Owings Mills, MD) to obtain aerosol measurements before and
after the HEPA filter. A TDA-5B aerosol generator (ATI, Owings Mills, MD) was used
to generate the alpha-olefin (Emery 3004) aerosols for the in-place leak tests. Filter
leak measurements were made by electronically setting the photometer upstream
concentration to 100% and reading the downstream concentration directly. The
in-place leak test yielded a leak of 0.01% for the test shown in Figure 1.

Following the initial test on the new HEPA filter, we made a single pinhole in
the filter medium using a 0.025 inch diameter needle and repeated the in-place
penetration and in-place leak tests. Additional pin holes were then made in the
filter, and the filter was retested each time for penetration and leakage. The test
results for the in-place penetration measurements are shown in Figure 3 for the
filter having 0, 1, 2, and 6 pin holes and in Figure 4 for the filter having 9, 13, 19,
27, and 40 pin holes. The photometer measurements for each of the filter tests are
shown in Table 1 along with the designated number of pinholes. Table 1 also
shows the filter pressure drop and the penetration measured at 0.15, 0.3, and 0.7
pm diameter. Note that the pressure drop is not affected by the pin holes, whereas
the laser penetration and photometer leaks show large increases with increasing
number of pin holes.

The agreement between the HS-LAS and the LASAIR laser counters is very good
over the overlapping size range as seen in Figures 3 and 4. The HS-LAS and LASAIR
data are indicated by the open and closed data points, respectively. Both laser
counters also yield the same value at the maximum filter penetration. However,
the maximum penetration for the LASAIR occurs in the first size channel (0.1-0.2
l.tm), which will not allow verification of maximum penetration when the LASAIR
instrument is used alone. This is not a serious problem since the maxi mum
penetration occurs at 0.15 ym diameter for filters with and without pin holes. The
preferred laser counter should have several measurements between 0.1 and 0.2 pm
to verify that the maximum filter penetration is bracketed.

655

.-“,- I- ._..-
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

1.00 lo-s

8.00 1 O-’

...............i:..........:......i:.....::::::
~...j...j..‘..+.~
.......... ....... ......i....:i....p.f& ......* . ...i.......... ...........\...f~..~.y
t i i : :gii: : ::: :: *:
60”
. 4A-4

f..
................
...........
I.........
.............................. ...(..-..-
>...( 7! .$................
.y...l..f..! ...!......
1...... ..i‘ : :
.....t ...!...!
i i.. I .i...;...i..i..~ : &y$‘~
.....
?.....................
t’::: j i ; i / i* ...I..
....
i..
...i...i...j..i.i
.i
/........
4.00 1 0 ...............
,4
.I......I.....g...& .j................
. .......
@I- .i:::z::J:~:lilI:::::11:::~::: i......
...... i....
i...l...j..i.i.;
j.........
i&i ..i.............. i....Jy&jqj.. ..............
i /......
......... I____
:...+.;..j. .!
j..........
............... jf.......i.....
:...j.
i:::j;b49’o.&*.
.. ! : i ! : : : : ::
..........
............... i......
j iii i.. ...
i...i...i..i.i
...............
1:.......... ..>.
i......\....;...~

0 0.1 1 10
Diameter, pm
Figure 3. Penetration of DOS aerosols as a function of aerosol diameter for the
same HEPA filter having 0, 1, 2, and 6 pin holes produced with a 0.025 inch needle.
HS-LAS, open points, LASAIR, closed points.

I
0.1 i 10
Diameter, pm
Figure 4. Penetration of DOS aerosols as a function of aerosol diameter for the
same HEPA filter having 9, 13, 19, 27, and 40 pin holes produced with a 0.025 inch
needle. HS-LAS, open points, LASAIR, closed points.

656

--. -.---- -_I-_~~.

 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

Table 1. Penetration and leak measurements on a HEPA filter

with varying pin holes .

We have plotted the threee different penetration measurements versus the

photometer measurements from Table 1 in Figure 5 to examine the correlation
between the various measurements.

1
....................
.............
j..........................&..~...~..... ................
i........i......+....+...+...; ....
....................
.............
i........zi-.
..... ..........;...;...:~
.. ... ... ............
......... i....... ....+..~...~...:r
....................
>............ : :.. . ..........
f............I........ . .....i.....;....+...;...’
ii........i..
j ....(.. ...<....j....)...,
;/{i;
....................
............. ..............:.........::............+...+...,
................j.....j....;...+.;.. . .............. i
::: ....
:I

0.1

0.01

-_. -... ..j. +...+..+..:

--@--0.15 um i i i / j
* 0.3 um ..:
..i_....
.....*...1...:...:
&-i-l...:,

0.0001
0.01 0.1 1
In-Place Leak Test (Photometer), %

Figure 5. Correlation of laser penetration test with in-place leak test.

