FOREMOST IN AIR FILTRATION - pdfs.semanticscholar.org€¦ · HEPA Filters and Filter Testing ... Note: Either of the two test methods or an alternative method may be used for filter

A Wholly Owned Subsidiary of Flanders Corporation

HEPA Filters and Filter TestingPB-2007-1203

HEPA Filters and Filter TestingA Comparison of Factory Tests and In-Service Tests

®

FOREMOST IN AIR FILTRATION

Quality Assurance

Any industry that has dangerous process orexhaust gases and/or particulates has a vitalconcern for the health and safety of personnel.In addition to corporate concern, the UnitedStates Government has dictated that safetyequipment meet minimum safety standards. Anyequipment sold to meet these minimum standardshas to be manufactured using accepted QualityControl procedures.

Flanders Corporation has developed a QualityAssurance program to assure that the product orservice provided meets these standards. Thisprogram addresses the entire range of Flandersinvolvement, including the purchase of rawmaterials, the shortage of these raw materials,incorporation of these materials into a product orservice, testing this product or service, and thenshipping it to its destination.

The program of Flanders has been audited manytimes, and each time the program has beenacceptable. An uncontrolled copy of the programmanual is available with each request for QualityAssurance information. Like any dynamicdocument, the program is continually beingrevised to include recent issues of standards andspecifications in order that Flanders/CSC mayuse the latest state-of-the-art methods in providingits products and services.

The Quality Assurance Program at FlandersCorporation has been audited and approvednumerous times by the Nuclear UtilitiesProcurement and Inspection Committee, NUPIC.This committee was established by nuclearelectric utilities to ensure that suppliers of goodsand services can meet all applicable regulatoryand quality requirements.

Notes:

1. As part of our continuing program toimprove the design and quality of allour products, we reserve the right tomake such changes without notice orobligation.

2. Flanders, through its limited warranty,guarantees that the products de-scribed herein will meet all specificationsagreed to by the buyer and the seller.

3. ASME N509 Nuclear Power Plant Air-Cleaning Units and Components.

4. ASME N510 Testing of Nuclear AirTreatment Systems.

5. ASME AG-1 Code on Nuclear Air andGas Treatment

© Copyright 2004 Flanders/CSC Corporation7013 Hwy 92E - PO Box 3

Bath, NC 27808

HEPA Filters and Filter Testing: Quality Assurance

Quality Assurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inside Front Cover

Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Important Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

GeneralPhotograph: In-Place Test Injection Section, HEPA and PrecisionScan

In-Place Test SectionPhotograph: Vertical Flow HEPA Filter CeilingChart: Recommended Test and Minimum Rating for Filter Types A - FFigure 1: Flanders Industrial Grade Filter Label

Q 107 Penetrometer Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Figure 2: Challenge Aerosol Test

Scan Testing or the “Cold” Challenge Aerosol Test . . . . . . . . . . . . . . . . . . . 7

Figure 3: Cold Aerosol TestFigure 4: Laskin NozzleFigure 5: Challenge Aerosol GeneratorFigure 6: Flanders Laminar Flow Grade Filter Label

Two-Flow Efficiency Testing and Encapsulation . . . . . . . . . . . . . . . . . . . . . 11

Figure 7: Flanders Nuclear Grade Filter LabelTwo Flow Efficiency Tested, Encapsulated and Scan-Tested FiltersFigure 8: Filter Test Portion of the Q-107Figure 9: Flanders Filter LabelFigure 10: Two-Flow Efficiency TestFigure 11: System using Calibrated Dual-Laser Spectrometer SystemFigure 12: Flanders Filter Label

In-Service & In-Place Tests for HEPA Filters . . . . . . . . . . . . . . . . . . . . . . . . . . 14

In-Service Tests for HEPA FiltersIn-Place Testing - HEPA Filter BanksFigure 13: Test of Ventilating System with Single Bank of HEPA FiltersFigure 14: The Ductwork and Plenums in HVAC Systems

Clean Room Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Figure 15: Scan Testing Clean Room Ceiling

In-Place Testing Housings for Efficiency Testing . . . . . . . . . . . . . . . . . . . . 19

In-Place Testing Housings for Scan Testing . . . . . . . . . . . . . . . . . . . . . . . . . 20

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Figure 16: Factory Test Specifications, Field Test Specifications, Applications for HEPA and VLSI ® Filters

Table of ContentsHEPA Filters and Filter Testing

1

Imp

ort

an

t M

essag

e

NOTICE . . . Compliance with installation andoperation standards must be met to ensurequality performance.

HEPA filters are factory tested to meetthe requirements of IEST RP-CC001.3for Type A, B, C, D, E or F filters:

• Industrial Grade• Nuclear Grade• Laminar Flow Grade• Bio/Hazard Grade HEPA• VLSI• ULPA• Pharmaceutical Grade

Test results appear on both the filterlabel and upon the filter carton label.An additional quality assurance testreport is kept on file and is availableupon request.

Flanders recommends that all HEPAfilters be tested in place byqualified personnel to ensure that thefilters have been correctly installed.

Flanders service personnel areavailable for installations, supervisionof installation, testing and certificationof compliance to industry andgovernment standards and instructionof the owner’s personnel in testing andmaintenance procedures.

Flanders does not guarantee that itsequipment will operate at theperformance levels given onthe identification labels or in thecatalog specifications under allconditions of installation and use, nordoes Flanders/CSC guarantee thesuitability of its product for thepar ticular end use which may becontemplated by the buyer.

For best results, it is recommendedthat the buyer supply completeinformation about the operatingconditions of the ventilation system toFlanders/CSC for evaluation.

When the system components aresupplied to the buyer or his agentfor final installation and assemblyin the field, it should be underthe supervision of factory trainedpersonnel.

Failure to adhere to this recommenda-tion or failure of the buyer to havefilters timely retested and serviced willnullify or limit any warranties whichmight otherwise apply and may resultin a compromised installation.

