09/19/1983 VOC110919831

Category: 11 – Stage I Vapor Recovery Roof Tanks

MEMORANDUM

SUBJECT: Status of Several New EPA Reference Test Methods for Measurement of VOC Leaks DATE: September 19, 1983

FROM: Nancy D. McLaughlin, Field Testing Section Emission Measurement Branch, ESED (MD-13)

TO: John Hannish Air Programs Branch, EPA Region I

The purpose of this letter is to inform you about the status of severalnew Environmental Protection Agency reference test methods for measurement ofvolatile organic compound emissions and leaks, and the new source performancestandards (NSPS) for bulk gasoline terminals. These test methods and standardshad been proposed on December 17, 1980. During the past months, you hadrequested information from me relating to this subject.

The final rulemaking for the NSPS for bulk gasoline terminals waspublished in the Federal Register on August 18, 1983 (48 FR 37578). Alongwith this, the final versions of six new reference methods were published(Methods 2A, 29, 25A, 25B, 27, and 21). For your information, I am attachingcopies of these Federal Register notices.

If you have any further questions regarding this area, please contact meat (919) 541-5543.

2 Attachments

Attachment

ERRATA SHEET

Federal Register Notice:Standards of Performance for NewStationary Sources; Bulk Gasoline Terminals

1. "is available"

2. Delete sentence.

3. "75,700"

4. "Summary of Environmental, Energy, and Economic Impacts

5. "same basic control device"

6. "NSPS"

7. "Section 60.15"

8. "replaced"

9. "CA and TO"

10. "limit, while"

11. "include"

12. "Bulk Gasoline Terminals"

13. "60.8(a)"

14. "during the performance test."

15. K = density of calibration gas, mg/m 3, at standard conditions

= 1.83 x 10 6 for propane

= 2.41 x,10 6 for butane

16.

17. "60.502 (e)(1)"

18. "=2 percent"

19. Insert "Where" before definition of terms.

EM

L

e1

n

= =∑ ii

2

ERRATA SHEET (continued)

20. " Ym " _21. " Ym " _22. Vms = O.3858 Ym (Vmf - Vmi) (Pb + Pa) / Tm

23. Qs = Vms/Theta

24. "1. Applicability and Principle"

25. "Organic Analyzers (2)."

25. "(stainless steel or equivalent)"

27. "HCi"

28. K = Calibration gas factor

= 2 for ethane

29. Theta = sample run time, min.

30. Parenthetical statement should read "(Instead, the CO2 concentrationin the ambient air nay be measured during the test period usingan NDIR.)"

31. "concentrations of"

32.

33. "analyzer"

34. "to the analyzer"

35. "to the analyzer"

36. "+/- 2 percent"

37. Delete parenthetical remark.

38. "7. Emission Measurement Test Procedure"

39. "The apparatus is the same"

40. Delete last word of the line "are".

V VK(HC )

K(HC ) CO CO 300es isi

e 2e e

=+ + −

⎛

⎝⎜

⎞

⎠⎟

3

ERRATA SHEET (continued)

41. "calibration gas value."

42. "6.2. Location of Sample Probe."

43. "Same as in Method 25A, Section 6.5."

44. First sentence of Section 5.1.2 should not be in italics.

45. "+/- 12.5 mm H2O"

46. " delta v"

Wednesday,December 17, 1980

Part III

EnvironmentalProtection AgencyStandards of Performance for NewStationary Sources; Emissions Limitationof Volatile Organic Compounds FromGasoline Tank Truck Loading Racks atBulk Gasoline Terminals

I I

M

_

m m

|

m

83126 Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

ENVIRONMENTAL PROTECTIONAGENCY

40 CFR Part 60

[AD-FRL-1634-4]

Standards of Performance for NewStationary Sources; Bulk GasolineTerminalsAGENCY: Environmental ProtectionAgency (EPA)., ACTION: Proposed Rule and Notice ofPublic Hearing.

SUMMARY: The proposed standardswould limit emissions of volatile organiccompounds (VOC) from new, modified,and reconstructed gasoline tank truckloading racks at bulk gasoline terminals:

The proposed standards implementSeciori 111 of the Clean Air Act and arebased on the Administrator'sdetermination that bulk gasolineterminals contribute significantly to airpollution that may reasonably beanticipated to endanger public health orwelfare. The intent is to require new,modified, and reconstructed bulkgasoline terminals to use the besttechnological system of continuousemission reduction, considering costs,non-air quality health, andenvironmental and energy impactswhich has been adequatelydemonstrated.

A public hearing will be held to.provide interested persons anopportunity for oral presentation ofdata, views; or arguments concerningthe proposed standards.DATES: Comments. Commehts must bereceived on or before February 17, 1981.

Public Hearing. A public hearing willbe held on January 21, 1981 (about 30days after proposal) beginning at 9 a.m.

Request to Speak at Hearing. Personswishing to present oral testimony mustcontact EPA by January 14, 1981 (1 weekbefore hearing).ADDRESSES: Comments. Commentsshould be submitted (in duplicate ifpossible) to: Central Docket Section (A-130), Attention: Docket Number A-79-52, U.S. Environmental ProtectionAgency, 401 M Street, SW, Washington,D.C. 20460.

Public Hearing. The public hearingwill be held at E.R.C. Auditorium, R.T.P.,•North Carolina 27711. Persons wishingto present oral testimony should notifyMrs. Naomi Dur Kee, EmissionStandards and Engineering Division(MD-13], U.S. Environmental ProtectionAgency, Research Triangle Park, NorthCarolina 27711, telephone number (919)541-5271.

,Background Information Document.The-Background Information Document(BID) for the proposed standards may beobtained from the U.S. EPA Library(MD-35), Research Triangle Park, NorthCarolina 27711, telephone (919) 541-2777. Please refer to "Bulk GasolineTerminals-Background Information forProposed Standards," EPA-450/3-80-038a.

Docket. Docket No. A-79-52,containing supporting information usedin developing the proposed standards, isavailable for public inspection andcopying between 8:00 a.m. and 4:00 p.m.,Monday through Friday, at EPA'sCentral Docket Section, West TowerLobby, Gallery 1, Waterside Mall, 401 MStreet, SW, Washington, D.C. 20460. Areasonable fee may be chargedd'orcopying.FOR FURTHER INFORMATION CONTACT:Ms. Susan R. Wyatt, Emission Standardsand Engineering Division (MD-13), U.S.Environmental Protection Agency,Research Triangle Park, North Carolina27711, telephone number (919) 541-5477.SUPPLEMENTARY INFORMATION: ABackground Information Document hasbeen prepared that contains informationon tank truck loading operations at bulkgasoline terminals; the available control

. technologies for VOC emissions; andanalysis of the environmental, energy,

.economic, and inflationary impacts of-regulatory alternatives. The informationcontained in this document issummarized in this preamble. Allreferences used for the informationcontained in the preamble can be foundin this document.

Proposed Standards

The proposed standards would limitvolatile organic compound (VOC)emissions from new, modified, andreconstructed gasoline tank truckloading racks at bulk gasoline terminals.Specifically, the proposed standardswould require the installation of vaporcollection equipment at the terminal forthe purpose of collecting the VOCemissibns displaced during loading ofliquid product into gasoline lrank trucks,and would limit these emissions fromthe collection system to 35 milligrams ofVOCs per liter of gasoline loaded.

Additionally, a termifial owner oroperator would be required to restrictgasoline tank.truck loadings to thosetank trucks which had passed an annualvapor-tight test. Written documentation,in the form of tank truck test resultswould be kept on file at the bulkgasoline terminal in a permanent formavailable for inspection.

Five new Reference Methods areproposed with these standards to

measure vapor processor outlet VOCmass emissions, and to test gasolinedelivery tanks for vapor tightness.Methods 2A and 2B measure gas flowrates in pipes and small ducts, and invapor incinerator exhausts, respectively,Methods 25A and 25B measure VOCconcentration by two detectionmethods. Method 27 is a pressure/-vacuum (vapor-tight) test for gasolinedelivery tanks. Terminal vapor handlingequipment would be monitored for leaksprior to each performance'test usingMethod 21, which has been proposedwith Standards of Performance for VOCFugitive Emission Sources in theSynthetic Organic ChemicalsManufacturing Industry.

Summary of Environmental, Energy, andEconomic Impacts

The proposed standards would reducethe projected nationwide 1985 VOCemissions from affected facilities byabout 6,600 megagrams per year, or 70percent.

Emissions of carbon monoxide andoxides of nitrogen from thermaloxidation systems would total up to 10Mg/yr and 4 Mg/yr, respectively, Itn thefifth year of the standards. Thisrepresents a relatively small airpollution impact.

Water is not used as a direct controlmedium by any of the available controltechniques. Existing separation and *handling systems could accommodatethe small amount ofwastewaterdischarged by some types of controlprocessors. The proposed standardswould have a negligible impact on waterquality.

Because all of the VOC emissions areincinerated or returned to storage asliquid product, there would be no directsolid waste impacts under the proposedstandards. Some solid waste would begenerated indirectly due to disposal ofactivated carbon from carbonadsorption units after the useful life ofthe carbon had expired. Even the worstcase situation would produce minimalimpacts on solid waste.

The proposed standards would havenegligible impacts on noise, spacerequirements, and availability offesources.

Because all of the available vaporprocessors, except the thermal oxidizer,recover energy in the form of gasoline,the proposed standards would result Ina net energy savings equivalent toapproximately 9 million liters (2.4million gallons) of gasoline per year inthe fifth year of the standards.

The proposed standards would resultin a total nationwide capital cost forVOC control during the first five yearsafter the effective date of the standards

Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

of about $25.3 million. The proposedstandards would also result in a totalnationwide annualized cost in the fifthyear of about $4.3 million.

Under the worst case situation, themaximum increase in the retail price ofgasoline resulting from the proposedstandards would be less thaii 0.6 percentdue to bulk terminals and less than 0.7percent due to the independent tanktruck industry.

Rationale

Selection of SourceThe EPA Priority List (40 CFR 60.16,44

FR 49222, August 21,1979) lists, in orderof priority for standards development,various source categories in terms ofquantities of nationwide pollutantemissions, the mobility and competitivenature of each source category, and theextent to which each pollutantendangers health and welfare. ThePriority List reflects the Administrator'sdetermination that emissions from thelisted source categories contributesignificantly to air pollution that mayreasonably be anticipated to endangerpublic health or welfare, and is intendedto identify major source categories forwhich standards of performance are tobe promulgated. PetroleumTransportation and Marketing isincluded as Number 23 on the PriorityList.

Bulk gasoline terminals are animportant part of the gasoline deliverychain and are usually the first linkbetween the refinery and the ultimateend-user. There are presently anestimated 1,511 bulk terminals handlinggasoline in the U.S. terminals typicallyreceive gasoline from the refinery bypipeline, ship, or barge, store thegasoline in large aboveground tanks,and redfstribute the gasoline to smallerfacilities in the marketing chain (i.e.,bulk plants and service stations).Gasoline is loaded into delivery tanktrucks at-the terminal loading racks andis transported to the next link in thedelivery chain. A typical bulk gasolineterminal has a gasoline throughput of950,000 liters (250,000 gallons) per day,three loading rack positions for gasoline,and four aboveground tanks for gasolinewitha combined storage capacity of24,000 m3 (150,000 bbl). Bulk gasolineterminals are normally found in or ,around urban areas since the demandfor gasoline is higher in these locations.

It is estimated that ten new terminalswill be built in the next ten years. Thisrelatively small growth rate is areflection of the small increase in •gasoline consumption projected for thenext ten years. Current industry trendsare toward consolidation of existing

terminals rather than the construction ofnew terminals. Estimates, based uponan industry survey, indicate that theremay be as many as 100 modified andreconstructed sources In the next tenyears.

Gasoline loading racks at terminalscurrently contribute approximately300,000 megagrams per year (Mg/year)of VOC emissions, which Isapproximately 2 percent of the totalnationwide VOC emissions. After fullimplementation of proposed Stateregulations on bulk gasoline terminals,expected by 1982, total VOC emissionsfrom loading racks are expected to bereduced to about 140,000 Mg/year.

Selection of Pollutants and AffectedFacilities. VOC is the only pollutant which isemitted during the loading of liquidproduct into tank trucks at bulk gasolineterminals. Consequently, the proposedstandards would regulate only VOCemissions from the loading operations atterminals.

Volatile organic compounds are anyof the organic compounds thatparticipate in atmosphericphotochemcial reactions. Ozone,produced in these reactions, results in avariety of adverse impacts on healthand welfare, including impairedrespiratory function, eye irritation,necrosis of plant tissue, anddeterioration of certain materials, suchas rubb6r. Further information on theseeffects can be found in the U.S.Environmental Protection Agency (EPA)document entitled "Air Quality Criteriafor Ozone and Other PhotochemicalOxidants" (EPA-00/6.78-04).

The two major sources of VOCemissions at terminals are the storagetanks and the tank truck loading racks.Storage tanks are currently regulatedunder Federal standards (40 CFR 60,Subpart Ka-Standards of Performancefor Storage Vessels for PetroleumLiquids). Those standards cover storagetank emissions caused by atmosphericchanges (breathing losses) andemissions due to filling and emptyingthe storage tank (working losses).Therefore, storage tanks would not beregulated by these proposed standards.

Loading racks consist of the piping,pumps, meters, and loading arms thatare necessary to transfer liquidpetroleum products from storage tanksto delivery tank trucks. Emissions fromthe loading racks are generated duringthe loading of liquid product intodelivery tank trucks when the liquidproduct being loaded displaces VOCvapors contained in the delivery tanks.These vapors consist of evaporatedgasoline components which fill the air

space above the liquid product. VOCemissions from the loading operationcan vary due to the loading method(splash loading causes more emissionsthan submerged loading) and due to theVOC concentration of the vapors in thedelivery tank truck prior to loading. ThisVOC concentration can varysignificantly depending upon the type ofproduct carried, temperature, pressure,vapor tightness of the delivery tank, andwhether vapors were transferred backto the delivery tank when the last loadof liquid product was unloaded (vaporbalanced).

Loading rack facilities in the bulkterminal industry can vary videly in thetypes and quantities of productshandled. In addition to gasoline, largequantities of fuel oil, diesel, and jet fuelmay be handled by a gasoline terminal.The amount of each product handled isdue to the different demands for eachproduct in the vicinity of the terminal.VOC emisisons from fuel oil, diesel, andjet fuel are very small compared to thosefrom gasoline. Consequently, only VOCemissions from gasoline would becovered by the proposed standards.

At many terminals, "switch loading"of delivery lank trucks is practiced.Switch loading involves the transport, ina single tank compartment onsuccessive deliveries, of variousproducts in addition to gasoline.Gasoline vapors can be displaced eitherby incoming gasoline or by any otherliquid product when a previous load ofgasoline left vapors idthe delivery tank.As an example, fuel oil loaded into atank compartment which had carriedgasoline on the previous load woulddisplace gasoline vapors, and thusproduce VOC emissions. For thepurposes of the proposed standards, thedelivery vehicle in both cases is referredto as a "gasoline tank truck."

Because gasoline vapors can beemitted from a tank truck loading aproduct other than gasoline,vswitchloading was taken into account indesignating the affected facility to beregulated under the proposed standards.Consequently, the proposed standardswould affect both the loading of gasolineinto delivery tanks and the loading ofany liquid product into delivery tankswhich contain gasoline vapors. Anydelivery tank carrying gasoline on theimmediately previous load is assumed tocontain gasoline vapors. The costs ofcontrolling switch loading facilities(loading racks) are not significantlygreater than the costs to control onlygasoline lo~ding racks. Since the samevapor processor is used to control all ofthe loading racks, the primary additionalcost would be for the vapor piping

83127

83128 Federal Register /'Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

connecting each loading rack to themain vapor line to the processor. Insome cases, d larger, more expensiveprocessor might be specified in order tohandle increased vapor flow, but theadditional cost would amount to a smallpercentage of the total cost. Many bulkterminals are already controlling allloading racks which have the potentialto displace gaoline vapors.

Since only small quantities of gasoline(less than 2 percent) are delivered byrail cars, vapor controls on theseloadings were not investigated, and theproposed standards would apply only tothe loading of liquid product intogasoline tank trucks.

For the purposes of the proposedstandards, gasoline is defined as anypetroleum distillate or petroleumdistillate/alcohol bled with a Reidvapor pressure of 27.6 kilopascals C4pounds per square inch) or greaterwhich is used as a fuel for internalcombustion engines. The addition of thedistinction for petroleum distillate/alcohol blend in the definition is toinclude gasohol fuels which haveexperienced increased consumption in

-recent months.Because vapor leakage from the tank

trucks being loaded can represent asignificant proportion of the total bulkterminal VOC emissions occurringduring liquid product loading, theproposed standard has been written tocontrol such vapor leakage as well asemissions from the loading racks. EPSbelieves that Section 111 wouldauthorize this regulation to take any oneof the three alternative forms describedbelow; the Agency solicits comments onall issues associated with eachapproach.

Under the first approach, thestandards would apply only to the bulkterminal. The terminal owner oroperator would be required to use vaporcollection equipment on loading racksservicing gasoline tank trucks, and torestrict loadings to vapor-tight tanktrucks. The affected facility under thisapproach would include only the loadingracks servicing gasoline tank trucks.Operators of gasoline tank truckswishing to load-at the teminal wouldneed compatible loading and vapor.recovery equipment, and vapor-tightdelivery tanks. This approach wouldconsolidate responsibility for controllingemissions without resulting in anexces'sive burden for the terminal owneror operaftor.

Under the second approach, standardswould apply directly to both theterminal and the tank trucks. Thestandards would require the terminalowner or operator to install vaporcollection equipment and the tank truck

operatbr to have compatible equipmentand vapor-tight tank trucks. Under thisappproach, the affected facility wouldconsist of the combination of the loadingrack and the truck-mounted tank, with asingle stafidard'covering the hybridloading rack/tank facility. The secondapproach could result in several ownersor operators (of the terminal and of thetank trucks)'at the'same terminal beingregulated under a single standard. Thiscould create enforcement difficultiesand problems in determining liability.

The third approach would involvedesignating two affected facilities, oneconsisting of the loading racks servicinggasoline tank trucks and the otherconsisting of the truck-mounted tanks,and applying a separate standard toeach facility. It would not be practical todirectly regulate gasoline tank trucksunder a separate standard because theVOC emissions being regulated occuronly during product loading at theterminal; and a situation of twostandards regulating the same source ofemissions would result. Furthermore, inthe case of new tank trucks loading atan existing uncontrolled bulk terminal,only the tank trucks would be regulated,and VOC emissions would be displacedto the atmosphere uncontrolled since theterminal would have no vapor collectionor control equipment to process the -vapors. Thus, separate standards wouldnot be effective in these circumstances.

After considering the issues involvedwith each of these approaches todesignating the affected facility, theAdministrator selected the firstapproach as the most practicaldesignation. This places directresponsibility under the proposedstandards on the owner or operator ofthe bulk terminal only, eliminates thepotential for enforcement problemsassociated with an impermanentaffected facility under the secondapproach, and eliminates the situationof regulating the same operation withtwo standards under the third approach.

The selected approach, whichconsiders only the bulk terminal loadingracks. as the affected facilitydesignation, presents severalpossibilities, Two potential affectedfacility designations considered underthis approach were (1) each individualloading rack, and (2) the combination ofall the loading racks at the terminalwhich service gasoline tank trucks.

In choosing the affected facility, EPAmust decide which piece or group ofequipment is the'appropriate unit (the"source"] for separate emissionstandards in the particular industrialcontext involved.'The Agency must dothis by exaniiing the situation in lightof the tei'ins andp iurose of Section I1n.

One major consideration in thisexamination is that the use of anarrower designation results in bringingreplacement equipment under the NSPSsooner. If, for example, an entire plant Isdesignated as the affected facility, nopart of the plant would be covered bythe standard unle'ss the plant as a wholeis "modified" or "recofnstructed." If onth e other hand, each piece of equipmentis designated as the affected facility,then as each piece is replaced, thereplacement piece will be a new sourcesubject to the standard. Since thepurpose of Section 111 is to minimizeemissions by application of the bestdemonstrated control technology at allnew and modified sources (consideringcost, other health and evxironmentaleffects, and energy requirements), thereis a presumption that a narrowerdesignation of the affected facility ispropei. This ensures that new emissionsources within plants will be broughtunder the coverage of the standards asthey are installed. This presumption canbe overcome, however, if the Agencyconcludes either that a) a broaderdesignation of the affected facilitywould result in greater emtssiondreduction than would a narrowdesignation; or b the other relevantstatutory factors (technical feasibility,cost, energy, and other environmentalimpacts) point to a broader designation.The application of these factors isdiscussed below.

While selection of a narrowerdesignation of affected facility generallyresults in greater emissions reduction byearlier coverage of replacementequipment, it appears that a broaderdesignation would result in greateremissions reduction in the bulk gasolineterminal industry. Replacement ofexisting racks in not expected to occurto any great extent, because properly,maintained racks do not generallyrequire replacement. In other words, theisolated replacement of a single rackdue to deterioration of that rack isexpected to occur rarely. Rather, EPAprojects that terminals will concentrateon additions of new racks to sets ofexisting racks rather than replacementof existing racks. EPA further projectsthat if replacement does occur, it willinvolve, a major change in the racksystem (such as conversion from top tobottom loading) and will'involve most orall of the racks at the terminal ratherthan just one rack. The reasons that atotal racks affected.facility designationis expected to result in greater emissionreduction than a single rack affectedfacility designation, in the situationsdescribed above, are explainedin thefollowing paragraphs.

83129 'Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

A modification, under 40 CER 60.14, isany physical or operational change to anexisting facilit which produces a netincrease in the emission rate from thatfacility. If a new rack were added to aterminal it would be an affected facilityunder a single rack designation, andonly that rack would be covered. Undera total racks designation, the addition ofa single new rack could result in amodification, in which case all of theracks would become an affected facility,resulting in greater emission reductionunder this designation, Even if theaddition of the new rack did not resultin a modification because there was noincrease in emissions (due to partialcontrol, for example), the total racksdesignation would still result in lessemissions. This is because the singlerack designation would still result in'asmall incremental emissions increaseeven if the rack were controlled.

In addition to modification, anexisting facility could becomereconstructed, under 40 CFR 60.15, if thefixed capital cost of replacing -components at that facility exceeded 50percent of the cost of a comparableentirely new facility. Under a single rackdesignation, this cost figure could beattained sooner for a given rack than itwould under a total racks designation,since total replacement cost for parts fora single rack would be less than for allthe-racks, and50 percent of the cost fora single new rack would be less than 50percent of the total cost for all newracks. However, under a total racksdesignation, all the racks at a terminalcould become affected facilities if theconversion cost exceeded 50 percent ofthe cost needed to build all new racks;although more racks would have to beconverted to attain this cost, more rackscould eventually be covered sooner thanthey would be under a single rackdesignation. Multiplerack conversionprojects of this type are the most likelytype of replacement at bulk terminals.

Whether the total racks designationwould actually result in more emissioiireduction than the single rackdesignation is somewhat uncertain. Thedesignation which would result in themost emission reduction depends ondecisions made by a terminal owner oroperator when replacing racks or addingnew racks to the terminal. It is difficultto accurately forecast what thesedecisions will be. For example, underthe single rack designation, even if onlyone rack had to be controlled, theterminal owner or operator may elect tocontrol all racks-instead of one rack.since it is common industry practice tocontrol all racks with one controldevice. If this occurred, the single rack

desiiation could result in control of allracks just as would the total racksdesignation.

In summary, considering that theaddition of new racks and the multiplereplacement of racks are expected to bemore likely occurrences In this industrythan single rack replacement, and thefact that more racks would come underthe standards under the total racksdesignation than under the single rackdesignation in these cases, It isprojected that the total racksdesignation would result in the greatestemission reduction. However, as statedpreviously, this depends on decisionsmade by a terminal owner or operator atthe time of construction.

After projecting that the total racksdesignation would result in the greatestemission reduction, the reasonablenessof the cost of this designation was thenevaluated. For terminals which do notalready have control devices (most newand existing terminals in attainmentareas), examination of the cost data hasindicated that the affected facilitydesignation of each individual rackwould generally result in lower capitalcosts than the designation of all theloading racks. However, the netannualized cost would be lower for thetotal racks approach, assuming that theterminal elected to use a control systemother than thermal oxidation, because ofthe greater liquid recovery cost creditsassociated with controlling all of theloading racks. For example, at anaffected terminal the capital cost toinstall controls on one loading rack(assuming 380,000 liters/day throughput)would be about $295,000. This costincludes the vapor processing system.installation, and piping. Under a totalracks designation, the capital cost toinstall controls on all loading racks(assuming 1,900,000 liters/daythroughput) would be about S345,000.Annualized costs, which include capitalcharges, labor, maintenance, utilities,and liquid recovery credits, indicate asmuch as an $80,000 per year differencein favor of the total racks controls (a netannualized cost of about $40,000 for asingle rack designation, and a netannualized cost savings of $40,000 for atotal racks designation]. The majorreason for this favorable annualizedcost, as stated earlier, Is the greaterrecovery cost credits associated withthe controls on all of the racks. Based onthis analysis, it was concluded that forterminals in attainment areas the costswhich would result from a total racksdesignation would be reasonable, andwould in fact be less expensive on anannualized basis provided that thecontrol system used recovered gasoline

from the collected vapors. Sj.stemswhich recover gasoline are expected tocomprise the majority of the systemswhich would be installed under theNSPS.

Another consideration regarding costsfor terminals in attainment areas is thatmost tank trucks serving bulk terminalsin attainment areas would not beequipped for vapor recovery under Stateregulations. In order to load at acontrolled loading rack, a tank truckwould have to be equipped with a vaporcollection system to route gasolinevapors to the terminal's control system.Besides having to retrofit vapor recoveryequipment, some tank trucks would alsohavelo convert to bottom loading fromtop loading, If the terminal switched tobottom loading in the course ofinstalling vapor control equipment. Inaddition, a vapor tightness requirementcould be in effect for tank trucks loadingat such a rack. However, under a singlerack designation. a terminal could endup with a mix of controlled racks anduncontrolled racks. In this case, tanktrucks would probably load at theuncontrolled racks so that the cost ofretrofitting could be avoided. Thus, aterminal owner or operator consideringa conversion of one ormore racks,which would result in those racksbecoming affected facilities, likelywould either be deterred from makingany changes or would convert all of theracks and all of his tank trucks in orderto prevent this situation from occurring.The result of this conversion would bethe same as under a total racksdesignation.

For terminals which already haveo control devices (the majority beingexisting terminals in non-attainmentareas), capital and annualized costscould both be lower for a total racksdesignation of the affected facility. Forexample, the case of an existingterminal which is modified by adding aloading rack and increasing emissionswas analyzed. If the limits of thestandard were more stringent than thoseunder which the existing control devicewas operating, then depending onwhether the device could meet the morestringent limit, the terminal owner oroperator might have to make anexpenditure in order to comply with thenew limits. Under a one rackdesignation for the affected facility, onlythe new rack would be required to meetthe limits of the standard. If a separatecontrol system was installed for the newrack, capital costs could be about$295,000 and annualized costs could beabout $70,000 per year. Under the totalracks designation of the affected facility,both the existing racks and the new rack

i II I I I Ill III I I I

B3130 Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

might be required to meet the limits ofthe standard under the modificationprovisions. One option would be toinstall an add-on system to the existingcontrol device if that control devicecould not meet the limits of thestandard. The capital cost for thisapproach could be about $100,000 andthe net annualized cost could be about$20,000 per year more than theannualized cost of operating the originalsystem. Another option would be toreplace the existing control device witha new device which could meet the,limits of the standard. The capital costfor this could be about $200,000 and theincremental net annualized cost couldamount to $50,000 per year. A third "option for terminal operators whosecontrol device could be altered toachieve a lower emission limit would beto upgrade the existing device throughdesign or operational changes. While the,cost for this approach would vary inindividual cases, it would beconsiderably less than the cost for either

*of the first two options. Finally, anypresently installed control device whichwas capable of complying with thelimits of a more stringent standardwould not have to be altered. Theoperator's decision to select any of theseoptions would depend on such factorsas the terminal's financial position andthe type, age, condition, and controlefficiency of the existing control device.Based on this analysis, it was concludedthat for terminals in non-attainmentareas, regardless of which option theterminal operator selected, the costsincurred under a total racks designationwould be reasonable, and in fact any ofthe options discussed would be lessexpensive than the costs under a onerack designation.

In addition to the emission reductionand cost considerations discussedabove, the single rack designation hastechnical complications. Performancetesting ofthis affected facility would bedifficult at terminals which already havesome means of vapor control installed(estimated to include about 70 percent ofthe existing terminals). If one rack werenewly installed or altered in such a wayas to become an affected facility undermodification or reconstructionprovisions (40 CFR 60.14 and 60.15) andwere required to meet a more stringentemission limit, the new rack-couldrequire controls different from theremainder of the reading equipmentSince the emissions from all of the racksare typically routed to the single vaporprocessor, it would be impossible todistinghish the vapor processor outletemissions originating from only the newloading rack. If an existing control

device were unable to meet a morestringent emission limit, a bulk terminaloperator could either install a separatevapor collection system and processorfor the new rack, or replace or upgradethe existing control device. The latterapproach is identical to what a totalracks designation of the affected facilitywould accomplish.

The foregoing discussion indicatesthat, based on the assumptions made,the total racks designation would resultin the greatest emission reduction.Furthermore, the total racks designationwould be the most consistent with theindustry practice of using one controldevice for all racks at a terminal. Thetotal racks designation, by causing allcollected vapors to be routed to a singlevapor processor, would result in the lessexpensive approach to achieving therequirements of the proposed standards,and the costs would be reasonable.Performance testing of this type ofaffected facility would be*straightforward because all loadingracks would be subject to the samestandards. Consequently, afterconsidering the emission reduction,technical, and cost impacts associatedwith each possible designation, theAdministrator selected the combinationof all the loading racks as the affectedfacility.

Comments are specifically invitedconcerning the -selection of the affectedfacility. In particular, comments arerequested on the question of whetherselection of the totalracks designationwould in fact result in greater emissionsreduction than would selection of theone rack designation, and the economic

-impact of this selection on' existingterminals. Comments are requested onthe factors considered and also on anyadditional factors which should beconsidered. Any comments submitted tothe Administrator on this issue shouldcontain specific information and datapertinent to an eviluation of themagnitude and severity of its impactand suggested alternative courses ofaction that would avoid this impact.

Selection of Basis'of PmposedStandards

Control Technology. Control systemscurrently being used atterminals consistof two main elements, the vaporcollection system and-the vaporprocessing system (or vapor processor].All of the vapor collection systems usedat terminals are somewhat similar. Theair-vapor mixture displaced during theloading of the delivery tank is containedand routed through vapor piping on thetank truck to the terminal vapor,collection piping. The terminal vaporcollection system, in turn, routes the air-

vapor mixture through piping to thevapor processing equipment. Knock-outtanks are commonly utilized betweenthe loading racks and the vaporprocessoi to remove liquid from thetransfer lines before it reaches the vaporprocessor. Liquid can enter the line dueto overfilling the delivery tank,entrainment of liquid droplets from theloading operation, or condensation ofvapor into liquid. Vapor holding tanksare also used in some vapor collectionsystems. Vapor holding tanks are usedto store a designated volume of air-vapor mixture and then release it to theprocessor to process the vapors on abatch basis. Fluctuations in VOCconcentration and air-vapor mixtureflow rate are minimized by using vaporholders.

Several vapor processing techniqueswere evaluated by EPA. These controltechniques included carbon absorption(CA], thermal oxidation, (TO),refrigeration (REF), compression-refrigeration-adsorption (CRA),compression-refrigeration-condensation[CRC), and lean oil absorption (LOA).These six techniques represent allof thecontrol methodologies commonlyemployed at bulk terminals. At least onesystem utilizing each of these controltechnologies was tested in an EPA-sponsored test program conductedbetween 1973 and 1978. The testprocedure used was the procedureoutlined in the draft bulk gasolineterminal Control Techniques Guideline(CTG) document, "Control ofHydrocarbons from Tank TruckGasoline Loading Terminals," datedMay 15, 1977. This test procbdure issimilar to the procedures in theproposed standards and in ReferenceMethods 2A, 28, 25A, and 25B. Althoughthe emission measurement methodswere the same as the proposedReference Methods 2A, 2B, 25A, and25B, the test procedure varied slightly, inthat the test period was longer than theperiod required in the proposedstandards. This difference is discussedin Appendix D of the BackgroundInformation Document, and would notaffect the achievability of the standard.

The test data considered In evaluatingthe six control technologies mentionedearlier represent terminals ranging ingasoline throughput from 190,000 litersper day (50,000 gal/day) to 5,700,000liters per day (1,500,000 gal/daj).Twenty-two tests were performed,totaling 61 days of testing. In addition,several tests performed by others usingthe same procedures were considered inevaluating the control technologies,Thus, these data are consideredrepresentative of the conditions at a

Federal Register [ Vol. 45, No. 244 / Wednesday, December 17. 1980 / Proposed Rules

wide range of terminal sizes and are anadequate basis for an evaluation of thebest systems of continuous emissionreduction-

Tank trucks have been demonstratedto be majorsources of vapor leakageduring product loading operations at.bulk terminals. Vaporleakage 'an varysignificantly from one tank to. another.The larger the tank truck leakage, thesmaller thivolume of air-vapor mixturethat enters the vapor collection system.To evaluate the results of the vaporcontrol system tests. the results from allof the tests were calculated on acomparable no-leak basis. Massemissions, in the form of milligrams ofVOC per liter of gasoline loaded,, wereused to compare the test results on acommon basis. These units were used. tobe consistent with the units utilized inthe test reports.

Carbon adsorption systems use bedsof activated carbon. to adsorb gasolinevapors from the air-vapor mixture. CAsystems at terminals typically consist oftwo carbon beds. One bed activelyadsorbs the vapors while the other bedis'being regenerated. After a set period"df time, the active bed is regenerated,with the air-vapor flow re-routed to theopposite bed.

Three EPA test, consisting of ninedays of testing, were performed on twocarbon adsorption, systems whichincorporated vacuum regenerationassisted by warm air purge. The dairyaverage emissions from these systemsranged from 1.8 milligrams of VOC perliter of gasoline loaded (mg/liter) to 11.0mg/liter. Two days of testing were notincluded in evaluating the systembecause of unit maladjustment andtesting irregularities. On one test day,the bed switching timer was setincorrectly, leading to excessive loadingon one carbon bed. On one day ofanother test, two tank trucks werepurposely loaded simultaneously,causing vapor loading to exceed thedesign capacity of the CA system. Thiswas done in order to determine theperformance limit of the system. Furtherdetails are contained-in Appendix C ofthe Background Information Document.C6ntrol unit efficiencies on theremaining seven days ranged from 9a.apercent to 99.6 percent. One of thesystems tested included a vapor holderin its design.

One test was performed in 1979. by theCalifornia Air Resources Board on acarbon adsorption system using vacuumregeneration. Mass emissions measuredat the system exhaust were not specifiedexactly, but were reported to be lessthan.Lz2mg/liter.

Thermal oxidizer systems do notrecover any product. Instead. the

gasoline vapors are oxidized in a burnerchamber. Many TO systems use vaporholders to. store air-vapor mixture fromthe loading rack. so that the system canprocess VO~vapors at a relativelyconstant concentration and flow.

Tests were performed on four thermaloxidizer control systems. Two of thecontrol systems incorporated vaporholders, while two systems operated onan on-demand basis. Emissions from thethermal oxidizer systems ranged from1.4 mg/liter to 107 mg/liter over the foursystems tested. Control efficlenciesvaried from 86.6 percent to 29.8 percent.Even though there was a widevariability in the test results all systemstested appeared to be operatingproperly. The two thermal oxidizersystems incorporating a vapor holderachieved an. average VOC emission rateof 13".3 mag/liter, while the two systemswithout vapor holders averaged46.4mg/liter.

Refrigeration systems, as with theremainder of the systems to bediscussed, recover gasoline vapors fromthe loading operation in the form of aliquid product. In the REFsystem, air-vapor mixture from the loading racks isrouted to: a condensation chamber andpassed over a series of cooling coils.Temperatures in the condensationsection can be as low as -15°F. Thegasoline vapors condense, with somewater vapor in the air, and areseparated in a gasoline/water separator.