657
.~-_--. .. ---,-
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

In general, there is poor correlation between the photometer leak and the laser
penetration measurements, even for the 0.7ym data, which is supposed to
represent the average size of the test aerosol in the photometer test. One of the
reasons for the poor correlation is the lack of sensitivity of the photometer for
penetration measurements less than 0.01%. However, the major reason for the
poor correlation between the photometer leak and the laser penetration
measurements is due to the fundamental difference between differential an d
integrated size measurements with heterodisperse aerosols. Bergman and
Biermann have shown that large variations in the photometer measurements are
possible compared to laser or condensation nuclei counters depending on the
degree of aerosol heterodispersion and the extent of filter leaks.(7’14) Figure 5 also
shows that the photometer measurements, although still not satisfactory, correlate
better with the maximum penetration measurements at 0.15 pm than with the
measurements at 0.3 or 0.7 ym. The lack of correlation between the in-place
penetration test and the in-place leak test illustrates that the present leak test
provides only an approximate measure of the sytem penetration.

V. Field Evaluation of In-Place Penetration and In-Place Leak Tests

We have conducted in-place penetration and leak tests on two typical HEPA
filter installations at LLNL, a single HEPA filter system and a two-stage HEPA filter
plenum, to evaluate the practicality of the in-place penetration test. The single
HEPA filter system located on the roof of a LINL building is shown in Figure 6 with
the HS-LAS laser counter on the HEPA filter, the LASAIR laser counter on the blower,
and the TSI aerosol diluter on the floor. The Laskin nozzle aerosol generator, not
shown, was placed inside a ventilation hood in one of the building laboratories.
After several in-place penetration tests were completed, the standard in-place leak
test was performed using a TDA-2GN aerosol photometer (ATI, Owens Mills, MD)
and a TDA-4A aerosol generator (ATI, Owens Mills, MD) with Emery 3004. The in-
place leak test indicated the HEPA filter system had 0.006% leakage.

Several in-place penetration tests were conducted on the single HEPA filter
system to determine the effect of challenge concentration and the repeatability of
the test results. The challenge concentration is an important factor in the in-place
penetration test because it affects the accuracy of the data and the duration of the
test. Higher aerosol concentrations result in shorter and more precise tests but also
result in instrument error due to coincidence counting. Counting errors due to
coincidence occur at higher concentrations when two or more particles are counted
as a single particle. Since filter penetration measurements involve two
measurements at significantly different concentrations, one upstream and one
downstream of the filter, separate optimizations are required for each
measurement. In theory, the challenge concentration is adjusted so the
downstream concentration after the filter is just below coincidence counting. The
upstream concentration then has to be diluted to avoid coincidence counting.
However, since the commercially available diluters have a fixed dilution ratio; e.g.
1OO:l for one stage dilution, 10,OOO:l for two stages of dilution; the challenge

658
_. .- -.- ..__“_”
___-
-.-_ ...--
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

concentration must be adjusted to avoid coincidence in both the upstream

(challenge) and downstream measurements.

Figure 6. Photograph of the in-place penetration test apparatus on a single HEPA

filter system using laser counters. The HS-LAS laser counter is on the HEPA filter, the
LASAIR laser counter on the blower, and the TSI aerosol diluter on the floor.

The available dilution ratios did not allow for optimization of the
concentration measurements as shown with the following illustration. Figure 7
shows the filter penetration curve derived from measurements using a 100: 1
dilution of the upstream (challenge) aerosols for a single HEPA filter system which
is similar to the system shown in Figure 6. The filter penetration curve is

659
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

extremely noisy, even with a 1 minute upstream and a 15 minute downstream

sample, because the low downstream aerosol concentration is at the background
level. This resulted from reducing the challenge concentration to avoid
coincidence counting. Increasing the sampling time did not help in this case
because the measurement of background aerosols also increased. Using a 10,000: 1
diluter on the upstream sample significantly improved the precision of the data
and also reduced the sampling time as seen in Figure 8. The upstream and
downstream sample times for that test were 2 and 6 minutes, respectively. An
optimized diluter between 1,000: 1 and 2,000: 1 would reduce the sample time t o
about 1 minute for each measurement. The optimized diluter and associated
calibration procedure must be developed before the in-place penetration method is
adopted for routine measurements.