HEPA Filters and Filter Testing

2

HEPA Filters and Filter Testing: Introduction

3

IntroductionHEPA filters, once known as absolute filters,were originally developed as the particulatestage of a chemical, biological, radiological(CBR) filtration/adsorber unit for use by the U.S.Armed Services. In the late 1940s the U.S.Atomic Energy Commission adopted them foruse for the containment of airborne radioactiveparticulates in the exhaust ventilation systemsof experimental reactors as well as for use inother phases of nuclear research. The periodfrom the mid 1950s to the present has seenthe emergence of many new industrial andscientific technologies requiring particulate freeair in order to produce more sensitive productssuch as microelectronic components, photo-products, parenteral drugs and dairy products.These technologies fostered the developmentof a wide range of specialized devices to houseHEPA filters to deliver clean air to productionareas. Uses for HEPA filters in hazardouscontainment applications have increased also,and they are more routinely used on the exhaustside of bio-hazard hoods, animal diseaseresearch laboratories and whenever airbornecarcinogens must be controlled.

Vertical Laminar Flow HEPA Filter CeilingThe many diverse applications for HEPA filtershave resulted in a large number of industrialand governmental specifications which oftenconflict with one another, principally because ofthe different methods and devices usedto test the performance of the filters, both atthe factory and in service. In 1968, theAmerican Association for ContaminationControl (AACC) addressed this problem byissuing the specification AACC CS-1T,Tentative Standard for HEPA filters, whichcategorized the filters as Type A, B or C.Following that, Flanders originated the termsIndustrial Grade, Nuclear Grade and LaminarFlow Grade for the Type A, B and C filters,respectively, to better relate them to theindustry or application in which the filter isprimarily used. The AACC organization ceasedto exist and the standards written under itsauspices were later adopted by the Instituteof Environmental Sciences (IES) for a lengthyinterim during which the standard becameIES CS-1T. In November of 1983, followingseveral years of committee work to update thematerial, the standard was reissued by theInstitute of Environmental Sciences as IES RP-CC-001-83-T (Recommended Tentative Practicefor Testing and Certification of HEPA Filters). Twomore filters were added, Types D and E, theequivalent of the Flanders VLSI® Filter and theFlanders Bio/Hazard Grade Filter. At present, a

In-Place test injection section, HEPA andPrecisionScan In-Place test section.

4


Recommended Test and Minimum Rating for Filter Types A—F

FlandersGrade

FilterType

Penetration Test

AerosolMethod

Scan Test(see note)

Comments MinimumEfficiency

RatingMethod Aerosol

Industrial A MIL-STD 282 Thermal None None 99.97% atDOP or PAO 0.3 µm

Nuclear B MIL-STD 282 Thermal None None Two-Flow 99.97% atDOP or PAO Leak Test 0.3 µm

Laminar C MIL-STD 282 Thermal Photometer Polydisperse 99.99% atFlow DOP or PAO DOP or PAO 0.3 µm

VLSI® D MIL-STD 282 Thermal Photometer Polydisperse 99.999% atDOP or PAO DOP or PAO 0.3 µm

Biological E MIL-STD 282 Thermal Photometer Polydisperse Two-Flow 99.97% atDOP or PAO DOP or PAO Leak Test 0.3 µm

ULPA F IES-RP-CC007 Open Particle Open 99.999% atCounter 0.1 to

0.3 µm

Type F filter has been added, which is theequipvalent to the Flanders ULPA Grade filter.By definition, a HEPA filter has a minimumefficiency of 99.97% when challenged with athermally generated dioctylphthalate (DOP)

aerosol whose particle size is 0.3 micrometers(homogeneous-monodisperse). This efficiencyis a manufacturing standard that the filterproducer must attain, although most FlandersHEPA filters average above 99.98%. Since afilter’s efficiency increases as it accumulatesparticulate matter, the initial efficiency is thelowest efficiency during the life of a filter. It isimportant to note that a filter’s initial (clean)efficiency represents the average initial efficiencyof that filter. Minute areas of greaterpenetration, either in the edge sealant betweenthe filter element and the filter’s integral frameor in the element itself, are often present. Whenthe filter is tested, these small penetrations arediluted by the greater amount of clean air pass-ing through the filter. These penetrations canbe tolerated as long as the overall penetrationthrough the filter does not exceed .03% (Note:100% - .03%=99.97%.)

Note: Either of the two test methods or an alternative method may be used for filter types C, D, E and F, ifagreed upon between the buyer and the seller. Equivalency of the alternative method should be determinedjointly by the buyer and the manufacturer.

Flanders Filters manufactures and testsits cleanroom filter products in its own

cleanroom.

5


Figure 1: Flanders Industrial Grade Filter Label - Type A FilterTypical test results are entered on the label. Originally, HEPA filter specifications called for a maximum pressure drop of 1” w.g. at1000CFM for a 24” x 24” x 11 1/2” size filter. Most nuclear specifications still require this. However, many filters perform better andmanufacturers have rated separator-type filters as high as 1200 CFM at 1” w.g. This difference between Test Flow and Rated Flo w hascaused some confusion in the industry.

The instrument used by manufacturers to testfilters for efficiency is commonly referred to asthe “hot” DOP machine because it usesthermally generated particles to challenge thefilter. The hot DOP test was a joint developmentof the U.S. Army and U.S. Navy and is performedon a forty foot long apparatus called a Q 107Penetrometer.

As shown in Figure 1, when a filter is tested onthe penetrometer, two values are taken: thepenetration reading and the pressure drop ata specified flow rate (Test Flow). Thesevalues are recorded on a bar-coded serializedlabel that is applied to each filter and aduplicate label appears on each filter carton.Rarely is the information alike on any twofilters. Filters larger than 24” x 12” x 5 7/8” areindividually packaged. A cer tification ofcompliance report listing the penetration andpressure drop values relative to the serialnumber and bar code on each filter can be sentto the buyer upon request.