Six refrigeration type vaporprocessing systems were tested in theEPA program, totaling 17 days of testing.The emissions in one test wereunusually high compared to those fromother REF tests. Problems with the testequipment and with the refrigerationsystem itself led to the highemissions inthis test. Since these test results werenot considered representative of thesystems performance, data from thistest were not included in the analysis ofthe REF system.- In another test. seriousleakage in the vapor collection systemprevented almost half of the air-vapormixture displaced from tank trucks fromreaching the refrigeration system. Datafrom this test were also not included inthe REF system performance evaluation.Appendix C of the BackgroundInformation Document contains furtherdetails on these tests. The daily averageresults from the fourremaining testsranged from 31.1 mg/liter to 103 mg/liter, and indicated a control efficiencyranging from -7.1 percent to 94.6percent. During two of these four tests,the refrigeration system was notlcoolingthe vapors to, the temperature for whichthe sytem was designed. Cooling sectiontemperatures were approximately 40'F

warmer than the design temperature -60'F instead of -100'fl. Emission ratesadjusted for system leakage from thesetwo tests averaged approximately5amg/liter It is notlinown how muchlower the emission ratewould havebeen if the design temperature of thecooling sections had been maintained.but improved emission rates areexpected for these systems ifdesigntemperatures are maintained. Most ofthe EPA testing performed since 1974 onREF system involved systems whichuse chilled methylene chloride "brine'

to cool the condenser section. Manynewer systems use direct expanisior ofrefrigerant for cooling, and recent testsindicate that thenew systems may becapable of improved performance andreliability when compared to the oldersystems. Three tests performed in l'aand 1979 by the California AirResources Board on the latest modelrefrigeration systems measured outletVOC mass emission rates of 48 mg/liter,36 mg/liter and 5 mg/liter.

In compression-refrigeration-absorption systems. the air-vapormixture from the loading racks is firstsaturated to bring the concentration ofthe gasoline vapors above the explosiverange and is then stored in a vaporholding tank When the volume limit isachieved, the air-vapor mixture is routedto the CRA processing unit. The air-vapor mixture is first compressed andthen passed to a cooler-condensersection where some liquid is condensedand recovered. The remaining mixture issent to an absorption section where thegasoline vapors are absorbed in chilledgasoline.

Six CRA-type vapor processingsystems were tested by EPA. totaling 16days of testing. All of the systems testedincorporated vapor holders in thesystem design. Average daily processoroutlet emission rates ranged from4l.5mg/liter to 914) mg/liter. Processorefficience ranged from 61.4 percent to94.apercenL

Compression-refrigera tiora-condensation systems are similar inoperation to the CRCsystems. The air-vapor mixture is saturated and thenstored in avaporholding tank. The air-vapor mixture is compressed and anycondensed liquid is collected. Theremaining mixture then passes through aseries of refrigerated condenser sectionsbefore exiting to the atmosphere.

Two CRC vapor processing systemswere tested, totaling five days of testing.One system tested had serious leakageproblems from the vapor holder. Nomethodwas available to estimate theleakage so the outlet emissions couldnot be adjusted for comparison to theother tests. Processor efficiency was

83131

83132 Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

also not calculated because the vaporleakage took place after the inletsampling location. The two test dayswith complete results indicated emissionrates of 48.4 mg/liter and 55.9 mg/liter,and pt'oceqsor efficiencies of 89.0percent and 91.5 percent.

The final system evaluated was thelean oi!.absorption.system. The LOAsystem is basically an absorptionsystem where the gasoline vapors,which are predominantly lightermolecular weight hydrocarbons, areabsorbed in a "lean oil," which is lean inlight ends. The air-vapor mixture isintroduced into the bottom of theabsorption column and passes -

countercurrent to lean oil, generated onsite, which is sprayed from the top of thetower. The recovered liquid is returnedto storage.

One lean oil absorber processingsystem was tested, consisting of threedays of testing. The daily averageemission rate ranged from 73.0 mg/literto 130.0 mg/liter with processingefficiency ranging from 74.1 percent to85.9 percent.,

Most of, the systems evaluated weredesigned and installed to meet an 80mg/liter (or about 90 perce4 efficiency)standard as required in 'many SIPs. Thetest data, however, reflect the potentialof most of the systems to-achieve ahigher control efficiency.

As mentioned earlier, vapor leakagefrom tank trucks can be a major sourceof VOC emissions during terminal 'loading operations. As the gasoline orother liquid product is being loaded,most of the displaced vapors arecollected and contained in the vaporcollection system, However , somevapors may leak but of the hatch covers,pressure-vacuum (P-V,) vents, and othervapor containment components on thetank truck to the atmosphereuncontrolled. Since, in this case, onlypart of the displaced vapor volume iscollected and controlled, the overallefficiency of the vapor control system isreduced. Vapor tightness requirementson-the delivery tank would reduce thefugitive emissions problem at theloading racks. No vapor tightnessrequirements were in effect at any of theterminals tested in the EPA control .system test program. Vapor leakage forindividual loadings varied from 0 to 100percent. The average vapor leakage wasapproximately 30 percent.

To control these emissions, amaintenance program would benecessary. Such a program wouldconsist of inspecting gaskets and sealsfor wear or crackiing, hatch covers forwarpage, and P-V vents to ensure thatthey seat properly. The maintenanceprogram would involve repairing or

replacing any items ii; the tank truckvapor containment equipment that mightallow vapors to escape. The ability ofthe delivery tank to maintain vaportightness is dependent upon the loadingmethod, maintenance program, and thetype of service to which the tank isexposed. Ih separate EPA-sponsoredprogram (EPA Report No. EPA-450/3-79-018), delivery tank trucks were testedin an area where a vapor tightnessprogram was implemented. in this area,delivery tank trucks were required topass an annual certification test whichverified the vapor tightness of the tank.In this program, the annual averagevapor leakage from the tested deliverytank trucks was reduced to about 10percent.

Emissions through leaking tank truckscan be increased by improperlydesigned loading rack vapor collectionsystems. For example, if two trucks areloading simultaneously, vapor collectedfrom one truck may pass through thevapor piping to another rack and escapethrough a non-vapor-tight truck. Theleaking tank represents the path of leastresistance to the atmosphere for thevapors in the loading rack collectionsystem. This has been observed inseveral terminal tests. This problem canbe eliminated by the installation ofcheck valves or similar devices in thevapor collection system which wouldnot allow vapors to pass from oneloading rack to another. This design hasbeen used at several terminals.

Regulatory Alternatives. Regulatoryalternatives were developed whichrepresent technically feasible levels ofcontrol for reducing VOC emissions'from bulk gasoline terminals. The Unitsof milligrams per liter (mg/liter) wereused to compare the vapor processingsystems tested and were, therefore, usedto distinguish between emissionreductions achiveable by each of thealternatives.

Based on review of the technicalsupport data, four regulatoryalternatives were selected. UnderAlternative I no standards would bedeveloped. Instead, the StateImplementation Plans (SIP's) would berelied upon to control VOC emissionsfrom bulk gasoline terminals. SIPregulations for VOC generally requirecontrols only in the areas which do notmeet National Ambient Air QualityStandards for ozone (non-attainmentareas): However, 17 States are expectedto require SIP controls statewide by

- 1982. A typical SIP regulation forgasoline loading at bulk terminals wouldrequire the routing of vapors to-a vaporprocessing system and would limit theemission rate from the processor outlet

to 80 mg/liter. The emission rate of 80mg/liter is roughly equivalent to 90percent control efficiency. The typicalSIP would also contain a requirement -for gasoline delivery tanks to pass anannual vapor-tight test to minimize

- fugitive vapor losses at the loadingterminal. It .was estimated that SIPregulations on tank truck loadings andvapor tightness would affectapproximately 70 percent of new andexisting terminals by 1982.

The remaining three alternativesreflect two basic levels of control forvapor processing equipment installed atterminals, but represent three levels ofoverall emission reduction. Each ofthese three alternatives would requirethat affected facilities be equipped withvapor collection equipment. Emissionlimits would be met using vaporprocessing equipment similar to thesystems tested by EPA. Emission testresults Indicate that most of the vaporprocessing systems which are now beinginstalled to meet existing emission limitscould, in fact, meet a more stringentstandard.

Alternative II would require the SIPemission limit of 80 mag/liter and wouldalso require that liquid product loadingsinto gasoline tank trucks be restricted totrucks which were vapor-tight. Emissionreductionsfrom the baseline would beexperienced under Alternative II sinceall new, modified, or reconstructedterminals not covered by the SIPs wouldbe regulated by Alternative 11. Theseterminals would include those inattainment areas not regulated by SIPs,The test data indicate that the carbonabsorption, thermal oxidation,refrigeration, CRA, and CRC vaporprocessing systems could meet theemission limits of Alternative II.

Alternative III would set a VOCemission limit based on 35 mg/liter asdetermined from the available test data.This alternative would have no specifictank truck vapor-tight requirements andwould rely on the SIPs to control tanktruck fugitive emissions in non-attainment areas. Tank truck fugitiveemissions in attainment areas wouldremain uncontrolled under AlternativeIl. Inspection of the test data revealedthat the carbon adsorption system andthermal oxidation system with vaporholder were roughly equivalent, givingthe most consistent results in reducingVOC emissions. Emission rates fromcarbon adsorption systems ranged from1.8 to 11.0 mg/liter, averaging 5.9 mg/liter. Thermal oxidizer systems using avapor holder produced emissionsranging from 1.4 to 29.4 mg/liter, for anaverage of 13.3 mg/liter. Althoughaverage emissions from carbon

Federal Register / VoL 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules 83133

adsorption systems were slightly lower.the cost analysis indicated that thethermal oxidation system, using a vaporholder is the most cost-effective systemfor small terminals.

This is important since'about half ofthe affected facilities are expectect to bein the smallest model plant size.

The highest adjusted daily emissionrate from the applicable tests on- carbonadsorption and thermal oxidizer withvapor holder systems wasapproximately 29 mag/liter. This adjustedemission rate represents the calculatedrate whichwould have occurred in avapor-tight collection system, based onactual measured emissions and anadjustment factor based onmeasurements of tank truck vaporleakage during testing. In order to allow-a small margin above the highestadjusted.emission rate from the testedsystems, a level of 35 mg/liter wasselected as the emission limit for theregulatory alternative. It appears thatrefrigeration, as well as carbonadsorption and thermal oxidation, hasthe capability to achieve this limit,although some operational or designmodifications might be required forspecific systems.

Alternative IV is similar toAlternative M in that it would limit thevapor collection system emissions to 35mg/liter. This emission limit is based onthe same control technologies as inAlternative . In addition, AlternativeIV would require that liquid productloadings into gasoline tank trucks berestricted to vapor-tight trucks.

Model plants were developed for new,modified, and reconstructed terminals inorder to analyze and compafe theenvironmental, energy, and economicimpacts of each regulatory alternative.Four model plants were selected torepresent the-cross-section of dailygasoline throughputs found in the bulkgasoline terminal industry. Gasolinethroughputs selected for comparison andanalysis were 380,000 liters/day (100,000gallons/day], 950,000 liters/day (250,000gallons/day), 1,900,000iters/day(500,000 gallons/day), and 3,800,000liters/day (1,000,000 gallons/day. Newterminals are best represented by thethree larger model plant sizes whileexisting terminals are best representedby the three smaller model plant sizes.

Impacts of RegulatoryAlternatfves.Under Alternative 1, in the absence ofadditional, standards of performance.there would be no VOC emissionreduction beyond the reductions due tothe SIPs, which will result in a 1982baseline VOC emission level of 140,000megagrams per year (Mg/year].

Under Alternative 11, the 1982 baselinelevel would be reduced by 5,750 Mg/

year by 1985. This represents areduction of about 60 percent. from 9.150Mg/3ear to 3,410 Mg/year. in the VOCemissions from all new, modified, andreconstructed terminals.

Under Alternative II, nationwideVOC emissions would be reduced by4,510 Mg/year by 1985. Emissions fromaffected. terminals would be reduced byabout 50 percent from 9,150 Mglyear to4,650 Mv/year by 1985. The loweremission reduction of Alternative Mwhen compared to Alternative I1illustrates thesignificance of tank truckvapor leakage. Even though processoroutlet emissions under Alternative Mwould be reduced from 80 mg/liter to 35mg/liter, the absence of a requirementthat terminals restrict loadings ofgasoline tank trucks to vapor-tighttrucks more than offsets the additionalVOC reduction.

Under Alternative IV, nationwideVOC emissions by 1985 would decreaseby 6,620 Mg/year. The reduction in VOCemissions from affected terminals duringthis period would be about 70 percent,from 9,150 Mg/year to 2,540 Mg/year.

The regulatory alternatives wouldapply to all new, modified, orreconstructed terminals but would affectterminals in non-attainment areasdifferently than terminals in attainmentareas. New and existing terminals innon-attainment areas would beregulated by SIP requirements andwould therefore have some type ofvapor control system already installed.Most terminals in attainment areaswould not be controlled by SIPs andwould experience the greatest effects ofthe regulatory alternatives. Terminals inattainment areas would experiencedifferent effects depending upon thetype of loading currently used atexisting terminals or that which wouldhave been used by a new terminal in theabsence of additional standards. Thereare two basic methods by whichdelivery tanks can be loaded at bulkterminals, top loaded through thehatchways on top of the tanks, orbottom loaded through adapters at thebottom of the tanks. Top splash loadinginvolves inserting a nozzle into the*hatchway and splashing the incomingproduct onto the surface of the productin the tank. Attaching a fixed orextensible downspout to the loading armallows product to be introduced belowthe liquid surface (submerged loading).Bottom loading can also be considered aform of submerged loading. Generally,top splash loading results in greaterVOC emissions than submerged loading.Thus, greater emission reductions wouldbe achieved under any of the regulatoryalternatives when controlling top splash

loading terminals compared to terminalsusing submerged loading.

VOC emissions for model plantswould vary for each alternative. UnderAlternative IL new. modified orreconstructed terminals in non-attainment areas would experience noVOC emissions reduction. The emissionlimits for existing terminals in non-attainment areas under SIP regulationsare identical to the limits underAlternative IL In an attainment area,under Alternative 11. a terminal with agasoline throughput of 950,000 liters perday which previously used submergedloading would experience a VOCemission reduction of'137 Mg/year (from194 Mg/yr to 57 Mg/yr). For the samethroughput terminal which previouslyused top splash loading, a VOCemisgion reduction of 408 Mg/yr (from465 Mg/yr to 57 Mg/yr] would beexperienced under Alternative IL

Under Alternative Iff. a 950,000 liter!day terminal in a non-attainment areawould experience an VOC emissionreduction of 19 Mg/yr (from 57 Mg[yr to38 Mglyr]. For a submerged loadingterminal in an attainment area. a VOCemission reduction of 87 Mg/yr (from194 Mg/yr to 107 Mg/yr) would beexperienced. If splash loading were usedprior to control, this same terminalwould experience a VOC emissionreduction of 358 Mg/yr (from 465 Mg/yrto 107 Mg/yr) under Alternative I1.

For a 950,000 liter/day terminal in anon-attainment area, emissionreductions under Alternative IV wouldbe the same as the emission reductionachieved under Alternative 11. 19 Mg/yr(from 57 Mg/yr to 38 Mg/yr). For a950.000 liter/day terminal in anattainment area, emission reductionsachieved under Alternative IV would be156 Mg!yr (from 194 Mglyr to 38 Mglyr]for a terminal which used submergedloading and 427 Mg/yr (from 465 Mg/yrto 38 Mg/yr] for a terminal whichpreviously used splash loading.

Thermal oxidation systems emit .carbon monoxide (CO) and oxides ofnitrogen (NOJ] during the combustion ofVOC vapors. A thermal oxidationsystem at a 930,000 liter/day terminalwould emit approximately 0.8 Mg/yr ofCO and 0.3 Mg/yr of NO.. A worst casesituation would be one in which 25percent of the 50 modified orreconstructed terminals by 1985 were toinstall thermal oxidation systems, andall of these facilities had gasolinethroughputs of about 950,000 liters perday. In this case. nationwide COemissions would increase by 10 Mg/yrand'nationwide NO2 emissions wouldincrease by 4 MSfyr. For both of thesepollutants, the emission increases

83134 Federal Register / Vol. 45, No. 244 / Wedpesday, December 17, 1980 / Proposed Rules

represent a small adverse nationwideair pollution impact.

Impacts on water pollution from anyof the alternatives considered would be-

- negligible, since none of the controlsystems considered uses water as acollection medium. Carbon adsorptionsystems using'steam in the regenerationmode would have the greatest impact onwater pollution. All of the steam wouldbe condensed and any gasoline presentin the condensedliquid would be ,separated in the terminal's gasoline/water separator.'However, there are nosteam-regenerated carbon adsorptionsystems currently installed at bulkterminals, and the vacuum-regeneratedsystems currently in use will representthe primary carbon adsorption

c technology at bulk terminals into theforeseeable future. All other systemswhich chill or condense the air-vaporm ixture use a gasoline/water separatorintegrated into their design, anddischarge a small amount of condensedwater vapor. The impacts on waterpollution are the same for each of theregulatory alternatives b6causeessentially the same control equipmentis being assdmed for each alternative.

Because all of the VOC emissions areincinerated or returned to storage asliquid product, there would be no directsolid waste impacts under the regulatoryalternatives. Some solid waste could begenerated indirectly dtle to disposal ofactivated carbon from carbonadsorption units after the useful life ofthe carbon had expired. The worst casefor solid waste impact would occur if allaffected facilities were to use carbonadsorption units for control and had todispose of the activated carbon every 10years. This would result in only about50,000 kilograms (55 tons) of solid wasteannually. Even in this worst case, theimpact of the alternatives on solid wastewould be negligible. In practice, not allaffected facilities are expected tochoose carbon adsorption for control,and in many cases the carbon May lastlonger than 10 years or may betransported off-site for regeneration andreuse.

The energy impacts were derived byassuming that all VOC emissionsreduction was recovered as liquidproduct, and that one liter of this liquidproduct was equivalent to one liter ofgasoline. It is assumed for a vapor-tightvapor collection system, that no part ofthe recovered product is lost to theatmosphere on the way to the storagetank. In addition, although the VOCliquid may not have the exactcomposition of gasoline, the liquid isreturned to the storage tank where eachliter becomes absorbed and is available

for loading into tank trucks as gasoline.The energy required to operate thevapor processing equipment wassubtracted to determine the net energyimpact of each alternative.

A net energy savings would resultfrom-each of the regulatory alternatives.A net energy savings for eachalternative is projected even though It isassumed that as many as half of thesmall new, modified, or rec6nstructedterminals may install thermal oxidizersystems, which do notrecover energyand have a small net energy loss..Alternative II would accomplish a netfuel savings of 8 million liters (2.1million gallons) of gasoline per year inthe fifth year of the standard.Alternative III would recoveg 6 million

'liters (1.6 million gallons) of gasoline peryear in the fifth year.

Because it results in the greatestrecovery of VOC, Alternative IV wouldresult in the greatest net energy savings.Alternative IV would recover-9 millionliters (2.4 million gallons) of gasoline peryear in the fifth year of the standard.

A net energy savings would resultfrom each of the model plant sizes forany of the vapor control systems exceptthermal oxidizer systems. Energysavings would range from an average of144,000 liters per year (38,000. gallons peryear) of gasoline for the smallest modelplant (gasoline throughput 380,000 liters/day) to an average of 1,540,000 liters peryear (407,000 gallons per year] ofgasoline for the largest model plant(gasoline throughput 3,800,000 liters/day). A net energy loss ranging from

-2,600 liters of gasoline per year forthesmall model plants to 22,000 liters ofgasoline per year for the largest modelplants would result through the use ofthermal oxidizer systems.* The total capital and annualized costs

tothe bulk gasoline terminal industrywere determined for each regulatoryalternative. Capital costs include thepurchase and installation of vaporcollection and processing systems,retrofit of tank trucks to bottom loadingand vapor recovery configurations, andconversion of top loading racks tobottom loading. Annualized costsinclude capital charges, utilities,maintenance and repairs, and routineoperating labor. Alternatives II and IV-would require an additional cost toperform an annual vapor-tight test andsubsequent repairs on tank trucks.

In addition to the incremental costsincurred by bulk terminals under theregulatory alternatives, there would be acost impact -on owners of the "for-hire"tank trucks operating at terminals. For-hire tank trucks are those trucks ownedby independent companies, whichtransport products from bulk terminals

to other distribution points. For-hiretrucks are estimated to constitute about70 percent of the tank trucks at bulkterminals. Companies operating for-hiretank trucks would have to installcompatible loading and vapor recoveryequipment on their tank trucks whichserve affected bulk terminals. Sinceseveral configurations of adapters arepossible, the regulation would requirecompatible equipment to ensure thattank truck and terminal vapor collectionsystems could be connected duringproduct loading. All trucks not alreadyhaving bottom loading and vaporrecovery provisions would be retrofittedwith this equipment, and thus therewould be a cost impact on thesecompanies as a result of the proposedstandards. It is estimated that 390, or 2percent, of the estimated 18,000 for-hiretank trucks would be affected in the firstfive years. Approximately 85 of thesewould have to convert to botton loadingand incorporate vapor recoveryprovisions, at $6,400 per tank truck, and305 would have to add vapor recoveryprovisions only, at $2,400 per tank truck.Annualized costs to the for-hire tanktruck industry would include the cost ofmaintaining the vapor recoveryeguipment and, under Alternatives IIand IV, the cost of performing an annualvapor-tight test on each gasoline tanktruck.

The total capital cost to the bulkgasoline terminal Industry for theinstalled vapor control equipmentnecessary to meet Alternative II on the55 new, modified, or reconstructedterminals expected through the first fiveyears of the standard would beapproximately $23.0 million. Theterminal industry annualized cost in1985 would be $3.3 million. The totalcapital cost to the for-hire tank truckindustry through the first five yearswould be approximately $1.3 million.Annualized cost to the tank truckindustry in 1985 would totalapproximately $0.7 million, due toincremental maintenance and testingrequirements. The overall annualizedcost-effectiveness in 1985 expectedunder Alternative II would be $696/Mg($632/ton) of VOC controlled.

Under Alternative III, the total capitalcost for vapor control equipmentnecessary through the first five yearswould be approximately $24.0 million.The industry annualized cost in 1985would be $4.1 million. The total capitalcost to the for-hire tank truck industrythrough 1985 would be about $1.3million, and the annualized cost in 1905would total about $0.6 million. TheIndustry annualized cost-effectivenessin 1985 expected under Alternative III

Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

would be $1,042/Mg ($946/ton) of VOCcontrolled.

-Under Alternative IV, the total capitalcost to the terminal industry for vaporcontrol equipment necessary through thefirst five years of the standard would beapproximately $24.0 million. Theterminal industry annualized cost in1985 would be $3.6 million. As inAlternative II, the total capital cost tothe for-hire tank truck industry through1985 would be about $1.3 million, andthe annualized cost in 1985 would totalabout $0.7 million. The overallannualized cost-effectiveness in 1985expected under Alternative IV would be$650/Mg (590/ton) of VOC controlled.

A mix of current control technologiesbeing installed to achieve 80 mg/literwas used to establish the-capital costfigures for Alternative II. Three of thefour technologies-used could, in fact,meet an emission limit.of 35 mg/liter.These three technologies were then usedto establish the capital costs forAlternatives Ell and IV. Because thecosts of all the control technologiesconsidered are similar and becausemuch the same equipment was used toestablish the capital costs for eachalternative, the resultant capital costsfor Alternatives II, III, and IV are similarTo meet the emission limits of any of thealternatives, the vapors from tank truckswould have to be collected and routedto the terminal collection system.Therefore, the costs.for tank truckretrofitting would be the same for eachalternative. The differences in netannualized cost among the alternativesresult from differing product recoverycost credits at affected terminals, andthe inclusion of tank truck-vapor-tightrequirements under Alternatives II andIV.

For Alternative II, cost assessmentswere performed on the carbonadsorption, thermal oxidizer,refrigeration- and compression-refrigeration-absorption vaporprocessing systems for each model plantsize. Operating costs for thecompression-refrigeration-condensationsystem are considered comparable tothe CRA system. The lean oil absorptionsystem was not included in the analysisbecause the test data indicated that thesystem tested would not be able to meetthe requirements of Alternative I. ForAlternatives III and IV. costassessments were performed on thecarbon adsorption, thermal oxidation,and refrigeration vapor processingsystems only. The test data indicatedthe inability of the CRA or CRC systemsto meet a 35 mg/liter limit. For all of thealternatives, the thermal oxidizer systemwould be competitive with the other

units only at the smaller model plantsizes. All other systems evaluated arecomparable when consideringannualized costs.

A secondary, or add-on, vaporprocessor could lie chosen by the owneror operator of an existing SIP-controlledfacility which became an affectedfacility under the proposed standards.Add-on systems, primarily carbonadsorption and thermal oxidation, havebeen used at bulk terminals to increasethe control efficiency of existingprocessing systems. Selecting an add-onsystem to process part of the vaporswould be an option to replacing theexisting system with a more efficientsystem designed to handle the entireload. The add-on option would requireoperating and maintaining twoprocessors, whereas the option ofreplacing the system entirely means thatexpenses would be incurred for just oneprocessor. The incremental netannualized cost for the add-on option isvirtually independent of terminal size,amounting to ap-proximately $20,000 peryear for an add-on carbon adsorptionsystem, and $45,000 per year for an add-on thermal oxidizer system. Theincremental net annualized cost of areplacement carbon adsorption systemwould average approximately $37,400for any terminal size. Due to the loss ofgasoline recovery cost credits, a thermaloxidizer system replacing a vaporrecovery system would cost from $43,000per year to $300,000 per year more thanthe costs incurred due to the originalsystem. As a result, add-on orreplacement thermal oxidizer systemsare likely to be selected for use only atthe smallest bulk terminals.

An economic analysis performed oneach of the regulatory alternativesinvestigated impacts for new terminalsand for modified or reconstructedterminals. New terminals constructed inpreviously regulated (non-attainment]areas will incur no additional costs as aresult of any of the regulatoryalternatives because the collection andprocessing systems being installed tomeet SIP requirements are essentiallyidentical to those systems which wouldbe considered under the regulatoryalternatives. New terminals inattainment areas would incur varyingcontrol costs depending on their sizeand type of loading. Control costs wouldnot vary significantly among theregulatory alternatives, although theimproved product recovery cost creditsunder Alternative IV lead to the lowestnet annualized cost of anyalternative.Industry information indicates that nonew 380.000 liter/day bulk terminals areplanned in the first five years of the

proposed standards, because thepotential rate of return on smallerterminals is not sufficient to encouragetheir growth. Terminals in the 950,000liter/day size category are consideredmarginally profitable, even withoutadditional control costs. Therefore, onlyone new terminal of this size is expectedto be constructed in an attainment areaIn the first five years of standards. Thetwo largest model plant sizes, 1,900,000and 3,800,000 liters/day, are consideredattractive investment possibilities, andone new 1,900,000 liter/day terminal isexpected to be constructed in anattainment area in the first five years.The construction of these two terminalsshould not be hindered under any of thealternatives. The 950,000 liter/dayterminal would have to.pass throughmost of the control costs to remain areasonable investment. The necessarydegree of cost pass-through appearspossible.

Approximately 50 existing bulkterminals are expected to be modified orreconstructed in the five year periodcovered by this assessment, with 30 ofthese being in attainment areas. Theaffected terminals in attainment areaswould be likely to experience the impactof installing a complete new vaporcollection and processing system wherenone existed previously. The remaining30 affected terminals in non-attainmentareas would experience a lesser impactbecaise a system to satisfy SIPrequirements would probably already bein place. Such a system would satisfythe requirements of Alternative II, butmay require upgrading or partialreplacement under Alternative Ill or IV.

Existing terminals of the smallestmodel plant size (380,000 liters/day)would have to pass through essentiallyall of the control costs in order tomaintain an acceptable rate of returnunder any regulatory alternative.Existing top loaded 950,000 liter/dayterminals in attainment areas would bein a similar situation because theywould experience the full impact ofconverting the loading racks to bottomloading and installing vapor collectionand processing systems. It is estimatedthat 25 of the former case and two of thelatter case will occur In the first fiveyears of the proposed standards. Ingeneral, full cost pass-through would beunlikely due to competition from otherexisting terminals and from consumerpressure as indicated by currentconservation patterns. However, it islikely that most of the control costs willbe able to be passed through, allowingmost of the 50 modified or reconstructedterminals to experience acceptable post-control returns on investmenL Capital

83135

8313. Federal Register I VoL 45, No. 244 '/ Wednesday, December 17, 1980 / Proposed Rules

availability, would not be adverselyaffected by any of the regulatoryalternatives. The larger terminals shouldnot encounter any difficulty in meetingthe control costs resulting'frommodifications or reconstructions underany alternative. Terminals in non-attainment areas which replace,upgrade, or add onto an existing controlsystem should also be able to maintainan- acceptable return on investment. Itshould! be noted that the control costsunder any regulatory alternative aresimilar to those being borne by a largenumber of terminals as a, result of StateVOC regulations.

The current trend tovfard theconsolidation of existing facilities of'marginal profitability can be expected, to.continue under the proposed standards-,but the analysis does notindicate any-additional closures. The costpass-through analyses for both new andexisting terminals indicate thatmaximum price increases of less thano.Q percent would result from anyregulatory alternative. It should be,noted that this increase would not affectnationwide gasoline prices, butrepresents a worst case situation withinthe bulk terminal industry due tocomplete costpass-through.

The regulatory alternatives- woul'd..affect the independent tank truck

industry. with minor impacts. Theprofitability, of the, firms, in the industry,would not be impacted significantlysince regulatory cost absorption would,be minimal. Most of the. regulatory', costswould bepassed through. to, theconsumer; causing: a maximum increaseIn retail gasolhne-prices of less than 0.07percent for any of the alternatives. Itshould be noted that this increase. wouldnot affect nationwide gasoline prices,but represents a worst case situationwithin, the. independert tank truckIndustry due to, complete cost pass-through. Additionally, no, closures, or'dislocations of tank truckfirms areexpected to result from any of theregulatory alternatives.

Section 111 of the CleanAir Actrequires, that, standards of performance,be based on the degree of emissionreduction. which the Administratorjudges to, be achievable- throughapplication of the best technologicalsystem of continuous emissionreduction;, considering costs,, non-air.quality health and' environmentalimpacts, and energy requirements.which has been adequatelydemonstrated; Therefore-,In selecting"the basis of the proposed standards,, theAdministrator first examined-Alternative IV,. which would achieve thegreatest reduction in, VOC emissions. -

This alternative would result in a netenergy savings of approximately 9million liters (2.4 i--illionr gallons) ofgasoline in the fifth year of the-standard,which is more than any otheralternative. .Water and solid wasteimpacts are essentially negligible underthis 'alternative. Both total capital andnet annualized costs to the hulk terminaland for-hire tank truck industries are notexcessive under Alternative IV., Smallbulk terminals, whichwould bear thegreatest economicimpact, would belikely to be able to pass through most of'the colitrol costs in order to remainviable. Even the product price increaseson the order of 1 percent which couldoccur iffull cost pass-through werepossible for terminals are consideredreasonable.The price increases due tocosts incurred by for-hifre tank truckfirms iyould be the same for anyalternative. Finally,, test data indicatethat systems have been demonstratedthat can achieve the emission limitationrequired by this alternative. Afterconsideration. of these factors, theAdministrator selected Alternative IV asthe basis for the proposed standards. Itis noted. that Alternative IV wouldachieve a greater VOC emissionreduction at less annualized cost than,Alternative I.Selection of Formaf of ProposedStandards,

Section 112 ofthe Clean AirActrequires the promulgation of standards'of performance, establishing allowableemission limitations for a category ofstationary sources, wheneverit isfeasible. o promulgate and enforcestaridards in such terms. Standards ofperformance areconsidered' not feasibleto promulgate oiienforce when either (1.)a pollutant or pollutants cannotbeemitted through a conveyance designed

.and constructed- to emit, or capture suchpollutant or (2).the application' ofmeasurement methodology to aparticular class of sources is.notpracticable due to technological oreconomic limitations. If theAdministrator judges that it is notfeasible to.prescribe or'enforce astandard of'performance Slectioa 111(h).allows the promulgation of a design,equipment, workpractfce, or operationalstandard; or combination of thesewhich reflects' the besttechnologicalsystem, of continuous emission reduction(taking into, consideration the cost ofachieving suck emission reduction, and,any, non-afr quality health andenvironmental impact and energyrequirements} which has. beenadequately demonstrated

As disbussed earlier, VOC'emissions'at tank truck loading racks are

generated when the incoming productdisplaces air-vapor mixture from tbetruck-mounted tanks. At an uncontrolledloading rack, the entire quantity ofmixture is emitted directly to theatmosphere through open hatch coversor vents. The vapor control systemscurrently being used at bulk gasolineterminals collect the air-vapormixturedisplaced from tank trucks and route themixture to a vapor processing system.The mixture is conveyed to. theprocessor through vapor collectionsystems installed on the tank trucks andon the bulk terminal's loading racksystem. Even in controlled systems,VOC emissions may occurfroim theloading: operation due to vapor leakagefrom closed gasoline tank trucks durln6product loading. These- VOC leakageemissions originate at various. points onthe tank, such as leaking pressure-vacuum vents and defective hatchcovers and seals.,Due to the fugitivenature of these emissions, It Is notfeasible to. collect the escaping vaporsand route them through a conveyahce.Since tank leakage measurements, at theloading racks do not provide aquantitative measurement of total VOCconcentration, flow rate. or massemissions, an enclosure around aloading tank truck would be necessaryin order to, trap emissions formeasurement. An enclosure andconveyance to, accomplish this Is nottechnologically or economicallypracticable. Due to these considerations,the Administrator determined that astandard of performance, in the form ofa numerical emission limit. could not beset, and that a work practice standardwould be appropriate for controlling.tank truck vapor leakage emissions.

Two methods of defining tank truckvapor tightness and regulating leakageemissions under a work practicestandard were considered. The firstmethod would'require the use of aportable combustible gas detectorduring product loading to: detect leaks.Any measurement in excess of aspecified limit would, define a leakingtank. However, the terminal, owner oroperatormay not have control ever themaintenance of all trucks loading at histerminal Also, many terminals useautomated billing equipment whichallows the tank truck driver to load thetank without any interaction withterminal personnel. At these terminals, arequirement that each loading bemonitored would represent an excessive'burden. For these reasons, theregulatory format requiring leakmonitoring of each gasoline tank truckduring.product loadingwas not selectedby the Administrator.

Federal Register I Vol. 45, No. 244 / Wednesday, December 17, 1980 1 Proposed Rules

The second method would require theterminal owner or operator to restrictloadings of gasoline tank trucks to thosewhich had passed an annual vapor-tighttest. This test would be a pressure testof the delivery tank itself and wouldyield a quantitative measure of tankleakage. Test data -how that annualtesting and subsequent leakage repaircan reduce the average, annual tanktruck emissions from 30 percent beforerepair-to 10 percent of the vapors

- displaced during product loading. Thisworkpractice standard format wouldconsist of a requirement that the owneror operator of an affected facilityrestrict product loadings of gasolinetank trucks to those for which hepossessed documentation that the tankhad passed the vapor-tight test withinthe 12 preceding months. This formatwould provide some control overleakage emissions and would notimpose an excessive burden on bulkterminal owners or operators. No direct

-requirements-would be placed onoperators of for-hire tank trucks whichload at affected loading racks. Becausethis format is the most practical means"of controlling emissions from tanktrucks, it was selected by theAdministrator as the format for the workpractice standard. ,

As discussed previously; in order toset a numerical emission limit f6r theloading operation at regulated loadingracks, the total VOC emissions wouldhave to be measurable, so that acomparison with this emission limitcould be made. Since the small portionof the displaced vapors which may leakfrom the tank trucks cannot bequantitatively measured, accuratemeasurements of total VOC emissionsfrom tank truck loading are not possible.However, the major portion of thedisplaced vapors can be measured afterthe vapors are collected at the loadingrack. Vapor collection systems typicallyinclude the equipment at the loadingrack used to contain and routeemissions, and generally consist ofhoses or arms, manifolding, piping, andcheck valves. This type of system isconsistent with the current state-of-the-art collection systems in use at manyexisting bulk terminals. Because of itsdemonstrated control effectiveness, andbecause it is not possible to set astandard of performance for the'totalemissions from the loading operation, anequipment standard requiring a vaporcollection system at each loading rackwas selected by the Administrator asthe format for controlling VOCemissions at the loading racks.

If there were leaks in the terminal'svapor collection system, some or all of

the displaced vapors would not reachthe vapor processor and would escapeto the atmosphere uncontrolled. Leak'sources can include flanges and otherconnections, valves, and pressure reliefdevices (such as those used In vaporholders). Leakage in excess of 80percent of the displaced vapors hasbeen found in some EPA tests at bulkterminals. In order for control measuresto be effective, leakage from the vaporcollection and processing equipmentmust be minimized. Section 111(h)(1) ofthe Clean Air Act directs theAdministrator to include as part of anyequipment standards promulgated under§ 111(h) "such requirements as willassure the proper maintenance of anysuch... equipment." Periodic visualmonitoring of the equipment required bythese proposed standards and repair ofobserved leaks would minimize VOCleakage without imposing anunreasonable burden on terminalowners and operators. Therefore, theproposed standards include such aninspection-and-repar requirementaimed at enhancingthe effectiveness ofthe proposed standards.