AP=l .l O"H20
6.00

5.00

4.00

3.00

2.00

1 .oo

0.00
0.01 0.1 1 10

Diameter, km

Figure 7. Filter penetration as a function of aerosol diameter for a single HFPA filter
system using the in-place penetration measurement with a 1OO:l diluter. Open
data was generated with HS-LAS, closed data with LASAIR. In-place leak test with a
photometer was 6 x 10e5.

660

_._ - ..-- ,
_.” .-. I_
.--“. ___.-. -_-.“.-.--“e.--....
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

AP=l.l O"H20

0.1 1 10
Diameter, pm
Figure 8. Filter penetration as a function of aerosol diameter for a single HEPA filter
system using the in-place penetration measurement with a 1,OOO:l diluter. Open
data was generated with HS-LAS, closed data with LASAlR. In-place leak test with a
photometer was 6 x 10m5.

A detailed comparison of the time requirements for the in-place leak and the
in-place penetration test is given in Table 2. The increased time to carry the
penetration equipment was due to the additional laser counter, the diluter an d
pumps and miscellaneous items. After the in-place penetration equipment and
procedure is finalized, the time for carrying the equipment will be the same for
both in-place tests. The much longer test time for the penetration test can be
reduced to be comparable to the leak test once the optimum diluter is developed.

Table 2. Comparison of time requirements for in-place leak and in-place

penetration measurements on a single HEPA filter installation.

Task Leak Test Penetration Test

Carry equipment to roof 2 min. 10 min.
Set up equipment 2 min. 3 min.
Set up generator 8 min. 8 min.
Test filter 2 min. 12 min.
Total 14 min. 33 min.

661

-.--.--

-,--.-” . ..--. “-,-

--,, -+- ,-1-., _--.,*“-
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

We repeated the in-place penetration test two additional times to assess the
repeatability of the test. Figure 9 shows the three in-place penetration tests on the
single HEPA filter installation are very repeatable.

0.1 1 10
Diameter, pm

Figure 9. Measurements of the filter penetration obtained with the in-place

penetration test apparatus shown in Figure 6 repeated three times.

The second field evaluation of the in-place penetration measurement was in a

two-stage HEPA filter plenum using the shroud sampling method. The shroud
sampling method allows individual HEPA filters to be leak tested independent of
the other HEPA filters in a filter bank. This is done by placing shrouds on the
upstream and the downstream side of individual HEPA filters to effectively isolate
the HEPA filter from all others in the filter bank. Each shroud is a sheet metal duct
that is held against the HEPA filter or frame on one end and has a reduced 1’ x 1’
section on the other end. The upstream shroud is used for injecting aerosols, an d
the downstream shroud is used for sampling the downstream aerosols. Figure 10
shows the front (A) and rear (B) sides of the upstream shroud, that is used to
expose a HEPA filter to a uniform aerosol concentration. Figure 10 B shows the
rear side of the upstream shroud with the 9 point aerosol injection manifold. The
aerosols are then mixed by a baffle plate seen in Figure 10 A and B and further
dispersed by a screen seen in Figure 10 B. The upstream shroud also has a sample

662

_-__ -.- . ..--.

port for sampling the challenge concentration. The downstream shroud, shown in
Figure 11, has a 9 point sampling manifold and no internal mixing devices. The
filter leak or penetration is obtained by simultaneously placing the upstream and
downstream shrouds against the HEPA filter or frame as shown in Figures 12 and 13
respectively.

10. Upstream shroud for exposing individual HEPA filters in a filter plenum to
challenge aerosols. (A) shows the front side, (B) shows the rear side.

663
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

Figure 11. Downstream shroud for sampling filter penetration or leak. Nine point
sampling manifold is seen from the inlet side facing the HEPA filter.

664

..-.^.. -_.._.---_,.
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

Figure 12. Downstream shroud for Figure 13. Upstream shroud for
sampling aerosol penetration from generating challenge aerosols.
individual filter. In-place penetra- Laskin nozzle aerosol generator
tion equipment used in this test. used in this test.