The specification, Mil-Std-282, DOP SmokePenetration and Air Resistance of Filters,describes the operating procedure for testingfilters with the “Hot” DOP penetrometer and isreferenced throughout industry. In order tocomply with the definition of a HEPA filter, eachfilter is required to be tested for resistance andefficiency. The Institute of Environmental Sci-ences and Technology, IEST-RP-CC001.3

states, “HEPA Filter. . .having minimum particlecollection efficiency of 99.97% for 0.3 micronthermally-generated dioctylphthalate (DOP)particles or specified alternative aerosol.Another challenge aerosol is polyalphaolefins(PAO) which provide appropriate testingcharacteristics. Further, a maximum cleanpressure drop of 1.0-inch water gage [or 1.3,depending upon the type of HEPA filter]. . .” Amanufacturer cannot certify that a filter is a HEPAfilter unless he owns a penetrometer and hashad it NIST (National Institute of Standards andTechnology) calibrated according to industry-accepted standards. The Type A filter, per IEST-RP-CC-001-3, is “One that has been tested foroverall penetration at rated flow with thermallygenerated DOP smoke. . .” This is theequivalent of the Flanders Industrial GradeHEPA Filter.

The DOP Test (Figure 2) begins with themanufacture of particles that are homogeneousin size (0.3µm) to form a nearly monodispersedaerosol, because not 100% of all particles areexactly 0.3µm size. To test a filter at 1000 CFMon the Q 107 Penetrometer, outside air is drawninto a duct at 1200 CFM and then dividedthrough three parallel ducts at 85, 265 and 850CFM (200 CFM is eventually exhausted throughan alternate exhaust path). As shown in Figure2, the top duct contains three banks of heatersand a challenge aerosol oil reservoir with a fourth

6

HEPA Filters and Filter Testing: Q 107 Penetrometer Instrumentation

Figure 2: Challenge Aerosol Test

heating element beneath the reservoir. Thecenter duct contains a cooling coil and a bankof heating elements. The air passing throughthe top duct is heated to approximately 365° Fand is then impinged through an orifice onto thechallenge aerosol oil in the reservoir. Theheating causes the challenge aerosol oil toevaporate and it is then carried forward to theconfluence of the top and center ducts where itis quenched with the cooler air from the centerduct. The 0.3 micrometer particle size iscontrolled here by maintaining the temperatureat 72° F. By increasing or decreasing thetemperature, the particle size can be increasedor decreased.

Next, the combined airflow from the upper twoducts is mixed with the remaining 850 CFM fromthe bottom duct. A series of baffles mixes theaerosol (smoke) thoroughly into the airstreamto distribute the aerosol uniformly prior tochallenging the filter. A similar set of bafflesis located on the exhaust side of the filterbeing tested to thoroughly mix the effluent.An upstream sample is taken and, when theaerosol concentration reading is between 80and 100 milligrams per liter, that value isaccepted as a 100% challenge. Next, areading (% concentration) is taken downstreamof the filter (downstream of the baffles so that

any leakage is thoroughly mixed into theeffluent) and is compared to the upstream value.This is read as a penetration, that is, if the down-stream concentration is .04% that is the percent-age of the upstream value that has penetratedthe filter. When subtracted from the 100% value,the filter would have an efficiency of 99.96% andwould be rejected.

Los Alamos National Laboratories developed analternate test method in the 1980’s undercontract to the U.S. Department of Energy(DOE). It is often referred to as the HFATS test(High Flow Alternative Test System). It wasdeveloped specifically to test filters rated at air-flows higher than 1100 CFM, but it can be usedfor lower flows. It is only limited by the size ofthe system fan and the aerosol generator out-put. This method was later standardized in thepublication of a recommended practice, IEST-RP-CC007.1, Testing ULPA Filters, published bythe Institute of Environmental Sciences andTechnology. Currently, ASME AG-1 Section FCallows for testing by this method. The filter is chal-lenged with an acceptable polydispersed oilaerosol and the penetration through the filter ismeasured with a Laser Particle Counter. TheParticle Counter counts and sizes individualdroplets in a size range from 0.1 to 3.0 microme-ters in diameter. The ratio of the downstreamcounts to the upstream counts in each size range

7

HEPA Filters and Filter Testing: Scan Testing —“Cold” Challenge Aerosol Test

is the penetration. Although this value is notequal to the penetration measured by theQ-107, research performed by Loa AlamosNational Laboratory verified it to be very similarand the method to be an acceptable alternativeto the penetration measured by Mil-Std-282 TestMethod.

Since this system measures the penetration ineach size range, and a HEPA filter penetrationvaries with particle size, the maximum allowablepenetration is 0.03% for the most penetrationparticle size (MPPS). FFI can use this system totest filters that are rated at flows higher than 1100CFM, if specified by the customer.

Q 107 Penetrometer Instrumentation1. Temperature Controllers

a. Hot Air @ (approx.) 365° F

b. Oil Reservoir @ 390° F

c. Quench Air @ 72° F

2. Mechanical Analyzer

This enables the operator to determinewhen he has the correct particle size.Smoke is drawn through a chamber andin between two photomultiplier tubes.The operator reads the particle size onthe Size Indicator.

3. Linear Photometer (.0001% to 100%)

This is used for reading the upstreamand downstream samples and compar-ing them. The downstream value as apercentage of the upstream value is thePenetration.

4. Manometer

Determines the pressure drop across thefilter at the test flow.

5. Averaging Pitot Tube SystemsEnables the operator to determine thevolume of airflow through the filter.

Scan Testing or the Cold ChallengeAerosol TestWhen individual filters cannot be tested, mostcontainment requirements are satisfied byachieving average filter bank efficiencies of99.95% or greater. A single pass through acorrectly installed and field-tested filter or bankof filters is sufficient to accomplish this efficiencyalthough most nuclear facilities, because ofadditional safety related considerations such asfire protection and redundancy, can have twoor more banks of HEPA filters in series on theexhaust of their HVAC systems. As previouslystated, the areas of greater penetration that canoccur on HEPA filters, frequently called “pinholeleaks,” are tolerated as long as the overallpenetration does not exceed .05%.

This is not the case in laminar flow systems(clean work stations, clean rooms, downflowhoods) where the HEPA filters are located at theboundary of the air supply entering the cleanroom or work area. In order to dilute the pinholeleak with the greater volume of clean airpassing through the filter, either a considerabledistance or some method of agitating such as abaffle would be pointless in a laminar flow cleanroom. Therefore, it could happen that theproduct or process requiring particulate free airduring its manufacture or assembly could belocated directly downstream of a pinhole leak.