Because emissions from the vaporcollection system can be measured.standards of performance in the form ofa numerical emission limit can beapplied to the vapor collection system.Several formats for these standards ofperformance are possible. Three formatsconsidered for limiting emissions fromthe vapor collection system include aconcentration standard, a controlefficiency standard, and a massemissions standard. It Is assumed that avapor processing system would be usedunder any of these formats to achievethe required emission limit

A format expressed in terms ofconcentration would limit the VOCconcentration in the exhaust from thevapor processing system. The advantageof the concentration format Is that a testmethod to determine VOC concentrationdoes not require flow measurements.These data are required to convertconcentration measurements to massemission measurements. There are,however, several disadvantages to aconcentration format. The test dataindicate a variation in exhaust gas flowrates and'concentrations among thevarious systems. Flow rates are highthrough the thermal oxidizer system,which uses large amounts of combustionair, and are low through the refrigerationand CRA systems, which use no outsideair in their operation. In addition, thevacuum regenerated carbon adsorptionsystem uses warm purge air to enhancethe desorption. These variations inamount of dilution air would require

adjustments to compare the systems onan equal basis. The outletconcentrations also vary from system tosystem and between similar systemsmanufactured by different companies.Separate concentration limits might berequired for each type of control sytsternat each affected terminal If aconcentration format were selected.

Information from the manufacturersand reiEflts from the testing programindicate that the control efficiencies ofthe processing systems are dependenton the inlet concentration to theprocessor. The test data further indicatethat concentrations at the inlet of theprocessor vary considerably fromterminalto terminal. This variation iscaused by many factors which caninclude temperature, pressure, vaportightness of tank trucks, loading method,and whether vapor balancing of tanktrucks is used. Vapor balancing consistsof routing the vapors, displaced duringloading of the customer tank, back to thedelivery tank truck. Because of the manyfactors which may affect the vaporprocessor inlet concentration,adjustment calculations to compare allterminals on an equal Inletconcentration would be very difficult

Two forms of a mass standard formatwere considered. The first of these massformats was an adjusted mass emissionlimit. An adjusted limit methodestimates the volume of vapor loss dueto tank truck leakage and assumes thatthis vapor loss is controlled by the vaporprocessor at the same efficiency as thatmeasured during the source test. Theseestimated "processed" truck leakageemissions are then added to theemissions actually measured at thevapor processor outlet to arrive at theadjusted emission rate. This adjustmentmethod, therefore, calculates theemissions from the tank truck loadingoperation assuming there is no i'aporleakage in the vapor collection system.The adjusted emissionsmethod wasused to normalize the test results fromall the terminals tested so that allcontrol systems could be compared onan equivalent leak-free basis.

The disadvantages of the adjustedmass emission format include: (1] acomplex and expensive test procedure,and (2) the mathematical adjustment ofan accurately measured value (actualVOC mass emissions from the processoroutlet) to obtain the emission limit. Thetest procedure required to determine theadjusted limits would be identical to theprocedure used in the EPA emissiontesting program. The test procedurerequires three days of testing andrequires measurements to be taken atthe processor outlet, at the vapor

83137III I I I I III I I I I I

83138 Federal Register / Vol! 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

collection system inlet at the loading-rack, and at the tank truck hatches fordetection of leaks. This test wouldtypically cost $15,00U to conduct.

The second of the mass formats, amass standard'based upon. the. vaporprocessor outlet emissions, wouldinvolve a simpler, less expensive, andmore straightforward- test procedure.The vapor processor outlet testwouldrequire measurement of theVOC massemissions at the prdcessor outley only.The emission test procedure, therefore,would not require any mathematickladjustments of the measured VOC massemissions. The test procedure would befurther simplified by requiring only oneday of testing. It is estimated thatthistype of test would cost-from $5,000 to$10,000, depending, on the type ofprocessor being tested.

The difference between the. processoroutlet mass emission format and theadjusted mass emission format is thatthe variable of fugitive tank truckemissions due to leakage is,not takeninto account under the outletmassemissions format. However, sinceneither approach would actually controlthe fugitive emissions, additional testingcomplexity and cost are considered tobe unwarranted.Due to theseconsiderations, a mass emission format,based' on measurements at the outlet ofthe vapor processor only,,Was selected.Selection ofNumerical Em isasii Limfts;

As discussedpreviously in the section.entitled "Regulatory Alternatives," thenumerical limit for RegulatoryAlternative IV, which was selected torepresent the performdnce of the bestsystems testedjby EPA at bulkterminals. Although measured emissionsfrom all' types of processing systems arehighly variable, two, of the controltechnologies achieved consistently lowemissions. Three tests on carbonadsorption systems, nd two tests onthermal oxidation systems using a vaporholder to-release accumulated vapors tothe processor on a batch basis,indicated that these two types ofsystemsrepresented the best control'technology foi this application. Thesetwo types of control systems wereselected: to represent the beittechnological system of continuousemission? reduction, as.requiredbySection, 111 of the Clean Air Act.

The highest adjusted. daily' emiissionrate of 29 mag/liter for these two, types ofsystems led to the selection of 35 mg/liter as the emission limit for theproposed standards. It should.be notedthat any system: capable of achievingthis limit wouldJbe acceptable. Some ofthe testdata andcomments frommanufacturers of vapor processors

indicate that several types of systemscould be designed. to achieve anemission: limit of 35 mg/liter. Designvariables could include equipmentsizing, increased utilities consumption,and improved system reliability.

The vapor processors designed forVOC control at bulk gasolirle terminalsrequire regular maintenance attention inorder to consistently' achieve theemission limit for which they aredesigned Proper maintenance for theseunits generally includes frequent (atleast daily) visual inspections in, order tomonitor competent operation, fluidlevels, warning lights, pressures,temperatures, presence of leaks, andother miscellaneous items.Manufacturers frequently supplyinspection checklists to facilitate theseroutine checks, and some terminals, havedeveloped'individual lists for their ownuse. Most terminals. incorporate suchinspections intor the normal duties oftheir'maintenance personnel, whichinclude routfnelcheckT of loading racks,storage tanks, pumps, and other terminalequipment. Of course, the inspectionsthemselves do not maintain the properoperation, of vapor processors, but anynecessary repairs indicated throughatypical readings; sounds, etc., can be-implemented, rapidly to. minimizedowntime.

Each type of vaporprocessor has;different maintenance requirements dieto varying system size and complexity,types of c6mponen', and'operating' thneand sequencing. Refrigeration systemsrequire daily checks of several-subsystems. and- components; Defrostsysteni pumppressure-. as, wel as flufd.levels and temperatures, should' bechecked regularly. Oil levels, pressures,and temperatures, inthe precooler and;refrigeration systems require regularinspection. Li'quid recovery meters andcondenser coil' temperature records on.some units, indicate the level ofperformance of the units. Maintenanceon carbon adsorption systems includeschecks of cycle timing and bed vacuumand temperatures. Elapsed systemoperation time meters on some systemsprovide. an indication of proper systemoperation and can indicate maintenanceintervals. Maintenance of thermaloxidation systems may include dailyobservation, of the activation sequence'and inspection of pilots and burners.Sight ports are generally provided so.that the condition of the flame can beobserved. Vapor holders in these,systems shouldbe frequently inspectedfor leaks, and the higli and low level.switches checked for proper operation. "

All vapor processors, are provided withindicator panels towarn of '

malfunctions, and most have automaticshutdown or interlock systems. Thesesystems provide automatic indicationthat maintenance attentionmay berequired. The annual costs. to maintainvapor processing systems, includingroutine inspections and the expectedtypical repair costs, have been,considered in determining the costimpact on affected terminals.

The vapor-tight test: for gasoline tanktrucks, Reference Method 27,.wouldrequire applying a pressure of 4,500pascals (450 millimeters of water) to thedelivery tank and require that the tanksustain a pressure loss of not more than750 pascals (75 millimeters of water) in 5minutes from the'initial pressure level.The applied pressurevalue of 4,500.pascals represents the pressure at whichtank P-V vents begin to open to relievetank pressure. Thus, this value wasselected for the test' limit used todetermine tank vapor tightness. This testhas been used successfully-in Californiasince 1977. Note that only the pressuretest, and not the vacuum test, ofReference Method.27'would beapplicable under the proposedstandards. Only the pressure test isrequired because lank truck vaporleakage during product loading occursonly when the delivery tank Is underpositive pressure (product displacingvapors, out of the'tank). These limits forthe tank truck vapor-tight test representa vapor containment efficiency of 99percent after testing. However, the tanksdo not remain vapor-tight all, year. Leakscan occur in the vapor containmentequipment due to wear and-tear duringloading, lodging of foreign material onvalve seats, or equipment shock during

•over-the-road travel-Tests show that theaverage annual containment efficiencyof leak-tested tanks decreases to about90.percent;

Back pressure from, the vaporcollection and processing equipmentshould not exceed the pressure limit ofthe tank truck vapor-tight test. If theback pressure exceeds this pressurelimit, leaks may occur even from tankswhich have passed the vapor-tight, test.

Therefore, to eliminate'the problem ofsystem back pressure causing leaks inthe delivery tanks during loading, thevapor collection and processing systemsmust be designed so that the systemback pressure, measured at the loadingrack, will alwais be less than thepressure limit of the tank truck pressuretest. This is accomplished in practice byspecifying theproper piping diameterand length, minimizing the-number offlow control components such as checkvalves, and selecting a vapor processorwhich is properly sizedl to match the

Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

loading activities at the terminal.Therefore, the proposed standardswould require that the terminal'scollection and loading systems bedesigned so that the test pressure limitof 4,500 pascals (450 mnn of water) willnot be exceeded in the delivery tankduring product loading. -

The pressure-vacuum (P-V) ventscommonly used in bulk terminal vapor

-- collection systems are designed to open-to relieve any system pressure whichexceeds a predetermined value. Thesevents should not open at any pressurevalue which may occur in a normallyoperating system. Since system backpressure may reach the pressure limit ofthe tank truckpressure test, the P-Vvents must not begin to open at anypressure less than this pressure limit.Vents opening at a lower pressure couldunnecessarily allow uncontrolled VOCemissions to escape to the atmosphere.Therefore, the proposed regulationwould require that these vents have thecapacity to contain the vapors in thesystem under the operating pressurerange of the system.

Modification/lReconstructionConsiderations

Modification, as defined in § 60.14 ofChapter I, Title 40, of the Code ofFederal Regulations (CFR], occurs whenany physical or operational change to anexisting facility results in an increase inthe emission rate to the atmosphere ofany pollutant to which a standardapplies.

Investigation of the bulk gasolineterminal industry indicated that thereare several changes at a bulk terminalwhich could constitute a modificationunder § 60.14. The criteria fordetermination of modification would beapplied to the entire affected facility,which is designated as the total of allthe loading racks which service gasolinetank trucks. For example, any loadingrack conversion resulting in a netincrease in the emission rate to theatmosphere from an existing facilitycould be considered a modification, andthe existing facility would becoire anaffected facility. A second examplewould be a physical change to anexistingfacility which resulted inincreased product throughput. However,according to § 60.14(e)(2), such a changewould not be considered a modificationunless it required a capital expenditure,as defined in § 60.2. For example, theaddition of a new loading position,which would require a capitalexpenditure, to an existing facility witha resulting increase in throughput and innet emission rate would be considered amodification.

Reconstruction, as defined in § 60.15of Chapter I, Title 40 of the CFI, occurswhen the fixed capital cost ofreplacement components of an existingfacility exceeds 50 percent of the fixedcapital cost that would be required toconstruct a comparable entirely newfacility, and it is shown that it istechnically and economically feasible tomeet the applicable standards. The 50percent capital cost figure forreconstruction is a cumulative value ofthe replacement components for theexisting facility. Upon replacement ofcomponents, the Administrator woulddetermine, on a case-by-case basis,whether a reconstruction had takenplace and whether the existing facilitywould become an affected facility underthe standards.

As in the case of modification, thedetermination as to whetherreconstruction had taken place would bemade by applying the criteria to theentire affected facility, which isdesignated as the total of all the loadingracks which service gasoline tanktrucks. Again, investigation of the bulkgasoline terminal industry has indicatedcertain component repairs andreplacements which would beconsidered under the reconstructionprovisions. Top to bottom loadingconversions of the loading racks, forexample, usually exceed the 50 percentfixed capital cost criterion. If so, theseconversions would be reviewed underthe reconstruction provisions. TheAdministrator reviews theseconversions on a case-by-case basisand, as specified in § 60.15(f), his-decision is based upon the following: (1)the fixed capital costs of thereplacement components, (2] theestimated life of the facility, (3] theextent to which the components beingreplaced cause or contribute to theemissions from the facility, and (4) anyeconomic or technical limitations oncompliance with applicable standards ofperformance which are inherent in theproposed replacements. Considering theabove items, the Administrator wouldthen determine if the top to bottomloading conversion would constitute areconstruction.

Replacement or unscheduled majorrepairs of such items as loading arms,pumps, or meters may not by themselvesexceed the 50 percent replacement costof a new facility.However, since the 50percent replacement cost is a cumulativefigure, these unscheduled major repairsand replacements would be included Inreaching the 50 percent criterion.

Normal maintenance items are notincluded in this determination of the 50percent replacement cost. Normal

scheduled maintenance items includepump seals, meter calibrations, gasketsand swivels in loading arms, couplergaskets, and overfill sensor repairs.Items which typically requirereplacement under a normalmaintenance program include vaporhoses and grounding cables at theloading rack.

Selection of Performance Test MethodsThe VOC concentrations in the vapor

processor exhaust would be determinedusing either EPA Reference Method 25Aor 25B. Method 25A. "Determination ofTotal Gaseous Organic ConcentrationUsing a Flame Ionization Analyzer,"applies to the measurement of totalgaseous organic concentration of vaporsconsisting of alkanes, alkenes, and/orarenes (aromatic hydrocarbons]. Theconcentration is expressed in terms ofpropane (or other appropriate organiccompound) or in terms of organiccarbon.

A sample is extracted from the sourcethrough a heated sample line and glass'fiber filter and routed to a flameionization analyzer (FIA). Results arereported as concentration equivalents ofthe calibration gas organic constituent,carbon, or other organic compound.

Method 25B, "Determination of TotalGaseous Organic Concentration Using aNondispersive Infrared Analyzer," issimilar to Method 25A and applies to themeasurement of total gaseous organicconcentration of vapor consistingprimarily of alkanes. The concentrationIs expressed in terms of propane or interms of organic carbon. The sample isextracted as described in Method 25Aand is analyzed with a nondispersiveinfrared analyzer (NDIR). Results arereported as propane equivalents or ascarbon equivalents.

Volumetric flow rate of the exit gasesfrom the vapor processor outlet wouldbe measured using EPA ReferenceMethod 2A or 2B. Method 2A. "DirectMeasurement of Gas Volume ThroughPipes and Small Ducts," applies to themeasurement of gas flow rates in pipesand small ducts, either in-line or atexhaust positions, within thetemperature range of 0 to 50C. A gasvolume meter is used to directlymeasure gas flow. Temperature andpressure measurements are made tocorrect the volume to standardconditions.

Method 2a, "Determination of ExhaustGas Volume Flow Rate from GasolineVapor Incinerators," applies to themeasurement of exhaust volume flowrate from incinerators that processgasoline vapors consisting generally ofalkanes, alkenes, and/or arenes "(aromatic hydrocarbons). It is assumed

83139

83140 Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

that the amount of auxiliary fuel isnegligible. The incinerator exhause flowrate is determined by carbon balance.Organic carbon concentration andvolume flow rate are measured at theincinerator inlet. Organic carbon, carbondioxide, and carbon monoxideconcentrations are measured at theoutlet. The ratio of total carbon at theincinerator inlet and outlet is multipliedby the inlet volume flow rate todetermine the exhaust flow rate.

Methods 2A, 2B, 25A, and 25B areessentially the same methods used onexisting bulk gasoline terminals toestablish the majority of the data baseused in the development of the proposedstandards. The tests conducted toestablish the data base used three 8-hour test repetitions to average outenvironmental effects on the vapor-to-liquid volume (V/L) measurements,because temperature and pressurevariations in the vapor collection systemcan affect the vapor volume measured atthe inlet to the processor. These V/Lvalues were used to adjust the measuredmass emissions to account for leakage.The proposed test procedures wouldmeasure the processor outlet only anddo not require any adjustments.However, the owner or operator mayadjust the bmission results to excludemethane and ethane, which areconsidered negligibly photochemicallyreactive and do not appreciablycontribute to the formation of ozone, apolicy announced in EPA's"Recommended Policy on the Control ofVolatile Organic Compounds," 42 FR35314 (July 8,1977). No referencemethods have been prpmnulgated by EPAfor specific measurement of methaneand ethane. However, these compoundscan be measured by gaschromatographic analysis, or any othermethod approved by the administrator.Since no V/L measurements are requirefor adjustment, the proposed testprocedures incorporate one 6-houraveraging period. The test period isconsidered to represent the performanceof the vapor processing systems. Aminimum of 300,000 liters of gasolinewould have to be loaded in order for thetest period to be valid. This volume ofgasoline represents 7 to 10 truckloadings, which is considered to be theminimum number required to allowsystem performance to be adequatelyevaluated. Conducting a performancetest using these procedures would cost afacility between $5,000 and $10,000,depending on the type of processorbeing tested.

At many terminals, switch loading ispracticed, as discussed in the sectionentitled "Selection of Pollutants and

Affected Facilities." There are twomajor types of switch loading of concernwith regard to the testing of VOCemissions generated during tank truckloading. First, gasoline may be loadedinto a tank which has carred a non--volatile product, such as diesel fuel, onthe previous load. This tank wouldcontain essentially no VOC vapors, sothe VOC emissions during loadingwould be negligible. Second, a productsuch as diesel fuel may be loaded into atank which has carried gasoline on theprevious load. The VOC vapors from theprevious load of gasoline would bedisplaced by the incoming product.

At a particular terniinal the tank truckpopulation is static over the short term,and each tank truck operates at just thatone terminal. Therefore, the frequencyof each of the two types of switchloading discussed above would be aboutequal, and the quantity of VOCemissions could be accounted for byconsidering only the volume of gasolinedisplensed during a given time period.This approach to determining emissionsat a terminal would simplify the testprocedure. If the liquid volume of allproducts dispensed into gasoline tanktrucks during the performance test wereconsidered, then the liquid volume notdisplacing gasoline -vapors would haveto be subtracted form the total volumeloaded in order to correlate the VOC'mass emitted with the correspondingliquid volume. This procedure wouldrequire that each driver be asked whichproduct was carried on the previousload. Based on the information obtained,only the loadings displacing gasolinevapors would be added to obtain thetotal volume to be used in thecalculations. However, since theaccuracy of this information woulddepend on the knowledge of severalindividuals who may not know the facts,and because it may require extra testpersonnel to question the drivers, thisprocedure is not considered to be themost practical method of conducting theperformance test., The procedure which considers onlythe volume of gasoline loaded during thetest relies on a known quantity whichcan be obtained directly from dispensingmeters, instead of relying on uncertaindata. The two cases of switch loadingessentially cancel each other in terms oftheir effects on the test results.Therefore, the proposed standardswould require emissions to becalculated in terms of the total volumeof gasoline dispensed during theperformance test. Since excessivepractice of switch loading has thepotential to affect the test results byincreasing the apparent emission level,

especially if there were extraunbalanced instances of nonvolatileproduct loadings into tanks containinggasoline vapors, it is recommended thatswitch loading be minimized during theperformance test.

If there are leaks in the vaporcollection system, part of the displacedvapors will escape to the atmosphereand not be controlled by the vaporprodessor. In order for the emissionlimitation from the collection system tobe effective, any leakage in the systemshould be repaired as soon as possible.For this reason, the proposed standardswould require that the vapor collectionand processing systems, as well as theaffected loading racks, be visuallyinspected for liquid or vapor leaks on amonthly basis. The costs presented forthe proposed standards include costs forinspection of the control equipment toensure proper operation andmaintenance. Visual inspections forleaks would be part of these inspectionsand would impose no costs in additionto those already reported. Suchinspections would require perhaps onehour to accomplish, and would notimpose an unreasonable burden on aterminal owner or operator. In fact suchinspections are already q routinepractice at many bulk terminals. Underthe proposed regulation, a summary ofthe findings during the inspectionswould be required as part of thequarterly written report of excessemissions required by the GeneralProvisions, § 60.7(c). The repair Interval,i.e., the length of time allowed betweenthe detection of a leak and repair of theleak, selected for the leak Inspectionrequirement is 15 days. This repairinterval would allow effective VOCemission reduction to be maintained,while not being burdensome to theterminal operator.

In addition to the monthly inspection,Potential sources of vapor leaks wouldbe monitored immediately prior to asystem performance test using EPAReference Method 21, which applies todetermination of VOC leaks fromorganic liquid and vapor processingequipment. A portable inistrument isused to detect VOC leaks fromindividual sources. All leaks would haveto be repaired before the test wasconducted. This ensures that the vaporprocessing system is processing the totalflow of air-vapor mixture while theperformance of the system is beingevaluated.

The terminal operator should acceptvapor tightness test documentation onlyfor gasoline delivery tank truck testingconducted according to EPA ReferenceMethod 27, "Determination of Vapor

Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

Tightness of Gasoline Delivery TanksUsing Pressure-Vacuum Test" Thismethod is applicable for thedetermination of vapor tightness of agasoline delivery tank which isequipped with vapor collectionequipment The cost to perform thisannual test would be about $100, plus anaverage additional repair cost of $50.Variations on this test method areacceptable only with the approval of theAdministrator.

Selection of Monitoring Requirements

There are presently no demonstratedcontinuous monitoring systemscommercially available which monitorvapor processor exhaust VOC emissionsin the units of the proposed standid(mg/liter). This monitoring wouldrequire measuring not only VOCexhaust concentration, but also exhaustgas volume flow rate, volume of productdispensed, temperature, and pressure.

"Therefore, continuous monitoring inunits of the standard would not berequired at this time. -

Monitoring equipment is available tomonitor the operational variables

-associated with vapor processingsystem operation. Monitoring ofoperations indicates -whether the vaporprocessing system is being properlyoperated and maintained, and whetherthe processor is continuously reducingVOC emissions to an acceptable level.The variable which would yield the bestindication of system operation is VOCconcentration at the processor outlet.Extremely accurate measurementswould not be required since. the purposeof the monitoring would not be todetermine the exact outlet emissions butrather to indicate operational andmaintenance practices regarding thevapor processor. Monitors for this typeof continuous VOC measurementtypically cost about $6,000. To achieverepresentative VOC concentrationmeasurements at the processor outlet,the concentration monitoring deviceshould be installed in the exhaust ventat least two equivalent-stack diametersfrom the exit point, and protected fromany interferences due to wind, weather,or other processes.

For some vapor processing systems,monitoring of a process parameter mayyield as accurate an indication of .system operation as the exhaust VOCconcentration. For example, temperaturemonitoring in the case of thermaloxidation or refrigeration systems mayindicate proper operation and-maintenance of these systems.Parameter monitoring pquipment wouldtypically cost about $3,000. Becausecontrol system design is constantlychanging and being upgraded in this

industry, all acceptable processparameters for all systems cannot bespecified. In generil, the regulationallows for substituting the monitoring ofvapor processing system processparameters for monitoring of exhaustVOC concentration if it can bedemonstrated to the Administator'ssatisfaction that the value of the processparameter is indicative of properoperation of the processing system andis related to the exhaust VOC content.Monitoring of these parameters wouldbe approved by the Administrator on acase-by-case basis. Continuousmonitoring systeins which are a part of avapor processor's design may substitutefor the requirement to install a separatesystem, with the approval of theAdministrator.

For any system installed to monitoroperations, a recording device must alsobe installed so that a permanent timerecord of the measured parameter isproduced.

EPA has not yet developedperformance specifications for thesemonitors, but a program is underway todevelop these specifications.Consequently, until EPA proposed andpromulgated monitor performancespecifications, owners and operatorssubject to the requirement to install avapor processor continuous monitoringsystem will not be required to do so.

For purposes of excess emissionsreports required under § 60.7(c), theperiod of time selected as the averagingtime is a 6-hour clock period. This timeinterval was selected to coincide withthe time interval specified in theperformance test. The VOCconcentration or parameter limit for theexcess emissions report would bedetermined during the performance test.After EPA establishes and promulgatesmonitor performance specifications, themonitoring equipment must be operatingduring the performance test to establishthe average VOC concentration orprocess-parameter value. This averagevalue from the monitoring devicebecomes the limit for the excessemissions report. The quarterly excessemissions report would indicate theamount of time during periods of vaporprocessing system operation that theaverage value of the VOC concentrationor process parameter value exceededthe average value of the parameterestablished during the performance test.It is possible that each installation mayhave a different monitoring limit.

Impacts of Reporting Requirements

The proposed standards for bulkgasoline terminals would require theterminal operator to keep on filedocumentation that all gasoline delivery

tank trucks loading at the terminal hadpassed an annual vapor-tight testperformed according to Method 27. Thedocumentation would include the nameof the tester, the test location and date,and the test results. These recordswould be kept on file at the terminal in apermanent form available for inspection,and would be updated at least once peryear to reflect current information. Theother typQ of report required under theproposed standards would be asummary report reflecting the findingson the monthly leak inspection. Thepreparation and filing of this reportwould represent only a modest increasein a bulk terminal's reportingrequirements. These reports would besubmitted quarterly with each report ofexcess emissions required under theGeneral Provisions.

The General Provisions require threeadditional types of reports. First, thereare notification requirements whichwould enable the Agency to keepabreast of facilities subject to thestandards of performance. Second, therewould be reporting of performance testresults which would show that a facilityis meeting the standards initially. Third,there would be quarterly reports ofexcess emissions which would bequarterly reports of excess emissionswhich would permit the Agency to

,determine whether the emissiont controlsystem installed to comply with thestandards is being properly operatedand maintained.

The resources needed by the industryto maintain records and to collect,prepare, and use the reporting throughthe first five years after proposal of thestandard would be about 26 man-years.

Public Hearing

A public hearing will be held todiscuss the proposed standards inaccordance with Section 307(d)(5) of theClean Air Act. Persons wishing to makeoral presentations should contact EPAat the address given in the ADDRESSESsection of this preamble. Oralpresentations will be limited to 15minutes each. Any member of the publicmay rle a written statement before,during, or within 30 days after thehearing. Written statements should beaddressed to the Central Docket Sectionaddress given in the ADDRESSESsection of this preamble.

A verbatim transcript of the hearingand written statements will be availablefor public inspection and copying duringnormal working hours at EPA's CentralDocket Section in Washington, D.C. (seeADDRESSES section of this preamble).

83141

83142 Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

DocketThe docket is an organized and

complete file of all the informationsubmitted to or otherwise considered inthe development of this proposedrulemaking. The principal purposes ofthe docket are (I)'to allow interestedparties to readily identify and locatedocuments so that they can intelligentlyand effectively participate in therulemaking process, and (2) to serve asthe record in case of judicial review.

MiscellaneousAs prescribed by Section.111,

establishment of standards ofperformance for bulk gasoline terminalswas preceded by the Administrator'sdetermination (40 CFR 60.16, 44 FR49222, dated August 21, 1979) that thesesources contribute significantly to airpollution which may reasonably beanticipated to endanger public health orwelfare. In accordance with Section 117of the Act, publication of this proposalwas preceded by consultation withappropriate advisory committees,independent experts, and Federaldepartments and agencies. TheAdministrator will welcome commentson all hspects of the proposedregulation, including economic andtechnological issues, monitoringrequirements, and proposed testmethods.

It should be noted that standards ofperformance for new sourcesestablished under Section 111 of theClean Air Act reflect:

* * * application of the best technologicalsystem of continuous emission reductionwhich (taking into consideration the cost ofachieving such emission reduction, and anynonair quality health and environmentalImpact and energy requirements) theAdministrator determines has beenadequately demonstrated [Section 111(a)(1)].

Although there may be emissioncontrol technology available that canreduce emissions below those levelsrequired to comply with standards ofperformance, this technology might notbe selected as the basis of standards orperformance due to costs associatedwith its use. Accordingly, standards ofperformance should not be viewed asthe ultimate in achievable emissioncontrol. In fact, the Act requires (or hasthe potential for acquiring) theimpositi6n of a more stringent emissionstandard in several situations.

For example, applicable costs do notnecessarily play as prominent a role indetermining the "lowest achievableemission rate" for new or modifiedsources locating in non-attainmentareas; i.e., those areas where statutorily-mandated health and welfare standards"

are being violated. In this respect,Section 173 of the Act requires that newor modified sources constructed in anarea where ambient pollutant .concentrations exceed the NationalAmbient Air Quality Standards(NAAQS) must reduce emissions to the'level that reflects the "lowestachievable emission rate" (LAER), asdefined in Section 171(3), for suchcategory of source. The statute definesLAER as-that rate of emissions based onwhichever of the following is morestringent:

(A) the most stringent emissionlimitation which is contained in theimplementation plan of any State forsuch class of category of source, unlessthe owner or operator of the proposedsource deomonstates that suchlimitations are not achievable, or

(B) the most stringent emissionlimitation which is achieved in practiceby such class or category of source.

In no event may the emission rate'exceed any applicable new sourceperformance standard [Section 171(3)].

A similar situation may arise underthe prevention of significantdeterioration of air quality provisions ofthe Act (Part C). These provisions 'require that certain sources [referred toin Section 169(1)] employ "bestavailable control technology" (BACT) asdefined in Section 169(3) for allpollutants regulated under the Act. Bestavailable control technology must bedetermined on a case-by-case basis,taking energy, environmental, andeconomic impacts and other costs intoaccount. In no event may the applicationof BACT result in emissions of anypollutants which exceed the emissionsallowed by any applicable standardestablished pursuant to Section 111 (or112) of the Act

In all events, State ImplementationPlans (SIPs) approved or promulgatedunder Section 110 of the Act mustprovide for the attainment andmaintenance of NAAQS designed-toprotect public health and welfare. Forthis purpose, SIPs may in some casesrequire greater emission reductions thanthose required-by standards ofperformance for new sources.

Finally, States are free under Section116 of the Act to establish even morestringent emission limits than thoseestablished under Section 111 or thosenecessary to attain or maintain theNAAQS under Section 110. Accordingly,new sources-may in some cases besubject to limitations more stringentthan standards of performance underSection 111, and prospective owners andoperators of new sources should beaware of this possibility, in planning forsuch facilities.

This regulation will be reviewed fouryears frdm the date of promulgation asrequired by the Clean Air Act. Thisreview will include an assessment ofsuch factors as the need for integrationwith other programs, the existence ofalternative methods; enforceability,improvements in emission controltechnology, and reporting requirements.The reporting requirements in thisregulation will be reviewed as requiredunder EPA's sunset policy for reportingrequirements in regulations.,

Section 317 of the Clean Air Actrequires the Administrator to prepare aneconomic impact assessment for anynew source standard of performancepromulgated under Section 111(b) of theAct. An economic impact assessmentwas prepared for the proposedregulations and for other regulatoryalternatives. All aspects of theassessment were considered in theformulation of the proposed standardsto ensure that the proposed standardswould represent the best system ofemission reduction considering costs,The economic impact assessment Isincluded in the Background InformationDocument.

Dated: December 8, 1980.Douglas M. Costle,Adzninlsltoor.

It is proposed that 40 CFR Part 00 beamended as follows:

1. By adding a new subpart as follows:Subpart XX-Standards of Performance forBulk Gasoline TerminalsSec.60.500 Applicability and designation of

affected facility.60.501 Defintions.60.502 Standards for volatile organic

compound emissions from bulk gasolineterminals.

60.503 Test methods and procedures,60.504 Monitoring of operations.60.505 Recordkeepng.

Authority: Sections 111 and 301(a) of theClean Air Act, as amended, (42 U.S.C. 7411,7601(a)), and additional authority'as notedbelow.

Subpart XX-Standards ofPerformance for Bulk GasolineTerminals

§ 60.500 Applicability and designation ofaffected facility.

(a) The affected facility to which theprovisions of this subpart apply is thetotal of all the loading racks at a bulk

.gasoline terminal which deliver liquidproduct into gasoline tank trucks.

(b) Each facility under paragraph (a)of this se'ction that commencesconstruction or modification after(date of publication in Federal Register

Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

is subject to -the provisions of thissubpaiL

(c) The provisions of § 60.504 will notapply until EPA has established andpromulgated performance specificationsfor the monitoringdevices. After thepromulgation of performancespecifications, these provisions willapply to Bach affected facility underparagraph (b) of-this section.

§ 60.501 Definitions.The terms used in this subpart are

defined in the Clean Air Act, in § 60.2 ofthis part, or in this section as follows:

"Bulk gasoline terminal" means anywholesale gasoline outlet whichreceives gasoline by pipeline, ship, orbarge.

"Continuous vapor processingsystem'' means a VOC vapor processingsystem-that treats VOC vapors collectedfrom gasoline tank trucks on a demandbasis without intermpdiateaccumulation in a vapor holder.

."Gasoline" means any petroleumdistillate or petroleum distillate/alcoholblend having a Reid vapor pressure of'27.6 kilopascals or greater which is usedas a fuel for internal combustionengines.I "Gasoline tank truck" means adelivery tank truck used at bulk gasolineterminals which is loading gasoline or-which has loaded gasoline on theimmediately previous load.

"Intermittentvapor processingsystem" means a VOC vapor processingsystem that employs an intermediatevapor holder to accumulate the collectedvapors from gasoline tank trucks, andtreats the accumulated vapors onlyduring automatically controlled cycles.

"Loading rack" means the loadingarms, pumps, meters, shutoff valves,relief valves, check valves, electricalgrounding, and lighting necessary to filldelivery tank trucks.

"Vapor collection system" means anyequipment used for containing VOCvapors displaced during the loading ofgasoline tank trucks.

"Vapor processing system" means anyequipment -used for recovering oroxidizing VOC vapors.

"Vapor-tight gasoline tank truck"means a gasoline tank truck which hasdemonstrated within the 12 precedingmonths that its product delivery tankwill sustain a pressure change of notmore than 750 pascals (75 mm of water)within 5 minutes after it is pressurizedto 4,500 pascals (450 mm of water). Thiscapability is to be demonstrated usingthe pressure test procedure specified in

. Reference Method 27."Volatile organic compound (VOC)"

means any organic compound whichparticipates in atmospheric -

photochemical reactions; or which ismeasured by Reference Methods 25A,25B, and 21.

§ 60.502 Standard for Volatile OrganicCompound (VOC) emissions from bulkgasoline terminals.

On and after the date on which the-performance test required under by§ 60.8 ii completed, the owner oroperator of a bulk gasoline terminalcontaining an affected facility shallcomply with the requirements of thissection.

(a) Each loading rack which loadsgasoline tank trucks shall be equippedwith a vapor collection system designedto collect the VOC vapors displacedfrom tank truck vapor collection systemsduring loading.

(b) The bulk gasoline terminal's vaporcollection system shall be designed toprevent any VOC vapors collected atone loading rack from passing toanother loading rack.

(c) The emissions to the atmospherefrom the bulk gasoline terminal's vaporcollection system due to the loading ofliquid product into gasoline tank trucksare not to exceed 35 milligrams of VOCper liter of gasoline loaded.

(d) Loadings of liquid product intogasoline tank trucks shall be restrictedto vapor-tight gasoline tank trucks only.

(e) Loadings of liquid product intogasoline tank trucks shall be restrictedto those equipped with vapor recoveryequipment that is compatible with thebulk gasoline terminal's vapor collectionsystem.

(f) The bulk gasoline terminal's andthe tank truck's vapor collectionsystems shall be connected during eachloading of a gasoline tank truck.

(g) The vapor collection and liquidloading equipment shall be designed andoperated to prevent gauge pressure inthe delivery tank from exceeding 4,500pascals (450 mm of water). This level isnot to be exceeded when measured bythe procedures specified in § 60.503(b).

(h) No pressure-vacuum vent in thebulk gasoline terminal's vapor collectionsystem shall begin to open at a systempressure less than 4,500 pascals (450 mmof water).

(i) Each calendar month, the vaporcollection system, the vapor processingsystem, and each loading rack handlinggasoline shall be visually inspectedduring the loading of gasoline tanktrucks for liquid or vapor VOC leaks.Each detection of a leak shall berecorded and the source of the leakrepaired within 15 calendar days after Itis detected. A summary of each set ofthree consecutive inspection recordsshall be submitted with the nextquarterly report required under § 60.7(c).

§60.503 Test methods and procedures-

(a) For the performance tests, § 60.8(f)does not apply.

(b) For the purpose of determiningcompliance with the pressure regulationof § 60.502(g), the following proceduresshall be used.

(1) Calibrate and install a liquidmanometer, or equivalent, capable ofmeasuring up to 500 mm of water gaugepressure with ±2.5 mm of waterprecision.

(2) Connect the manometer to apressure tap in the terminal's vaporcollection system, located as close aspossible to the connection with thedelivery tank.

(3) During the performance test, readand record the pressure every 5 minuteswhile a delivery tank is being loaded.

(c) For the purpose of determiningcompliance with the VOC massemission limitation of § 60.502(c), thefollowing reference methods shall beused:

(1) For the determination of volume atthe exhaust venL

(I) Method 2B for combustion vaporprocessing systems.

