Sie sind auf Seite 1von 3

Main feature

HEPA Filter testing – difficult to test filters,


Part 1: Volumetric testing
Neil Stephenson

Introduction fan and HEPA filter would normally be Extract filters are thus classed as
HEPA filters are designed to remove protected from gross dirt by pre-filters difficult to test filters. They are difficult
particulate from the air and provide installed upstream of the fan. The to scan test and in many cases very
clean particulate free air after the filters. positive pressure makes access to the difficult to test volumetrically. Other
They are used in a variety of critical clean side of the filter possible for filters fitting this category would be any REPRINTED
applications for:
• Product protection
scanning, without disrupting the
normal flow through the filter. (The fan
filter which cannot be accessed due to
lack of space or for any other reason.
FROM:
speed control would be in manual for
• Personal protection
gaining access to the filters) This Volumetric and scan testing
Clean Air and
• Environmental protection assumes that there is physically Probably for the first time in a standard, Containment
It follows that if a HEPA filter is
sufficient space provided for access in
the design. Generally source filters will
ISO EN BS 14644-3 provides a
procedure for volumetric testing. The
Review
installed in a system then it should be
be face scanned when tested. arrangement is shown in Figure 3. The Issue 7,
tested. Unfortunately, all too often,
HEPA filters are found in systems which
The extract filter installation is basic ‘acceptance criterion (B.6.4) is that July 2011
shown in Figure 2. The extract filter fan “none of the penetrations should be
are not designed with testing in mind.
is found on the downstream side of the higher than five times the specified
The standard that governs the testing
filter – the clean side. The fan is pulling nominal MPPS (Most penetrating
of HEPA filter is ISO 14644-3:2005 – Test
air through the filter as opposed to the Particle Size)* penetration of the filter.
Methods. All percentage penetrations or
fan in the source filter installation which However for photometers this
efficiencies quoted in this article are on
is pushing air through the filter. The penetration should not be greater than
the basis that the test being carried out is
negative pressure created makes it 10-4 (0.01%)”. Other acceptance criteria
a traditional aerosol filter test (DOP Test)
nearly impossible to access the clean may be agreed between the customer
in accordance with section B.6 of this
side of the filter without disrupting the and the supplier. 0.01% penetration is
standard using an aerosol photometer
air flow through the filter. Opening an exactly the same acceptance criterion as
and thermal aerosol generator.
access panel between the filter and fan that given for a single local leak when
Two widely used test methods are:
will cause the fan air to be drawn in scan testing. Whilst the standard warns
• the procedure for installed filter
through the access panel and no air will that volumetric testing “is much less
system leakage scan test with an
pass through the filter. For this reason sensitive for finding leaks than” scan
aerosol photometer as described in
most extract filters have traditionally testing methods, it does not emphasise
ISO14644-3:2005 section B.6.2. This
been volumetrically tested as opposed or demonstrate how much less sensitive
procedure is commonly known as
to scan tested. it really is.
scan testing.
• the procedure for overall leak test of
filters mounted in ducts or air
handling units as described in ISO
14644-3:2005 section B.6.4. This
procedure is commonly known as
volumetric testing.
The volumetric testing is much less
sensitive for finding leaks than scan
Figure 1: Source or supply air filter installation
testing and this article explains and
quantifies the difference.

Source filters and extract filters


Installations (systems) containing
HEPA filters may be divided into two
categories, source filter installations and
extract filter installations.
Source filter installations, as shown
in Figure 1, will have the fan located on
the dirty side of the filter and provide a
positive pressure to the installation. The Figure 2: Extract filter installation

4  Clean Air and Containment Review | Issue 7 | July 2011 www.cleanairandcontainment.com

ACR7_aw.indd 4 12/09/2011 20:38


Main feature

* MPPS (Most Penetrating Particle


Size) is defined as “the particle size at
which the minimum of the particle
size efficiency curve occurs” (EN
1822-1:2009 – High efficiency air filter
(HEPA and ULPA) – Part 1:
Classification, performance, testing,
marking). It is a characteristic of a
HEPA filter that the penetration is at
its maximum at a certain particle size,
the MPPS, and is less for particles
both larger and smaller than the
MPPS particle. This is due to the
HEPA filter having a range of
filtration mechanisms, from ‘sieving’
which obviously removes the largest
particles to ‘electrostatic’ which
removes the very smallest of
particles. For a given filter, the MPPS,
Figure 3: Diagrammatic representation of volumetric testing
which is typically between 0.1 µm
and 0.2 µm, varies slightly with
airflow velocity, and the penetration
or efficiency varies significantly with
airflow velocity. HEPA filters must
therefore be operated at their rated
airflows in order to achieve their
rated efficiencies.

Scan testing of a filter installation,


shown diagrammatically in Figure 4, is
by far the safest test that can be carried
out on a filter. It will find a local leak at
the filter face which is essentially
undiluted and a direct measurement of
the size of the leak. There are many
opinions about scan testing circular and
“wedge” (also known as “W” filters or
Figure 4: Diagrammatic representation of scan testing
“V” filters or Delta filters due to their
filter pack shape) cannot be face
scanned. To face scan one of these filters
we assume that the area immediately at
the exit of the filter is the filter face. This
area is then scanned as if it were the
filter media face. Even this provides a
much safer test than the volumetric test.
The volumetric test combines all the
leaks in the filter installation and allows
these leaks to mix over a minimum
length of 20 duct diameters. A SINGLE
sample measured 20 duct diameters
from the filter then determines the
penetration through the filter.
The 20 duct diameter length is the
minimum distance required in a straight
duct to achieve enough mixing for a
homogenous sample. Figure 5 shows
Figure 5: Sketch showing straight line mixing
how the concentration from a single
A single sample measurement in the duct before the 20 duct diameter distance could show ‘no
point leak ‘spreads’ until mixing and leak’ as it is measuring in a clean air zone. Such a sampling point is shown as ‘Single sample’ in
dilution are complete – in theory – at 20 the figure.