Realizing this early researchers into clean roomtechniques developed a procedure to scan orprobe the downstream face of a bank of filtersin a laminar flow system, not only to locatepinhole leaks in the filter element, but todetermine whether the filters were sealed to theirmounting frames. A challenge aerosol, with aparticle size range of 0.1 to 3.0 micrometers(polydispersed), is generated and introducedupstream of the filter bank while the system isin operation. The downstream side is probedwith a por table forward light scatteringphotometer. Pinhole leaks and filter-to-frame areidentified and patched.


8

Figure 3: The “Cold” Aerosol Test -- the entire filter face is scanned for pinhole leaks.

Photometer

Procedures Manual

HEPA Filters

Scanning Probe

Probe Tubing

Hood

Anemometer

Aerosol Generator

HEPA filter manufacturers, confronted with theprospect of failing a field test that could locatedefects escaping detection in the overallefficiency test with the Q 107 Penetrometer,began to factory probe filters destined forlaminar flow clean rooms. In time, this additionaltest requirement became an industry standard.The Type C filter, per IEST-RP-CC001.3is “One that has been tested for overallpenetration. . .and in addition has been leaktested using air-generated challenge aerosolsmoke. . .” This is the equivalent of the FlandersLaminar Flow Grade HEPA Filter.

As shown in Figure 3, there are three majorcomponents used to perform the cold challenge

aerosol test; the challenge aerosol generator,the test box (plenum) with motor/blower and thelight scattering photometer. (The vernaculardescription cold challenge aerosol test frequentlyis used to distinguish between the polydispersedDOP aerosol generated at ambienttemperatures and the thermally generatedmonodispersed aerosol.)

In this case, the challenge aerosol is generatedby forcing air at 20-25 psi into liquid challengeaerosol contained in a reservoir. A sufficientchallenge of 10-20 micrograms per liter can bemaintained by using one Laskin nozzle per 500CFM of air or increment thereof.


9Figure 4: Laskin Nozzle

A single Laskin nozzle is illustrated in Figure 4.There are two sets of holes in the nozzle, oneset of four holes is located directly beneath thecollar around the bottom of the tube. Thesecond set of four is located in the collar witheach hole being positioned directly above thecorresponding hole at the tube. The air flowingout of the holes in the tube causes the challengeaerosol oil to be drawn through the holes in thecollar, fragmenting the liquid into an aerosol.Unlike the homogeneous, monodispersedparticles generated by the hot challengeaerosol test, the cold challenge aerosol isheterogeneous of polydispersed having aparticle size distribution as follows:

99% less than 3.0 micrometers95% less than 1.5 micrometers92% less than 1.0 micrometers50% less than 0.72 micrometers25% less than 0.45 micrometers10% less than 0.35 micrometers

Although test plenums vary somewhat in sizeand design, the arrangement shown in Figure 3is typical of the type used at Flanders. Theessential purpose of the plenum is to mix anddisperse the air/aerosol upstream of the filter toprovide a uniform challenge to the filter. Animportant feature of the test equipment isthe hood or baffle that is located on the air-leaving side of the filter. This device preventsthe intrusion of particles from the room air ontothe downstream face of the filter and isessential to obtain valid results. During the test,the filter is clamped in place between the hoodand the test plenum. In older photometers, theoperator set the needle of the meter to read atthe zero point while holding the probe at thefilter face and sampling the effluent from thefilter. (Current photometers contain their ownfilter for setting the zero reading, but there is nospecification requiring their use.) Next, anupstream reading is taken through an orificein the plenum upstream of the filter. If thechallenge is insufficient, an adjustment is madeby increasing the air pressure into the genera-tor and checking the upstream concentration

reading until the correct limit is attained.

The photometer probe is connected by flexibletubing to the intake of the light-scatteringchamber of the photometer. To test the filter,the operator scans the perimeter of the packand then, using slightly overlapping strokes,probes the entire face of the filter. Mostphotometers sample at 1.0 CFM. Air is drawnthrough the chamber and any entrained particlespresent in the sampled air deflect the lightsource onto the sensitive area of the photomul-tiplier tube. This causes the needle on the meterto move, indicating the size of the leak by themeter reading. If the photometer reading isgreater than .01%, the leak is unacceptable andthe spot must be repaired. Thus, a leak may notpass smoke in greater proportion than 1:10000relative to the clean air that surrounds it.

Scan tested filters are frequently and errone-ously described as “zero probe” or “99.99” filterswith the inference that they have a higherefficiency rating than the minimum efficiency of99.97% required for Industrial and NuclearGrade filters.


10

Figure 6: Flanders Laminar Flow Grade Filter Label - Type C Filter Indicating that the filter has beentested for efficiency and has been scan tested.

Figure 5: Challenge Aerosol Generator

Manufacturers who do not own a Q 107Penetrometer to perform the overall efficiencytest depend upon this misinformation to justifythe minimal expense required to own theequipment required to perform cold DOP test-ing only. The probe test is described as a morestringent test with the implication that it is there-fore better, when it is, in fact, unrelated to theoverall efficiency test. At best, the probe test isa supplement, not a substitute, to the overallefficiency test. As described above, theprocedure for probe testing includes setting themeter at zero while sampling the effluent fromthe filter being tested. This procedure could befollowed just as easily using a 95% efficientfilter! As stated above, some photometers arenow equipped with HEPA filters which are usedas the reference filter, but there is no industry-wide specification requiring their use.

A HEPA filter performance rated at 99.99% oncold DOP is one that has no pinhole, cracks orimperfections showing an indicated leak greaterthan .01% at specific location relative to theupstream concentration. It is pointless tocompare this test to the overall efficiency ratingobtained with the hot DOP test since there areso many differences, including the particlesize(s) and concentration of the challenge. Afilter which has passed the scan test can havean overall efficiency of 99.97%.

With a multiplicity of Laskin nozzles as shown. Thegenerators used to test individual filters at Flandershave at least three nozzles. More are required totest larger filter systems.

11

HEPA Filters and Filter Testing: Two-Flow Efficiency Testing & Encapsulation

Two-Flow Efficiency Testing and EncapsulationPinhole leaks in filter media result in greaterpenetration at lower velocities because theconstriction of air flow through a pinhole is afunction of the square of the air velocity when-ever the constriction of air flow through the filtermedia is close to a linear function of velocity.Therefore, at 20% or 1/5 of full test flow, apinhole leak shows up as approximately 25 timesgreater in proportion to total flow than it does atfull flow. Also, at higher velocities particlesimpact upon the fibers of the filter elementwhereas at lower velocities Brownian motioncauses them to meander and they are morelikely to “find” the leak.