(ii) Method 2A for all other vaporprocessing systems.

(2) For the determination of VOCconcentration at the exhaust vent,Method 25A or 25B. The calibration gasshall be either propane or butane.

(d) Immediately prior to a -performance test required fordetermination of compliance with§ 60.502(c) and (g, all potential sourcesof vapor leakage in the terminal's vaporcollection system equipment shall bemonitored for leaks using Method 21. Areading of greater than or equal to 10,000ppmv as methane shall be considered aleak. All leaks shall be repaired prior toconducting the performance test.

(e) The test procedure for determiningcompliance with § 60.502(c) and g) is asfollows:

(1) The time period for a performancetest shall be as follows:

(i) For continuous vapor processingsystems, not less than 6 hours, duringwhich at least 300,000 liters of gasolineare loaded.

(ii) For intermittent vapor processingsystems, not less than 6 hours, duringwhich at least 300,000 liters of gasolineare loaded and at least two full cycles ofoperation of the vapor processingsystem occur. The end of theperformance test shall coincide with theend of a cycle of operation.

(2) All testing equipment shall beprepared and installed as specified inthe appropriate test methods.

(3) For intermittent vapor processingsystems, the system shall be manuallystarted and allowed to process vapors

83143

83144 Federal Register / Vol. 45, No. 244 /Wednesday, December 17, 1980 / Proposed Rules

already in the vapor holder until thelower automatic cutoff is reached. Thisshould be done immediately prior to thebeginning of testing.

(4) An emission testing interval duringthe performance test shall consist ofeach 5 minute period orlincrementthereof, while the vapor processingsystem is operating; and each 15 minuteperiod or increment thereof, while thevapor processing system is notoperating.

(5) For each testing interval:(i) The reading from each

measurement instrument shall berecorded, and

(ii) The volume exhausted and theaverage VOC concentration in theexhaust vent, as specified in theappropriate test method, shall be;determined.

(6) The volume of gasoline dispensedduring the performance test period at allloading racks whose vapor emissionsare controlled by the processing systembeing tested shall be determined. Thismay be determined from terminalrecords or from gasoline dispensingmeters at.each loading rack.

(7) The mass emitted for each testinginterval shall be calculated as follows:M,=10-6K V Cewhere:M0=mass of VOC emitted at the

exhaust vent, mg.V,.=volume of air-vapor mixture

exhausted, m3 at standardconditions.

C=VOC concentration (as measured)at the exhaust vent, ppmv.

K=density of calibration gas, mg/m, atstandard conditions

--1.83 X106 for propane=2.41 X 10 ; for butane.

s= standard conditions, 20°C and 760mm HSg.

(8) The VOC emissions shall becalculated as follows:

n1. M

E=

where:E=mass of VOC emitted per volume of

gasoline loaded, mg/l.L= total volume of gasoline loaded, 1.M, 1=mass of VOC emitted f6r each

testing interval i, mg.n =number of testing intervals.

(f) The owner or operator may adjustthe emissJon results to exclude themethane and ethane contbnt in theexhaust vent by any method approvedby the Administrator.(Sec. 114 of the Clean Air Act as amended (4ZU.S.C. 7414))

§ 60.504 Monitoring of operations.(a) The owner or operator of each

affected facility shall install, calibrate,operate, and maintain a monitoringsystem to continuously measure theVOC concentration of the exhaust ventstream, of the vapor processing systemto determine the proper operation ofeach system.

(b] Upon application to theAdministrator, monitoring of a. vaporprocessing system process parametermay be substituted for the measurementof the exhaust vent VOC content, if itcan be demonstrated to theAdministrator's satisfaction that thevalue of the process parameter isindicative of proper operation of the.system. and is related to the exhaustvent VOC content. Monitoring ofprocess parameters must be approvedon a case-by-case.basis by theAdministrator.

(c] Each monitoring device shall beinstalled, calibrated, operated, andmaintained according to' acceptedpractices and the manufacturer'sspecifications.

(d) The VOC concentration monitoringdevice shall be installed in a locationthat is representative of the VOCconcentration in the exhaust vent, atleast two equivalent stack diametersfrom. the exhaustpoint, and protectedfrom any interferences due to wind,weather. or other processes.

(e) Each monitoring device shall beequipped with a recordin& device so thata permanent time record of themeasured process parameter isproduced-

(f) The exhaust vent VOC.concentration or approved processparameter shall be continuouslymeasured and recorded during theperformance test required under § 60.8.

(g) For the purposes of reportsrequired under § 60.7(c), periods ofexcess emissions are-defined as any 6-hour clock periods during which theaverage value of the exhaust vent VOCconcentration or measured processparameter, during periods of vaporprocessing system operation, differs.from the'average value measured duringthe performance test required under§ 60.8.

(h) The owner or operator of eachaffected facility shall install and operateall monitoring equipment before

conducting the performance testrequired under § 60.8.(Sec. 114 of the' Clean Air Act as amended (42U.S.C. 7414))

• 60.505 Recordkeeplng.(a) The owner or operator of each

bulk gasoline'terminal containing anaffected facility shall keep on filedocumentation that each gasoline tanktruck loading at that terminal is a vapor-tight.gasoline tank truck. Thisdocumentation shall be kept on file atthe terminal in a permanent formavailable for inspection.

(b) The documentation file for eachgasoline tank truck shall be updated atleast once per year to reflect current testresults as determined by Method 27.This documentation shall Include, as aminimum, the following information:

(1) Test Short Title: Gasoline DeliveryTank Pressure Test-EPA Test Method27.

(2) Tank Owner and Address.(3) Tank ID Number.(41 Testing Location.(5) Date of Test.(6) TesterName and Signature.(7) Witnessing Inspector, if any:

Name, Signature, and Affiliation.(8) Test Results: Actual Pressure'.

Change in 5 minutes, m of water(average for 2 runs).

(c) The owner or operator of each bulkgasoline terminal containing an affectedfacility shall keep on file at the terminala record of each monthly leak inspectionrequired under § 60.502(i). Inspectionrecords shall include, as a minimum, thefollowing information:

(1) Date of nspection.(2) Findings (may indicate no leaks

discovered; or location, nature, andseverity of each leak).

(31 Corrective Action (date each leakrepaired; reasons for any repair intervalin excess of 15 days).

(4) Inspector Name and Signature.(Sec. 114 of the Clean Air Act as amended (4ZU.S.C. 7414))

2. By adding five new ReferenceMethods (Method 2A, Method,2B,Method 25A, Method 25B, and Method27) to Appendix A as follows:

Appendix A-Reference Methods

Method 2A. Direct Measurement of GasVolume Through Pipes and Small Ducts1. Apphlcability and Principe

1.1 Applicability. This method applies tothe measurement of gas flow rates In pipesand small ducts,'either in-line or at exhaustpositlors. Vvithin the teterature range of 0to 50C. .

1.2 Principle. A gas volume motor is usedto directly measure gasi volume. Temperature

Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

and pressure measurements are made to"correct the volume to standard conditions.2. Apparatus

Specifications for the apparatus are givenbelow. Any other apparatus that has beendemonstrated (subject to approval of heAdministrator) to be capable of meeting thespecifications will be considered acceptable.

2.1 Gas Volume Meter. A positivedisplabement meter, turbine meter, or otherdirect volume measuring device capable ofmeasuring volume to within 2 percent. themeter shall be equipped with a temperaturegauge (2 percent of the minimum absolute-temperature) and a pressure gauge (±h2.5 mmHg). The manufacturer's recommendedcapacity of the meter shall be sufficient forthe expected maximum and minimum flowrates at the sampling conditions,Temperature, pressure, corrosivecharacteristics, and pipe size are factorsnecessary to consider in choosing a suitablegas meter.

2.2 Barometer. A mercury, aneroid, orother barometer capable of measuringatmospheric pressure to within, 2.5 mm Hg. Inmany cases, -the barometric reading may beobtained from a nearby national weatherservice station, in which case the stationvalue (which is the absolute barometricpressure) shall be requested, and anadjustment for elevation differences betweenthe weather station and the sampling point

- shall be applied at a rate of minus 2.5 mm Hgper 30-meter elevation increase, or vice-versafor elevation decrease.

2.3 Stopwatch. Capable of measurementto within I second.3.-Procedure -

_3.1 Installation As there are numeroustypes of pipes and small ducts that may be-subject to volume measurement, it would bedifficult to describe all possible installationschemes. In general, flange fittings should beused for all connections wherever possible.

* Gaskets or other seal materials should beused to assure leak-tight connections. The.volume meter should be located so as toavoid severe vibrations and other factors thatmay affect the meter calibration.

3.2 Leak'Test A volume meter installed ata location under positive pressure may beleak-checked at the meter connections byusing a liquid leak detector solutioncontaining'a surfactant. Apply a smallamount of the solution to the connections. If aleak exists, bubbles will form, and the leakmust be corrected.

A volume meter installed at a location* undernegative pressure is very difficult to

t~st for leaks without blocking flow at theinlet of the line and watching for metermovement If this procedure is not possible,visually check all connections ind assuretight seals. i ' ' , I

3.3 Volum Meastirement.-3.3.1 For sources with continuous, steady

emission flow rates, record the initial metervolume reading, meter temperature(s), meterpressure, and start the stopwatch.Throughout the test period, record the metertemperature(s) and pressure so'that averagevalues can be determined. At the end of thetest stop the timer and recoid thb'elapsedtime, the.final volume reading, meter

temperature(s), and pressure. Record thebarometric pressure at the beginning and endof the test run. Record the data on a tablesimilar to Figure 2A-1.

3.3.2 For sources with noncontinuous,non-steady emission flow rates, use theprocedure in 3.3.1 with the addition of thefollowing. Record all the meter parametersand the start and stop times corresponding toeach process cyclical or noncontinuous event.4. Calibration

4.1 Volume Meter. The volume meter iscalibrated against a standard reference meterprior to its initial use In the field. The treference meter Is a spirometer or liquiddisplacement meter with a capacityconsistent with that of the test meter.Alternative references may be used uponapproval of the Administrator.

Set up the test meter in a configurationsimilar to that used in the field Installation(i.e, in relation to the flow moving device).Connect the temperature and pressure gaugesas they are to be used in the field. Connectthe reference meter at the inlet of the flowline, if appropriate for the meter, and begingas flow through the system to condition themeters. During this conditioning operation,check the system for leaks.

BILUNG CODE 6S60-26-M

83145

Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

Plant

Date

Sample Location-

Barometric Pressure mi Hg

Operators.

Meter Number

Run Number____________

Start Finish-

Meter Calibration Coefficient.

Last Date Calibrated_

StaticTime Volume pressure

Meter TemperatureRun/clock reading mm H q 0C OK

Average

BILUNG CODE 6560-26-C

Figure 2A-I. Volume flow rate measurement data.

I8qim;!

Federal Register /Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

The calibration shall be run over at leastthree different flow rates. The calibrationflow rates shall be about 0.3. 0.6. and 0.9times the meter's rated maximum flow rate.

For each calibration run, the data to becollected include: reference meter initial andfinal volume readings, the test meter initial

SVrf - V ri)(t r + 273,)

iM _(Vf- V -)(t + 273)

Where:Y,,=Test volume meter calibration

coefficient, dimensionless.Vr=Reference meter volume reading, ma.V =Test meter volume reading, m.t--Reference meter average temperature. *C.t. =Test meter average temperature, *C.Pb=Barometric pressure, mmHg.P,=Test meter average static pressure, in

Hg.f=Final reading for rum.i=Initial reading for run.

Compare the three Y. yalues at each ;f theflow rates tested and determine themaximum and minimumvalues. Thedifference between the maximum andminimum values at each flow rate should beno greater than 0.030. Extra runs may berequired to complete this requirement. If thisspecification cannot be met in six successiveruns, thetest meter is not suitable for use. Inaddition, th6 meter coefficients should bebetween 0.95 and 1.05. If these specificationsare met at all the flow rates, average all theY,,, values for an average meter calibrationcoefficient, fi'.

The procedure above shall be performed atleast once for each volume meter. Therefore,an abbreviated calibration check shall becompleted after each field test. Thecalibration of the volume meter shall bechecked by performing three calibration runsat a single, intermediate flow rate (based onthe previous field test] with the meterpressure set at the average value encounteredin the field test. Calculate the average valueof the calibration factor. If the calibration haschanged by more than 5 percent recalibratethe meter over the full range of flow asdescribed above. Note: If the volume metercalibration coefficient values obtained before

5.2 Volume.

and final volume reading, mettemperature and pressure, barpressure, and run time. Repeaeach flow rate at least three ti

Calculate the test meter callcoefficient, Y,,, for each run as

P b"(Pb' . P

er averageometrict the runs atmes.brationfollows:

Eq. 2A-1

and after a test series differ by more than 5percent. the test series shall either bo voided.or calculations for the test series shall beperformed using whichever meter coefficientvalue (i.e. before or after) gives the greatervalue of pollutant emission rate.

4.2 Temperature Gauge. After each testseries, check the temperature gauge atambient temperature. Use an ASTM mercury-in-glass reference thermometer, or equivalent,as a reference. If the gauge being checkedagrees within 2 percent (absolutetemperature] of the reference, thetemperature data collected in the field shallbe considered valid. Otherwise, the test datashall be considered invalid or adjustments ofthe test results shall be made, subject to theapproval of the Administrator.

4.3 Barometer. Calibrate the barometerused against a mercury barometer prior to thefielrd tes L5. Calculations

Carry out the calculations, retaining atleast one extra decimal figure beyond that ofthe acquired data. Round off figures after thefinal calculation.

5.1 Nomenclature.Pb=Barometric pressure, mn Hg.P,=Average static pressure in volume meter,

mmHg.%,=Gas flow rate. m3/min, standard.

conditions.T,=Average absolute meter temperature, OK.V,=Meter volume reading. ml.

"Y=Meter calibration coefficient,dimensionless.

f=Final reading for run.i=Initial reading for run.s=Standard conditions, 20' C and 760 nun

Hg.O=Elapsed run time, min.

0. References0.1 United States Environmental

Protection Agency. Standards of Performancefor New Stationary Sources, Revisions toMethods 1-8. Title 40, part 60. Washington,D.C. Federal Register Vol. 42; No. 10. August18.1977.

02 Rom. Jerome J. Maintenance.Calibration. and Operation of IsokineticSource Sampling Equipment U.S.Environmental Protection Agency. ResearchTriangle Park. N.C. Publication No. APTD-0576. March 1972.

6.3 Wortman.Martin. R.Vollaro, and P.R.Westlin. Dry Gas Volume Meter Calibrations.Source Evaluation Society Newsletter. VoL 2,No. 2. May 1977.

6.4 Westlin. P.R. andR. T. Shigehara.Procedure for Calibrating and Using Dry GasVolume Meters as Calibration Standards.Source Evaluation Society Newsletter. VoL 3.No. 1. February 1976.

Method 2B-Determination of Exhaust GasVolume Flow Rate From Gasoline VaporIncinerators1. Applica bilitand Prnc ple

1.1 Applicability. This method applies tothe measurement of exhaust volume flow ratefrom incinerators that process gasolinevapors consisting of generally non-methanealkanes, alkenes, and/or arenes [aromatichydrocarbons). It is assumed that the amountof auxiliary fuel is negligible.

1.2 Principle. The incinerator exhaustflow rate is determined by carbon balance.Organic carbon concentration and volumeflow rate are measured at the incineratorinlet. Organic carbon, carbon dioxide tC0a=,and carbon monoxide (COl concentrationsare measured at the outlet. Then the ratio oftotal carbon at the incenerator inlet andoutlet is multiplied by the inlet volume todetermine the exhaust volume and volumeflow rate.2. Apparatus

2.1 VolumeMeter. Equipment describedIn Method 2A.

PR + PVms 2 .35 -m (mf " mi ) ( Tm

5.3 Gas Flow Rate.

Qs - Vmss e

Eq. 2A-2

Eq. 2A-3

83147

83148 Federal Register / Vol. 45,'No. 244 / Wednesday, December 17, 1980 / Proposed Rules

2.2 Organic Analyzers (2). Equipmentdescribed in Method"25A or 25B.

2.3 CO Analyzer.-Equipment described inMethod 10.

2.4 CO 2 Analyzer. A nondispersiveinfrared (NDIR) CO2 analyzer and supporting,equipment described in Method 10.

3. Procedure

3.1 Inlet Installation. Install a volumemeter in the vapor line to incinerator inletaccording to the procedure in Method 2A. Atthe volume meter inlet, install a sample probeas described in Method 25A. Alternatively, asingle opening probe may be used so that agas sample is collected frbm the centrallylocated 10 percent area of the vapor linecross-section. Connect to the probe a leak-tight, heated (if necessary to preventcondensation) sample line (stainless steel orequivalent) and an organic analyzer systemas described in Method 25A or 25B.

3.2 Exhaust histallation. Three analyzersare required for the incinerator exhaust-CO2 , CO, and organic. A sample manifoldwith a single sample probe may be used.Install a sample probe as described Method25A or, alternatively, a single opening probepositioned so that a gas sample is collectedfrom the centrally located 10 percent area ofthe stack cross-section. Connect a leak-tightheated sample line to the sample probe. Heatthe sample line sufficiently to prevent anycondensation.

3.3 Recording Requirements. The outputof each analyzer must be permanentlyrecorded on an analog strip chart, digitalrecorder, or other recording device. The chartspeed or number of readings per time unitmust be similar for all analyzers so that datacan be correlated, The minimum data.recording requirement for each analyzer isone measurement value per minute during theincinerator test period.

3.4 Preparation. Prepare and calibrate allequipment and analyzers according to the'procedures in the respective methods. Allcalibration gases must be'introduced at theconnection between the probe and thesample line. If a manifold system is used forthe exhaust analyzers, all the analyzers andsample pumps must be operating when thecalibrations are done. Note: For the purposesof this test, methane should not be used as anorganic calibration gas.

3.5 Sampling. At the beginning of the testperiod, record the initial parameters for theInlet volunme meter according to theprocedures in Method 2A and mark all of therecorder strip charts to indicate he start ofthe test. Continue recording inlet organic andexhaust CO., CO. and organic concentrationsthroughout the test. During periods of processinterruption and halting of gas flow, stop thetimer and mark the recorder strip charts sothat data from this interruption are notincluded in the calculations. At the end of the

test period, record the final parameters forthe inlet volume meter and mark the end onall of the recorder strip charts.

3.6 Post Test Calibrations. At theconclusion of the sampling period, introducethe calibration gases as specified in therespective reference methods. If analyzeroutput does not meet the specifications of themethod, invalidate the test data for thatperiod. Alternatively, calculate the volumeresults using initial calibration data and usingfinal calibration data and report bothresulting volumes. Then, for emissionscalculations, use the volume measurementresulting in the greatest emission rateconcentration.

4. Calculations

Carry out the calculations, retaining at -least one extra decimal figure beyond that ofthe acquired data. Round off figures after thefinal calculation.

4.1 NomenclatureCO.-Mean carbon monoxide concentration

in system exhaust, ppmv.C0 2,-Mean carbon dioxide concentration in

system exhaust, ppmv.HC-Mean organic concentration in system

exhaust as defined by the calibrationgas, ppmv.

HCr-Mean organic concentration in systeminlet as defined by the calibration gas,ppmv.

K-Calibration gas factor=2 for othanocalibration gas.

=3 for propane calibration gas.=4 for butane calibration gas.

V&-Exhaust gas volume, ms.Vt,-Inlet gas volume, in.Q~c-Exhaust gas volume flow rate, m3/mln.Ql-Inlet gas volume flow rate, m3/mIn.0-Sample run time, min.s-Standard Conditions: 20'C, 760mm 1Hg.300-Estimated concentration of ambient

CO2 ppmv. (CO. colncentation in theambient air may be measured during thutest period using an NDIR and the meanvalue substituted into the equation.)

4.2 Concentrations. Determine meanconcentrations of inlet organics, outlet CO2,CO, and outlet organics according to theprocedures in the respective me.hods and thanalyzers' calibration curves, and for thetime intervals specified In the applicableregulations. Concentrations should bedetermined on a parts per million by volume(ppmv) basis.

4.3 Exhaust Gas Volume. Calculate theexhaust gas volume as follows:

K(HCi) ...

es is K(HCe) + CO +CO - 300 Eq. 28-1

4.4 Exhaust Gas Volume Flow Rate. Calculate the exhaust

gas volume flow rate as follows:

q - Veses 9

5. References

5.1 Measurement of Volatile OrganicCompounds. U.S. Environmental Protection

- Agency. Office of Air Quality Planning andStandards. Research Triangle Park, N.C.27711. Publication No. EPA-450/2-78-041.October 1978. p. 55.

5.2 Method 10-Determination of CarbonMonoxide Emissions from StationarySources. U.S. Environmental ProtectionAgency. Code-of Federal Regulations. Title40, Chapter 1. part 60, Appendix A.Washington, D.C. Office of the FederalRegister. March 8,1974.

5.3 Method 2A-Deterinination of GasFlow Rate in Pipes and Small Ducts.Tentative Method. U.S. Environmental

Eq. 28-2

Protection Agency. Office of Air QualityPlanning and Standards. Research TrianglePark, N.C. 27711. March 1980,

5.4 25A-Determination of Total GaseousOrganic Compounds Using a Flamelonizdtion Analyzer. Tentative Method. U.S.Environmental Protection Agency. Office ofAir Quality Planning and Standards.Research Triangle Park, N.C. 27711. March1980.

5.5 Method 25B--Detormination of TotalGaseous Organic Compounds Using aNondispersive Infrared Analyzer. TentativeMethod, U.S. Environmental ProtectionAgency. Office of Air Quality Planning andStandards. Research Triangle Park, N,C,277.11. March 1980.

Federal Register / Vol. 4, No. 244 [ Wednesday, December 17, 1980 / Proposed Rules

Method 25A-Determination of TotalGaseous Organic Concentration Using aFlame Ionization Analyzer1. Applicability and Principle

1.1 Applicability. This method applies tothe measurement of total gaseous organicconcentration of vapors consisting ofnonmethane alkanes, alkenes, and/or arenes(aromatic hydrocarbons). The concentrationis expressed in terms of propane (or otherappropriate organic compound) or in terms oforganic carbon.

1.2 Principle. A gas sample is extractedfrom the source, through a heated sampleline, if necessary, and glass fiber filter to aflame ionization analyzer (FIA]. Results arereported as concentration equivalents of thecalibration gas organic constituent, carbon, orother organic compound.2. Definitions

2.1 Measurement System. The totalequipment required for the determination ofthe gas concentration. The system consists ofthe following major subystems:

2.1.1 Sample Interface. That portion of thesystem that is used for one or more of thefollowing: sample acquisition, sampletransportation, sample conditioning, orprotection of the analyzer from the effects ofthe stack effluent.

2.1.2 Organic Analyzer. That portion ofthe system that senses organic concentrationand generates an output proportional to thegas concentration.

2.2 Span Value. The upper limit of a gasconcentration measurement range that isspecified for affected source categories in the-applicable part of the regulations. Forconvenience, the span value shouldcorrespond to 100 percent of the recorderscale.

2.3 Calibration Gas. A knownconcentration of a gas in an appropriatediluent gas.

2.4 Zero Drift. The difference in themeasurement system output regdings beforeand after a stated period of operation duringwhich no unscheduled maintenance, repair,or adjustment took place and the inputconcentration at the time of the

- measurements were zero.2.5 Calibration Drift. The difference in the,

measurement system output readings beforeand after a stated period of operation durirgwhich no unscheduled maintenance, repair,or adjustment took place and the inputconcentration at the time of themeasurements was a mid-level value.3. Apparatus

A schematic of an acceptable measurementsystem is known in Figure 25A-1. Theessential components of the measurementsystem are described below:

3.1 Organic Concentration Analyzer. Aflame ionization analyzer (FIA) capable ofmeeting or exceeding the specifications inthis method.

3.2 Sample Probe. Stainless steel, orequivalent, three-hole rake type. Sampleholes shall be 4 mmr in diameter or smallerand located at 16.7.50, and 83.3 percent of theequivalent stack diameter.

3.3 Sample Line. Stainless steel orTeflon'tubing to fransport the sample gas to theanalyzers. The sample line should be heated,if necessary, to prevent condensation In theline.

3.4 Calibration Valve Assembly. A three-wivy valve assembly to direct the zero andcalibration gases to the analyzers Isrecommended. Other methods, such as quick-connect lines, to route calibration gas to theanalyzers are applicable.

3.5 Particulate Filter. An In-stack or anout-of-stack glass fiber filter Is recommendedif exhaust gas particulate loading Issignificant. An out-of-stack filter should beheated to prevent any condensation.

3.6 Recorder. A strip-chart recorder.analog computer, or digital recorder forrecording measurement data. The minimumdata recording requirement Is onemeasurement value per minute. Notem Thismethod is often applied in highly explosiveareas. Caution and care should be exercisedin choice of equipment and installation.

4. Calibration and Other GasesGases used for calibrations, fuel, and

combustion air (if required) are contained incompressed gas cylinders of stainless steel oraluminum. Preparation of calibration gasesshall be done according to the procedure InProtocol No. 1, listed in Reference 9.2. Thepressure in the gas cylinders is limited by thecritical pressure of the subject organiccomponent. As a safety factor, the maximumpressure in the cylinder should be no morethan half the critical pressure. Additionally.the manufacturer of the cylinder shouldprovide a recommended shelf life for eachcalibration gas cylinder over which theconcentration does not change more than :E2percent from the certified value.

Calibration gas usually consists of propanein air or nitrogen and Is determined in termsof the span value. The span value Isestablished in the applicable regulation andis usually 1.5 to 2.5 times the applicableemission limit. If no span value Is provided,use a span valueequivalent to 1.5 to 2.5 timesthe highest expected concentration. Organiccompounds other than propane can be usedfollowing the above guidelines and makingthe appropriate corrections for carbonnumber.

4.1 Fuel A 40 percent H.-/60 percent Ho or40 percent H2/60 percent N: gas mixture isrecommended to avoid an oxygen synergismeffect that reportedly occurs when oxygenconcentration varies significantly from amean value.

4.2 Zero Gas. High purity air with lessthan 0.1 parts per million by volume oforganic material (propane or carbonequivalent).

4.3 Low-level Calibration Gas. An organiccalibration gas with a concentrationequivalent to 25 to 35 percent of theapplicable span value.

4.4 Mid-level Calibration Gas. An organiccalibration gas with a concentrationequivalent to 45 to 55 percent of theapplicable span value.

'Mention of trade names on specific productsdoes not constitute endorsement by theEnvironmental Protection Agency.

4.5 High-level Calibration Gas. Anorganic calibration gas with a concentrationequivalent to 60 to 90 percent of theapplicable span value.5. Measurement System PerformanceSpecifications

5.1 Zero DrifL Less than. ± I percent ofthe span value.

52 Calibration Drifl Less than.±4 1percent of the span value.6. Protest Preparations

6.1 Selection of Sampling Site. Thelocation of the sampling site is generallyspecified by the applicable regulation orpurpose of the test: i.e. exhaust stack, inletline, etc. The sample port shall not be locatedwithin 1.5 meters or 2 equivalent diameters(whichever is less] of the gas discharge to theatmosphere.

6.2 Location of Sample Probe. Install thesample probe so that theprobe is centrallylocated In the stack, pipe, or duct and issealed tightly at the stack port connection.

.3 Measurement System Preparation.Prior to the emission test, assemble themeasurement system following themanufacturers written instructions inpreparing the sample interface and theorganic analyzer. Make the system operable.

FIA equipment can be calibrated for almostany range of total organics concentrations.For high concentrations of organics [>L0percent by volume as propane) mod&l iftionsto most commonly available analyzers arenecessary. One accepted method ofequipment modification is to decrease thesize of the sample to the analyzer through theuse of a smaller diameter sample capillary.Direct and continuous measurement oforganic concentration is a necessaryconsideration when, determining anymodification design.

&.4 Calibration. Immediately prior to thetest series. introduce zero gas andhigh-levelcalibration gasat the calibrationvalveassembly. Adjust the analyzer output to theappropriate levels. if necessary. Thenintroduce low-level and mid-level calibrationgases successively to the measurementsystem. Record the analyzer responses for allfour gases and develop apermanent record ofthe calibration curve. This curve shallbeused in performing the post-test drift checksand in reducing all measurement data duringthe test series. No adjustments to themeasurement system shall be conducted afterthe calibration and before the drift check(Section 7.3]. If adjustments are necessarybefore the completion of the test series,perform the drift checks prior to the requiredadjustments and repeat the calibrationfollowing the adjustments. If multipleelectronic ranges are to be used, eachadditional range must be checked with a mid-level calibration gas to verify themultiplication factor.7. Emission Measurement Test Procedure

7.1 Organic Measurement. Begin samplingat the start of the test period, recording timenotations and any required processinformation as appropriate. In particular, noteon the recording chart periods of processInterruption or cyclic operation.

7.2 Drift Determination. Immediatelyfollowing the completion of the test period, or

8,3149I

83150 Federal Register / V61. 45, No. 244 / Wednesday, December 17, i980 / Proposed Rules

if adjustments are necessary for themeasurement system during the test,reintroduce the zero and mid-level calibrationgases, one at a time, to the measurementsystem-at the calibration valve assembly.'(Make no adjustments to the measurement.systeri until after the drift checks are made.)Record the analyzer response. If the driftvalues exceed the specified limits, invalidatethe test run preceding the check and repeatthe test run following corrections to themeasurement system. Alternatively,recalibrate the test -neasurement system as inSection 6.4 and report the results'using thecalibration data that yield the highestcorrected emission concentration.8. Organic Concentration Calculations

Determine the average organicconcentration in terms of ppmv as propane orother calibration gas. The average shall bedetermined by the integration of the outputrecording over the period specified in theapplicable regulation.

If results are required in terms of ppmv ascarbon, adjust measured concentrations usingEquation 25A-1.C = K C Eq. 25A-1Where:C, = Organic concentration as carbon, ppmv.C.,., = Organic concentration as meaqured,

ppmv.K = Carbon equivalent correction factor.

K = 2 for ethane.K = 3 for propane.K = 4 for butane.

9. References9.1 Measurement of Volatile Organic

Compounds-Guideline Series. U.S.Environmental Protection Agency. ResearchTriangle Park, N.C. Publication No. EPA-450/2-78-041. June 1978. p. 46-54.

9.2 Traceability Protocol for EstablishingTrue Concentrations of Gases Used forCalibration and Audits of Continuous SourceEmission Monitors (Protocol No. 1). U.S.

" Environmental Protection Agency,Environmental Monitoring and SupportLaboratory. ResearchTriangle Park, N.C.June 1978. 10 pgs.

9.3 Gasoline Vapor Emission LaboratoryEvaluation-Part 2. U.S. EnvironmentalProtection Agency, Office of Air QualityPlanning and Standards. Research TrianglePark, N.C. Report No. 75-GAS-6. August1975. 32 pgs.BILUNG CODE 6560-26-M

Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 / Proposed Rules

wU

_j,

WUj= Ca

83151

83152 Federal Register / Vol. 45, No. 244 / Wednesday, December 17, 1980 1 Proposed Rules

Method 25B-Determination of TotalGaseous Organic Concentration Using aNondispersivo Infrared Analyzer1. Applicability and Principle

1.1 Applicability. This method applies tothe measurement of Total gaseous organicconcentration of vapors consisting primarilyof nonmethane alkanes. (Other organicmaterials may be measured using the generalprocedure in this method, the appropriatecalibration gas, and an analyzer set to theappropriate absorption bnd.) Theconcentration is expressed in terms ofpropane (or other calibration gas) or in termsof organic carbon.

1.2 Principle. A gas sample is extractedfrom the source, through a heated sample lineand glass fiber filter to a nondispersiveinfrared analyzer (NDIR). Results arereported as equivalents of the calibration gasor as carbon equivalents.2. Definitions

The ter ms and definitions are the same asfor Method 25A.3. Apparatus

The apparatus are the same as for Method25A with the exception of the following:

3.1 Organic Concentration Analyzer. Anondispersive infrared analyzer designed tomeasure alkane organics and capable ofmeeting or exceeding'the specifications inthis method.4. Calibration Gases

The calibration gases are thesaame as arerequired for Method 25A, Section 4. No fuelgas is required for an NDIR.5. Measurement System PerformanceSpecifications

5.1 Zero Drift. Less than ± Zpercent ofthe span value.

5.2 Calibration Drift. Less than ± 2percent of the span value.6. Pretest Preparations

6.1 Selection of Sampling Site. Same as inMethod 25A, Section 6.1.

6.2 Location of Sample Probe. Same as inMethod 25A, Section 6.2.

6.3 Meaburement System Preparation.Prior to the emission test, assemble themeasurement system following themanufacturer's written instructions inpreparing the sample interface and theorganic analyzer. Make the system operable.

6.4 Calibration. Same as in Method 25A.Section 6.4.7. Emission Measurement Test Procedure

Proceed with the emission measurementimmediately upon satisfactory completion ofthe calibration.

7.1 "OrganicMeasurement. Same as inMethod 25A, Section 7.1.

7.2 Drift Determination. Same as inMethod 25A, Section 7.2.8. Organic Concentration Calculations

The calculations are the same as in Method25A, Section 8.9. References

The references are the same as in Method25A. Section 9.

Method.27-Determination of VaporTightness of Gasoline Delivery Tank UsingPressure-Vacuum Test

1. Applicability and Principle

1.1 . Applicability. This method isapplicable for the determination of vaportightness of a gasoline delivery tank wlich'isequipped with vapor collection equipment.

1.Z Principle. Pressure and vacuum areapplied alternately to the compartments of agasoline delivery tankand the change inpressure or vacuum is recorded after aspecified period of time.

2. Definitions and Nomenclature

2.1 Gasoline. Any petroleum distillate orpetroleum distilfite/alcohol blend having a -

Reid'Vapor pressure of 27.6 kilopascals orgreater which is used as a fuel for internalcombustion engines.

2.2 Delivery tank. Any pontainer,including associated pipes and fittings, that isattached lo or forms a part of any truck orrailcar used for the transport of gasoline.

2.3 CompartmenLA liqud-tight divisionof a delivery tank.

2.4 Delivery tank vapor collectionequipment. Any piping, hoses. and devices onthe delivery tank used to collect and routegasoline vapors either from the tank to a bulkterminal vapor control system or from a bulkplant or service station into the tank.

2.5 Time period of the pressure or vacuumtest (t]. The time period of the test, asspecified in the appropriate regulation, duringwhich the change in pressure 9f vacuum ismonitored, in minutes.

2.6 Initial pressure (PI). The pressureapplied to the delivery tank at the beginningof the static pressure test, as specified in theappropriate regulation, in nm HO.

2.7 Initial vacuum (Vj. The vacuumapplied to the delivery tank at the beginningof the static vacuum test, as specified in theappropriate regulation, in nm H2 0.

2.8 Allowable pressure change'(Ap). Theallowable amount of decrease in pressureduring the static pressure test, -within the timeperiod t, as specified in the appropriate .regulation, in mm. H.O.- 2.9 Allowable vacuum change (Av). Theallowable amount of increase in vacuumduring the static vacuum test, within the timeperiod t, as specified in the appropriateregulation, in mm H20.

3. Apparatus

3.1 Pressure source. Pump or compressedgas cylinder ofaIr or inert gds sufficient topressurize the delivery tank to 500 mm H20above atmospheric pressure.

3.2 Regulator. Low pressure regulator forcontrolling pressurization of the deliverytank., 3.3 Vacuum source. Vacuum pumpcapable of evacuating the delivery tank to250 mm H2 0 below atmospheric pressure.

3.4 Pressure-vacuum supply hose.3.5 Manometer. Liquid manometer, or

equivalent instrument, capable of measuringup to 500 mm H2 0 gauge pressure with ±2.5mm H20 precision.

3.6 Pressure-vacuum relief valves. Thetest apparatus shall be equipped with an In-line pressure-vacuum relief valve set toactivate at 675 mm H20 above atmospheric .

pressure or 250 dim HO below atmospherlapressure, with a capacity equal to thepressurizing or evacuating pumps.

3.7 Test cap for vapor recovery hose. Thiscap shall have a tap for manometerconnection and a fitting with shut-off valvefor connection to the pressure-vacuum supplyhose.

3.8 Caps for liquiddelivery hoses.4. Pretest Preparations

4.1 Emptying of tank. The delivery tankshall be emptied of all liquid.

4.2 Purging of vapor. The delivery tankshall be purged of all volatile vapors by anysafe, acceptable method. One method 19 tocarry a load of non-volatile liquid fuel, suchas diesel or heating oil, Immediately prior tothe test, thus flushing out all the volatilegasoline vapors. A second method is toremove the volatile vapors by blowingambient air Into each tank campartment for

'at least 20 minutes. This second method Isusually not as effective and-often causesstabilization problems, requiring a muchlonger time for stabilization during thetesting.

