11

Emissions fromEmissions fromAboveground Storage TanksAboveground Storage Tanks

the basics the basics -- 20022002

Rob FerryRob Ferrythe TGB partnershipthe TGB partnershipHillsborough, NC Hillsborough, NC

Rob.Ferry@TGBpartnership.comRob.Ferry@TGBpartnership.com

22

STORAGE TANKSSTORAGE TANKSEmission EstimatesEmission Estimates

•• Fixed Roof Tank (no floating roof)Fixed Roof Tank (no floating roof)–– Standing (Breathing) LossStanding (Breathing) Loss–– Working (Filling) LossWorking (Filling) Loss–– ASSUMES STABLE LIQUIDSASSUMES STABLE LIQUIDS

•• Floating Roof TankFloating Roof Tank–– Standing LossStanding Loss–– Working (Clingage) LossWorking (Clingage) Loss–– ASSUMES THAT THE DECK IS FLOATINGASSUMES THAT THE DECK IS FLOATING

33

Background on EmissionBackground on Emission(Loss) Factors(Loss) Factors

•• Developed from over 20 years of testing.Developed from over 20 years of testing.Research sponsored by API, in cooperation with EPA.Research sponsored by API, in cooperation with EPA.

•• Published by both API and EPA.Published by both API and EPA.Chapter 19 of API’s MPMS;Chapter 19 of API’s MPMS;Section 7.1 of EPA’s APSection 7.1 of EPA’s AP--42.42.

•• Encoded in EPA’s TANKS software.Encoded in EPA’s TANKS software.Latest factors in TANKS 3.1 (DOSLatest factors in TANKS 3.1 (DOS--version);version);TANKS 4.0X is for Windows, with the same factors.TANKS 4.0X is for Windows, with the same factors.

44

API ResearchAPI Research

•• Pilot tank testing in the late 1970’s.Pilot tank testing in the late 1970’s.•• Yielded useful info., such as effect of ambient wind. Yielded useful info., such as effect of ambient wind. •• Not feasible for isolating components.Not feasible for isolating components.

•• Wind tunnel and bench scale tests.Wind tunnel and bench scale tests.•• Testing of individual deck fittings & rim seals.Testing of individual deck fittings & rim seals.•• Testing at various wind speeds.Testing at various wind speeds.•• ‘Zero wind speed’: frequent air changes (not still air).‘Zero wind speed’: frequent air changes (not still air).•• Bolted deck seams the most difficult to standardize.Bolted deck seams the most difficult to standardize.

55

API’s MPMS, Chapter 19API’s MPMS, Chapter 19Manual of Petroleum Measurement StandardsManual of Petroleum Measurement Standards

Chapter 19Chapter 19––Evaporative Loss Measurement.Evaporative Loss Measurement.19.1 Evaporative Loss from Fixed19.1 Evaporative Loss from Fixed--Roof Tanks.Roof Tanks.19.2 Evaporative Loss from Floating19.2 Evaporative Loss from Floating--Roof Tanks.Roof Tanks.19.3 Loss Factor Certification Program.19.3 Loss Factor Certification Program.

Parts AParts A--E are test methods.E are test methods.Parts FParts F--H are for administration of the tests.H are for administration of the tests.

19.4 RP for Speciation of Evaporative Losses.19.4 RP for Speciation of Evaporative Losses.

66

EPA’s APEPA’s AP--4242Compilation of Air Pollutant Emission FactorsCompilation of Air Pollutant Emission Factors

•• Volume I: Stationary Point & Area Sources.Volume I: Stationary Point & Area Sources.•• Contains emission factors for all sorts of sources. Contains emission factors for all sorts of sources.

•• Section 7.1 Section 7.1 –– Organic Liquid Storage Tanks.Organic Liquid Storage Tanks.•• This is where the TANKS factors are found.This is where the TANKS factors are found.

•• Fifth Edition, Supplement D Fifth Edition, Supplement D –– September ’97.September ’97.•• This version has the TANKS 3.1 & 4.0X loss factors.This version has the TANKS 3.1 & 4.0X loss factors.•• Major revision of loss factors and equations between Major revision of loss factors and equations between

TANKS 2 and TANKS 3.0; relatively minor TANKS 2 and TANKS 3.0; relatively minor differences between TANKS 3.0 and TANKS 3.1.differences between TANKS 3.0 and TANKS 3.1.

•• TANKS 4.0X is a WINDOWS version of TANKS 3.1.TANKS 4.0X is a WINDOWS version of TANKS 3.1.

77

APAP--42 vs MPMS 1942 vs MPMS 19The loss factors and equations used are The loss factors and equations used are

identical, with the following exceptions:identical, with the following exceptions:–– Liquid surface temperature.Liquid surface temperature.

Both use the same determination for fixedBoth use the same determination for fixed--roof tanks, roof tanks, but differences in the method for floatingbut differences in the method for floating--roof tanks roof tanks results in slightly higher vapor pressures from EPA.results in slightly higher vapor pressures from EPA.

–– Bolted deck seam loss factor.Bolted deck seam loss factor.EPA reduced this factor by about 60% (0.14 vs 0.34) in EPA reduced this factor by about 60% (0.14 vs 0.34) in

response to data from testing in the late 1990’s. API response to data from testing in the late 1990’s. API has not changed from the earlier value of 0.34.has not changed from the earlier value of 0.34.

88

Selected ConsiderationsSelected Considerations

•• Single versus Single versus multicomponentmulticomponent liquid. liquid. •• Type of tank.Type of tank.•• Special considerations for crude oil.Special considerations for crude oil.•• Special considerations for low volatility Special considerations for low volatility

stocks.stocks.•• Floating roof landing lossesFloating roof landing losses

99

Single Single vsvs MulticomponentMulticomponent

•• Single component means only one chemical.Single component means only one chemical.–– e.g., storing only benzene, or only MTBE.e.g., storing only benzene, or only MTBE.