The development of acceptance criteria for cleanroom components resulted in a greater aware-ness of the existence of pinhole leaks in HEPAfilters. This, in turn, led to a reevaluation ofthe filter test procedures for filters used forradioactive containment. Prior to the advent ofcommercial nuclear power stations, most ofthese filters were used either by the U.S.weapons program or in the field of nuclearresearch. The U.S. Department of Energy(formerly the Atomic Energy Commission)operates, through prime contractors, afilter test facility equipped with Q 107Penetrometers. HEPA filters purchased forweapons and nuclear research facilities areretested en route to these plants.

Not wishing to commit additional time to scantest HEPA filters at the retest stations or to raisethe purchase price for additional factory testing,a two-flow test was adopted wherein the HEPAfilters are tested at the flow rate specified inASME-AG-1, Section FC, HEPA Filters andagain at 20% of that flow rate. The 20% flowtest served to detect any gross pinhole leakageescaping notice in the full flow test. The test isnot effective in detecting all pinhole leaks nordoes it enable the operator to locate them, but ithas been found as an effective device to aid inimproving overall filter performance.

When the 20% flow test was added to theprocedure, a second modification was alsomade: the addition of the encapsulation hood,shown in Figure 8. Previously, filters tested forefficiency had only the element and the frametested. Experience has shown that HEPA filterscan have frame leakage, caused primarily byimproper sealing methods during manufacture,racking of the frame or leakage through theframe material (metal frame filters are particu-larly susceptible). By enshrouding the entirefilter, any leakage through the frame, joints orcorners is included in the overall efficiency ofthe filter.

Figure 7: Flanders Nuclear Grade Filter Label - Type B Filter Indicating that the filter has been tested forefficiency at two flows while encapsulated.

HEPA Filters and Filter Testing: Two-Flow Efficiency Testing & Encapsulated

12

Two-Flow Efficiency Tested, Encapsulated and Scan Tested FiltersSpecification requires filters used in air clean-ing systems involving chemical, carcinogenic,radiogenic, or hazardous biological particles begiven both a scan test (as is given to Type C,Laminar Flow Grade HEPA Filters) and Two-FlowEfficiency Testing and Encapsulation (as is per-formed upon Nuclear Grade HEPA Filters). This

Figure 8: Filter Test Portion of the Q 107 Shown on the left prior to modifications made for encapsulationand on the right following the addition of encapsulation components.

is described as a Type E Filter in the IEST-RP-CC001.3 Recommended Practice. In the early1980s, The National Institute of Health waspreparing specifications for filters used in theseapplications. This specification was planned asMIL-F-51477(EA). It will be described in Flandersliterature as a Bio/Hazard Grade HEPA Filter.

Figure 9: Flanders Biological Grade Filter Label - Type E Filter Indicating that the filter has been testedfor efficiency at two flows while being encapsulated and has been scan tested.

13

The invention of the Single Particle, Particle SizeSpectrometer with a laser light source has madeit possible to measure removal efficiencies onparticles smaller than 0.3 micron size particleswith convenience, accuracy and reproducibility.A system using a calibrated Dual LaserSpectrometer System and a dilution device isused to test the Flanders VLSI® Filter for bothefficiency and resistance to airflow.

A cold challenge aerosol is introduced into thesystem while the Dual Laser Spectrometersamples simultaneously on both the upstreamand downstream sides of the filter. The dilutiondevice permits an upstream challenge which issufficient for verification of filter efficiency. Thissophisticated test system provides the operatorwith efficiency by particle size in thirty-oneslightly overlapping bands from .12 to 3.0

Two-Flow Efficiency Tested, Encapsulated and Scan Tested Filters

Figure 11: System using Calibrated Dual-LaserParticle Counter. A dilution device is used to testthe Flanders VLSI ® Filter for both efficiency andresistance to airflow.

microns. A computerized printer produces ahistogram presenting the efficiency data in bothtabulated and graph form. VLSI® Filters have aminimum efficiency of 99.9995% on .12 micron-size particles. Because the Dual Laser Systemfar exceeds the sensitivity of the Q 107Penetrometer (used to test HEPA filters), it isthe only method which can be used to verify theperformance of these ultra high efficiency filters.

The breakthrough on VLSI® filtration is certainto have far-reaching effects in both containmentand clean room applications in the years tocome. At this time, industry and governmentare working together to develop industrystandards for testing these filters. It is expectedthat the Institute of Environmental Sciences willissue a Recommended Practice within the nextfew years.

Figure 10: System using Two-Flow EfficiencyTest

HEPA Filters and Filter Testing: Two-Flow Efficiency Testing & Encapsulation

Figure 12: Flanders Filter Label - Type D Filter Indicating that the filter has been tested for efficiency andhas been scan tested.

HEPA Filters and Filter Testing: In-Service & In-Place Tests for HEPA Filters

14

The most stringent factory tests for HEPA filtershave resulted from the requirements of boththe nuclear industry and the operators oflaminar- flow clean rooms. Experience hasdemonstrated that HEPA filters that havepassed these tests do not always arrive at theirdestination without mishap. Damage can occurduring shipping and handling by trainedpersonnel. Once installed, the filter-to-mount-ing-frame seal or leaks in the mounting frameitself can contribute to a loss of the efficiency ofthe bank. Consequently, it is not surprising thatthe nuclear industry and those industries requir-ing laminar- flow clean rooms also requireverification of in-service performance of bothNuclear Grade and Laminar-Flow Grade HEPAFilters and their supporting frameworks.