4.3 Location of test site. The delivery tankshall be tested where it will be protectedfrom direct sunlight.5. Test Procedurb

5.1 Preparations.5.1.1 Open and close each dome cover.5.1.2 Connect static electrical ground

connections io tank. Attach the liquiddelivery and vapor return hoses, remove thtiliquid delivery elbows, and plug the liquiddelivery fittings.

5.1.3 Attach the test cap to the end of thevapor recovery hose.

5.1.4 Connect the pressure-vacuum supplyhose and the pressure-vacuum relief valve tothe shut-off valve. Attach a manometer to thepressure tap.

5.1.5 Connect compartments of the tankinternally to each other If possible. If notpossible, each compartment must be testedseparately, as if it were an individualdelivery tank.

5.2 Pressure Test.5.21 Connect the pressure source to thu

pressure-vacuum supply hose.5.2.2 Open the shut-off valve In the vapor

recovery hose cap. Applying air pressureslowly, pressurize the tank to P1, the Initialpressure specified in the regulation.

5.23 Close the shut-off valve and allowthe pressure in the tank to stabilize, adjustingthe pressure if necessary to maintainpressure of Pi. When the pressure stabilizes,record the time and initial pressure.

5.24 At the end of t minutes, record thetime and final pressure.

5.25 Repeat steps 5.2.2 through 5,2.4 untilthe change in pressure for two consecutiveruns agrees with ±10mm 1.0. Calculate thuarithmetic average of the two results.

5.2.6 Compare the average measuredchange in pressure to the allowable pressurechange, Ap, as specified in the regulation. Ifthe delivery tank does not satisfy the vaportightness criterion specified In the regulation,repair the sources of leakage, and repeat thepressure test until the criterion is met.

5.2.7 Disconnect the pressure source fromthe pressure-vacuum supply hose, and slowly

Federal Register / Vol. 45, No. 244 / Wednesday; December 17, 1980 / Proposed Rules 83153

open the shut-off valve to bring the tank toatmospheric pressure.

5.3 Vacuum Test.5.3.1 Connect the vacuum source to the

pressure-vacuum supply hose.5.3.2 Open the shut-off valve in the vapor

recovery hose cap. Slowly evacuate the tankto Vi, the initial vacuum specified in theregulation..

5.3.3 Close the shut-off valve and allowthe pressure in the tank to stabilize, adjustingthe pressure if necessary to maintain avacuum of Vi. When the pressure stabilizes,record the time and initial vacuum.

5.3.4 At the end of t minutes, record thetime and final vacuum.

5.3.5 Repeat steps 5.3.2 through 5.3.4 untilthe change in vacuum for two consecutiveruns agrees within ± 10 mm 1120. Calculatethe arithmetic average of the two results.

5.3.6 Compare the average measuredchange in vacuum to the allowable vacuumchange, Av, as specified in the regulation. Ifthe delivery tank does not satisfy the vaportightness criterion specified in the regulation,repair the sources of leakage, and repeat thevacuum test until the criterion is met

,5.3.7 Disconnect the vacuum source fromthe pressure-vacuum supply hose, and slowlyopen the shut-off valve to bring the tank toatmospheric pressure.

5.4 Post-test clean-up. Disconnect all testequipment and return the delivery tank to itspretest condition.6.Altemative Procedures

Techniques other than specified above maybe used for purging and pressurizing adelivery tank, if prior approval is obtainedfrom the Administrator. Such approval willbe based upon demonstrated equivalencywith the above method.[MRDo. &-3= Filed 12-10- &45 am)

BI.UNG CODE 6560-26-MA

ThursdayAugust 18, 1983

Part II

EnvironmentalProtection AgencyStandards of Performance for NewStationary Sources; Bulk GasolineTerminals

mm

ii

m m mm m mI m, mmummm. mm mmmm wmmt m

u

I

" 'w

ii __-- --i

rmi mm mm mm /m 1r m Iromp n m,mm m mm m iBmm m

im

i

37578 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

ENVIRONMENTAL PROTECTIONAGENCY

40 CFR Part 60

IAD-FRL-224-6]

Standards of Performance for NewStationary Sources; Bulk GasolineTerminals

AGENCY: Environmental ProtectionAgency (EPA).ACTION: Final rule.

SUMMARY: Standards of performance forbulk gasoline terminals were proposedin the Federal Register on December 17,1980 (45 FR 83128). This actionpromulgates standards of performancefor bulk gasoline terminals. Theseatandards implement Section 111 of theClear Air Act and are based on theAdministrator's determination thatpetroleum transportation and marketingcause, or contribute significantly to, airpollution which may reasonably beanticipated to endanger public health or'welfare. The intended effect of thesestandards is to require all new,modified, and reconstructed facilities atbulk gasoline terminals to controlemissions to the level achievablethrough use of the best demonstratedsystem of continuous emissionreduction, considering costs, nonairquality health, and environmental andenergy impacts.EFFECTIVE DATE: August 18, 1983.

Under Section 307(b)(1) of the CleanAir Act, judicial review of this newsource performance standard isavailable only by the filing of a petitionfor review in the U.S. Court of Appealsfor the District of Columbia Circuitwithin 60 days of today's publication ofthis rule. Under Section 307(b)(2) of theClean Air Act, the requirements that arethe subject of today's notice may not bechallenged later in civil or criminalproceedings brought by EPA to enforcethese requirements.ADDRESSES:

Background Information Document.The background information document(BID, Volume II) for the promulgatedstandards may be obtained from theU.S. EPA Library (MD-35), ResearchTriangle Park, North Carolina 27711,telephone number (919) 541-2777. Pleaserefer to "Bulk Gasoline Terminals-Background Information forPromulgated Standard," EPA-450/3-80-038b. BID, Volume II, contains (1) asummary of all the public commentsmade on the proposed standards and theAdministrator's response to thecomments, (2) a summary of the changesmade to the standards since proposal

and (3) the final environmental impactstatement which summarizes theimpacts of the standards.

Dbcket Docket No. A-79-52,containing information considered byEPA in developing the promulgatedstandards, in available for publicinspection and copying between 8:00a.m. and 4:00 p.m., Monday throughFriday, at EPA's Central Docket Section,West Tower Lobby, Gallery 1,Waterside Mall, 401 M Street, SW.,Washington, D.C. 20460. A reasonablefee may be charged for copying.FOR FURTHER INFORMATION CONTACT:.For information concerning thebackg-oand information supporting thepromulgated standards contact Mr.James F. Durham, Chemicals, andPetroleum Branch, Emission Standardsand Engineering Division (MD-13), U.S.Environmental Protection Agency,Research Triangle Park, North Carolina27711, telephone number (919) 541-5671.For futher information concerning thepromulgated standards contact Mr.Gilbert H. Wood, StandardsDevelopment Branch, EmissionsStandards and Engineering Division(MD-13), U.S. Environmental ProtectionAgency, Research Triangle Park. NorthCarolina 27711, Telephone number (919)541-5578.SUPPLEMENTARY INFORMATION:

Summary of Promulgated Standards

Standards of performance for newsources established under Section 111 ofthe Clean Air Act reflect:

. * application of the best technologicalsystem of continuous emission reductionwhich (taking into consideration the cost ofachieving such emission reduction, and anynonair quality health and environmentalimpact and energy requirements) theAdministrator determines has beenadequately demonstrated [Section 111(a)(1J.

For convenience, this will be referredto as "best demonstrated technology" or"BDT,"

The promulgated standards ofperformance limit volatile organiccompound (VOC) emissions from eachaffected facility on which construction.modification, or reconstructioncommenced after December 17, 1980(after August 18, 1983, forreconstructions necessitated by State orlocal regulations). The affected facilityis the total of all the loading racks at abulk gasoline terminal which delivereither gasoline into ahy delivery tanktruck or some other liquid product intotrucks which have loaded gasoline onthe immediately previous load.

The promulgated standards requirethe installation of vapor collectionequipment at the terminal to collect total

organic compounds vapors displacedfrom gasoline tank trucks during productloading. The standards limit emissionsfrom the collection system to 35milligrams of total organic compoundsper liter of gasoline loaded, unless thefacility has an existing vapor processingsystem (construction or refurbishmentcommenced before December 17, 1980).In this latter case, the standards limitemissions from the vapor collectionsystem to 80 mg/liter.

The Agency has concluded that it isquite costly in light of the resultingemission reduction for an owner whoseexisting facility becomes subject toNSPS (e.g., through modification orreconstruction) to meet 35 mg/liter whenthe facility already has a systemcapable of meeting 80 mg/liter.

To control tank truck leakageemissions during loading, thepromulgated standards require thatloadings be made only into gasolinetank trucks tested for vapor tightness.The terminal owner or operator isrequired to obtain the identificationnumber and test documentation for eachgasoline tank truck loading at thefacility. In accordance with Section111(h)(3) of the Clean Air Act, theAdministrator may approve alternativeprocedures that assure that loading willbe limited to vapor-tight trucks.

These standards are based on the useof carbon adsorption and thermaloxidation type vapor processors for the35 mg/liter limit, which represent thebest demonstrated technology. Test datashow the ability of these systems ofcontinuous emission reduction toachieve the 35 mg/liter emission limit ofthe standards of performance. Althoughonly some of the refrigeration systemstested met 35 mg/liter (all the systemstested were designed to meet the Stateimplementation plan (SIP) limit of 80mg/liter), test data and engineeringcalculations also support the ability ofrefrigeration systems to achieve the 35mg/liter emission limit of the standards.In addition, the major manufacturer hasstated that all currently manufacturedrefrigeration systems can be specified tooperate at 35 mg/liter. In selecting thesestandards, the Agency considered costs,nonair quality health and environmentalimpacts, and energy requirements.

The proposed section on continuousmonitoring of operations, § 60.504, hasbeen reserved pending development ofmonitor performance specifications.Monthly system leak inspections arerequired under § 60.502(j), butsubmission of leak inspection records isnot required in the final regulation.However, under § 60.505(c), theserecords are required to be kept at the

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

terminal for at least 2 years. Therequirement for quarterly reports ofexcess emissions under § 60.7(c) of theGeneral Provisions is deleted under§ 60.505(e).

Summary of Major Changes SinceProposal

Several changes of varyingimportance have been made to thestandards since proposal. Most of thechanges were made in response tocomments, but some of them were madefor the sake of clarity or consistency.One of the most significant of thechanges dealt with proposed § 60.502(d),which required loadings of gasoline tanktrucks to be restricted to vapor-tighttanks only, as evidenced by an annualvapor tightness test. Most of thecomments on this requirementconcerned the terminal operator'sapparent liability for the condition oftank trucks owned by other parties.Several commenters felt that terminalswould have to provide extra personnelat the loading racks to enforce thisrestriction. Section 60.502(d) [now§ 60.502(e)] was expanded to delineateclearly the terminal owner or operator'sresponsibilities and to clarify that on-the-spot monitoring of product loadingswould not be necessary. A terminaloperator need only compare a tankidentification number against the file ofvapor tightness documentation within 2weeks after a loading of that tank tookplace. If a terminal owner or operatorchecked his files and found that anonvapor-tight truck was loadedwithout vapor tightness documentation,he would then be required to take stepsassuring that no further loading into thattank truck took place until the propervapor tightness documentation wasreceived by the terminal. Thus, the finalstandard clarifies that a terminal owneror operator can comply with this part ofthe standard by cross-checking files anddoes not have to monitor loadings.

One paragraph about facilities withexisting vapor processing equipmentwas added to § 60.502. The Agency hasconcluded that it is quite costly in lightof the resulting emission reduction foran owner whose existing focilitybecomes subject to NSPS (e.g., throughmodification or reconstruction) to meet35 mg/liter when the facility already hasa system capable of meeting 80 mg/liter,but not 35 mg/liter. For this reason, EPAhas added § 60.502(c), which permitsaffected facilities with such vaporcontrol equipment to meet 80 mg/liter ifconstruction or substantial rebuilding(i.e., "refurbishment") of that equipmentcommenced before the proposal date,December 17, 1980. This is based on theAdministrator's judgment that BDT for

these facilities is no further control,while BDT for facilities with vaporprocessing systems on whichconstruction or refurbishmentcommenced after proposal is thereplacement or add-on of technologythat would enable the facility to achieve35 mg/liter.

Several commenters objected to therequirement for excess emissionsreports and to using an averagemonitored value as the basis for anexcess emissions determination. Section60.504, Monitoring of Operations, hasbeen reserved pending the developmentand promulgation of performancespecifications for continuous monitoringdevices. Therefore, specific commentsconcerning the proposed continuousmonitoring requirements cannot beaddressed at this time. The Agency iscurrently investigating several types ofsimple, low-cost monitors for varioustypes of vapor processors. Afterspecifications have been selected, theywill be proposed in a separate action inthe Federal Register for public comment.

A new § 60.500(c) has been added tochange the applicability date from thedate of proposal to the date ofpromulgation for existing facilitiescommencing component replacementprior to the promulgation date for thepurpose of complying with State or localregulations. Such facilities are notsubject to the standards by means of thereconstruction provisions. New § 60.506was added in response to commenters'concerns about the burden ofaccumulating records of componentreplacements at an existing source overthe lifetime of the source for the purposeof determining reconstruction. Section60.506(b) limits the time period fordetermination of reconstruction to 2years and § 60.506(a) excludesfrequently replaced components forconsideration in applying thereconstruction provisions to bulkgasoline terminals.

In response to industry comments, asize cutoff by gasoline throughput wasadded to the definition of "bulk gasolineterminal" (only facilities handling morethan 76,700 liters, or 20,000 gallons, perday are covered), to clarify that bulkplants served by ship or barge are notcovered by these standards. Also, theword "wholesale" has been removedbecause the throughput cutoff shouldexclude retail outlets (service stations)from possible applicability.

The terminology used in the emissionlimits in the standard has changed sinceproposal. The emission limits are nowexpressed in terms of total organiccompounds rather than VOC (VOC isthe proportion of the organic compounds

that is regarded as photochemicallyreactive). This change does not changethe effect on stringency of the standard,but it does make the standard betterreflect the intent behind the standardand the data base and test proceduresused in establishing the standard.

The standard is intended to reduceemissions of VOC through theapplication of best demonstratedtechnology (BDT) (considering costs andother impacts), and the emission limitsin the standard are designed to reflectthe performance of BDT. The bestdemonstrated technologies applicable tobulk terminals do not selectively controlVOC, but rather they control all organiccompounds. Furthermore, the emissionlimits in the standard are based on testdata and test procedures that measuretotal organic compounds, and the testmethods used to determine compliancewith the standard measure total organiccompounds. Therefore, to reflectaccurately the performance of thetechnologies selected as BDT and to beconsistent with the data base and testmethods upon which the emission limitsare based, the emission limits in theproposed standard should have beenexpressed in terms of total organiccompounds. To reflect the applicabletechnology and test methods, theemission limits in the promulgatedstandard are expressed in those terms.EPA is relying on control of total organiccompounds as the best demonstratedsurrogate for controlling volatile organiccompounds, which react to form ozonein the atmosphere.

However, the test procedures in theproposed standard gave the owner oroperator the option to subtract methaneand ethane in determining compliancewith the standard. Because the testprocedures were proposed in this wayand because the relative quantity ofthese compounds is expected to besmall, the promulgated standard retainsthis option in the test procedures. Theowner or operator may invoke thisoption only by using a method approvedby the Administrator.

Summary of Environmental, Energy,Economic Impacts

The promulgated standards willreduce projected 1986 VOC emissionsfrom affected bulk terminals from about8,300 megagrams per year (Mg/yr) toabout 2,600 Mg/yr, or 68 percent.

The promulgated standards are basedon the use of carbon adsorption (CA)and thermal oxidation (TO) type vaporprocessors for the 35 mg/liter emissionlimit. TO systems emit a small quantityof carbon monoxide (CO) and oxides ofnitrogen (NO1), but since few oxidation

37579

37580 . Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

systems are expected to be installed,total emissions of CO and NO. will benegligible.

Neither of these control systems useswater as a direct control medium, andso the water pollution impact will beminimal. Refrigeration (REF) systems,which may also be used to meet thestanddrds, discharge a small amount ofwater which condenses in the systemdue to the humidity of the air. Organicsare separated from the condensed waterin an oil-water separator on therefrigeration unit. The excess water issubsequently handled by the bulkterminal's existing drainage system.

There will be no solid effluent fromany of the control systems. CA systemsmay produce a small quantity of solidwaste if the activated carbon must bereplaced due to a loss in workingcapacity of the carbon beds. The worst-case nationwide waste production isestimated at about 50,000 kilograms (kg)per year, which represents a small solidwaste impact.

All of the vapor processors consideredin setting the standards consumeelectricity in the course of theiroperation, to power fans, dampers,pumps, compressors, valves, timers, andother miscellaneous components.However, all of the processors, exceptthe thermal oxidizer, recover energy inthe form of liquid gasoline. Therefore,while the power costs to operate controlequipment to comply with thepromulgated standards average about 25percent higher than the power costs tocomply with a typical SIP at a 950,000liter per day terminal, the extra productrecovery realized under these standardsmeans that this terminal will experiencea net energy savings which is equivalentto about 15,000 liters of gasoline peryear greater than the SIP. The total netenergy recovery experienced by the bulkterminal industry in the fifth year of thestandards will be about 7.0 million litersof gasoline equivalent.

Compliance with these standards willresult in net annualized costs in the bulkgasoline terminal industry of about $1.6million by 1986. Cumulative capital costsof complying with the promulgatedstandards will amount to about $10.8million by 1986. Net annualized andcumulative capital costs to the for-hiretank truck industry will total about $0.9million and $1.4 million, respectively, bythe fifth year of the standards. The totalannualized cost for this standard wouldthen be $2.5 million. This annualizedcost, coupled with the estimatedemmission reduction of 5,700 Mg/yr,results in a cost per unit emissionreduction of $440/Mg. The percentincrease in the price of gasolinenecessary to offset costs of compliance

with the promulgated standards willrange from zero for certain largerterminals up to about 0.48 percent forthe smallest terminals. The overallimpact on national gasoline prices willbe negligible. The environmental,energy, and economic impacts arediscussed in greater detail in the BID,Volume I. Also discussed are all of thecommenters' suggested changes in theimpact calculations and the rationale formaking some of these changes and notothers.

The nationwide impact numberspresented here include a composite ofimpacts for new, modified, andreconstructed facilities in locationswhere States require the level of controlrecommended in the control techniquesguideline document (CTG) and inlocations where States have no controlrequirements. If an average size bulkterminal (950,000 liters/day gasolinethroughput) subject to the standards dueto modification or reconstruction werelocated in an area with Staterequirements equivalent to the levelrecommended by the CTG and theterminal had an existing vaporprocessing system which met theseState requirements, no additionalcontrols would be required. For thesame size new terminal, the incrementalannualized cost for a terminal using CAor TO would be negligible because thesame basis control device could be usedto meet either set of requirements. If anew, modified, or reconstructed terminalof the same size were located in an areawith no State requirements, theuncontrolled emissions would bereduced by about 160 Mg/yr at anannualized cost of about $38,000, whichis less than $240/Mg of VOC reduced.

Public ParticipationPrior to proposal of the standards,

interested parties were advised bypublic notice in the Federal Register (45FR 30686, May 9, 1980), of a meeting ofthe National Air Pollution ControlTechniques Advisory Committee(NAPCTAC) to discuss the bulk gasolineterminal standards recommended forproposal. This meeting was held on June5, 1980. The meeting was open to thepublic and each attendee was given anopportunity to comment on therecommended standards. The standardswere proposed and published in theFederal Register on December 17, 1980,(45 FR 83126). The preamble to theproposed standards discussed theavailability of the backgroundinformation document, "Bulk GasolineTerminals--Background Information forProposed Standards," EPA-450/3-80-038a (BID, Volume I), which described indetail the regulatory alternatives

considered and the impacts of thosealternatives. Public comments weresolicited at the time of proposal and.when requested, copies of the BID,Volume I, were distributed to interestedparties. To provide interested personsthe opportunity for oral presentation ofdata, views, or arguments concerningthe proposed standards, a public hearingwas held in two sessions, on January 21and 28, 1981, at Research Triangle Park,North Carolina. The hearings were opento the public and each attendee wasgiven an opportunity to comment on theproposed standards. The publiccomment period was from December 17,1980, to March 20, 1981.

Forty-two comment letters werereceived and six interested partiestestified at the public hearingsconcerning issues relative to theproposed standards of performance for.bulk gasoline terminals. The commentshave been carefully considered and,where determined to be appropriate bythe Administrator, changes have beenmade in the proposed standards.

Major Comments on the ProposedStandards

Comments on the proposed standardswere received from bulk gasolineterminal owners and operators, Federalagencies, State and local air pollutioncontrol agencies, trade associations, andair pollution control equipmentsuppliers. A detailed discussion of thesecomments and Agency responses can befound in the background informationdocument for the promulgated standards(BID, Volume II), which is referred to inthe ADDRESSES section of thispreamble. The summary of commentsand responses in the BID, Volume I1serves as the basis for the revisionswhich have been made to the standardsbetween proposal and promulgation.The major comments and responses aresummarized in this preamble.

Need for Standard

Several commenters recommendedthat the proposed standards be canceledand that no additional regulation beadopted. Instead, the Stateimplementation plans (SIP's) should berelied upon to control VOC emissionsfrom bulk gasoline terminals. Onereason given was that gasoline demandis projected to stabilize or decline in thefuture, so that emissions from new,modified, or reconstructed sourcesshould not present any increasingenvironmental hazard.

Other commenters felt that theadditional emission reduction achievedunder Alternative IV (35 mg/liter fromprocessor plus vapor-tight tank trucks)

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

as opposed to Alternative 11 (80 mg/literfrom processor plus vapor-tight tanktrucks) would be insignificant. Thecommenters stated that the control limitof 80 mg/liter required by many SIP'shas already reduced VOC emissions by90 percent; the proposed 35 mg/literlimit would reduce nationwide bulkterminal VOC emissions by the fifthyear by only an extremely smallpercentage. Due to these smallreductions, these commenters felt thatstandards had been proposed simplybecause they are "technically feasible."Thus, the commenters felt EPA had notdemonstrated, as required by Section111, that new terminals will present asignificant air pollution problem.

The Agency proposed these standardsof performance under the authority ofSection 111 of the Clean Air Act (42U.S.C. 7411) as amended. Section111(b)(1) requires the Administrator toestablish standards of performance forcategories of new, modified, orreconstructed stationary sources whichin the Administrator's judgment cause orcontribute significantly to air pollutionwhich may reasonably be anticipated toendanger public health or welfare.

The Ageney's listing of PetroleumTransportation and Marketing 23rd onthe Priority List, as required underSection 111(f) (40 CFR 60.16, 44 FR 49222,August 21, 1979), reflects theAdministrator's determination that thissource category contributes significantlyto air pollution. Before arriving at thisdecision, the Administrator consideredthe projected rate of growth in thenumber of facilities in this industry, theemission rates at uncontrolled facilities,and the emissions allowed under typicalSIP's. EPA used the emissions forecastsin the BID, Volume I, and cited by thecommenters, in analyzing these factors,and the Administrator has found noreason to alter the conclusions based onthat analysis.

It is important to note that VOC isemitted by a wide variety of sourcecategories. The emissions contributionfrom many categories with VOCemissions that appear small incomparison with the total VOC emittedby all source categories is nonethelesssignificant to ozone formation. This isbecause failure to control these sourcesto the level achievable by the bestdemonstrated technology would serve toundermine the Congressionallymandated effort to prevent furtherdeterioration of air quality caused byadditional ozone formation. Emissionreductions from this source categoryalso appear small because the projectednumber of affected facilities is only asmall percentage (less than 5 percent) of

the total number of terminalsnationwide.

The Agency accounted for theprojected demand for gasoline in thecoming years in estimating the emissionreduction achievable through the NSPS.Despite a leveling off or reduction ingasoline demand, there will still be asignificant number of affected terminalswhich will result in significant emissionsreduction under these standards.Although the small number of newterminals (five in the next 5 years)reflects this leveling off in productdemand, the current industry trend istoward the consolidation of existingterminals rather than the construction ofnew terminals. AS a result, estimatesindicate that there will be as many as 50modified or reconstructed terminals inthe next 5 years.

Regulatory alternatives, reflectingdifferent levels of control technology,were evaluated for these 55 affectedfacilities, and it was determined that thecontrol technology was available, at areasonable cost, to control emissionsfrom new, modified, and reconstructedterminals. Relying only on the SIP's forthis category would mean that manysources, in areas not requiring controlsunder SIP's will remain uncontrolled. Itappeared reasonable, therefore, torequire additional controls, for theaffected facilities in both controlled anduncontrolled areas, that weretechnologically demonstrated to be bothreadily achievable and economicallyreasonable.

Standards of performance have otherbenefits in addition to achievingreductions in emissions beyond thoserequired by a typical SIP. They establisha degree of national uniformity, whichprecludes situations in which someStates may attract new industries as aresult of having relaxed air pollutionstandards relative to other States.Further, standards of performanceprovide documentation which reducesuncertainty in case-by-casedeterminations of best available controltechnology (BACT) for facilities locatedin attainment areas, and lowestachievable emission rates (LAER) forfacilities located in nonattainmentareas. This documentation includesidentification and comprehensiveanalysis of alternative emission controltechnologies, development of associatedcosts, an evaluation and verification ofapplicable emission test methods, andidentification of specific emission limitsachievable with alternate technologies.The costs are utilized in an economicanalysis that determines theaffordability of controls in an unbiased

study of the economic impact of controlson an industry.

The rulemaking process thatimplements a performance standardassures adequate technical review andpromotes participation ofrepresentatives of the industry beingconsidered for regulation,representatives from government, andthe public affected by that industry'semissions. The resultant regulationrepresents a balance in whichgovernment resources are applied in awell publicized national forum to reacha decision on a pollution emission levelthat allows for a dynamic economy anda healthful environment.

The promulgated standards reflectapplication of the best demonstratedtechnology for new, modified, andreconstructed sources in the bulkterminal subcategory. While technicalfeasibility is a fundamental criterion forstandard-seting, EPA consideredadditional factors, including cost, energyrequirements, and other impacts, beforearriving at the final standard. Basedupon these factors, the Agency selectedat proposal a control alternative whichreflects Alternative IV. As explained inthe preamble section on "Modificationand Reconstruction," the Agency hasrevised the standard in response tothese and other comments; thestandards are now based upon acombination of Alternatives II and IV.

Several commenters were concernedthat a number of their smaller loadingfacilities, typically considered as bulkplants, would be included under thedefinition of a terminal for purposes ofthis standard. These commenters felt athroughput cutoff should be added to thedefinition of a terminal.

To clarify the intended applicability ofthe NSPS, a definition of bulk terminaldependent upon a throughput cutoff hasbeen included in § 60.501. The purposeof this definition is to exclude thesmaller bulk plant. With this intention, abulk terminal has been defined to havea gasoline throughput greater than75,700 liters per day. The gasolinethroughput shall be the maximumcalculated design throughput as may belimited by compliance with anenforceable condition under Federal,State, or local law. Reference to anenforceable condition allows a source tolimit its maximum design throughput bylimiting its hours of operation, or bycontrolling any other operatingparameter. The only requirements arethat this limitation be a part of anenforceable document and that thesource maintain compliance with it. Thisdocument could be issued by anygovernment entity as long as it was

37581

37582 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

discoverable by both EPA and anycitizen as contemplated in Section 304 ofthe Clean Air Act. By obtaining suchdocumentation, which would reflect asource's maximum expected actualthroughput, ambiguities as to how onewould determine throughput areeliminated. For example, a bulk plantwhich receives gasoline by barge, with astatement (documented in anenforceable permit) that they will notexceed a throughput of 15,140 liters/day(4,000 gal/day), would not bemisconstrued as a bulk terminal.

Modification and Reconstruction

Several commenters were concernedthat conversions being made toterminals to satisfy SIP controlrequirements, such as top-to-bottomloading conversions and installation ofvapor control equipment, could subjectthese terminals to the more stringentrequirements of these standards throughthe reconstruction provisions of 40 CFR60.15. Also, the economic impact wouldbe significant for these terminals sincethey have already made commitmentstoward complying with SIP limitations.It was suggested by some of thecommenters that these conversionsshould be exempted from thereconstruction provisions (40 CFR 60.15].

The section entitled "Impacts ofRegulatory Alternatives" in thepreamble to the proposed standardsdiscussed the environmental, costs, andeconomic impacts on bulk terminalfacilities complying with therequirements of those standards.Included in the discussion were impactson new, modified, and reconstructedfacilities. The impacts estimated for thestandards did not include anyreconstructions resulting fromapplication of State or local air pollutionrequirements. However, as severalcommenters pointed out, a large numberof terminal facilities that the Agency didnot project as affected could indeedbecome subject to thq standards in theprocess of complying with suchrequirements. Thus, the preamblediscussion suggested that existingfacilities commencing componentreplacement in response to State orlocal requirements would not be subjectto 40 CFR 60.15.

The Agency believes that thissuggestion introduced some doubt as tothe otherwise straightforwardapplication of the reconstructionprovisions to existing facilitiesundergoing such changes. Consequently,owners and operators making plans toinstall control systems at these facilitiesmay have been misled to believe thatstricter NSPS requirements might notapply, and may therefore not have

considered the stricter NSPSrequirements when designing theirsystems.

For this reason, the Administrator hasdetermined that any facility that hascommenced substantial componentreplacement in response to State orlocal emission standards after theapplicability date (the proposal date-December 17, 1980) but prior to the dateof promulgation will not be subject tothese NSPS requirements by operationof the reconstruction provisions of 40CFR 60.15. Under § 60.500(c), anycomponent replacement programcommenced (as defined in Section 60.2)before today's date, and determined bythe Administrator to be necessitated byState or local bulk terminal regulations,will not subject a bulk terminal facilityto the NSPS by means of thereconstruction provisions.

It should be noted, however, that 40CFR 60.15 applies by straightforwardapplication to any existing facilityundergoing component replacement.Neither the language nor the purposes ofthat provision and the definition of "newsource" in Section 111 supportsexemptions based on the owner's intentin performing construction on thefacility.

Because this preamble corrects themisimpression that Section 60.15 doesnot apply to facilities undergoing SIPcomponent replacement, the Agency isapplying that provision to SIPcomponent replacement programscommenced after today's date. Ofcourse, owners or operators performingreconstruction for other purposes, ormodifications or new construction forany purpose, are still governed by theapplicability date of December 17, 1980,contained in § 60.500(b).

Commenters also felt that EPA hadgreatly underestimated the number ofexisting terminals which would beaffected by the modification andreconstruction provisions. At least 30SIP's will contain bulk terminal vaporrecovery requirements, and it wasbelieved that conversion workperformed at affected facilities wouldsubject those facilities to the provisionsof these standards:

Since most State of local regula tion-related construction programs at bulkterminals will have commenced by thepromulgation date, the change in theapplicability date, in effect, excludesthese terminals from the standards.Therefore, EPA's estimate at the time ofproposal of 55 new, modified, orreconstructed terminals in 5 years is stillconsidered a reasonable projection. Theestimate of 5 new facilities and 50modified or reconstructed facilities was

based primarily on information obtainedfrom oil companies through responses toSection 114 letter requests. Telephoneconversations with several controlagencies, oil companies, and terminalconstruction engineering firms providedsupplementary information.

Many of the commenters stated thatthe interpretation of "reconstruction" isan unwarranted extension of EPA's pastprocedure in defining this provision andan illegal extension of EPA's authorityunder Section 111. They felt that thereconstruction provisions were meant tobe applied to each capital constructionproject as it occurs, and not applied on acumulative basis over an unlimited timeperiod. The commenters felt that underthe present interpretation ofreconstniction every existing loadingrack, including those in attainmentareas, would, through ordinarymaintenance and replacement ofcomponents, become a new source longbefore the end Of its useful life. Theyconcluded that the use of cumulativecosts would be a tremendousadministrative burden on the industryand EPA.

The Agency promulgated thereconstruction provisions to ensure thatessentially new facilities due toreconstruction would be subject to "newsource" performance standards. Thereconstruction provisions werepromulgated in 1975 (40 FR 5846), andEPA has applied these provisionsconsistently since that time. Further, theAgency's authority to subjectreconstructed sources to new sourcestandards of performance has not beenquestioned in any court decision.

If one considers the 50 percent costfactor which triggers reconstructionstrictly on a project-by-project basis, awide variety of interpretations can ariseas to what a "project" entails. Forexample, a terminal with three toploading racks may convert one rack tobottom loading, and then 6 months laterconvert a second loading rack to bottomloading. If the two conversions wereinterpreted as separate projects, neitherone would likely exceed the 50 percentreplacement cost to triggerreconstruction. If, however, it was theterminal owner's original intent toconvert both loading racks, the twoconversions would be interpreted as oneproject and would probably constitute areconstruction. In many cases, it wouldnot be possible to determine the originalintent of the terminal owner or operator.In order to reduce the number ofsubjective determinations concerningintent in these cases, the reconstructionprovisions will be applied on a basiswhich considers the expenditures made

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

toward a facility over a fixed timeperiod.

To eliminate the ambiguity in thecurrent wording of § 60.15 and furtherthe intent underlying Section 111 (asdescribed above), the Agency isclarifying the meaning of "proposed"component replacements in § 60.15.Specifically, the Agency is interpreting"proposed" replacement componentsunder § 60.15 to include componentswhich are replaced pursuant to allcontinuous programs of componentreplacement which commence (but arenot necessarily completed) within theperiod of time determined by theAgency to be appropriate for theindividual NSP involved. The Agency isselecting a 2-year period as theappropriate period for purposes of thebulk gasoline terminal NSPS(§ 60.506(b)). Thus, the Agency willcount toward the 50 percentreconstruction threshold the "fixedcapital cost" of all depreciablecomponents (except those describedbelow) replaced pursuant to allcontinuous programs of reconstructionwhich commence within any 2-yearperiod following proposal of thesestandards. In the administrator'sjudgment, the 2-year period provides areasonable, objective method ofdetermining whether an owner of bulkgasoline terminal facilities is actually'proposing" extensive componentreplacement, within the Agency'soriginal intent in promulgating § 60.15.

The administrative effort to keep therequired records should not be a burdenon the industry. The recordkeepingrequired under this interpretation ofreconstruction is the same as therecordkeeping that would be requiredunder a strictly project-by-projectinterpretation. In either case, the dollaramount of the component replacementstaking place at the facility must bedetermined and recorded. Section 6.15defines the "fixed capital cost" ofreplacement components as the capitalneeded to provide all the "depreciable"components. By excludingnondepreciable components fromconsideration in calculating componentreplacement costs, this definitionexcludes many components that arereplaced frequently to keep the plant inproper working order. There may,however, be some depreciablecomponents that are replaced frequentlyfor similar purposes. In the Agency'sjudgment, maintaining records of therepair or replacement of these itemsmay constitute an unnecessary burden.Moreover, the Agency does not considerthe replacement of these items anelement of the turnover in the life of the

facility concerning Congress when itenacted Section 111. Therefore, inaccordance with 40 CFR 60.15(g), thesestandards (§ 60.506) will exempt certainfrequently replace components, whetherdepreciable or nondepreciable, fromconsideration in applying thereconstruction provisions to bulkgasoline terminal facilities. The costs ofthese components will not be consideredin calculating either the "fixed capitalcost of the new components" or the"fixed capital costs that would berequired to construct a comparableentirely new facility" under § 60.15. Inthe Agency's judgment, these items arepump seals, loading arm gaskets andswivels, coupler gaskets, ovei fillsensors, vapor hoses, and groundingcables.

One commenter felt that if theproposed standards further limitedallowable total organic compoundsemissions from 80 mg/liter to 35 mg/literof gasoline loaded, then over half of histerminals would experience "immediateoperational constraints," since they areequipped with vapor processing units ofthe compression-refrigeration-absorption (CRA) or lean oil absorption(LOA) type, which EPA data indicatecannot meet the proposed 35 mg/literlimit.

The existing facilities described bythe commenter would not be subject tothe standards unless modification orreconstruction were commenced afterthe proposal date of December 17, 1980.For those facilities with existing vaporprocessing systems which becomeaffected facilities under modification orreconstruction, the Administratorconcluded that it was not reasonable forthe owner or operator to replace orperform costly upgrading on existingvapor processing systems, in order toachieve the small incremental emissionreduction which reflects the change from80 mg/liter to 35 mg/liter. As anexample, emissions from a 950,000 liter/day terminal would decrease about 15Mg/year in the change from 80 to 35 mg/liter, at a net annualized cost of about$50,000 for replacement or add-oncontrols. In the Administrator'sjudgment, however, it is unreasonablycostly to require such a facility to installthe add-on technology that will achieve35 mg/liter only if the facility beganconstructing or substantially rebuilding(i.e., "refurbishing") the control systembefore receiving notice December 17,1980, that BDT for those facilities, werethey later to come under NSPS, wouldlikely be equipment capable of meeting35 mg/liter.