The result of the in-place penetratia measurement on one filter in the plenum
is shown in Figure 14. We were unable to generate the required high concentration
of challenge aerosols to use the 10,OOO:l diluter because the compressor shown in
Figure 13 could not supply sufficient pressure to the Laskin nozzle aerosol
generator. As a result, we used the 1OO:l diluter with a lower aerosol
concentration. This resulted in lower precision and a longer sampling time than
would be required with a higher aerosol concentration and a 10,OOO:l diluter. The
upstream and downstream sample times were 2 and 8 minutes, respectively. The
equipment used for the in-place penetration measurement using the shroud
method was the same as previously described for the single filter test.
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

Since the shroud method only measures the penetration or leaks through the
filter, and not around gasket leaks, a separate leak test is performed on each filter.
This is done by directing a concentrated aerosol challenge around the perimeter of
the upstream side of the filter using a long tube. Another person samples the
perimeter of the downstream side of the filter using a long probe that is moved in
syncronization with the upstream challenge tube. If the downstream leak is
greater than 0.03% of the upstream concentration, then the filter is replaced. This
traverse leak test is far more conservative than the leak or penetration
measurement through the filter because no significant air volume passes through
the leak compared to that flowing through the filter. Since the air flow through a
gasket leak parth is not known, the traverse leak test is not quantitative, but rather
a qualitative test. When using the laser counter in this leak test, the counter
output is set to the concentration mode and not the count mode.

The conventional in-place leak test indicated the filter in Figure 14 had a leak
of 2 x 10-4. We used a TDA-2EN photometer and a TDA-5B aerosol generator, both
from ATI, for the in-place leak test. The test aerosol for the in-place leak test was
Emery 3004.

PLENUM AP=0.70"H20

Diameter, pm

Figure 14. In-place penetration measurement of a HEPA filter in a plenum using

the shroud sampling method. Open data obtained with HS-LAS, closed data with
LASAR

666
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

We performed a detailed analysis of the time requirement for the in-place leak
and the penetration test using the shroud method and tabulted the results in
Table 3. The time requirements for all of the tasks except for the downstream
measurements are comparable for the two in-place tests. As noted before, the long
downstream sampling times was primarily due to the inability to generate a
sufficient concentration. We anticipate that the in-place penetration
measurement would not require much more time than the in-place leak test once
the experimental test system is optimized.

Table 3. Comparison of time requirements for in-place leak and in-place

penetration measurements on a HEPA filter bank using the shroud method.

Task Leak Test Penetration Test

Equipment set up 10 min. 15 min.
Equipment warm up 10 min. 10 min.
Upstream meas./filter (5 min.) (4 min.)
Upstream bank (16 filters) 80 min. 64 min.
Downstream meas./filter (0.3 min.) (10 min.)
Downstream bank(l6 \
filters) I
5 min. 160 min.
Tear down 30 min. 30 min.
Total 135 min. 279 min.

VI. Conclusions

We have demonstrated the feasibility of conducting in-place penetration tests

on high efficiency particulate air (HEPA) filters as installed in nuclear ventilation
systems. The in-place penetration test, which is designed to yield equivalent
penetration measurements as the standard DOP penetraton test, is based o n
measuring the aerosol penetration of the filter installation as a function of particle
size using a portable laser particle counter.(‘) Additional tests are required to
confirm that the in-place penetration test yields identical results as the standard
DOP penetration test for HEPA filters with controlled leaks in the filter and gasket
and duct by-pass leaks. Further development of the procedure is also required to
reduce the test time before the in-place penetration test is practical.

VII. Acknowledpments

We gratefully acknowledge the assistance Mr. Wayne Krause and Mr. Donald
Beason in the shroud tests.
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

VII. References

1 . MIL-STD-282 Filter Units, Protective Clothing, Gas-Mask Components and

Related Products: Performance-Test Methods, Method 105.9, Military Standard
MIL-STD-282, Commanding Officer, Frankford Arsenal, Navy Department, ATTN:
SMUFA-N 1100, Philadelphia, PA, 19 137, ( 1974)

2. American Society of Mechanical Engineers, “Testing of Nuclear Air Treatment

Systems”, ASME Standard N510-1989, The American Society of Mechanical
Engineers, 345 47th Street, New York, N.Y. 10017, (1989)

3. American Society of Mechanical Engineers , Code on Nuclear Air and Gas

Treatment, ASME AGl-1994, The American Society of Mechanical Engineers, 345
47th Street, New York, N.Y. 10017, (1989)

4. Burchsted, CA, Kahn, JE, and Fuller, AB, “Nuclear Air Cleaning Handbook”, ERDA
76-21, National Technical Information Service, 5285 Fort Royal Rd, Springfield, VA
22161, (1976).