www.cleanairandcontainment.com Clean Air and Containment Review | Issue 7 | July 2011  5

CACR7_aw.indd 5 12/09/2011 20:38


Main feature

duct diameters. This distance may be ISO 14644-3 using a DOP aerosol The best way to understand
less in practice if there are bends in the photometer. A volumetric test would environmental loading is to consider an
duct or special arrangements made to hope for no leaks in the system and example in which
achieve mixing in a shorter distance. therefore the volumetric penetration the mass of a product which could be
(This will not be covered in this article) expected would be the same as the deposited into the environment through
There are some US practitioners now media penetration – 0.0005%. an apparently compliant filter system
advocating 25 duct diameters. In the UK A failure of such a system under test will be calculated.
nuclear standards 10 duct diameters is would be a volumetric value that exceeded The system configuration which will
recommended. Whatever the mixing the media penetration. We could define a be considered consists of six 610 x 610 x
distance, the fact remains that it is often failure to be, say, twice the value of the 292 mm filters in one bank with
not possible to find an accessible point media penetration. This would represent a common plenums inlet and outlet. The
at that distance. For example, the value of 0.001% in the example given media penetration for the filters is
location could be in an extract duct on above. Unfortunately this is unsatisfactory 0.0005% when tested according to |ISO
the outside of a building. in a large filter system. A 0.001% 14644-3 DOP aerosol photometer. The
volumetric leak in the 10,000 m3/hr permissible maximum volumetric
Understanding the numbers installation would represent a 3% single penetration is 0.01%. The filter bank is
and a new approach to leak at the filter face. The conclusion subjected to a continuous product
volumetric testing would be that volumetric testing is unsafe loading of 0.001 mg/m3. A volumetric
A % penetration measured using the at best, dangerous at worst. test of the filters reveals a volumetric
volumetric test is wholly related to Extract filters protect our penetration of 0.009% thus the system is
volume flow. A single local leak of 3% at environment from toxic pollution. That in compliance.
the face of the filter will have a different is why they are used. They should be the The mass of material which will be
volumetric result when the leak is most rigorously tested filters but in deposited into the environment through
measured in a 500 m3/hr installation practice they are the worst tested filters this compliant system can now be
compared to one which is in a 10,000 m3/ as is shown above. calculated. The answer is 43 grams per
hr installation (e.g. a large bank of We have so far not looked at the long year which would accumulate to 430
filters). Thus the 3% local leak in the 500 term environmental loading effect. With grams over a ten year period!
m3/hr installation will give a volumetric potential leaks of the sizes discussed This may not appear to be a huge
leak reading of 0.01% whereas in the this can be substantial. mass of material. However if it is
10,000 m3/hr installation it will produce a dangerous to people and animals, then
volumetric leak reading of 0.0005%. Long-term environmental loading it poses a significant danger
This is clearly unsatisfactory. Similar The long term loading caused by the Alternative methods of testing filters
calculations show that a single local face continuous discharge to atmosphere that get over this problem will be
leak of 59% would produce a reading for through apparently compliant filters described in the next part of this paper
volumetric penetration of 0.01% if has, to the author’s knowledge, never in the next issue.
diluted with 10,000 m3/hr of clean air. been properly considered.
There are filter scanning solutions
which eliminate the need to test
volumetrically. Notwithstanding such Neil Stephenson graduated as an Industrial Electronic
solutions there will always be reasons to Engineer from the Natal Technikon in South Africa in 1971.
test volumetrically. Thus we need to He started his career with 13 years in the Scientific and
approach the numbers in a far more Industrial Division at Philips, gaining extensive knowledge of
cautious manner. all types of analytical instrumentation. He moved to the
To deal with the large dilution United Kingdom in 1983 and soon afterwards founded
problem we first need to know what the Steptech Instruments and DOP Solutions, focussing on
media penetration of the filter(s) is. This instrumentation for the cleanroom industry. DOP Solutions is the only
is quite easily achieved by sampling company outside the USA undertaking research and design of instruments for
underneath a number of filters in the aerosol filter testing. Today, as Technical Director at DOP Solutions, Neil
installation being tested. This can provides consultancy, instrumentation design and technical turnkey solutions
normally be achieved through the for clean air industry customers. As a recognised expert on particle counting,
differential pressure sample port for the airflow and calibration standards, he sits on the BSI LBI/30 Committee and the
filters which is generally about 100 mm ISO Working Groups which were responsible for writing the original ISO 14644
beyond the filter on the clean side. The Parts 1, 2 and 3 and are currently re-writing Parts1 and 2. He also sits on the
measuring probe may be a tube 8 mm BSI group MCE/21/3 responsible for EN1822 (HEPA and ULPA filters) and ISO
OD, 6 mm ID and approximately 500 TC142 WG1 (Cleaning equipment for air and other gases – Terminology). He
mm long. The ID should never be less has developed comprehensive training courses on HEPA Filter Testing and
than 5 mm. Particle Counting to ISO standards and is a lead trainer for the Academy for
Filters used in extract systems would Cleanroom Testing and the Cleanroom Testing and Certification Board,
normally have a media penetration of lecturing on the Irish Cleanroom Society and R3 Nordic training programmes.
0.0005% or better (standard UK nuclear He is a member of the PDA, IEST, SEE, S2C2 and ICS.
extract filter) when tested according to

6  Clean Air and Containment Review | Issue 7 | July 2011 www.cleanairandcontainment.com

CACR7_aw.indd 6 12/09/2011 20:38

Das könnte Ihnen auch gefallen