By contrast, EPA considers itreasonable to apply the 35 mg/liter limit

to a facility whose owner commencedconstruction or refurbishment of acontrol system not capable of meeting 35mg/liter despite having received thisnotice. It is reasonable to expect such anowner to avoid the high cost of goingfrom 80 mg/liter to 35 mg/liter simply byconstructing or refurbishing the facility'scontrol system with technology thatwould meet EPA's proposed 35 mg/literlimit and make later retrofitunnecessary. This is reasonable torequire even of facilities with existingcontrol systems constructed orrefurbished after December 17, 1980, forthe purpose of meeting an 80 mg/literState limit.

For these reasons, EPA has added§ 60.502(c), which permits affectedfacilities with such vapor controlequipment to meet 80 mg/liter ifconstruction or substantial rebuilding(i.e., "refurbishment") of that equipmentcommenced before the proposal date,December 17, 1980. This is based on theAdministrator's judgment that BDT forthese facilities is no further control,while BDT for facilities with vaporprocessing systems on whichconstruction or refurbishmentcommenced after proposal is thetechnology that would enable thefacility to achieve 35 mg/liter.

Definitions for "existing vaporprocessing system" and "refurbishment"were added to the regulation to indicatethat if in any 2-year period following thedate the facility becomes an affectedfacility the fixed capital cost ofimprovements or changes to an existingvapor processing system exceeds 50percent of the cost of a comparableentirely new vapor processing system,the altered vapor processing systemmust then meet the 35 mg/liter limit.Consequently, refurbishment appliesonly to those systems which becomeextensively rebuilt over this period.

Several commenters felt that theinterpretation of "modification" isoverly broad because it may includealtered facilities from which the overallemissions have not increased. Aclarification was sought so thatreplacement of needed components thatimprove loading efficiencies would notbe considered modifications unless theyresulted in an increase in the averagedaily emissions. For example, thereplacement of worn-out pumps withnew higher capacity pumps would allowfaster loading, increasing emissions on akg/hour basis during peak loadingperiods, but not on a mg/liter basis,which is the measurement of thestandard. In fact, the number of tanktrucks loaded during a day would not

37583

37584 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

necessarily increase due to a fasterloading rate.

Section 60.14(e)(2) was purposelyincluded in the General Provisions toexclude from consideration under themodification provisions increases inemissions due to relatively smallchanges. If a change increasesproduction capacity and yet does notresult in a "capital expenditure" asdefined in the definitions in the GeneralProvisions, the change would not beconsidered a modification.

Economic Impact

Some of the commenters stated thatmany of the costs of compliance toindustry presented in BID, Volume I,were seriously underestimated. Tworeasons provided were that controlsystems necessary to achieve theproposed standard of performancewould cost more than systems capableof meeting only the less stringent SIPemission limit, and the actual numberfor affected facilities would be greaterthan the estimate due to conversionsresulting from SIP requirements.

Many control systems being installedunder SIP programs are capable ofcontrolling emissions below 35 mg/liter,the limit of the promulgated standards ofperformance. Test data show that, intheir normal operating mode, carbonadsorption (CA) and thermal oxidation(TO) units can consistently operate wellbelow the 35 mg/liter limit. Therefore,for CA and TO units there are noadditional costs involved in meeting 35mg/liter versus the units currently beinginstalled to meet 80 mg/liter.

Test results on current refrigeration(REF) units show that only some ofthese units meet the 35 mg/liter limit.However, these systems were installedto comply with a limit at or near the 80mg/liter limit contained in most SIP's.The major manufacturer of thesesystems has indicated that adjustmentsto operating parameters can be madewhich will increase the controlefficiency of individual systems (docketitem IV-E-32). Such adjustments wouldbe likely to increase electrical costs.Cost increases of up to 50 percent werereported by the manufacturer (docketitem IV-F-3). The assumption that costswould not increase in the case of CAand TO units in order to meet 35 mg/liter is still considered valid. However,since data show that state-of-the-artREF technology can meet the standard,at somewhat increased capital andoperating cost levels from the averagecurrent system, and since a largesegment of industry is presently usingthis form of control (approximately 25percent of existing units arerefrigeration units), the potential cost

impact to industry, if current usepatterns are maintained, was examined.

As discussed under the preamblesection "Modification andReconstruction," the vast majority ofconversions necessary to comply withState or local regulations will havecommenced before the revisedapplicability date, and, therefore, not beregulated under these standards. Onlythose few State or local regulation-related conversions which commenceafter the promulgation date will beaffected. Thus, the estimate of 55facilities affected in 5 years is stillbelieved to represent a reasonableapproximation, based on Section 114letter responses from industry. Theupdated industry costs were used torecalculate the nationwide cost impact,with the costs of purchasing andoperating continuous monitors added tothese estimates. By 1986, the terminaland independent tank truck industrieswill spend about $12.2 million in capitalinvestment, and the net annualized costin the fifth year will be $2.5 million. Thecapital and annualized cost estimateshave decreased since the originalevaluation mainly because of re-analysis of loading rack top-to-bottomloading conversion costs and changes inthe requirements for existing vaporprocessing systems. In the previousanalysis, presented in the BID, Volume I,the costs for the top-to-bottom loadingconversions were attributed to thestandards for all affected top loadingterminals in the nationwide costdetermination. However, in the revisedevaluation, the cost of top-to-bottomloading conversions (not as a result ofvapor control requirements) whichwould trigger reconstruction were notincluded in the costs to comply with thepromulgated standards. These costswould be incurred by the terminalowner regardless of the standards sincethe conversions were performedvoluntarily.

One commenter felt that even thesmall cost per gallon of productnecessary to comply with the standardswould discourage an owner or operatorfrom investing in conversion workwhich might make a terminal subject tothe standards, and that this could maketerminal closures more prevalent. Inresponse to this and similar comments,the economic analysis which supportedthe proposal was reviewed and manycost estimates were updated. The resultsof both the original and revisedeconomic analyses showed that for thetwo smallest model plants the standardscould, in the worst case, have asignificant negative impact onprofitability in the unlikely absence ofcomplete control cost pass-through.

In the original analysis on existingfacilities both the 380,000 liter/day and950,000 liter/day model plants (modelplants I and 2) would encounter returns-on-investment (ROI's) of less than 11percent, taken to be the minimumacceptable level. The revised analysisindicates that only a 380,000 liter/daytop-loaded facility (projected to be only2 or 3 affected facilities per year) wouldexperience a significant decrease inprofitability, with a post-control ROIrange of 7.7 to 8.0 percent. A 950,000liter/day terminal would still maintain amarginal profitability level with a post-control ROI range of 10.6 to 11.0 percent.However, the preceding impacts areworst-case scenarios and very unlikelyto occur. Since the price increasenecessary to offset the control costs isless than 0.5 percent, the most likelyscenario will involve an impact withmost of the control costs passed throughand very little cost absorption. Underthis scenario no existing terminals areexpected to close. Industry profiles doforecast a trend away from new smallbulk terminals to larger terminals;however, this is a result of previoustechnological advances and economiesof scale and is not a result expected tobe accelerated by the implementation ofthese standards.

Some commenters questioned the BID,Volume I, cost estimates associated withpurchasing, installing, operating, andmaintaining vapor control systems. Inparticular, most CA system costs andsome REF system costs were pointed outas being underestimated.

Most carbon adsorption units arecurrently being produced by twomanufacturers. The purchase costs usedin the original cost analysis werereceived from one major manufacturerat the time the analysis was performed.After proposal, estimated costs wereupdated through contacts with bothmanufacturers. The average cost ofinstalling a vapor processor wasestimated as 85 percent of the initialpurchase price of the unit, based on 14actual installations. Values used tocompute the average installation costranged from 37 to 147 percent. Since notrend in this percentage as a function ofpurchase cost or unit type was noted, asingle value representing the averagewas selected. Consequently, some unitinstallation costs will be higher andsome lower than those presented in theanalysis. Installation costs submitted byone commenter averaged about 115percent of the purchase price of theprocessor, which is consistent with therange of values considered in derivingEPA's 85 percent figure. Anothercommenter submitted data showing that

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

the typical installation cost for a REFunit at his terminals was $90,000, or 55percent of the $165,000 purchase price.Again, this percentage falls within therange of values considered previouslyby the Agency.

Operating costs for all controltechnologies considered in developingthe standards were calculated usingelectrical consumption data supplied bythe system manufacturers. The REF unitpurchase cost and electricalconsumption figures used to developimpacts of the proposed standardsapplied to systems used to achieve theSIP limit of 80 mg/liter. The data havesubsequently been reassessed usingmore current costs. The manufacturer ofessentially all of the current REF unitswas contacted to obtain presentpurchase and operating figures whichwould be reflected for a system to meetthe emission limit of 35 mg/liter. Unitmodels were selected for application tothe four model plants, based on theparameter suggested by themanufacturer, peak hourly productloading. Models were selected withconsiderable excess capacity, so thatcost estimates would be conservative.The power costs for current CA systemswere calculated in the same way asthose for REF systems, using updatedmanufacturers' information. The limitedavailable field data on the operatingcosts of installed units generallycorrelate well with the calculatedfigures.

Emission Control Technology

Several commenters remarked thatthe technology to achieve the 35 mg/literemission limit has not beendemonstrated, because only a few short-term tests have been performed. Thesecommenters stressed the necessity fordata on continuous performance, and onthe ability of the considered systems toachieve the emission limit over the longterm.

Since the beginning of the standardsdevelopment, the Agency has sought themost recent results of tests performed byoil companies and State agencies, inorder to collect the best possible database. Since all of the tested systemswere installed in response to SIPlimitations at or near the 80 mg/literlimit, oil company and systemmanufacturer technical representativeswere consulted to determine theassumed design conditions for theinstalled systems and the collectionpotential of the various controltechnologies. Emission test results onseveral CA units tested between 1979and 1981, representing over 30 days oftesting, were received after proposalfrom four State agencies and one control

system manufacturer. Outlet totalorganic compounds mass emissionsmeasured in these tests ranged from 0.34to 17.9 mg/liter, with 28 of the daily testvalues below 10 mg/liter. Three REFunits owned by a single oil company intwo States were tested in 1980 and 1981.Daily average emissions in these testswere 21.9, 22.6, and 41.8 mg/liter. Theseresults support the observation thatcurrent REF units perform at variouslevels with respect to the 35 mg/literlimit. Since total organic compoundsmass emissions are related to thecondenser temperature maintained inthese units, setting the thermostaticcontrols at different levels can producea range of emission levels from the samecontrol equipment. The currentgeneration of REF units can be adjustedto maintain the low temperatures(approximately -84°Ci or -120°F)required to achieve 35 mg/literconsistently. Recent tests of TO systemsverify the ability of oxidation units tolimit emissions to levels considerablybelow 35 mg/liter.

Even though the tests did not followEPA procedures exactly, the recent testdata collected since proposal of thesestandards demonstrate the ability of thebest systems to achieve the requiredlevel of 35 mg/liter. The continuingability of these systems to achieve thislimit depends on their proper operationand maintenance. The costs of operatingand maintaining CA, TO, and REF typevapor processors were considered inassessing the economic impact of thepromulgated standards. As discussedearlier, the 80 mg/liter limit applied tofacilities with existing vapor processorsshould be able to be met by any of thecontrol equipment which was installedunder SIP requirements.

Some commenters stated that it hadnot been shown by EPA that theproposed standards would beachievable under all the variableoperating conditions that may existthroughout the industry. However, thesecommenters did not identify any specificvariable operating conditions whichthey felt may affect emission levels, norwas any technical information includedwith the comments. The typicalperformance test on bulk terminalcontrol systems does not monitoroperating conditions and their possibleeffect on emissions, because generallyall that is required in this test procedureis the measurement of outlet massemissions over several hours. However,data were collected during the EPA-sponsored test program and variables(gasoline composition, vaporconcentration, and peak loading levels)have been identified as having a

possible effect on the mass emissionlevel or control efficiency of the controltechnologies considered capable ofachieving the limit of the standard.

Gasolines with different Reid vaporpressures (RVP) are marketed indifferent seasons of the year, in order tomaintain approximately constant actualvapor pressure as the mean ambienttemperature changes. Under winterconditions, therefore, mass emissionsmay be higher for some systems becauseof increased light ends in the inletvapors. If CA and REF units are sizedwith sufficient collection area to meetthe emission limit in winter, emissionsin summer will then be well below thelimit. TO systems are often designed tohandle saturated streams stored invapor holders, and should not beaffected by the variable RVP. Tests ofCA to TO units considered by theAgency show that the emission limitwas achieved at various times of theyear and, therefore, under variousgasoline compositions.

Both CA and TO systems have beentested under a range of inlet VOCconcentrations returned from tanktrucks, and the test results indicate theability of these technologies to achievethe limit of the standards under highinlet concentrations. Also, theoreticalestimations and analyses for CA andREF systems have indicated that thesesystems will collect efficiently, andexhibit outlet emissions below 35 mg/liter, throughout the range ofconcentrations which will beexperienced at new bulk terminals[docket items IV-A-2, IV-D--36, IV-D-38). Efficiencies, in fact, are likely toincrease with increasing inletconcentration. TO systems are easilydesigned to handle saturated inletstreams.

Most control systems are designed forpeak loading hours at a terminal, ratherthan daily throughput, because of thefluctuation in loading activitythroughout the day. Thus, a properlysized unit that can handle peak periodsshould have improved performanceduring the remainder of the day.

It was concluded that the operationalvariables at a terminal are merelydesign variables which affect theselection and sizing of the vaporprocessor. No variables have beenidentified which would prevent thesestandards from being met on aconsistent basis.

Several commenters felt that theproposed emission limit of 35 mg/literfor new vapor processors is too stringentfor the current generation of vaporprocessors in use at bulk terminals.Some of the commenters stated that

37585

37586 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

certain types of processors would beunable to achieve this limit, wholeothers felt that the limit wasunnecessarily stringent for any of theexisting technologies. Alternate limits of55 mg/liter and 80 mg/liter weresuggested.

Standards of performance, in the formof numercial emission limits, areintended to reflect the degree ofemission limitation achievable throughapplication of the best adequatelydernonstated technological system ofcontinuous emission reduction, takinginto consideration the cost of achievingsuch emission reduction, any nonairquality health and environmentalimpacts, and energy requirements.Carbon adsorption vapor processorsmanufactured by both of the majorsuppliers have demonstrated thecapability to achieve emission levelsbelow 35 mg/liter on a regular basis.Also, thermal oxidation units haveshown the capability to achieve 35 mg/liter, although some TO systems mayrequire a vapor holder to achieve thislimit reliably. Compression-oxidationhybrid systems have been found toachieve the same high controlefficiences as the straight TO systems.In addition, test data, computermodeling, and the manufacturer's claimssuggest the REF systems can bedesigned and operated to meet 35 mg/liter.

Based on a number of emission tests,EPA has identified carbon adsorptionand thermal oxidation as the bestdemonstrated technologies (BDT) forcontrolling vapors from gasoline loadingracks. Section 111 requires EPA to setnumerical emission limits achievablethrough application of BDT (consideringthe statutory factors), even if by doingso the Agency precludes the use of lesseffective systems. Owners arenonetheless free to use any technologythat will achieve the limit.

Some commenters referred to carbonbed temperature excursions at severalCA unit installations during the summerof 1980. Due to the resulting extendedshutdowns, one commenter felt thatdoubt had been cast on the ability ofcurrently designed systems to maintainhigh efficiency consistently. Contactswere made by EPA with systemmanufacturers and oil industryrepresentatives, to determine theapparent reasons for the six reportedoccurrences of carbon bed overheating.Discussions indicated that theoverheating incidents were primarily theresult of improper flow distribution andImproper startup procedures resulting inthe insufficient preloading of the virgincarbon in some new, larger units.

Precautionary measures to preventoverheating including: (1) Completeconditioning of the virgin carbon toensure that an adequate heel has beenplaced on the carbon to minimizesubsequent high adsorption heatreleases, and (2) sizing the unit tomaintain proper vapor velocity and flowdistribution through the carbon beds.According to the system manufacturers,overheating should not occur if theseprecautionary measures are employed(docket item IV-D-36).

Industry representatives haveaddressed the carbon bed overheatingissue by incorporating emergencyshutdown measures and bed coolingdevices on newer systems. Twoadditional oil industry representativesindicated that, on any new carbonsystem ordered (and possibly retrofittedto existing systems), they will specifycooling provisions and additionaltemperature sensors. Since only 6temperature excursion occurrences havebeen Identified in the approximately 200operating carbon systems, theoverheating problem does not appear tobe widespread. EPA agrees with themanufacturers and with industryrepresentatives that an effort should bemade to follow carefully therecommended startup and operationalprocedures to minimize the conditionswhich may tend to promote'temperatureexcursions. The added costs ofemergency shutdown and bed coolingprovisions on the newest CA units havebeen incorporated in the revised costanalysis in estimating the control cost ofthe standards to the bulk terminalIndustry.

Two commenters felt that CA systemshave several general operationalproblems and that this technology is stillin the developmental stages. The firstcarbon adsorption system for bulkterminal vapor recovery was installed inNovember of 1976, and today the marketis shared by two*manufacturers withapproximately 200 units in operation.Most types of vapor processors can beconsidered to be under development inthe sense that continual designimprovements are being made. Someproblems with vacuum valve actuatorsand vacuum pump seals have occurred,as well as problems related to extremelyco!d weather operation. Many of theseproblems have been solved (docket itemIV-E-53), and EPA has not been madeaware of any remaining operationalproblems which would affect the abilityof CA systems to comply with thepromulgated standards.

Comments on refrigeration unitsconcerned the ability of this technologyto achieve the proposed standard of

performance. Some commenters agreedthat REF units could be designed andoperated to achieve 35 mg/literconsistently, but felt that the addedcosts over current units would not beeconomically practical. The promulgatedemission limit of 35 mg/liter wasselected to reflect the performance ofthe best control systems, which test datashowed to be the CA and TOtechnologies. The most currentrefrigeration systems have generallybeen installed to meet the 80 mg/literlimit and have achieved 35 mg/liter inonly some instances, with emissionsfrom most units slightly above the 35mg/liter limit. Indications are that theseunits can be specified and operated tomeet 35 mg/liter, at increased capitaland operating costs over most currentunits. The capital costs for most sizes ofREF units fall between the costs for TOand CA type units. Electrical costs forREF units are comparable to those forTO and CA units, except for the smallerbulk terminal sizes, where they areslightly higher. Detailed costs arepresented in Appendix B of BID, VolumeII.

Tank Truck Issues

Several commenters questioned EPA'slegal authority to impose restrictions,i.e., retrofitting and vapor tightnesstesting, on gasoline tank trucks. Theyfelt that trucks do not fall within thecategory of a stationary source and,therefore, cannot be regulated underSection 111. The commenters furtherstated that EPA could not regulate amobile source directly or indirectlyunder Section 111. One commenteroharacterized the regulation of tanktruck emissions as constituting "thetaking of private property without cause,compensation, or due process."

For purposes of this NSPS, thestationary source, or affected facility, isthe total of all bulk terminal loadingracks loading liquid product intogasoline tank trucks. Those loadingracks are essential to carrying out theactivity known as product loading.While product loading involves both theaffected facility and mobile equipment,including the tank truck, it is clearly astationary activity, since it requires nomovement from the affected facility site.Among the pollutants created byproduct loading are vapors forced fromthe tank truck as a direct result of thepumping of liquid product into the tanktruck. Since escape of these vapors iscaused by stationary activities at astationary facility, they are "stationarysource" emissions subject to regulationunder Section 111-even though the tanktrucks from which they escape during

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

that activity have the capability tomove.

As indicated above, the tank truck isnot included in the designation of the"affected facility" under thesestandards. The standards placeresponsibility on the terminal owneronly, requiring the owner to restrictloadings to vapor-tight tank trucksequipped with compatible vaporrecovery equipment. The regulationwould not directly require either new orold tank trucks to be vapor-tight orequipped with certain types ofhardware.

Section 111(a)(2) defines "stationarysource" as any "building, structure,facility, or installation which emits ormay emit any air pollutant." EPAidentifies the "stationary source" ascertain specified stationary equipment(termed the "affected facility") that"emits" a pollutant. In theAdministrator's view, stationaryequipment "emits" a pollutant if itcauses that pollutant to enter theatmosphere.'

In the Administrator's view, affectedfacility emissions subject to regulationunder Section 111 include all pollutantsthat enter the atmosphere as a result ofthe stationary industrial activities at theaffected facility, even those that enterthe atmosphere after contactingequipment with mobility. Stateddifferently, the test for whetheremissions are "stationary source"emissions subject to regulation underSection 111 is whether the emissions arecaused by a stationary facility duringactivities that require no movement fromthe facility, not whether the emissionsescape to the atmosphere withouttouching equipment having thecapability to move.

Interpreting "stationary source"emissions to include emissions resultingfrom stationary activities in which boththe affected facility and some mobileequipment take part serves the intent ofthe statute. Congress enacted Section111 for the "overriding purpose" of"prevent(ing) new pollution problems."S. Rep. No. 91-1196, 1970 Leg. Hist. at416. The Senate Report states thatSection 111 seeks to attain this goal byrequiring control of new commercial andindustrial establishments "to themaximum practicable degree regardlessof their * * * industrial operations." Id.Similarly, the Report states that"maximum use of available means of

'EPA's authority to define the term "emits" in thisway derives from Section 301 of the Act, asinterpreted in the cases (see, e.g.. Alabama Power v.Castle, 636 F.2d 323 (D.C. Cir. 1979)). In accordancewith this provision, the Agency is interpreting theterm "emits" broadly, to serve the broad purposesof Section 111 (described in the text below).

preventing and controlling air pollutionis essential" to the attainment of thegoals of Section 111. Id. The legislativehistory thus indicates that Congressintend Section 111 to address emissionsfrom all stationary operations atindustrial establishments when theAgency can identify the maximumpracticable degree of control for theseemissions. To interpret Section 111(a)(2)so that emissions resulting from certainstationary activities involving thestationary source would not constitute"stationary source" emissions simplybecause those emissions pass throughsome equipment with the capability tomove would be incompatible with thatintent.

The Agency recognizes thatpromulgation of standards regulatingloading racks as "stationary sources"may significantly affect tank truckowners and other segments of thepetroleum marketing and transportationindustry. The fact that standards withinan agency's statutory authorityindirectly affect nonregulated entities,however, does not in and of itselfdiminish the authority to set thestandards. Nothing in the statute or itshistory indicates that, in the case athand, the indirect impact that regulationof emissions from loading racks willhave on certain tank truck ownersdeprives the Agency of its clearauthority to set new source performancestandards for this source category.

In fact, it is likely that most newsource standards affect to some degreeindustries other than that to which thestandards directly apply. The standardsfor electric utility steam generators, forinstance (40 CFR 60.40a-49a, SubpartDa), significantly affect the coal miningand railroad industries. The impact ontank trucks of a requirement that certainbulk terminals load only into vapor-tighttrucks equipped with compatibleequipment does not differ in kind fromthe indirect impacts resulting fromSubpart Da and other new sourceperformance standards. Bulk terminalsdeal extensively with delivery vehicles.As a result, it is to be expected thatregulation of bulk terminals would affectdelivery vehicles in some manner.

The potential effect of the standardson tank truck owners does not amountto a denial of due process or anunconstitutional taking of property.Because the commenter did notelaborate on the specific bases for theseclaims of unconstitutionality, theAgency can respond only generally. TheClean Air Act reflects a congressionaldetermination that air pollution has asubstantial effect on interstatecommerce and therefore may be

regulated by Congress (and, throughproper delegation, EPA] under thecommerce clause. District of Columbiav. Train, 521 F.2d 971, 988 (D.C. Cir.1975). It is unreasonable to suggest thatregulation of emissions forced from thetank truck during loading bears norational relationship to protection ofpublic health and welfare, and thusviolates the due process clause of theFifth Amendment. There is a rationalrelationship between escape of thesevapors and the public health andwelfare, because these emissionscontribute to ozone formation. SierraClub v. EPA, 540 F.2d 1114, 1139 (D.C.Cir. 1976). There is also a properlegislative purpose underlying therequirements aimed at controlling theseemissions. Moreover, the means theAgency has chosen, as discussed above,are reasonable and appropriate. Id., at1139 n.80 [citing Heart of Atlanta Motel,Inc. v. United States, 379 U.S. 241, 258-59 (1964)].

Nor do these standards transgress thetakings prohibition in the Constitution.Given the substantial public interest inpreserving clean air, tight restrictionsmay constitutionally be imposed onprivate property. South Terminal Corp.v. EPA, 504 F.2d 646, 678-80 (1st Cir.1974). While this NSPS indirectly limitsthe uses of tank trucks, the limitation isnot so extreme as to constitute anappropriation of the vehicles. SierraClub v. EPA, supra, at 1140. Thisregulation affects only one of the tanktruck uses available to the truckowner-loading at affected facilities.The right to use nonvapor-tight tanktrucks at other facilities is neitherextinguished nor transferred to someoneelse.

Several commenters felt that theterminal owner or operator should nothave any responsibility for the vapor-tight status of for-hire tank trucks. Thecommenters felt that the terminaloperator should not be required to policethe testing and use of tank trucks whichare owned by others.

Fugitive, or leakage, VOC emissionsfrom tank trucks which occur duringloading can be a significant emissionsource. Test data indicate that, on theaverage, a nonvapor-tight tank couldlose 30 percent of the potential vaportransferred through leaks in domecovers and pressure-vacuum vents. Thedata further show that, by requiring thetanks which handle gasoline to pass anannual vapor tightness test, the averagevapor loss due to leakage during theyear between tests can be reduced to 10percent of the potential vaporstransferred. Fugitive VOC losses fromtank trucks not only increase the

37587

37588 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

pollution problem but decrease theamount of product that can be reclaimedin vapor recovery equipment. Theterminal owner or operator could lose asmuch as $2 in recovered product perloading into nonvapor-tight trucks. For asmall 380,000 liter/day (100,000 gallon/day) terminal this could represent adaily loss of over $25. For a large3,800,000 liter/day (1,000,000 gallon/dayterminal the losses could be over $250/day. Bulk terminal industryrepresentatives agree that the vaportightness requirement for tank trucks isa necessary provision of the regulation(docket items IV-E-19, IV-F-3).

The objections from the bulk terminalindustry arise regarding theresponsibility for assuring loadings areinto vapor-tight tanks. The industry feelsthe responsibility should be on the tanktruck operator, who in fact may be theterminal operator or oil company, or anindependent who operates for-hire tanktrucks. However, in order for theresponsibility under new sourcestandards to be on an independent tanktruck operator, the tank truck wouldhave to be part of the affected facility.The feasibility of including the tanktruck as part of the affected facility wasreviewed in the preamble to theproposed standards. It was determinedthat the best approach to controllingfugitive tank truck leakage was to makethe standards applicable only to bulkterminals, with a requirement thataffected terminals load only into truck-mounted tanks that have passed a vaportightness test. Because tank trucks loadprimarily with equipment owned by theterminal owner, and on the property ofthe terminal owner, EPA be!ieves it isreasonable to presume, for the purposeof this regulation, that these owners canexercise sufficient control over thesource to justify making themresponsible for the emissions therefrom.

EPA did not intend for terminalpersonnel to man the racks 24 hours perday, or actually observe the loading ofevery tank truck to verify that eachtruck had passed an annual vaportightness test. EPA felt that requiringdocumentation on file that gasoline tanktrucks operating out of the terminal hadpassed a vapor tightness test wouldprovide a sufficient means of promotingloadings into vapor-tight tanks. Industryopposition is centered around theliability on the terminal owner for tanktrucks he does not own. At unmanned,automated terminals, the terminaloperator is usually not present andcannot determine which trucks areloading. The Agency realizes theselimitations but believes that the vapor

tightness requirement is necessary inorder for these standards to be effective.

Changes to the vapor tightnessrequirement have been incorporatedinto the promulgated regulation toclarify that the standards do not requirethe terminal operator to monitor eachtank truck loading. A requirement to logthe tank identification number of allgasoline tank trucks loading at affectedfacilities has been incorporated into thefinal regulation. Since the quantity ofproduct which passes through theterminal and its corresponding worth isvery large, there is already considerablepaperwork involved in tracking theproducts in and out of the terminal. Thetruck identification information could berecorded by the truck driver as part ofthe normal paperwork which alreadyaccompanies each loading. If the tankidentification number is logged eachtime the tank is loaded, the owner canperiodically cross-check the tankindentification number with the vaportightness documentation on file at theterminal. This cross-checking is requiredwithin 2 weeks of the loading. If theterminal discovers that an unauthorizedtank truck has received gasoline, theterminal operator notifies the tankowner, and takes steps to assure thatthe nonvapor-tight truck does not reloadat the terminal until proper vaportightness documentation is obtained.This notification must be documentedand kept on file at the terminal. Methodsof achieving this are available to the.terminal owner or operator and couldinclude revocation of loading privilegesor contractural agreements between theterminal owner or operator and thetruck owner or operator. However, EPAhas not specified any particular method,to allow the terminal owner or operatorthe flexibility to meet the requirements,with minimum disruption to terminaloperations. Section 111(h)(3) of theClean Air Act provides that if theterminal owner, after notice andopportunity for public hearing,"establishes to the satisfaction of theAdministrator that an alternative meansof emission limitation will achieve areduction in emission * * * at leastequivalent to the reduction in emissionsof such air pollutant" achieved underthe tank truck vapor tightnessrequirement, the Administrator "shallpermit the use of such alternatives* * " Thus, the terminal owner is free,with EPA approval under Section111(h)(3), to develop a different strategyfor controlling fugitive emissions fromtank trucks.

One commenter felt that anadministrative burden would be createdby a requirement to keep vapor

tightness documentation for as many as400 to 500 transport trucks using a giventerminal. Several other commentersgenerally argued that the tank truckcontrols would represent anadministrative burden, as well as beingcostly and inequitable.

The testing and maintenance of tank 1trucks for vapor tightness has beenshown to have a significant effect inreducing total emissions during loading.Thus, this procedure has a veryimportant function in bulk terminal VOCemissions limitation. The administrativeburden of keeping the documentation onfile would be minimal since theinformation would in most cases besupplied by the owner of for-hire tanktrucks and the terminal would simplyfile the data. Cross-checking these fileswith tank identification numbers loggedduring loading should be a simpleprocess and would not be an excessiveburden. Furthermore, this filing andcross-checking would represent muchless of a burden then the in-personmonitoring by terminal personnel ofeach loading as it occurred.

Docket

The docket is an organized andcomplete file of information submitted,or otherwise considered, in thedevelopment of this rulemaking. Thedocket is a dynamic file. since materialis added throughout the rulemakingdevelopment. The docketing system isintended to allow members of the publicand industries involved to identify andlocate documents readily so that theycan effectively participate in therulemaking process. Along with thestatement of basis and purpose of theproposed and promulgated standardsand EPA responses to significantcomments, the contents of the docket,except for certain interagency reviewmaterials, will serve as the record incase of judicial review [Section307(d)(7)(A)].

Miscellaneous

The effective date of this regulation isAugust 18, 1983. Section 111 of the CleanAir Act provides that standards ofperformance or revisions thereofbecome effective upon promulgation andapply to affected facilities, constructionor modification of which wascommenced after the date of proposal(December 17, 1980).

As prescribed by Section 111, thepromulgation of these standards waspreceded by the Administrator'sdetermination (40 CFR 60.16, 44 FR49222, dated August 21, 1979) that thesesources contribute significantly to airpollution which may reasonably be

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

anticipated to endanger public health orwelfare. In accordance with Section 117of the Act, publication of thesepromulgated standards was preceded byconsultation with appropriate advisorycommittees, independent experts, andFederal departments and agencies.

This regulation will be reviewedwithin 4 years from the date ofpromulgation as required by the CleanAir Act. This review will include anassessment of such factors as the needfor4ntegration with other programs, theexistence of alternative methods,enforceability, emission controltechnology, and reporting requirements.

Section 317 of the Clean Air Actrequires the Administrator to prepare aneconomic impact assessment for anynew source standard of performanceunder Section 111(b) of the Act. Aneconomic impact assessment wasprepared for this regulation and forother regulatory alternatives. Allaspects of the assessment wereconsidered in the formulation of thestandards to ensure that cost wascarefully considered in determiningBDT. The economic impact assessmentis included in the backgroundinformation documents for the proposedand promulgated standards (BID,Volumes I and II).

The Paperwork Reduction Act (PRA)of 1980 (Pub. L 96-511) requiresclearance from the Office of Mangementand Budget (OMB) of reporting andrecordkeeping requirements that qualifyas an "information collection request"under the PRA. For the purposes ofOMB's review, and analysis of theburden associated with the reportingand recordkeeping requirements of thisregulation has been made. During thefirst 2 years of this regulation, theaverage annual burden of the reportingand recordkeeping requirements wouldbe 4.8 person-years, based on anaverage of 11 respondents per year.Information collection requirementscontained in this regulation (§§ 60.502,60.503, 60.505) have been approved bythe OMB under the provisions of thePaperwork Reduction Act of 1980, 44U.S.C. 3501 et seq. and have beenassigned OMB control number 2060-0006.

The Regulatory Flexibility Act of 1980(RFA) requires that differential impactson small businesses resulting from allFederal regulations be identified andanalyzed. The RFA does not by its termsapply to regulations proposed prior toJanuary 1, 1981. Consequently the RFAdoes not impose any requirements in theAgency's development of the bulkgasoline terminal NSPS (proposedDecember 17, 1980). However, the

Agency has considered the economicimpact of the standards on relativelysmall terminals and tank truck firms,and the economic analysis has Fncebeen reviewed in reference to the RFA.The definition of a small business in thebulk terminal industry (SIC 5171),according to the criterion to qualify forSBA loans, is a firm with less than $22million in annual receipts.Approximately 50 to 60 percent of thebulk terminal industry can beconsidered as small businessesaccording to this criterion. In the for-hiretank truck industry (SICs 4212, 4213, and4214), a small business is defined as afirm with less than $6.5 to $7 million inannual receipts. Approximately 60percent of the for-hire tank truckindustry can be considered as smallbusinesses according to this criterion.The RFA further stipulates that theanalysis must be prepared if 20 percentof the small businesses are significantlyaffected.

Five new terminals are expected to beconstructed in the first five years, andapproximately 50 facilities will becomeaffected through modification orreconstruction. Of the 55 affectedfacilities, 15 terminals, a 27 percentshare, can be considered small businessentities (assuming Model Plant 1approximates a small business), and sothe 20 percent criterion is exceeded. Theanalysis concluded that significantimpact for small business entities wouldoccur only under the worst-caseassumption of complete cost absorption.Under a more likely scenario, furtheranalysis revealed no significant impact.Since the impact on small bulk terminalbusinesses is not expected to besignificant, no Regulatory FlexibilityAnalysis is required for this industrysector.

Thirty-four model firms in the for-hiretank truck industry are expected to beaffected by 1985. Twenty-three affectedfirms are expected to be small businessentities, representing a 68 percent share,which exceeds the 20 percent criteion.The potential exists for a significantimpact to occur in worst-case scenario ifcontrol costs are completely absorbed.The results from the return-on-transportation investment analysis notonly suggested as significant worst-caseimpact, but that the impacts are moresevere for the largest model truckingfii'ms. A more likely scenario wasanalyzed and no significant economicimpact was found. This scenario wasbased on the realistic assumption thatmost of the control costs will be passedthrough with very little cost absorptionaffecting the ROTI. Even under completecost pass-through the price of gasoline

increases at most by 0.03 percent. Sincethe impact on small independent tanktruck firms is not expected to besignificant, no Regulatory FlexibilityAnalysis is required for this industrysector.

Under Executive Order 12291, EPA isrequired to judge whether a regulation isa "major rule" and therefore subject tocertain requirements of the Order. TheAgency has determined this regulationwill result in none of the adverseeconomic effects set forth in Section I ofthe Order as grounds for finding aregulation to be a "major rule." The netannualized costs through the first 5years of implementation, includingdepreciation and interest, are projectedto be considerably below the thresholdcost for defining a "major rule." Onlynegligible increases in the price ofgasoline attributable to implementationof these standards are expected. TheAgency has therefore concluded thatthis regulation is not a "major rule"under Executive Order 12291. Inaddition to the economic analysis, theAgency carefully examined the cost ofvarious technical alternatives in termsof the emission reductions achieved.This was done for the range ofconfigurations and facility sizes whichare anticipated to be affected by thestandard and, as described under thepreamble section "Modification andReconstruction," led to relaxation of theproposed standard for sources with SIPlevel controls in-place. The incrementalcost of the final standard in terms of theincremental emission reductionachieved would range from a savings atcertain medium to large size plants to acost of approximately $1,100/Mg for atypical small facility. The total cost perunit of VOC emission reductionassociated with this regulation is $440/Mg. This cost is consistent with that ofother new source performancestandards some of which cost $1,000/Mgto $2,000/Mg of VOC emissionreduction, or higher.

List of Subjects in 40 CFR Part 60

Air pollution control, Aluminum,Ammonium sulfate plants, Asphalt,Cement industry, Coal copper, Electricpower plants, Glass and glass products,Grains, Intergovernmental relations,Iron, Lead, Metals, Metallic minerals,Motor vehicles, Nitric acid plants; Paperand paper products industry, Petroleum,Phosphate, Sewage disposal, SteelSulfuric acid plants, Waste treatmentand disposal, Zinc, Tires.