5. Parrish, EC and Schneider, RW, “Review of inspection and testing of installed

high-efficiency particulate air filters at ORNL”, in Treatment of Airborne
Radioactive Wastes, International Atomic Energy Agency, Vienna, pp 243-264,
(1968)

6. W. Bergman, A. Biermann, W. Kuhl, B. Lum, A. Bogdanoff, H. Hebard, M. Hall,

D. Banks, M. Mazumder, and J. Johnson, “Electric Air Filtration: Theory, Laboratory
Studies, Hardware Development, and Field Evaluations”, LLNL Report, UCID- 19952.
January 9, (1984)

7. W. Bergman and A. Biermann, “Effect of DOP Hetero-dispersion on HEPA Filter

Penetration Measurements,” in Proceedings of 18th DOE Nuclear Airborne Waste
Management and Air Cleaning Conference, Baltimore, MD, Aug. 12-16, 1984, p p .
327-437, NTIS, Springfield, VA, CONF-840806, (1985)

8. Scripsick, R.C., Smitherman, R.L., and McNabb, S.A. “Operational evaluation

of the High Flow Alternative Filter Test System” Proceedings of 19th DOE/NRC
Nuclear Air Cleaning Conference, CONF-860820, National Technical Information
Service, Springfield, VA 22161, pp 863-889, (1987).

9. Schusster, B and Osetek, D, “The use of a single particle intra-cavity laser

particle spectrometer for measurements of HEPA filters and filter systems” 14th
ERDA Air Cleaning Conference, p.528, CONF-760822, NTIS, Springfield, VA, (1977).

668
10. Ortiz, J, “In-place testing of multiple stage filter systems without disruption
of plant operations in the plutonium facility at Los Alamos, 18th DOE Nuclear
Airborne Waste Management and Air Cleaning Conference, CONF-840806,
NTIS,Springfield, VA, p. 209, (1985).

11. Ortiz, J, Biermann, A, and Nicholson, R, “Preliminary test results of a round

robin test program to evaluate a nulti-stage HEPA filter system using single
particle size counts” in Advances in Filtration and Separation Technology, 4, 2 13,
American Filtration Society, Gulf Publishing Co., Houston, Texas (1991).

12. ASTM Standard Test Method for “Air cleaning performance of a high-
efficiency particulate air filter system”, F-1471-93, ASTM, 1916 Race St.
Philadelphia, PA, 19 103, (1993).

13. Martinez, V. “‘Procedure for in-place filter testing” Los Alamos National
Laboratory, Operating Manual, (1995).

14. Biermann, AH and Bergman, W, “Filter penetration measurements using a

condensation nuclei counter and an aerosol photometer” J. Aerosol Sci., Vol. 19,
No. 4, pp 471-483, (1988).

669
 24th DOE/NRC NUCLEAR AIR CLEANING AND TREATMENT CONFERENCE

DISCUSSION

FRANKLIN: On the DOP feed of the sampler, was that downstream of the grid? I could not tell
where you were sampling the concentration of the DOP feed.

BERGMAN: The upstream sample was measured through a single probe that was inserted through a
porthole in the shroud just before the HEPA filter. A single probe is sufficient at this point because the
Emory 3009 aerosols were well mixed. The aerosols are injected through a nine-port manifold within a
one foot by one foot area. The aerosols then hit a baffle that covers about 70% of the open area and is
further mixed by a grid. The sample is then taken after the grid.

KOVACH,% I have one question . When you tested with the shroud method I understand how you
calculated the average leak with the shroud, but how did you add the gasket leakage to the efficiency?

BERGMAN; The shroud method only allows you to measure the average leak if the HEPA filter is
mounted on the upstream, dirty side of the frame. In this case, the shroud is placed against the frame
on the downstream side. Penetrations through gasket leaks and the filter are measured together as a
single measurement. If the HEPA filter is mounted correctly on the downstream, clean side of the
frame, it is not possible to quantify the leak through the gasket. The shroud is placed over the HEPA
filter and only measures the leak through the filter. Leaks through the gasket are determined in a
separate test in which a probe is traversed around the perimeter of the gasket. If the aerosol penetration
at any point exceeded 0.03% the filter was removed.

670
---.--- - .---- .--__
 NUREG/CP-0153
CONF-960715

Proceedings of the
24th DOEINRC Nuclear
Air Cleaning and Treatment
Conference
Held in Portland Oregon
July 15-18, 1996

Date Published: Aug,",st 1997

Edited by
M. W. First

Sponsored by
Office of Environmental Guidance
U.S. Department of Energy
Washington, DC 20585

Office of Nuclear Regulatory Research

U.S. Nuclear Regulatory Commission
Washington, DC 20555-0001

International Society of Nuclear Air

Treatment Technologies, Inc.
4400 Glen Willow Lake Lane
Batavia, OH 45103

Harvard School of Public Health

The Harvard Air Cleaning Laboratory
665 Huntington Avenue
Boston, MA 02115-6021

Proceedings prepared by The Harvard Air Cleaning Laboratory

NRC Job Code G6543

View publication stats