By comparison, Industrial Grade HEPA Filtersare generally used in industries less structuredby either government regulations or industrystandards and usually where the user is notconcerned with conducting additional field test-ing to verify performance. This is not to say thatIndustrial Grade Filters cannot pass certainin-service tests, but they are less likely to do sofor two reasons: First, by electing to not test thefilters in service, the owner and/or the designergenerally issue a less stringent procurementspecification, with the result that the mountingframeworks of filter housings are of mediocrequality, not designed to pass an in-service test.If the owner should decide at a later time toupgrade the HEPA filter installation and addin-service testing, it may be necessary to modifyor replace frames. A major reason for thefailure of these filter banks, assuming thatundamaged filters have been installed, is thefilter-to-frame seal. Bypass leakage haslong been a principal cause of improperlyinstalled filters. This can be compounded byframe leaks in welds or caulking or by poorquality workmanship by the installer of theframes. Second, Industrial Grade Filters arenot probe tested at the factory. Since allspecifications for laminar-flow installations and

In-Service Tests for HEPA Filtersdevices call for probe testing in the field, acertain percentage will fail.

Exclusive of particle monitoring techniques,there are two kinds of field tests used to verifyin-service performance: In-Place Testing, thetesting of filter banks in nuclear air-cleaningsystems; and, Probe or Scan Testing, discussedpreviously, required for clean rooms andlaminar-flow devices.

In-Place Testing — HEPA Filter Banks

HEPA filters installed in nuclear air-cleaningsystems are required to be tested in-servicefollowing each filter change and periodicallyduring the life of the filters. Since eachventilation system differs in design, a standardor uniform test throughout industry is aproblem. Therefore, depending upon thearrangement of a par ticular system, theprocedure can vary. The in-place leak test isfrequently called an efficiency test of anindividual filter.

The test by the factory of quality assurance sta-tion is an efficiency determination using amonodeispersed challenge. The total filter is chal-lenged at one time and a single reading of pen-etration is obtained. ASME, AG-1, Section FC,“HEPA Filters” provides requirements for the per-formance, design, construction, acceptance test-ing and quality assurance for Nuclear Grade HEPAfilters.

In-place field test of installed HEPA filters are madewith a polydispersed aerosol, and do not showthe efficiency of the filter but only reveal the pres-ence of leaks in the filter bank. The in-place fieldleak test is not an efficiency test and should notbe so considered. ASME, AG-1, Section TA, “FieldTesting of Air Treatment Systems” provides pro-cedural guidelines for HEPA filter bank in-placeleak testing.

*The authors have attempted to clarify the differ-ence between factory testing an individual filterfor efficiency using a monodisperse aerosol fromthe in-place field test conducted in service thatuses a polydisperse challenge under a multiudeof varying conditions.


15

Under ideal conditions, and with a well-designedfilter system, an efficiency test on a bank offilters using a cold aerosol challenge can beperformed, but such a system would prove tobe the exception and not the rule. (Figure 13illustrates such a system.) The operator mustfirst generate the cold aerosol challenge by thesame method and apparatus as describedabove for probe testing HEPA filters at thefactory. To gain an accurate reading from thebank of filters being tested, the challenge mustbe uniform across the upstream face of the filterbank. In most cases, this can be achieved byintroducing the challenge aerosol into thesystem at least ten duct diameters upstream ofthe bank to obtain thorough mixing of the air/aerosol. However, a uniform challenge is alsodependent upon a balanced flow through the

bank; therefore, careful attention should begiven by system designers to this requirement.

Cold challenge aerosol is first generated andintroduced into the system. A single sample istaken along a linear plane upstream of the bankto verify uniformity. The upstream concentra-tion required is relative to the sensitivity of theinstrument, but an acceptable minimum for mostphotometers is approximately 50 micrograms perliter. This value is accepted as 100% and asample is taken downstream of the bank andcompared to the upstream reading. The down-stream reading should be taken far enough fromthe bank to allow thorough mixing between thebank and the sample. Usually a minimum of tenduct diameters is sufficient. The penetrationcannot exceed .05%.

Figure 13: Test of Ventilating System with a Single Bank of HEPA Filters


16

In performing an in-place test, the filter bank istreated as a single filter. Just as individualpinhole leaks are diluted by the greater volumeof clean air and escape detection in theefficiency tests performed at the factory, bypassleaks, frame leaks and holes in the medium canescape detection in the in-place test providedthe overall penetration does not exceed .05%.Certain holes that are undetected in large filterbanks could not accurately be described aspinhole leaks. They can be quite large. For thisreason, a visual examination of the filters is agood practice whenever possible.

In-place testing is almost always complicated bysystem design or location. Components suchas prefilters, entrainment separators, gas ad-sorbers and secondary HEPA filter banks arefrequently arranged in close proximity to oneanother and to the bank being tested. The ideal

system described previously and illustrated inFigure 13 has several essential features: afrusto-converging transition piece both upstreamand downstream of the bank and ten ductdiameters on either side of the bank,unobstructed by other system components. Theoperator stands an excellent chance ofobtaining a balanced flow through this systemas well as good mixing both upstream anddownstream of the bank. Figure 14 shows atypical air cleaning system with prefilters, HEPAfilters, carbon adsorbers and a second bank ofHEPA filters in series. The HEPA filters in thefirst bank could not be tested because thoroughmixing of the air/aerosol would not take place ifthe challenge aerosol were injected between theprefilter and the HEPA banks. Further, it wouldbe impossible to take a representative samplein the limited space between the HEPA filter andthe adsorber. In this system, all the components

Figure 14: The Ductwork and Plenums in HVAC SystemsFrequently designed for best use of building space. Filter systems having multiple stages of filters are groupedtogether for the same purpose, but also to facilitate mainenance operations. Unfortunately, this frequentlyresults in HEPA filter banks that are difficult or impossible to test.

HEPA Filters and Filter Testing: Clean Room Testing

17

are located too close to each other for a goodtest and the housing is too small for man entry.Figure 14 also illustrates another commonproblem in filter banks that must be testedin-place: the relationship of the duct andplenum design and its effect upon a uniform flowthrough the filters. Certainly the arrangement,as shown, would have high and low velocitypoints across the face of the prefilter bank, andalthough the prefilters themselves would assistin balancing the flow ahead of the HEPA filters,the only way to test this particular bank of HEPAfilters would be to pull both the prefilters andone bank of HEPA filters out of the system.