37589

37590 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

Dated: August 4, 1983.William D. Ruckelsbaus,Administrator.

PART 60-[AMENDED]

40 CFR Part 60 is amended as follows:1. By adding a new subpart as follows:

Subpart XX-Standards of Performance forBluk Gasoline Terminals

Sec.60.500 Applicability and designation of

affected facility.60.501 Definitions.60.502 Standards for Volatile Organic

Compound (VOC) emissions from bulkgasoline terminals.

60.503 Test methods and procedures.60.504 [Reserved.]60.505 Reporting and recordkeeping.60.506 Reconstruction.

Authority: Sections 111 and 301(a) of theClean Air Act, as amended [42 U.S.C. 7411,7601(a)], and additional authority as notedbelow.

Subpart XX-Standards ofPerformance for Bulk GasolineTerminals

§ 60.500 Applicability and designation ofaffected facility.

(a) The affected facility to which theprovisions of this subpart apply is thetotal of all the loading racks at a bulkgasoline terminal which deliver liquidproduct into gasoline tank trucks.

(b) Each facility under paragraph (a)of this section, the construction ormodification of which is commencedafter December 17, 1980, is subject to theprovisions of this subpart.

(c) For purposes of this subpart, anyreplacement of components of anexisting facility, described in paragraph§ 60.500(a), commenced before August18, 1983 in order to comply with anyemission standard adopted by a State orpolitical subdivision thereof will not beconsidered a reconstruction under theprovisions of 40 CFR 60.15.

[Note: The intent of these standards is tominimize the emissions of VOC through theapplication of best demonstratedtechnologies (BDT). The numerical emissionlimits in this standard are expressed in termsof total organic compounds. This emissionlimit reflects the performance of BDT.]

§ 60.501 Definitions.The terms used in this subpart are

defined in the Clean Air Act, in § 60.2 ofthis part, or in this section as follows:

"Bulk gasoline terminal" means anygasoline facility which receives gasoline,by pipeline, ship or barge, and has agasoline throughput greater than 75,700liters per day. Gasoline throughput shallbe the maximum calculated designthroughput as may be limited bycompliance with an enforceable

condition under Federal, State or locallaw and discoverable by theAdministrator and any other person.

"Continuous vapor processingsystem" means a vapor processingsystem that treats total organiccompounds vapors collected fromgasoline tank trucks on a demand basiswithout intermediate accumulation in avapor holder.

"Existing vapor processing system"means a vapor processing system[capable of achieving emissions to theatmosphere no greater than 80milligrams of total organic compoundsper liter of gasoline loaded], theconstruction or refurbishment of whichwas commenced before December 17,1980, and which was not constructed orrefurbished after that date.

"Gasoline" means any petroleumdistillate or petroleum distillate/alcoholblend having a Reid vapor pressure of27.6 kilopascals or greater which is usedas a fuel for internal combustionengines.

"Gasoline tank truck" means adelivery tank truck used at bulk gasolineterminals which is loading gasoline orwhich has loaded gasoline on theimmediately previous load.

"Intermittent vapor processingsystem" means a vapor processingsystem that employs an intermediatevapor holder to accumulate total organiccompounds vapors collected fromgasoline tank trucks, and treats theaccumulated vapors only duringautomatically controlled cycles.

"Loading rack" means the loadingarms, pumps, meters, shutoff valves,relief valves, and other piping andvalves necessary to fill delivery, tanktrucks,

"Refurbishment" means, withreference to a vapor processing system,replacement of components of, oraddition of components to, the systemwithin any 2-year period such that thefixed capital cost of the newcomponents required for suchcomponent replacement or additionexceeds 50 percent of the cost of acomparable entirely new system.

"Total organic compounds" meansthose compounds measured according tothe procedures in § 60.503.

"Vapor collection system" means anyequipment used for containing totalorganic compounds vapors displacedduring the loading of gasoline tanktrucks.

"Vapor processing system" means allequipment used for recovering oroxidizing total organic compoundsvapors displaced from the affectedfacility.

"Vapor-tight gasoline tank truck"means a gasoline tank truck which has

demonstrated within the 12 precedingmonths that its product delivery tankwill sustain a pressure change of notmore than 750 pascals (75 mm of water)within 5 minutes after it is pressurizedto 4,500 pascals (450 mm of water). Thiscapability is to be demonstrated usingthe pressure test procedure specified inReference Method 27.

§ 60.502 Standard for Volatile OrganicCompound (VOC) emissions from bulkgasoline terminals.

On and after the date on which§ 60.8(b) requires a performance test tobe completed, the owner or operator ofeach bulk gasoline terminal containingan affected facility shall comply withthe requirements of this section.

(a) Each affected facility shall beequipped with a vapor collection systemdesigned to collect the total organiccompounds vapors displaced from tanktrucks during product loading.

(b) The emissions to the atmospherefrom the vapor collection system due tothe loading of liquid product intogasoline tank trucks are not to exceed 35milligrams of total organic compoundsper liter of gasoline loaded, except asnoted in paragraph (c) of this section.

(c) For each affected facility equippedwith an existing vapor processingsystem, the emissions to the atmospherefrom the vapor collection system due tothe loading of liquid product intogasoline tank truckb are not to exceed 80milligrams of total organic compoundsper liter of gasoline loaded.

(d) Each vapor collection system shallbe designed to prevent any total organiccompounds vapors collected at oneloading rack from passing to anotherloading rack.

[e) Loadings of liquid product intogasoline tank trucks shall be limited tovapor-tight gasoline tank trucks usingthe following procedures:

(1) The owner or operator shall obtainthe vapor tightness documentatio-ndescribed in § 60.505(b) for eachgasoline tank truck which is to beloaded at the affected facility.

(2) The owner or operator shallrequire the tank identification number tobe recorded as each gasoline tank truckis loaded at the affected facility.

(3) The owner or operator shall cross-check each tank identification numberobtained in (e)(2] of this section with thefile of tank vapor tightnessdocumentation within 2 weeks after thecorresponding tank is loaded.

(4) The terminal owner or operatorshall notify the owner or operator ofeach nonvapor-tight gasoline tank truckloaded at the affected facility within 3weeks after the loading has occurred.

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

(5) The terminal owner or operatorshall take steps assuring that thenonvapor-tight gasoline tank truck willnot be reloaded at the affected facilityuntil vapor tightness documentation forthat tank is obtained.

(6) Alternate procedures to thosedescribed in (e)(1) through (5) of thissection for limiting gasoline tank truckloadings may be used upon applicationto, and approval by, the Administrator.

(f) The owner or operator shall act toassure that loadings of gasoline tanktrucks at the affected facility are madeonly into tanks equipped with vaporcollection equipment that is compatiblewith the terminal's vapor collectionsystem.

(g) The owner or operator shall act toassure that the terminal's and the tanktruck's vapor collection systems areconnected during each loading of agasoline tank truck at the affectedfacility. Examples of actions toaccomplish this include training driversin the hookup procedures and postingvisible reminder signs at the affectedloading racks.

(h) The vapor collection and liquidloading equipment shall be designed andoperated to prevent gauge pressure inthe delivery tank from exceeding 4,500pascals (450 mm of water) duringproduct loading. This level is not to beexceeded when measured by theprocedures specified in § 60.503(b).

(i) No pressure-vacuum vent in thebulk gasoline terminal's vapor collectionsystem shall begin to open at a systempressure less than 4,500 pascals (450 mmof water).

(j) Each calendar month, the vaporcollection system, the vapor processingsystem, and each loading rack handlinggasoline shall be inspected during theloading of gasoline tank trucks for totalorganic compounds liquid or vaporleaks. For purposes of this paragraph,detection methods incorporating sight,sound, or smell are acceptable. Eachdetection of a leak shall be recorded andthe source of the leak repaired within 15calendar days after it is detected.

(Approved by the Office of Management andBudget under control number 2060-0006)

§ 60.503 Test methods and procedures.(a) Section 60.8(f) does not apply to

the performance test proceduresrequired by this subpart.

(b) For the purpose of determiningcompliance with § 60.502(h). thefollowing procedures shall be used:

(1) Calibrate and install a pressuremeasurement device (liquid manometer,magnehelic gauge, or equivalent

instrument), capable of measuring up to500 mm of water gauge pressure with_2.5 mm of water precision.

(2) Connect the pressure inva ::w urnentdevice to a pressure tap in the terminal'svapor collection system, located as closeas possible to the connection with thegasoline tank truck.

(3) During the performance test,record the pressure every 5 minuteswhile a gasoline tank truck is beingloaded, and record the highestinstantaneous pressure that occursduring each loading. Every loadingposition must be tested at least onceduring the

(c) For the purpose of determiningcompliance with the mass emissionlimitations of § 60.502(b) and (c), thefollowing reference methods shall beused:

(1) For the determination of volume atthe exhaust vent:

(i) Method 2B for combustion vaporprocessing systems.

(ii) Method 2A for all other vaporprocessing systems.

(2) For the determination of totalorganic compounds concentration at theexhaust vent, Method 25A or 2513. Thecalibration gas shall be either propaneor butane.

(d) Immediately prior to aperformance test required fordetermination of compliance with§ 60.502[b), (c), and (h), all potentialsources of vapor leakage in theterminal's vapor collection systemequipment shall be monitored for leaksusing Method 21. The monitoring shallbe conducted only while a gasoline tanktruck is being loaded. A reading of10,000 ppmv or greater as methane shallbe considered a leak. All leaks shall berepaired prior to conducting theperformance test.

(e) The test procedure for determiningcompliance with § 60.502(b) and (c) is asfollows:

(1) All testing equipment shall beprepared and installed as specified inthe appropriate test methods.

(2) The time period for a performancetest shall be not less than 6 hours,during which at least 300,000 liters ofgasoline are loaded. If the throughputcriterion is not met during the initial 6hours, the test may be either continueduntil the throughput criterion is met, orresumed the next day with anothercomplete 6 hours of testing. As much aspossible, testing should be conductedduring the 6-hour period in which thehighest throughput normally occurs.

(3) For intermittent vapor processingsystems:

(i) The vapor holder level shall berecorded at the start of the performancetest. The end of the performance testshall coincide with a time when thevapor holder is at its original level.

(iH) At least two startups andshutdowns of the vapor processor shalloccur during the performance test. If thisdoes not occur under automaticallycontrolled operation, the system shall bemanually controlled.

(4) The volume of gasoline dispensedduring the performance test period at allloading racks whose vapor emissionsare controlled by the processing systembeing tested shall be determined. Thisvolume may be determined fromterminal records or from gasolinedispensing meters at each loading rack.

(5) An emission testing interval shallconsist of each 5-minute period duringthe performance test. For each interval:

(i) The reading from eachmeasurement instrument shall berecorded, and

(ii) The volume exhausted and theaverage total organic compoundsconcentration in the exhaust vent shallbe determined, as specified in theappropriate test method. The averagetotal organic compounds concentrationshall correspond to the volumemeasurement by taking into account thesampling system response time.

(6) The mass emitted during eachtesting interval shall be calculated asfollows:

Me = 10- sKVC.

where:M. 1=mass of total organic compounds

emitted during testing interval i. mg.V.=volume of air-vapor mixture exhausted,

m3, at standard conditions.Ce= total organic compounds concentration

(as measured) at the exhaust vent, ppmv.K=density of calibration gas, mg/m3, at

standard conditions =1.83X106, forpropane = 2.41 X loffor butane.

s= standard conditions, 20'C and 760 min Hg.

(7) The total organic compounds massemissions shall be calculated as follows:

"i mE-

L

where:

E=mass of total organic compounds emittedper volume of gasoline loaded, mg/liter.

M.,=rnass of total organic compoundsemitted during testing interval i. mg.

L= total volume of gasoline loaded, liters.n=number of testing intervals.

37591

37592 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

(f) The owner or operator may adjustthe emission results to exclu'de themethane and ethane content in theexhaust vent by any method approvedby the Administrator.

(Sec. 114 of the Clean Air Act as amended (42U.S.C. 7414)](Approved by the Office of Management andBudget under control number 2060-0006.)

§ 60.504 [Reserved].

§ 60.505 Reporting and recordkeeping.(a) The tank truck vapor tightness

documentation required under160.502(e)(1) shall be kept on file at theterminal in a permanent form availablefor inspection.

(b) The documentation file for eachgasoline tank truck shall be updated atleast once per year to reflect current testresults as determined by Method 27.This documentation shall include, as aminimum, the following information:

(1) Test Title: Gasoline Delivery TankPressure Test-EPA Reference Method27.

(2) Tank Owner and Address.(3) Tank Identification Number.(4) Testing Location.(5) Date of Test.(6) Tester Name and Signature.(7) Witnessing Inspector, if any:

Name, Signature, and Affiliation.(8) Test Results: Actual Pressure

Change in 5 minutes, mm of water(average for 2 runs).

(c) A record of each monthly leakinspection required under § 60.502(j)shall be kept on file at the terminal forat least 2 years. Inspection records shallinclude, as a minimum, the followinginformation:

(1) Date of Inspection.(2) Findings (may indicate no leaks

discovered; or location, nature, andseverity of each leak).

(3) Leak determination method.(4) Corrective Action (date each leak

repaired; reasons for any repair intervalin excess of 15 days).

(5) Inspector Name and Signature.(d) The terminal owner or operator

shall keep documentation of allnotifications required under§ 60.502(e)(4) on file at the terminal forat least 2 years.

(e) [Reserved].(f) The owner or operator of an

affected facility shall keep records of allreplacements or additions ofcomponents performed on an existingvapor processing system for at least 3years.

[Sec. 114 of the Clean Air Act as amended (42U.S.C. 7414)]

(Approved by the Office of Management andBudget under control number 2060-0006.)

§ 60.506 Reconstruction.For purposes of this subpart:(a) The cost of the following

frequently replaced components of theaffected facility shall not be consideredin calculating either the "fixed capitalcost of the new components" or the"fixed capital costs that would berequired to construct a comparableentirely new facility" under § 60.15:pump seals, loading arm gaskets andswivels, coupler gaskets, overfill sensorcouplers and cables, flexible vaporhoses, and grounding cables andconnectors.

(b) Under § 60.15, the "fixed capitalcost of the new components" includesthe fixed capital cost of all depreciablecomponents [except componentsspecified in § 60.506(a)] which are orwill be replaced pursuant to allcontinuous programs of componentreplacement which are commencedwithin any 2-year period followingDecembtr 17, 1980. For purposes of thisparagraph, "commenced" means that anowner or operator has undertaken acontinuous program of componentreplacement or that an owner oroperator has entered into a contractualobligation to undertake and complete,within a reasonable time, a continuousprogram of component replacement.[Sec. 114 of the Clean Air Act as amended (42U.S.C. 7414)]

2. By adding five new ReferenceMethods (Method 2A, Method 2B,Method 25A, Method 25B, and Method27) to Appendix A as follows:

Appendix A-Reference Methods

Method 2A. Direct Measurement of GasVolume Through Pipes and Small Ducts

1. Applicability and Principle.1.1 Applicability. This method applies to

the measurement of gas flow rates in pipesand small ducts, either in-line or at exhaustpositions, within the temperature range of 0to 50'C.

1.2 Principle. A gas volume meter is usedto measure gas volume directly. Temperatureand pressure measurements are made tocorrect the volume to standard conditions.

2. Apparatus.Specifications for the apparatus are given

below. Any other apparatus that has beendemonstrated (subject to approval of theAdministrator) to be capable of meeting thespecifications will be considered acceptable.

2.1 Gas Volume Meter. A positivedisplacement meter, turbine meter, or otherdirect volume measuring device capable ofmeasuring volume to within 2 percent. The

meter shall be equipped with a temperaturegauge (± percent of the minimum absolutetemperature) and a pressure gauge (±2.5 mmHg). The manufacturer's recommendedcapacity of the meter shall be sufficient forthe expected maximum and minimum flowrates at the sampling conditions.Temperature, pressure, corrosivecharacteristics, and pipe size are factorsnecessary to consider in choosing a suitablegas meter.

2.2 Barometer. A mercury, aneroid, orother barometer capable of measuringatmospheric pressure to within 2.5 mm Hg. Inmany cases, the barometric reading may beobtained from a nearby national weatherservice station, in which case the stationvalue (which is the absolute barometricpressure) shall be requested, and anadjustment for elevation differences betweenthe weather station and the sampling pointshall be applied at a rate of minus 2.5 mm Hgper 30-meter elevation increase, or vice-versafor elevation decrease.

2.3 Stopwatch. Capable of measurementto within I second.

3. Procedure.3.1 Installation. As there are numerous

types of pipes and small ducts that may besubject to volume measurement, it would bedifficult to describe all possible installationschemes. In general, flange fittings should beused for all connections wherever possible.Gaskets or other seal materials should beused to assure leak-tight connections. Thevolume meter should be located so as toavoid severe vibrations and other factors thaimay affect the meter calibration.

3.2 Leak Test. A volume meter installedat a location under positive pressure may beleak-checked at the meter connections byusing a liquid leak detector solutioncontaining a surfactant. Apply a smallamount of the solution to the connections. If aleak exists, bubbles will form, and the leakmust be corrected.

A volume meter installed at a locationunder negative pressure is very difficult totest for leaks Without blocking flow at theinlet of the line and watching for metermovement. If this procedure is not possible,visually check all connections and assuretight seals.

3.3 Volume Measurement.3.3.1 For sources with continuous, steady

emission flow rates, record the initial metervolume reading, meter temperature(s), meterpressure, and start the stopwatch.Throughout the test period, record the metertemperature(s) and pressure so that averagevalues can be determined. At the end of thetest, stop the timer and record the elapsedtime, the final volume reading, metertemperature(s), and pressure. Record thebarometric pressure at the beginning and endof the test run. Record the data on a tablesimilar to Figure 2A-1.

BILLING CODE 6560-60-U

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

Plant

Date

Sample Location

Barometric Pressure imn Hg

Operators

Meter Number

Run ,,umber

Start Finish

Meter Calibration Coefficient

Last Date Calibrated

Figure 2A-1. Volume flow rate measurement data.

37593

I II m I I l

37594 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

3.3.2 For sources with noncontinuous,non-steady emission flow rates, use theprocedure in 3.3.1 with the addition of thefollowing: Record all the meter parametersand the start and stop times corresponding toeach process cyclical or noncontinuous event.

4. Calibration.4.1 Volume Meter. The volume meter is

calibrated against a standard reference meterprior to its initial use in the field. Thereference meter is a spirometer or liquiddisplacement meter with a capacityconsistent with that of the test meter.

Alternately, a calibrated, standard pitotmay be used as the reference meter inconjunction with a wind tunnel assembly.Attach the test meter to the wind tunnel sothat the total flow passes through the testmeter. For each calibration run, conduct a 4-point traverse along one stack diameter at aposition at least eight diameters of straighttunnel downstream and two diameters.upstream of any bend, inlet, or air mover.Determine the traverse point locations asspecified in Method 1. Calculate the referencevolume using the velocity values followingthe procedure in Method 2, the wind tunnelcross-sectional area, and the run time.

Set up the test meter in a configurationsimilar to that used in the field installation(i.e., in relation to the flow moving device).Connect the temperature and pressure gaugesas they are to be used in the field. Conncetthe reference meter at the inlet of the flowline, if appropriate for the meter, and begingas flow through the system to condition themeters. During this conditioning operation,check the system for leaks.

The calibration shall be run over at leastthree different flow rates. The calibrationflow rates shall be about 0.3, 0.6, and 0.9times the test meter's rated maximum flowrate.

For each calibration run, the data to becollected include: reference meter initial andfinal volume readings, the test meter initialand final volume reading, meter averagetemperature and pressure, barometricpressure, and run time. Repeat the runs ateach flow rate at least three times.

Calculate the test meter calibrationcoefficient, Y.,' for each run as follows:

(Vd-VJ (t,+273) PbY. =

(V,,-Vm}(tm + 273) (b + P.)

Eq. 2A-1Ym=Test volume meter calibration

coefficient, dimensionless.V,= Reference meter volume reading, m.V.=Test meter volume reading, m3

.

t,=Reference meter average temperature,*C.

tm =Test meter average temperature, °C.Pb =Barometric pressure, mm Hg.P,=Test meter average static pressure, mm

Hg.f=Final reading for run.i =Initial reading for run.

Compare the three YI values at eachof the flow rates tested and determinethe maximum and minimum values. Thedifference between the maximum andminimum values at each flow rateshould be no greater than 0.030. Extraruns may be required to complete thisrequirement. If this specification cannotbe met in six successive runs, the test

meter it not suitable tar use. In addition,the meter coefficients should bebetween 0.95 and 1.05. If thesespecifications are met at all the flowrates, average all the Ym values fromruns meeting the specifications to obtainan average meter calibration coefficient,YV.

The procedure above shall beperformed at least once for each volumemeter. Thereafter, an abbreviatedcalibration check shall be completedfollowing each field test. The calibrationof the volume meter shall be checked byperforming three calibration runs at asingle, intermediate flow rate (based onthe previous field test) with the meterpressure set at the average valueencountered in the field test. Calculatethe average value of the calibrationfactor. If the calibration has changed bymore than 5 percent, recalibrate themeter over the full range of flow asdescribed above.

Note.-lf the volume meter calibrationcoefficient values obtained before and after atest series differ by more than 5 percent, thetest series shall either be voided, orcalculations for the test series shall beperformed using whichever meter coefficientvalue (i.e., before or after) gives the greatervalue of pollutant emission rate.

4.2 Temperature Gauge. After eachtest series, check the temperature gaugeat ambient temperature. Use anAmerican Society for Testing andMaterials (ASTM) mercury-in-glassreference thermometer, or equivalent, asa reference. If the gauge being checkedagrees within 2 percent (absolutetemperature) of the reference, (hetemperature data collected in the fieldshall be considered valid. Otherwise,the test data shall be considered invalidor adjustments of the test results shallbe made, subject to the approval of theAdministrator.

4.3 Barometer. Calibrate the barometerused against a mercury barometer prior to thefield test.

5. Calculations.Carry out the calculations, retaining at

least one extra decimal figure beyond that ofthe acquired data. Round off figures after thefinal calculation.5.1 NomenclaturePb=Barometric pressure, mm Hg.P,=Average static pressure in volume meter,

mm Hg.Q,=Gas flow rate, m3/min, standard

conditions.Tm=Average absolute meter temperature, *K.V.=Meter volume reading, m .

Y= Average meter calibration coefficient,dimensionless.

f=Final reading for test period.i=Initial reading for test period.s=Standard conditions, 20* C and 760 mm

Hg.0=Elapsed test period time, min.5.2 Volume.

(Pb +- P,)V_ 0.3853 Ym (Vr-VJ

T.

Eq. 2A-2

5.3 Gas Flow Rate.Vms

0

Eq. 2A-36. Bibliography.6.1 Rom, Jerome J. Maintenance,

Calibration, and Operation of IsokineticSource Sampling Equipment. U.S.Environmental Protection Agency. ResearchTriangle Park, N.C. Publication No. APTD-0576. March 1972.

6.2 Wortman, Martin, R. Vollaro, and P.R.Westlin. Dry Gas Volume Meter Calibrations.Source Evaluation Society Newsletter. Vol. 2,No. 2. May 1977,

6.3 Westlin, P.R. and R.T. Shigehara.Procedure for Calibrating and Using Dry GasVolume Meters as Calibration Standards.Source Evaluation Society Newsletter. Vol. 3,No. 1. February 1978.

Method 2B-Determination of Exhaust GasVolume Flow Rate From Gasoline VaporIncineratorsApplicability and Principle

1.1 Applicability. This method applies tothe measurement of exhaust volume flow ratefrom incinerators that process gasolinevapors consisting primarily of alkanes,alkenes, and/or arenes (aromatichydrocarbons). It is assumed that the amountof auxiliary fuel is negligible.

1.2 Principle. The incinerator exhaustflow rate is determined by carbon balance.Organic carbon concentration and volumeflow rate are measured at the incineratorinlet. Organic carbon, carbon dioxide (CO 2),and carbon monoxide (CO) concentrationsare measured at the outlet. Then the ratio oftotal carbon at the incinerator inlet and outletis multiplied by the inlet volume to determinethe exhaust volume and volume flow rate.

2. Apparatus.2.1 Volume Meter. Equipment described

in Method 2A.2.2 Organic Analyzer (2). Equipment

described in Method 25A or 25B.2.3 CO Analyzer. Equipment described in

Method 10.2.4 CO2 Analyzer. A nondispersive

infrared [NDIR) CO2 analyzer and supportingequipment with comparable specifications asCO analyzer described in Method 10.

3. Procedure.3.1 Inlet Installation. Install a volume

meter in the vapor line to incinerator inletaccording to the procedure in Method 2A. Atthe volume meter inlet, install a sample probeas described in Method 25A. Connect to theprobe a leak-tight, heated (if necessary toprevent condensation sample line (Stainlesssteel or equivalent) and an organic analyzersystem as described in Method 25A or 25B.

3.2 Exhaust Installation. Three sampleanalyzers are required for the incineratorexhaust: CO2, CO, and organic analyzers. Asample manifold with a single sample probemay be used. Install a sample probe asdescribed Method 25A. Connect a leak-tightheated sample line to the sample probe. Heatthe sample line sufficiently to prevent anycondensation.

3.3 Recording Requirements. The outputof each analyzer must be permanentlyrecorded on an analog strip chart, digitalrecorder, or other recording device. The chartspeed or number of readings per time unit

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

must be similar for all analyzers so that datacan be correlated. The minimum datarecording requirement for each analyzer isone measurement value per minute.

3.4 Preparation. Prepare and calibrate allequipment and analyzers according to theprocedures in the respective methods. For theCO2 analyzer, follow the proceduresdescribed in Method 10 for CO analysissubstituting CO2 calibration gas where themethod calls for CO calibration gas. The spanvalue for the CO2 analyzer shall be 15 percentby volume. All calibration gases must beintroduced at the connection between theprobe and the sample line. If a manifoldsystem is used-for the exhaust analyzers, allthe analyzers and sample pumps must beoperating when the calibrations are done.Note: For the purposes of this test. methaneshould not be used as an organic calibrationgas.

3.5 Sampling. At the beginning of the testperiod, record the initial parameters for theinlet volume meter according to theprocedures in Method 2A and mark all of therecorder strip charts to indicate the start ofthe test. Continue recording inlet organic andexhaust CO, CO, and organic concentrationsthroughout the test. During periods of processinterruption and halting of gas flow, stop thetimer and mark the recorder strip charts sothat data from this interruption are notincluded in the calculations. At the end of thetest period, record the final parameters forthe inlet volume meter and mark the end onall of the recorder strip charts.

3.6 Post Test Calibrations. At theconclusion of the sampling period, introducethe calibration gases as specified in therespective reference methods. If an analyzeroutput does not meet the specifications of themethod, invalidate the test data for theperiod. Alternatively, calculate the volumeresults using initial calibration data and usingfinal calibration data and report bothresulting volumes. Then, for emissionscalculations, use the volume measurementresulting in the greatest emission rate orconcentration.

4. Calculations.Carry out the calculations, retaining at

least one extra decimal figure beyond that ofthe acquired data. Round off figures after thefinal calculation.

4.1 NomenclatureCO.=Mean carbon monoxide concentration

in system exhaust, ppmv.C02.=Mean carbon dioxide concentration in

system exhaust, ppmv.HC.=Mean organic concentration in system

exhaust as defined by the calibrationgas, ppmv,

HC, =Mean organic concentration in systeminlet as defined by the calibration gas,ppmv.

K=Calibration gas factor=2 for ethanecalibration gas.

=3 for propane calibration gas.=4 for butane calibration gas.=Appropriate response factor for other

calibration gas.V.= Exhaust gas volume, M.Va=Inlet gas volume, M.Qe.=Exhaust gas volume flow rate, m3/min.Q =Inlet gas volume flow rate, m3/min.e = Sample run time, nin.s=Standard Conditions: o0 C, 760 mm Hg.300= Estimated concentration of ambient

C0 2 , ppmv. (CO2 concentration in theambient air may be measured during thetest period using an NDIR atd the meanvalue substituted into the equation.)

4.2 Concentrations. Determine m. jnconcentration of inlet organics, outlet COQ!,outlet CO, and outlet organics according tothe procedures in the respective methods andthe analyzers' calibration curves, and for thetime intervals specified in the applicableregulations. Concentrations should bedetermined on a parts per million by volume(ppmv) basis.

4.3 Exhaust Gas Volume. Calculate theexhaust gas volume as follows:

V_ K(HC)

KfHC) + COl

Eq. 2B-14.4 Exhaust Gas Volume Flow Rate.

Calculate the exhaust gas volume flow rateas follows:

V_Q, =-

Eq. 2B-25. Bibliography.5.1 Measurement of Volatile Organic

Compounds. U.S. Environmental ProtectionAgency. Office of Air Quality Planning andStandards. Research Triangle Park, N.C.27711. Publication No. EPA-450/2-T--041.October 1978. p. 55.

Method 25A-Determination of TotalGaseous Organic Concentration Using aFlame Ionization Analyzer

1. Applicability and Principle.1.1 Applicability. This method applies to

the measurement of total gaseous organicconcentration of vapors consisting primarilyof alkanes, alkenes, and/or arenes (aromatichydrocarbons). The concentration isexpressed in terms of propane (or otherappropriate organic calibration gas) or interms of carbon.

1.2 Principle. A gas sample is extractedfrom the source through a heated sample line,if necessary, and glass fiber filter to a flameionization analyzer (FIA). Results arereported as volume concentration equivalentsof the calibration gas or as carbonequivalents.

2. Definitions.2.1 Measurement System. The total

equipment required for the determination ofthe gas concentration. The system consists ofthe following major subsystems:

2.1.1 Sample Interface. That portion of thesystem that is used for one or more of thefollowing: sample acquisition, sampletransportation, sample conditioning, orprotection of the analyzer from the effects ofthe stack effluent.

2.1.2 Organic Analyzer. Thatportion of'the system that senses organic concentrationand generates an output proportional to thegas concentration.

2.2 Span Value. The upper limit of a gasconcentration measurement range that isspecified for affected source categories in theapplicable part of the regulations. The spanvalue is established in the applicableregulation and is usually 1.5 to 2.5 times theapplicable emission limit. If no span value isprovided, use a span value equivalent to 1.5to 2.5 times the expected concentration. Forconvenience, the span value shouldcorrespond to 100 percent of the recorderscale.

2.3 Calibration Gas. A knownconcentration of a gas in an appropriatediluent gas.

2.4 Zero Drift. The difference in themeasurement system response to a zero levelcalibration gas before and after a statedperiod of operation during which nounscheduled maintenance, repair, oradjustment took place.

2.5 Calibration Drift. The difference in themeasurement system response to a mid-levelcalibration gas before and after a statedperiod of operation during which nounscheduled maintenance, repair oradjustment took place.

2.6 Response Time. The time intervalfrom a step change in pollutant concentrationat the inlet to the emission measurementsystem to the time at which 95 percent of thecorresponding final value is reached asdisplayed on the recorder.

2.7 Calibration Error. The differencebetween the gas concentration Indicated bythe measurement system and the knownconcentration of the calibration gas.

3. Apparatus.A schematic of an acceptable measurement

system is shown in Figure 25A-1. Theessential components of the measurementsystem are described below:

HEATEb

PARTICULATEFILTER

SAMPL EPUMP

STACK

Fisre 25A 1. Otoani Concentralion Measarement Sysliem.

. ucpu379115

37596 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 1 Rules and Regulations

3.1 Organic Concentration Analyzer. Aflame ionization analyzer (FIA) capable ofmeeting or exceeding the specifications inthis method.

3.2 Sample Probe. Stainless steel, orequivalent, three-hole rake type. Sampleholes shall be 4 nn in diameter or smallerand located at 16.7, 50, and 83.3 percent of theequivalent stack diameter. Alternatively, asingle opening probe may be used so that agas sample is collected from the centrallylocated 10 percent area of the stack cross-section.

3.3 Sample Line. Stainless steel or Teflon*tubing to transport the sample gas to theanalyzers. The sample line should be heated,if necessary, to prevent condensation in theline.

3.4 Calibration Valve Assembly. A three-way valve assembly to direct the zero andcalibration gases to the analyzers isrecommended. Other methods, such as quick-connect lines, to route calibration gas to theanalyzers are applicable.

3.5 Particulate Filter. An in-stack or anout-of-stack glass fiber filter is recommendedif exhaust gas particulate loading issignificant. An out-of-stack filter should beheated to prevent any condensation.

3.6 Recorder. A strip-chart recorder,analog computer, or digital recorder forrecording measurement data. The minimumdata recording requirement is onemeasurement value per minute. Note: Thismethod is often applied in highly explosiveareas. Caution and care should be exercisedin choice of equipment and installation.

4. Calibration and Other Gases.Gases used for calibrations, fuel, and

combustion air (if required] are contained incompressed gas cylinders. Preparation ofcalibration gases shall be done according tothe procedure in Protocol No. 1, listed inReference 9.2. Additionally, the manufacturerof the cylinder should provide arecommended shelf life for each calibrationgas cylinder over which the concentrationdoes not change more than #2 percent fromthe certified value. For calibration gas valuesnot generally available (i.e., organics.between 1 and 10 percent by volume),alternative methods for preparing calibrationgas mixtures, such as dilution systems, maybe used with prior approval of theAdministrator.

Calibration gases usually consist ofpropane in air or nitrogen and are determinedin terms of the span value. Organiccompounds other than propane can be usedfollowing the above guidelines and makingthe appropriate corrections for responsefactor.

4.1 Fuel. A 40 percent H./60 percent He or40 percent H 2/60 percent N2gas mixture isrecommended to avoid an oxygen synergismeffect that reportedly occurs when oxygenconcentration varies significantly from amean value.

4.2 Zero Gas. High purity air with lessthan 0.1 parts per million by volume (ppmv)of organic material (propane or carbon

Mention of trade names or specific productsJoes not constitute endorsement by theEnvironmental Protection Agency.

equivalent] or less than 0.1 percent of thespan value, whichever is greater.

4.3 Low-level Calibration Gas. An organiccalibration gas with a concentrationequivalent to 25 to 35 percent of theapplicable span value.

4.4 Mid-level Calibration Gas. An organiccalibration gas with a concentrationequivalent to 45 to 55 percent of theapplicable span value.

4.5 High-level Calibration Gas. Anorganic calibration gas with a concentrationequivalent to 80 to 90 percent of theapplicable span value.

5. Measurement System PerformanceSpecifications.

5.1 Zero Drifl Less than ±3 percent ofthe span value.

5.2 Calibration Drift. Less than ±3percent of span value.

5.3 Calibration Error. Less than ±5percent of the calibration gas value.

6. Pretest Preparations.6.1 Selection of Sampling Site. The

location of the sampling site is generallyspecified by the applicable regulation orpurpose of the test; i.e., exhaust stack, inletline, etc. The sample port shall be located atleast 1.5 meters or 2 equivalent diameters(whichever is less) upstream of the gasdischarge to the atmosphere.

6.2 Location of Sample Probe. Install thesample probe so that the probe is centrallylocated in the stack, pipe, or duct and is •sealed tightly at the stack port connection.

6.3 Afeasurement System Preparation.Prior to the emission test, assemble themeasurement system following themanufacturer's written instructions inpreparing the sample interface and theorganic analyzer. Make the system operable.

FIA equipment can be calibrated for almostany range of total organics concentrations.For high concentrations of organics (>1.0percent by volume as propene) modificationsto most commonly available analyzers arenecessary. One accepted method ofequipment modification is to decrease thesize of the sample to the analyzer through theuse of a smaller diameter sample capillary.Direct and continuous measurement oforganic concentration is a necessaryconsideration when determining anymodification design.

6.4 Calibration Error Test. Immediatelyprior to the test series, [within 2 hours of thestart of the test] introduce zero gas and high-level calibration gas at the calibration valveassembly. Adjust the analyzer output to theappropriate levels, if necessary. Calculate thepredicted response for the low-level and mid-level gases based on a linear response linebetween the zero and high-level responses.Then introduce low-level and mid-levelcalibration gases successively to themeasurement system. Record the analyzerresponses for low-level and mid-levelcalibration gases and determine thedifferences between the measurement systemresponses and the predicted responses. Thesedifferences must be less than 5 percent of therespective calibration gas value. If not, themeasurement system is not acceptable andmust be replaced or repaired prior to testing.No adjustments to the measurement systemshall be conducted after the calibration and

before the drift check [Section 7.3). Ifadjustments are necessary before thecompletion of the test series, perform the driftchecks prior to the required adjustments andrepeat the calibration following theadjustments. If multiple electronic ranges areto be used, each additional range must bechecked with a mid-level calibration gas toverify the multiplication factor.

6.5 Response Time TesL Introduce zerogas into the measurement system at thecalibration valve assembly. When the systemoutput has stabilized, switch quickly to thehigh-level calibration gas. Record the timefrom the concentration change to themeasurement system response equivalent to95 percent of the step change. Repeat the testthree times and average the results.

7. Emission Measurement Test7.1 Orgaic Measurement. Begin sampling

at the start of the test period, recording timeand any required process information asappropriate. In particular, note on therecording chart periods of processinterruption or cyclic operation.

7.2 Drift Determination. Immediatelyfollowing the completion of the test periodand hourly during the test period, reintroducethe zero and mid-level calibration gases, oneat a time, to the measurement system at thecalibration valve assembly. (Make noadjustments to the measurement system untilafter both the zero and calibration driftchecks are made.) Record the analyzerresponse. If the drift values exceed thespecified limits, invalidate the test resultspreceding the check and repeat the testfollowing corrections to the measurementsystem. Alternatively, recalibrate the testmeasurement system as in Section 6.4 andreport the results using both sets ofcalibration data (i.e., data determined prior tothe test period and data determined followingthe test period].

8. Organic Concentration Calculations.Determine the average organic

concentration in terms of ppmv as propane orother calibration gas. The average shall bedetermined by the integration of the outputrecording over the period specified in theapplicable regulation.

If results are required in terms of ppmv ascarbon, adjust measured concentrations usingEquation 25A-1.

C,=K C.

Eq. 25A-1Where:C,=Organic concentration as carbon, ppmv.C..=Organic concentration as measured,

ppmv.K= Carbon equivalent correction factor,

K=2 for ethane.K=3 for propane.K=4 for butane.K=Appropriate response factor for other

organic calibration gases.9. Bibliography.9.1 Measurement of Volatile Organic

Compounds-Guideline Series. U.S.Environmental Protection Agency. ResearchTriangle Park, N.C. Publication No. EPA-450/2-78-041. June 1978. p. 46-54.

9.2 Traceability Protocol for EstablishingTrue Concentrations of Gases Used for

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

Calibration and Audits of Continuous SourceEmission Monitors (Protocol No. 1). U.S.Environmental Protection Agency,Environmental Monitoring and SupportLaboratory. Research Triangle Park, N.C.June 1978.

9.3 Gasoline Vapor Emission LaboratoryEvaluation-Part 2. U.S. EnvironmentalProtection Agency, Office of Air QualityPlanning and Standards. Research TrianglePark, N.C. EMB Report No. 75-GAS--6. August1975.

Method 25B-Determination of TotalGaseous Organic Concentration Using aNondispersive Infrared Analyzer

1. Applicability and Principle.1.1 Applicability. This method applies to

the measurement of total gaseous organicconcentration of vapors consisting primarilyof alkanes. (Other organic materials may bemeasured using the general procedure in thismethod, the appropriate calibration gas, andan analyzer set to the appropriate absorptionband.) The concentration is expressed interms of propane (or other appropriateorganic calibration gas) or in terms of carbon.

1.2 Principle. A gas sample is extractedfrom the source through a heated sample line,if necessary, and glass fiber filter to anondispersive infrared analyzer (NDIR).Results are reported as volume concentrationequivalents of the calibration gas or ascarbon equivalents.

2. Definitions.The terms and definitions are the same as

for Method 25A.3. Apparatus. The apparatus are the same

as for Method 25A with the exception of thefollowing:

3.1 Organic Concentration Analyzer. Anondispersive infrared analyzer designed tomeasure alkane organics and capable ofmeeting or exceeding the specifications inthis method.

4. Calibration Gases.The calibration gases are the same as are

required for Method 25A, Section 4. No fuelgas is required for an NDIR.

5. Measurement System PerformanceSpecifications.

5.1 Zero Drift. Less than L3 percent ofthe span value.

5.2 Calibration Drift. Less than h3percent of the span value.

5.3 Calibration Error. Less than ±:5percent of the calibration gas valve.

& Pretest Preparations,8.1 Selection of Sampling Site. Same as in

Method 25A, Section 6.1.62 Location of Sampling Probe. Same as

in Method 25A, Section 6.2.6.3 Measurement System Preparation.

Prior to the emission test, assemble themeasurement system following themanufacturer's written instructions inpreparing the sample interface and theorganic analyzer. Make the system operable.

6.4 Calibration Error Test. Same as inMethod 25A. Section 6.4.

6.5 Response Time Test Procedure. Sameas in Method 25A, Section 6.4.

7. Emission Measurement Test Procedure.Proceed with the emission measurement

immediately upon satisfactory completion ofthe calibration.

7.1 Organic Measurement. Same as inMethod 25A, Section 7.1.

7.2 Drift Determination. Same as inMethod 25A, Section 7.2.

8. Organic Concentration Cut It.'iirs.The calculations are the same as in Method

25A, Section 8.9. Bibliography.The bibliography is the same as in Method

25A, Section 9.

Method 27-Determination of VaporTightness of Gasoline Delivery Tank UsingPressure-Vacuum Test

1. Applicability and Principle.1.1 Applicability. This method is

applicable for the determination of vaportightness of a gasoline delivery tank which isequipped with vapor collection equipment.

1.2 Principle. Pressure and vacuum areapplied alternately to the compartments of agasoline delivery tank and the change inpressure or vacuum is recorded after aspecified period of time.

2. Definitions and Nomenclature.2.1 Gasoline. Any petroleum distillate or

petroleum distillate/alcohol blend having aReid vapor pressure of 27.6 kilopascals orgreater which is used as a fuel for internalcombustion engines.

2.2 Delivery tank. Any container,including associated pipes and fittings, that Isattached to or forms a part of any truck,trailer, or railcar used for the transport ofgasoline.

2.3 Compartment A liquid-tight divisionof a delivery tank.

2.4 Delivery tank vapor collectionequipment. Any piping, hoses, and devices onthe delivery tank used to collect and routegasoline vapors either from the tank to a bulkterminal vapor control system or from a bulkplant or service station into the tank.

2.5 Time period of the pressure orvacuum test (t). The time period of the test, asspecified in the appropriate regulation, duringwhich the change in pressure or vacuum ismonitored, in minutes.

2.6 Initial pressure (P). The pressureapplied to the delivery tank at the beginningof the static pressure test, as specified in theappropriate regulation, in nun H,O.

2.7 Initial vacuum (Vj. The vacuumapplied to the delivery tank at the beginningof the static vacuum test, as specified in theappropriate regulation, in num HO.

2.8 Allowable pressure change (Ap). Theallowable amount of decrease in pressureduring the static pressure test, within the timeperiod t, as specified in the appropriateregulation, in mim HO.

2.9 Allowable vacuum change (Av). Theallowable amount of decrease in vacuumduring the static vacuum test, within the timeperiod t, as specified in the appropriateregulation, in mm H.O.

3. Apparatus.3.1 Pressure source. Pump or compressed

gas cylinder of air or inert gas sufficient topressurize the delivery tank to 500 mm HOabove atmospheric pressure.

3.2 Regulator. Low pressure regulator forcontrolling pressurization of the deliverytank.

3.3 Vacuum source. Vacuum pumpcapable of evacuating the delivery tank to250 mm H,O below atmospheric pressure.

3.4 Pressure-vacuum supply hose.3.5 Manometer. Liquid manometer, or

equivalent instrument, capable of measuringup to 500 mm H2O gauge pressure with ±E2.5mm H.0 precision.

3.6 Pressure-vacuum relief valves. Thetest apparatus shall be equipped with an in-line pressure-vacuum relief valve set toactivate at 675 mm H O above atmosphericpressure or 250 mm H.0 below atmosphericpressure, with a capacity equal to thepressurizing or evacuating pumps.

3.7 Test cap for vapor recovery hose. Thiscap shall have a tap for manometerconnection and a fitting with shut-off valvefor connection to the pressure-vacuum supplyhose.

3.8 Caps for liquid delivery hoses.4. Pretest Preparations.4.1 Summary. Testing problems may

occur due to the presence of volatile vaporsand/or temperature fluctuations inside thedelivery tank. Under these conditions, it isoften difficult to obtain a stable initialpressure at the beginning of a test, anderroneous test results may occur. To helpprevent this, it is recommended that, prior totesting, volatile vapors be removed from thetank and the temperature inside the tank beallowed to stabilize. Because it is not alwayspossible to attain completely these pretestconditions a provision to ensure reproducibleresults is included. The difference in resultsfor two consecutive runs must meet thecriterion in Sections 5.2.5 and 5.3.5.

4.2 Emptying of tank. The delivery tankshall be emptied of all liquid.

4.3 Purging of vapor. As much as possible,the delivery tank shall be purged of allvolatile vapors by any safe, acceptablemethod. One method is to carry a load ofnon-volatile liquid fuel, such as diesel orheating oil, immediately prior to the test, thusflushing out all the volatile gasoline vapors. Asecond method is to remove the volatilevapors by blowing ambient air into each tankcompartment for at least 20 minutes. Thissecond method is usually not as effective andoften causes stabilization problems, requiringa much longer time for stabilization duringthe testing.

4.4 Temperature stabilization. As muchas possible, the test shall be conducted underisothermal conditions. The temperature of thedelivery tank should be allowed toequilibrate in the test environment. Duringthe test, the tank should be protected fromextreme environmental and temperaturevariability, such as direct sunlight.

5. Test Procedure.5.1 Preparations.5.1.1. Open and close each dome cover.5.1.2 Connect static electrical ground

connections to tank. Attach the liquiddelivery and vapor return hoses, remove theliquid delivery elbows, and plug the liquiddelivery fittings.

(Note.-The purpose of testing the liquiddelivery hoses is to detect tears or holes thatwould allow liquid leakage during a delivery.Liquid delivery hoses are not considered tobe possible sources of vapor leakage, andthus, do not have to be attached for a vaporleakage test. Instead, a liquid delivery hose

37597

37598 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

could be either visually inspected, or filledwith water to detect any liquid leakage.]

5.1.3 Attach the test cap to the end of thevapor recovery hose.

5.1.4 Connect the pressure-vacuum supplyhose and the pressure-vacuum relief valve tothe shut-off valve. Attach a manometer to thepressure tap.

5.1.5 Connect compartments of the tankinternally to each other if possible. If notpossible, each compartment must be testedseparately, as if it were an individualdelivery tank.

5.2 Pressure test.5.2.1 Connect the pressure source to the

pressure-vacuum supply hose.5:2.2 Open the shut-off valve in the vapor

recovery hose cap. Applying air pressureslowly, pressurize the tank to P1, the initialpressure specified in the regulation.

5.2.3 Close the shut-off valve and allowthe pressure in the tank to stabilize, adjustingthe pressure if necessary to maintainpressure of P1. When the pressure stabilizes,record the time and initial pressure.

5.2.4 At the end of t minutes, record thetime and final pressure.

5.2.5 Repeat steps 5.2.2 through 5.2.4 untilthe change in pressure for two consecutiveruns agrees within ±12.5 mm H.O. Calculatethe arithmetic average of the two results.

5.2.6 Compare the average measuredchange in pressure to the allowable pressurechange, Ap, as specified in the regulation. Ifthe delivery tank does not satisfy the vaportightness criterion specified in the regulation,repair the sources of leakage, and repeat thepressure test until the criterion is met.

5.2.7 Disconnect the pressure source fromthe pressure-vacuum supply hose, and slowlyopen the shut-off valve to bring the tank toatmospheric pressure.

5.3 Vacuum test.5.3.1 Connect the vacuum source to the

pressure-vacuum supply hose.5.3.2 Open the shut-off valve in the vapor

recovery hose cap. Slowly evacuate the tankto Vi, the initial vacuum specified in theregulation.

5.3.3 Close the shut-off valve and allowthe pressure in the tank to stabilize, adjustingthe pressure if necessary to maintain avacuum of V. When the pressure stabilizes,record the time and initial vacuum.

5.3.4 At the end of t minutes, record thetime and final vacuum.

5.3.5 Repeat steps 5.3.2 through 5.3.4 untilthe change in vacuum for two consecutiveruns agrees within 012.5 mm H1O. Calculatethe arithmetic average of the two results.

5.3.6 Compare the average measuredchange in vacuum to the allowable vacuumchange, Ap, as specified in the regulation. If.the delivery tank does not satisfy the vaportightness criterion specified in the regulation,repair the sources of leakage, and repeat thevacuum test until the criterion is met.

5.3.7 Disconnect the vacuum source fromthe pressure-vacuum supply hose, and slowlyopen the shut-off valve to bring the tank toatmospheric pressure.

5.4 Post-test clean-up. Disconnect all testequipment and return the delivery tank to itspretest condition.

6. Altemnative Procedures.6.1 The pumping of water into the bottom

of a delivery tank is an acceptable

alternative to the pressure source describedabove. Likewise, the draining of water out ofthe bottom of a delivery tank may besubstituted for the vacuum source. Note thatsome of the specific step-by-step proceduresin the method must be altered slightly toaccommodate these different pressure andvacuum sources.

6.2. Techniques other than specified abovemay be used for purging and pressurizing adelivery tank, if prior approval is obtainedfrom the Administrator. Such approval willbe based upon demonstrated equivalencywith the above method.IFR Ooc. 83-22380 Filed 8-17-3: 8:45 aml

BILLING CODE 6560-5O-U

40 CFR Part 60

[AD-FRL-2241-6a]

Addition of Reference Method 21 toAppendix A

AGENCY: Environmental ProtectionAgency (EPA).ACTION: Final rule.

SUMMARY: This action establishes a newreference method to be added toAppendix A of 40 CFR Part 60,standards of performance for newstationary sources. Reference Method 21will be used to determine volatileorganic compound (VOC) leaks fromprocess equipment such as valves,flanges and other connections, pumpand compressor seals, pressure reliefdevices, process drains, open-endedvalves, pump and compressor sealsystem degassing vents, accumulatorvessel vents, agitator seals, and accessdoor seals. This reference method willbe used in several air pollutionregulations for the limitation of fugitiveVOC emissions which are beingdeveloped for proposal andpromulgation.EFFECTIVE DATE: August 18, 1983.ADDRESSES: Docket. A docket, numberA-79-32, containing informationconsidered by EPA in development ofstandards of performance for fugitiveemission sources in the syntheticorganic chemical manufacturingindustry, and which also containsinformation considered in developmentof the promulgated reference method, isavailable for public inspection andcopying between 8:00 a.m. and 4:00 p.m.,Monday through Friday, at EPA'sCentral Docket Section (A-130), WestTower Lobby, Gallery 1, 401, M Street,SW., Washington, D.C. 20460. Areasonable fee may be charged forcopying.FOR FURTHER INFORMATION CONTACT:Mr. Winton Kelly, EmissionMeasurement Branch, EmissionStandards and Engineering Division

(MD-13), U.S. Environmental ProtectionAgency, Research Triangle Park, NorthCarolina 27711, telephone(919) 541-5543.SUPPLEMENTARY INFORMATION:

Summary of the Reference Method

Reference Method 21, "Determinationof Volatile Organic Compound Leaks" isused to detect VOC leaks fromindividual sources of fugitive emissions.This procedure is used to identify andclassify leaks only, and is not to be usedas a direct measure of mass emissionrates from individual sources. Aportable instrument is used to measurethe local organics concentration at thesurface of a potential leak source. If ameter reading equal to or greater than alimit specified in an applicableregulation is obtained, a VOC emission(leak) exists. The procedure can also beused to confirm that "no detectableemissions" are present. If the measureddifference between the local ambientconcentration and the concentrationpresent at the surface of the potentialleak source is less than a concentrationspecified in an applicable regulation,then there are no detectable emissions.

Background

On January 5, 1981, as an appendix tothe proposed standards of performancefor fugitive emission sources in thesynthetic organic chemicalmanufacturing industry, EPA proposedReference Method 21. This methodwould normally be promulgated withthose standards. However, the methodis being promulgated earlier becauseseveral additional regulations are beingdeveloped for promulgation in the nearfuture that specify that ReferenceMethod 21 be used. This earlypromulgation will ensure that thereference procedure will be promulgatedprior to being specified in promulgatedstandards of performance.

Under Executive Order 12291, EPAmust judge whether a regulation is"major" and, therefore, subject to therequirement of a regulatory impactanalysis. This regulation is not majorbecause it will not have an annual effecton the economy of $100 million or more;it will not result in a major increase incosts or prices; and there will be nosignificant adverse effects oncompetition, employment, investment,productivity, innovation, or on theability of U.S.-based enterprise tocompete with foreign-based enterprisesin domestic or export markets.

Pursuant to the provisions of 5 U.S.C.605(b), I hereby certify that the attachedrule will not have a significant economic

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

impact on a substantial number of smallentities.

Public ParticipationDuring the development of the test

method, trade and professionalassociations and individual companiessupplied comments on the methods.After proposal on January 5, 1981,comments were received from varioussources. A public hearing was held onMarch 3, 1981, to receive additionalformal comments. The formal commentperiod was extended from April 6, 1981.to July 31, 1981.

Public Comments and Changes Made tothe Proposed Reference Method

Numerous comments were received inresponse to proposal of the standards ofperformance for fugitive emissions fromsynthetic organic chemical industry.Most of the comments concerning thetest method were specifically related tothe selection of the definition of a leak,and not to the procedure alone. Thesecomments have been carefullyconsidered and, where determined to beappropriate, changes have been made inthe proposed test method. A detaileddiscussion of the comments and fullresponses will be included in Docket No.A-79-32.

The following discussion summarizesthe changes made to the referencemethod based on additional review andin response to public comment.

The promulgated reference methodhas been reorganized to improve theclarity of the description of theprocedures. A "Definitions" section hasbeen added to place all the definitionsin one section.

A change in the requirements forinstrument performance evaluation wasalso made to improve the quality controlof the procedure. Instead of requiring acalibration precision test consisting ofnine repetitions at 6-month intervals, thenew requirements specify a testconsisting of three repetitions at 3-month intervals. This change willprovide a quality control result everyquarter and will require less effort.

The definition of "no detectableemissions" has been changed to beconsistent with the instrumentspecification of scale readability. Theproposed procedure defined "nodetectable emissions" as 2 percent ofthe leak definition concentration, with aminimum scale readability of 5 percentof the leak definition. The definition of"no detectable emissions" has beenchanged to correspond to the minimumreadability specification and will bespecified in applicable regulations.

Several commenters noted that theinstruments used during screening

studies responded differently fordifferent chemicals. One commenterstated that the actual response factorwas poorly related to the theoreticalresponse factor and cited inconistentresponses for nonane and decane, aswell as no response for some chemicals,to support his claims. Anothercommenter suggested that the leakconcentration for the standards shouldvary according to the process unit sincea wide variability (0 to 571) in responsefactors has been determined for theindustry. And, another commenterstated that aromatic compounds such asbenzene, toluene, and xylenedemonstrate a nonlinear response closeto 10,000 ppmv. In response to thesecomments, Reference Method 21 givesspecifications for the instrument to beused in monitoring fugitive VOCemission sources. The technique isintended to classify leaks only, not toprovide a rigorous analyticalconcentration or mass emission rate ofVOC. A specific statement has beenadded to Method.,+?l to clarify theintention to classify leaks only. Thevariation in response factor due tocompound or instrument is not expectedto affect significantly the number ofleaks determined through screeningbecause screening values are usuallymuch greater than the leak definition forleaks and much less than the leakdefinition for nonleaks. Two industrycommenters concur with EPA in thisposition. However, to remove some ofthe wide variability, a definition.specification, and test procedure forresponse factors have been added toMethod 21. This specification willassure that the analyzer used willrespond to the compounds to bemeasured.

Another commenter suggested that thegas specification section be amended toinclude a turnover of calibration gasstandards every 3 months sincecalibration gases can deterioratesignificantly over time. A provision hasbeen added to the promulgatedReference Method 21 to require a shelf-life specification on calibration gasesand procedures to follow to ensure thatcalibration gas concentrations areaccurate.

Two comments concerned theinstrumentation requirements ofReference Method 21. The commenterstated that only two instruments on themarket today could be considered, andneither one would meet thespecifications of the reference methodentirely: the first instrument fails thecalibration accuracy, and the sectioninstrument does not meet the responsetime requirement. In response, althoughthere are only two instruments which

have been used to any great extent, thetechnical literature and productinformation suggest that there are otherswhich could be used for detecting leaks.The specifications included in theproposed reference method areachievable based on performance duringEPA studies.

One comment letter expressedconcern that no provision was made forthe use of new instruments orcalibration procedures which wouldprovide equivalent or more accurateresults. They asked that equivalencyprovisions be added for test methodsand procedures. In response, ReferenceMethod 21 gives specifications for themonitoring instrument that are generalenough so as not to preclude newanalytical developments. In addition,the General Provisions (40 CFR Part 60,Subpart A) allow for equivalent methodsand procedures to be used forperformance testing and monitoringwhen the results of the equivalentmethod have been demonstrated to beat least as accurate as results obtainedby the required methods.

One commenter suggested that use ofa windscreen upwind of the componentbeing screened would preventmeteorological effects on the instrumentreadings. During EPA studies, theselection of a measurement location atthe surface of the source was made tominimize meteorological effects. Duringthe data collection efforts, no furtherprovisions were found necessary toobtain repeatable screening values.Therefore, all of the field data werecollected without a windscreen. In viewof these facts, it seems unnecessary torequire that a windscreen be used.

An alternative screening procedurehas been added for those sources thatcan be tested with a soap solution.These sources are restricted to thosewith non-moving seals, moderatesurface temperatures, without largeopenings to atmosphere, and withoutevidence of liquid leakage. The soapsolution is sprayed on all applicablesources and the potential leak sites areobserved to determine if bubbles areformed. If no bubbles are formed, thenno detectable emissions or leaks exist. Ifany bubbles are formed, then theinstrument measurement techniquesmust be used to determine if a leakexists, or if no detectable emissionsexist, as applicable.

The alternative soap solutionprocedure does not apply to pump seals.sources with surface temperaturesgreater than the boiling point or lessthan the freezing point of the soapsolution, sources such as open-endedlines or valves, pressure relief value

37599

37600 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

horns, vents with large openings toatmosphere, and any source whereliquid leakage is present. The instrumenttechnique in the method must be usedfor these sources.

The alternative of establishing a soapscoring leak definition equivalent to aconcentration-based leak definition isnot included in the method and is notrecommended for inclusion in anapplicable regulation because of thedifficulty of calibrating and normalizinga scoring technique based on bubbleformation rates. A scoring techniquewould be based on estimated ranges ofvolumetric leak rates. These estimatesdepend on the bubble size andformation rate, which are subjectivejudgments of an observer. Thesesubjective judgments could only becalibrated or normalized by requiringthat the'observers correctly identify andscore a standard series of test bubbles.It has been reported that trainedobservers can correctly and repeatablyclassify ranges of volumetric leak rates.However, because soap scoring requiressubjective observations and since anobjective concentration measurementprocedure is available, a soap scoringequivalent leak definition is notrecommended for the applicableregulation. The alternate procedure thathas been included will allow more rapididentification of potential leaks for morerigorous instrumental concentrationmeasurement.

Miscellaneous

This final rulemaking is issued underthe authority of Sections 111, 114, and301(a) of the Clean Air Act, as amended[42 U.S.C. 7411, 7414, and 7691(a)].

List of Subjects in 40 CFR Part 60

Air pollution control, Aluminum,Ammonium sulfate plants, Asphalt,Cement industry, Coal copper, Electricpower plants, Glass and glass products,Grains, Intergovernmental relations,Iron, Lead, Metals, Metallic minerals,Motor vehicles, Nitric acid plants, Paperand paper products industry, Petroleum,Phosphate, Sewage disposal, Steel,Sulfuric acid plants, Waste treatmentand disposal, Zinc, Tires.

Dated: August 4, 1983.William D. Ruckelsbaus,Administrator.

Appendix A of 40 CFR Part 60 is amendedby adding Reference Method 21 as follows:

Appendix A-Reference Methods

Method 21. Determination of Volatile OrganicCompounds Leaks

1. Applicability and Principle.

1.1 Applicability. This method applies tothe determination of volatile organiccompound (VOC) leaks from'processequipment. These sources include, but are notlimited to, valves, flanges and otherconnections, pumps and compressors,pressure relief devices, process drains, open-ended valves, pump and compressor sealsystem degassing vents, accumulator vesselvents, agitator seals, and access door seals.

1.2 Principle. A portable instrument isused to detect VOC leaks from individualsources. The instrument detector type is notspecified, but it must meet the specificationsand performance criteria contained in Section3. A leak definition concentration based on areference compound is specified in eachapplicable regulation. This procedure isintended to locate and classify leaks only,and is not to be used as a direct measure ofmass emission rates from individual sources.

2. Definitions.2.1 Leak Definition Concentration. The

local VOC concentration at the surface of aleak source that indicates that a VOCemission (leak) is present. The leak definitionis an instrument meter reading based on areference compound.

2.2 Reference Compound. The VOCspecies selected as an instrument calibrationbasis for specification of the leak definitionconcentration. (For example:lf a leakdefinition concentration is 10,000 ppmv asmethane, then any source emission thatresults in a local concentration that yields ameter reading of 10,000 on an instrumentcalibrated with methane would be classifiedas a leak. In this example, the leak definitionis 10,000 ppmv, and the reference compoundIs methane.)

2.3 Calibration Gas. The VOC compoundused to adjust the instrument meter readingto a known value. The calibration gas isusually the reference compound at aconcentration approximately equal to theleak definition concentration.

2.4 No Detectable Emission. The localVOC concentration at the surface of a leaksource that indicates that a VOC emission(leak) is not present. Since background VOCconcentrations may exist, and to account forinstrument drift and imperfectreproducibility, a difference between thesource surface concentration and the localambient concentration is determined. Adifference based on meter readings of lessthan a concentration corresponding to theminimum readability specification indicatesthat a VOC emission (leak) is not present.(For example, if the leak definition in aregulation is 10,000 ppmv, then the allowableincrease in surface concentration versus localambient concentration would be 503 ppmvbased on the instrument meter readings.)

2.5 Response Factor. The ratio of theknown concentration of a VOC compound tothe observed meter reading when measuredusing an instrument calibrated with thereference compound specified in theapplication regulation.

2.6 Calibration Precision. The degree ofagreement between measurements of thesame known value, expressed as the relativepercentage of the average difference betweenthe meter readings and the knownooncentration to the known concentration.

2.7 Response Time. The time intervalfrom a step change in VOC concentration atthe input of the sampling system to the timeat which 90 percent of the corresponding finalvalue is reached as displayed on theinstrument readout meter.

3. Apparatus.3.1 Monitoring Instrument.3.1.1 Specifications.a. The VOC instrument detector shall

respond to the compounds being processed.Detector types which may meet thisrequirement include, but are not limited to,catalytic oxidation, flame ionization, infraredabsorption, and photoionization.

b. The instrument shall be capable ofmeasuring the leak definition concentrationspecified in the regulation.

c. The scale of the instrument meter shallbe readable to 5 percent of the specified leakdefinition concentration.

d. The instrument shall be equipped with apump sQ that a continuous sample is providedto the detector. The nominal sample flow rateshall be to 3 liters per minute.

e. The instrument shall be intrinsically safefor operation in explosive atmospheres asdefined by the applicable U.S.A. standards(e.g., National Electrical Code by the NationalFire Prevention Association).

3.1.2 Performance Criteria.a. The instrument response factors for the

individal compounds to be measured must beless than 10.

b. The instrument response time must beequal to or less than 30 seconds. Theresponse time must be determined for theinstrument configuration to be used duringtesting.

c. The calibration precision must be equalto or less than 10 percent of the calibrationgas value.

d. The evaluation procedure for eachparameter is given in Section 4.4.

3.1.3 Performance EvaluationRequirements.

a. A response factor must be determinedfor each compound that is to be measured,either by testing or from reference sources.The response factor tests are required beforeplacing the analyzer into service, but do nothave to be repeated as subsequent intervals.

b. The calibration precision test must becompleted prior to placing the analyzer intoservice, and at subsequent 3-month intervalsor at the next use whichever is later.

c. The response time test is required priorto placing the instrument into service. If amodification to the sample pumping systemor flow configuration is made that wouldchange the response time, a new test isrequired prior to further use.

3.2 Calibration Gases. The monitoringinstrument is calibrated in terms of parts permillion by volume (ppmv) of the referencecompound specified in the applicableregulation. The calibration gases required formonitoring and instrument performanceevaluation are a zero gas (air, less than 10ppmv VOC) and a calibration gas in airmixture approximately equal to the leakdefinition specified in the regulation. Ifcylinder calibration as mixture are used, theymust be analyzed and certified by themanufacturer to be within :L2 percent

Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

accuracy, and a shelf life must be specified.Cylinder standards must be either reanalyzedor replaced at the end of the specified shelflife. Alternately, calibration gases may beprepared by the user according to anyaccepted gaseous standards preparationprocedure that will yield a mixture accurateto within ±2 percent. Prepared standardsmust be replaced each day of use unless itcan be demonstrated that degradation doesnot occur during storage.

Calibrations may be performed using acompound other than the referencecompound if a conversion factor isdetermined for that alternative compound sothat the resulting meter readings duringsource surveys can be converted to referencecompound results.

4. Procedures.4.1 Pretest Preparations. Perform the

instrument evaluation procedures given inSection 4.4 if the evaluation requirements ofSection 3.1.3 have not been met.

4.2 Calibration Procedures. Assemble andstart up the VOC analyzer according to themanufacturer's instructions. After theappropriate warmup period and zero internalcalibration procedure, introduce thecalibration gas into the instrument sampleprobe. Adjust the instrument meter readout tocorrespond to the calibration gas value.

Note.-If the meter readout cannot beadjusted to the proper value, a malfunction ofthe analyzer is indicated and correctiveactions are necessary before use.

4.3 Individual Source Surveys.4.3.1 Type I-Leak Definition Based on

Concentration. Place the probe inlet at thesurface of the component interface whereleakage could occur. Move the probe alongthe interface periphery while observing theinstrument readout. If an increased meterreading is observed, slowly sample theinterface where leakage is indicated until themaximum meter reading is obtained. Leavethe probe inlet at this maximum readinglocation for approximately two times theinstrument response time. If the maximumobserved meter reading is greater than theleak definition in the applicable regulation,record and report the results as specified inthe regulation reporting requirements.Examples of the application of this generaltechnique to specific equipment types are:

a. Valves-The most common source ofleaks from valves is at the seal between thestem and housing. Place the probe at theinterface where the stem exists the packinggland and sample the stem circumference.Also, place the probe at the interface of thepacking gland take-up flange seat and samplethe periphery. In addition, survey valvehousings of multipart assembly at the surfaceof all interfaces where leak could occur.

b. Flanges and Other Connections-Forwelded flanges, place the probe at the outeredge of the flange-gasket interface andsample the circumference of the flange.Sample other types of nonpermanent joints(such as threaded connections) with a similartraverse.

c. Pumps and Compressors-Conduct acircumferential traverse at the outer surfaceof the pump or compressor shaft and sealinterface. If the source is a rotating shaft,position the probe inlet within 1 cm of the

shaft-seal interface for the survey. If thehousing configuration prevents a completetraverse of the shaft periphery, sample allaccessible portions. Sample all other jointson the pump or compressor housing whereleakage could occur.

d. Pressure Relief Devices-Theconfiguration of most pressure relief devicesprevents sampling at the sealing seatinterface. For those devices equipped with anenclosed extension, or horn, place the probeinlet at approximately the center of theexhaust area to the atmosphere.

e. Process Drains-For open drains, placethe probe inlet at approximately the center ofthe area open to the atmosphere. For covereddrains, place the probe at the surface of thecover interface and conduct a peripheraltraverse.

f. Open-Ended Lines or Valves-Place theprobe inlet at approximately the center of theopening to the atmosphere.

g. Seal System Degassing Vents andAccumulator Vents-Place the probe inlet atapproximately the center of the opening tothe atmosphere.

h. Access Door Seals-Place the probe inletat the surface of the door seal interface andconduct a peripheral traverse.

4.3.2 Type II- "No Detectable Emission".Determine the local ambient concentration

around the source by moving the probe inletrandomly upwind and downwind at adistance of one to two meters from thesource. If an interference exists with thisdetermination due to a nearby emission orleak, the local ambient concentration may bedetermined at distances closer to the source,but in no case shall the distance be less than25 centimeters. Then move the probe inlet tothe surface of the source and determine theconcentration described in 4.3.1. Thedifference between these concentrationsdetermines whether there are no detectableemissions. Record and report the results asspecified by the regulation.

For those cases where the regulationrequires a specific device installation, or thatspecified vents be ducted or piped to acontrol device, the existence of theseconditions shall be iisually confirmed. Whenthe regulation also requires that nodetectable emissions exist, visualobservations and sampling surveys arerequired, Examples of this technique are:

(a) Pump or Compressor Seals-Ifapplicable, determine the type of shaft seal.Preform a survey of the local area ambientVOC concentration and determine ifdetectable emissions exist as describedabove.

(b) Seal System Degassing Vents,Accumulator Vessel Vents, Pressure ReliefDevices-If applicable, observe whether ornot the applicable ducting or piping exists.Also, determine if any sources exist in theducting or piping where emissions couldoccur prior to the control device. If therequired ducting or piping exists and thereare no sources where the emissions could bevented to the atmosphere prior to the controldevice, then it is presumed that no detectableemissions are present. If there are sources inthe ducting or piping where emissions couldbe vented or sources where leaks couldoccur, the sampling surveys described in this

paragraph shall be used to determine ifdetectable emissions exist.

4.3.3 Alternative Screening Procedure. Ascreening procedure based on the formationof bubbles in a soap solution that is sprayedon a potential leak source may be used forthose sources that do not have continuouslymoving parts, that do not have surfacetemperatures greater than the boiling point orless than the freezing point of the soapsolution, that do not have open areas to theatmosphere that the soap solution cannotbridge, or that do not exhibit evidence ofliquid leakage. Sources that have theseconditions present must be surveyed usingthe instrument techniques of 4.3.1 or 4.3.2.

Spray a soap solution over all potentialleak sources. The soap solution may be acommercially available leak detectionsolution or may be prepared usingconcentrated detergent and water. A pressuresprayer or a squeeze bottle may be used todispense the solution. Observe the potentialleak sites to determine if any bubbles areformed. If no bubbles are observed, thesource is presumed to have no detactableemissions or leaks as applicable. If anybubbles are observed, the instrumenttechniques of 4.3.1 or 4.3.2 shall be used todetermine if a leak exists, or if the source hasdetectable emissions, as applicable.

4.4 Instrument Evaluation Procedures. Atthe beginning of the instrument performanceevaluation test, assemble and start up theinstrument according to the manufacturer'sinstructions for recommended warmup periodand preliminary adjustments.

4.4.1 Response Factor. Calibrate theinstrument with the reference compound asspecified in the applicable regulation. Foreach organic species that is to be measuredduring individual source surveys, obtain orprepare a known standard in air at aconcentration of approximately 80 percent ofthe applicable leak definition unless limitedby volatility or explosivity. In these cases,prepare a standard at 90 percent of thesaturation concentration, or 70 percent of thelower explosive limit, respectively. Introducethis mixture to the analyzer and record theobserved meter reading. Introduce zero airuntil a stable reading is obtained. Make atotal of three measurements by alternatingbetween the known mixture and zero air.Calculate the response factor for eachrepetition and the average response factor.

Alternatively, if response factors have beenpublished for the compounds of interest forthe instrument or detector type, the responsefactor determination is not required, andexisting results may be referenced. Examplesof published response factors for flameionization and catalytic oxidation detectorsare included in Section 5.

4.4.2 Calibration Precision. Make a total ofthree measurements by alternately using zerogas and the specified calibration gas. Recordthe meter readings. Calculate the averagealgebraic difference between the meterreadings and the known value. Divide thisaverage difference by the known calibrationvalue and mutiply by 100 to express theresulting calibration precision as apercentage.

37601

37602 Federal Register / Vol. 48, No. 161 / Thursday, August 18, 1983 / Rules and Regulations

4.4.3 Response Time. Introduce zero gasinto the instrument sample probe. When themeter reading has stabilized, switch quicklyto the specified calibration gas. Measure thetime from switching to when 90 percent of thefinal stable reading is attained. Perform thistest sequence three times and record theresults. Calculate the average response time.

5. Bibliography.5.1 DuBose D.A., and G.E. Harris.

Response Factors of VOC Analyzers at a

Meter Reading of 10,000 ppmv for SelectedOrganic Compounds. U.S. EnvironmentalProtection Agency, Research Triangle Park,N.C. Publication No. EPA 600/2-81-051.September 1981.

5.2 Brown, G.E., et al. Response Factors ofVOC Analyzers Calibrated with Methane forSelected Organic Compounds. U.S.Environmental Protection Agency, ResearchTriangle Park, N.C. Publication No. EPA 600/2-81-022. May 1981.

5.3 DuBose, D.A., et al. Response ofPortable VOC Analyzers to ChemicalMixtures. U.S. Environmental ProtectionAgency, Research Triangle Park, N.C.Publication No. EPA 600/2-81-110. September1981.[FR Doc. 3-99 Fied 8-17-83; 8.45 am]

BILLING CODE 6560-50-M