Factors that can contribute to the difficulties ofperforming an in-place test are as varied as thearrangements that the system planner candesign. The standards anticipate thesecommon problems and permit compromisesolutions (e.g. Bypass ducting between systemcomponents may be used for filter testpurposes.). In man entry housings where thebanks of HEPAs are too large to generateenough challenge aerosol, a portion of the bankmay be tested by shrouding adjacent sections.Scan testing techniques may also be used,averaging the results to compute a systemefficiency. None of these methods fully solvessuch problems as access into housings that aretoo small or too remote for man-entry or theproblem of uniform challenge where banks inseries are located too close together. For thisreason, in-place testing should be consideredas a quantitative rather than a qualitative test.For many years, Flanders has designed, proof-tested and manufactured housings withcomplete in-place test equipment built into thehousings, permitting testing of filters from out-side the system at a remote location and theidentification of the filter that might be leaking,because an individual efficiency leak testisperformed on each filter in the system. Thebuilt-in test systems are compatible with othercomponents in the housing and approximatelytwo feet is required between consecutive filterbanks, regardless of the size of the system.

Clean Room TestingNuclear and biological facilities usually have theirown safety or health physics personnel andmaintenance crews who can perform in-placetesting, thus enabling them to enforce industryspecifications and design criteria.

The same is not always the case in plantshaving laminar-flow clean rooms or cleanair devices. HEPA filters and ancillaryequipment are not routine components inHVAC system construction, and thetechniques of their installation and in-servicevalidation are unfamiliar to many facilityoperators and their mechanical contractors.Even when a contractor can demonstrateprevious experience in clean room construction,most of his employees are transitory, workingfor various contractors on a per job basis ratherthan for a single employer. As a result, manyHEPA filter systems are incorrectly installedbecause of inexperience. To overcome thisproblem, factory service personnel are availablefrom Flanders to supervise the contractor’sinstallation of the Flanders products duringthe critical construction phase. Once theinstallation is satisfactorily completed, theservice personnel test and certify the job toindustry standards.

The control of airborne particulates within aclean room or laminar flow device is dependentupon the efficiency of the filters and their

Figure 15: Probe Testing Clean Room Ceiling

HEPA Filters and Filter Testing: Clean Room Testing

18

supporting framework, but it is also dependentupon achieving parallel and turbulent free airvelocity patterns through the clean room or zone.The subject herein is the in-service testing andvalidation of the filter banks themselves, but it isimportant to stress that several other critical testsare conducted during the certification procedureby service personnel, including velocity profilesand particle monitoring at work surface levels.

Filters that have first been factory scan testedshould also be tested in service in a mannersimilar to the factory test by challenging theinstalled filter and its supporting frameworkwith a polydisperse challenge aerosolintroduced upstream of the filter bank and byscanning the downstream face of the bank. Asin the testing of nuclear systems in-place, auniform challenge is required. Standard cleanroom design criteria requries laminar-flow roomsand clean air devices to have a face velocityof 90 fpm, ±20 fpm, far below the velocity ofair required in conventional (e.g. non-laminarflow) ventilation systems. Consequently, it is notusually difficult in smaller laminar-flow systemsto generate sufficient amounts of challengeaerosol and to attain uniform disbursementupstream of the filters. However, as the size ofthe system increases, it becomes increasinglymore difficult to challenge the entire banksimultaneously. One alternate method oftesting is to isolate adjacent sections and testthe bank section by section. Where this is notpractical, the filters can be tested at the job siteprior to installation, using the same test rig thatis used at the factory for probe testing. This testverifies that there has been no damage incurredto the filters in transit or during unpacking andhandling. Following this test, the filters areinstalled immediately under close supervision.When all filters have been installed, the sealsbetween the filter and the supporting frameworkitself and the perimeter of the filter bank arescanned for bypass leakage. Considering thetime that it takes to scan a large HEPA filter wallor ceiling bank, there is a valid argument for aless time consuming test.

Notes: _________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

HEPA Filters and Filter Testing: In-Place Test Housings for Efficiency Testing

19

In-Place Test Housings for EfficiencyTesting

The standard in-place test procedure calls for asuitable aerosol to be generated and introducedinto the airstream at a sufficient distance up-stream of the filter bank to thoroughly mix thechallenge aerosol into the airstream before itreaches the filter bank. The recommended mini-mum distance is ten duct diameters upstreamof the bank. If this distance is not available, acollapsible baffle may be used in conjunction withthe injection port to assist in mixing the chal-lenge aerosol. The baffle can be located six orseven duct diameters from the bank.

Samples are taken at points on a linear planedirectly upstream of the filter bank to determineif there is a uniform challenge to the bank. Theuniformity can be affected by incorrect mixingprocedures or by uneven velocity of air throughthe bank. Once the tester is satisfied that thechallenge is uniform, a sample is taken at a pointdownstream of the filter bank, usually ten ductdiameters or more, to ensure that any leakagein the bank is thoroughly mixed with the cleanair passing through the bank.

HEPA filters that are installed in ventilationsystems to prevent the release of hazardous particulate matter into the atmosphere arerequired to be tested in-place following eachfilter change, as well as periodically during theservice life of the filters, to determine theefficiency of the filter bank and its supportingframework. Leakage through the filters, bypassbetween the filters and the supporting frame-work, or leakage through the framework itselfmay reduce the efficiency of the filter bankbelow the required 99.95%.

The in-place procedure appears to be relativelysimple; therefore, very often little considerationis given to in-place testing by the designers ofventilation systems that contain HEPA filters.Filter testing has historically been treated as ahealth physics function and not as a designobjective. Consequently, there are countless

filter banks that have been installed in air clean-ing systems that are impossible to test becausesystem design interferes with attaining mixingand dispersion on both the upstream and down-stream sides of the banks being tested.

To test HEPA filters in-place in the field underideal conditions, an operator has a minimum tenduct diameters upstream and downstream of hisfilter bank. He has a well-designed transitionpiece on the inlet and outlet of his housing andhas no obstructions such as other filters, adsorb-ers, etc., in his system. Figure 13 illustrates abank of filters in such a system. The two essen-tial factors for a satisfactory test in this situationare distance and design. the former ensuresmixing on both sides of the filter bank; whereas,it is the system design—e.g. the transition—thataids in balancing the flow through the bank andresults ina uniform challenge to the bank.

Even with the ideal conditions shown in Figure13, there are still certain impractical limitations:

• If the system, when tested, does not havea minimum efficiency of 99.95%, theoperator is requried to take correctiveaction.

• If the system is large enough forman-entry, a tester must suit up andenter the system on the downstream sideto scan the bank.

Exposure time of maintenance and testpersonnel in containment systems is anincreasing concern in facilities where toxicmaterials are present in the ventilation air.

The Flanders’ in-place test housings use twoidentical mixing devices on the upstream anddownstream side of the filter to achieve the samepurpose. A hinged diffuser is located on boththe air-entering and the air-leaving sides of thefilter.

HEPA Filters and Filter Testing: In-Place Test Housings for Scan Testing

20

Prior to testing the filter, both diffusers are movedto the test position to mix the challenge aerosolinto the airstream on the upstream side and tomix any leakage into the air on the downstreamside. The challenge aerosol is then introducedinto the system ahead of the first diffuser; up-stream and downstream samples are taken andthe results are compared to determine thepenetration through the filter and its supportingframework.

Each test module is designed so that the read-ings are not affected by adjacent filter. Sinceeach filter is tested individually, a penetrationreading for every filter in the system is obtained.

In-Place Test Housings for Scan Testing

The standard in-place scan (or pinhole) testprocedure calls for a suitable aerosol to begenerated and introduced into the airstream asufficient distance upstream of the filter bank tothoroughly mix the challenge aerosol into theairstream before it reaches the filter bank. Therecommended minimum distance is ten ductdiameters upstream of the bank. If this distanceis not available, a collapsible baffle may be usedin conjunction with the injection port to assist inmixing the challenge aerosol.

The baffle may be located six or seven ductdiameters from the bank. Samples are taken atpoints on a linear plane directly upstream of thefilter bank to determine if there is a uniform chal-lenge to the bank. The uniformity can be af-fected by incorrect mixing procedures or whenthe velocity of air through the bank is uneven.Once the tester is satisfied that the challenge isuniform, scan testing can begin downstream ofthe filter bank, through a field fabricated andinstalled access area, typically found in thedownstream ductwork. This arrangement re-quires test personnel to enter or reach into theairstream to find potential leaking HEPA filters.

This type in-place procedure also appears to berelatively simple; therefore, very often little con-sideration is given to in-place testing by the de-signers of ventilating systems that contain HEPA

filters for scan testing. There are filter banksthat have been installed in air cleaning systemsthat are impossible to test because systemdesign often interferes with attaining mixing anddispersion on both the upstream and down-stream sides of the banks being tested.

To scan test HEPA filters in-place under idealconditions, an operator has a minimum ten ductdiameters upstream and access to the down-stream of his filter bank. A well-designedtransition piece on the inlet and outlet of thehousing, with no obstructions such as otherfilters, adsorbers, etc., in the system, arerequired. The outlet transition must have ac-cess to allow scan testing of each HEPA filter.

Exposure time of maintenance and testpersonnel in containment systems is anincreasing concern in facilities where harmfultoxic and biological materials are present in theventilation air.

The Flanders PrecisionScan test housing usestwo devices on the upstream and downstreamside of the filter to achieve the same purpose.Access for scan testing is located on the air-leaving side of the filter.

Prior to testing the filter, the inlet diffuser is closedto mix the challenge aerosol into the airstreamon the upstream side. The challenge aerosol isthen introduced into the system ahead of thediffuser, an upstream sample is taken and theresults are determined by scanning the filter facearea and its supporting framework.

Each test housing is designed so the readingsare not affected by the adjacent filter. Becauseeach filter is tested individually, a penetrationreading for every filter in the system is obtained.

HEPA Filters and Filter Testing: Conclusion

21

ConclusionFigure 16 illustrates one of the many kinds oftests performed at the factory and in the field.Most filters purchased by the Department ofEnergy are shipped directly to the governmentretest facility where they are given a secondefficiency test with a Q 107 Penetrometer andare then reshipped to the buyer. Unfortunately,there have been no shortcuts devised to verifythe performance of HEPA filters, either at thefactory or in the field,. Typically, these filters aresusceptible to damage during shipping andhandling unless stringent precautions are taken.

Historically, HEPA filters have been difficultto seal into their supporting framework or hous-ings, a fact which inspired Flanders to developthe Filter-to-Frame Fluid Seal in the 1960s. Thisseal is now known as the gel seal, and has beenused in all applications of air filtration since itsdevelopment. Finally, they are unfamiliar itemsto most mechanical contractors and construc-tion workers. The matter-of-fact approach usedin the installation of standard building compo-nents should not be applied when installingHEPA filters.

Figure 16: Factory Test Specifications, Field Test Specifications, Applications for HEPA and VLSI ®

Filters

Represented by:

Flanders Corporation

Important Notice

For best results in the application of Flanders products, it is recommended that thebuyer supply complete information about the operating conditions of theventilation system to Flanders for prior evaluation. Flanders does not guarantee that itsequipment will operate at the performance levels given on the identification labels or inthe catalog specifications under all conditions of installation and use, nor does Flandersguarantee that suitability of its product for the particular end use which may becontemplated by the buyer. When the system components are supplied to the buyer orhis agent for final installation and assembly in the field, it should be under the supervisionof factory trained personnel who are equipped to test the installation and certify its performance and conformance to industry accepted specifications. Failure to follow theseprocedures may result in a compromised installation.

Flanders CorporationCorporate HeadquartersSt. Petersburg, FL 33713

Technical Inquiries:531 Flanders Filters RdWashington, NC 27889

Representatives of Flanders Corporationare located throughout the world.

Your closest representative’s office may befound by contacting our manufacturingand sales department.

Tel: 252-946-8081 Fax: 252-946-3425Email: [email protected] site: www.flandersffi.com Printed in USA 05/2004

The foremost designer and manufacturer of high efficiencyair filtration systems for science and industry.

Flanders Corporation continues to research and develop product improvements and reserves the right to change product designs and specifications without notice.

®

FOREMOST IN AIR FILTRATION

Documents

FOREMOST IN AIR FILTRATION - pdfs.semanticscholar.org€¦ · HEPA Filters and Filter Testing ... Note: Either of the two test methods or an alternative method may be used for filter