•• MulticomponentMulticomponent means a chemical mixture.means a chemical mixture.–– The crude oil that comes out of the ground may The crude oil that comes out of the ground may

be made up of dozens of chemicals, so most of be made up of dozens of chemicals, so most of the fuels made from crude oil are also mixtures the fuels made from crude oil are also mixtures of many chemicals. The concentrations of of many chemicals. The concentrations of individuals chemicals vary considerably individuals chemicals vary considerably (depending upon what the dinosaurs ate?) (depending upon what the dinosaurs ate?)

1010

Single Component LiquidsSingle Component Liquids

•• Individual chemicals.Individual chemicals.•• Properties are wellProperties are well--defined.defined.

–– Boiling point.Boiling point.–– Molecular weight.Molecular weight.–– Reid vapor pressure.Reid vapor pressure.–– True vapor pressure at a given temperature.True vapor pressure at a given temperature.

1111

Multiple Component LiquidsMultiple Component Liquids

•• Concentration of each component varies.Concentration of each component varies.–– Stuff that comes out of the ground is not all alike.Stuff that comes out of the ground is not all alike.

•• Physical properties vary.Physical properties vary.–– Use properties representative of the range.Use properties representative of the range.–– By definition, ‘average’ will be too high or too By definition, ‘average’ will be too high or too

low for any given tank, but deviations should low for any given tank, but deviations should cancel out for a large number of tanks. cancel out for a large number of tanks.

1212

Types of TanksTypes of Tanks

•• FixedFixed--roof tanks (without a floating roof).roof tanks (without a floating roof).–– There is no cover directly on the liquid surface, There is no cover directly on the liquid surface,

but there is a roof on the top of the tank.but there is a roof on the top of the tank.

•• FloatingFloating--roof tanks.roof tanks.–– A raftA raft--like cover floats on the surface of the like cover floats on the surface of the

liquid.liquid.

1313

FixedFixed--Roof TankRoof Tank(no floating roof)(no floating roof)

1414

FixedFixed--Roof TanksRoof Tanks

•• API 2518 (& APAPI 2518 (& AP--42/TANKS) emission factors:42/TANKS) emission factors:–– Expressly intended for stable liquids only.Expressly intended for stable liquids only.–– Will not account for flashing losses.Will not account for flashing losses.

•• To account for flash losses (unstable liquids):To account for flash losses (unstable liquids):–– VasquezVasquez--Beggs correlations, orBeggs correlations, or–– EC/R algorithms, orEC/R algorithms, or–– API/GRI E&P TANK computer modelAPI/GRI E&P TANK computer model

1515

Stable Stable vsvs Unstable LiquidUnstable Liquid

•• Stable implies all components are liquid.Stable implies all components are liquid.–– i.e., none of the components are trying to boil.i.e., none of the components are trying to boil.

•• Unstable implies a gaseous component.Unstable implies a gaseous component.–– i.e., one or more components wants to boil.i.e., one or more components wants to boil.–– Similar to fizz when opening a soft drink.Similar to fizz when opening a soft drink.–– Gas that was held in by pressure wants to Gas that was held in by pressure wants to

bubble out when the pressure is relieved.bubble out when the pressure is relieved.–– Typical of tanks in the oil patch.Typical of tanks in the oil patch.

1616

Crude Oil ProductionCrude Oil Production

Oil & Natural Gas Oil & Natural Gas pumped out of the pumped out of the

ground; under ground; under pressure until the pressure until the

first tank.first tank.

NaturalNaturalGasGas

“flashes”“flashes”(bubbles out)(bubbles out)

After flashing After flashing is complete, is complete,

the remainingthe remainingCrude OilCrude Oilis stable.is stable.

1717

External FloatingExternal Floating--Roof TankRoof Tank(floating roof, no fixed roof)(floating roof, no fixed roof)

1818

Internal FloatingInternal Floating--Roof TankRoof Tank(both a floating roof & a fixed roof)(both a floating roof & a fixed roof)

1919

FloatingFloating--Roof TanksRoof Tanks

•• Only suitable for stable liquids.Only suitable for stable liquids.–– Minimize evaporation by covering the surface.Minimize evaporation by covering the surface.–– Unstable liquids would bubble underneath.Unstable liquids would bubble underneath.

•• Vapors would build up under the floating roof until Vapors would build up under the floating roof until either finding a way out or capsizing the roof. either finding a way out or capsizing the roof.

•• Emission factors from 20 years of testing.Emission factors from 20 years of testing.–– API 2517 & 2519 and EPA APAPI 2517 & 2519 and EPA AP--42 & TANKS.42 & TANKS.

•• Tests sponsored by API; data shared with EPA.Tests sponsored by API; data shared with EPA.

2020

FloatingFloating--Roof TankRoof TankWithdrawal LossWithdrawal Loss

EvaporationEvaporationfrom thefrom thewet shellwet shell

(i.e., from liquid(i.e., from liquidleft behind onleft behind onthe shell afterthe shell after

the liquid levelthe liquid leveldrops)drops)

OutgoingOutgoingliquidliquid

2121

FloatingFloating--Roof TankRoof TankWithdrawal (Clingage) LossWithdrawal (Clingage) Loss

Equation for estimating withdrawal loss, Equation for estimating withdrawal loss, LLWW::LLWW = [(0.943 = [(0.943 QQ CC WWLL ) /) /DD] [1+(] [1+(NNFCFC FFCC //DD)])]Where:Where:

QQ = annual net throughput (barrels/year),= annual net throughput (barrels/year),CC = clingage factor (barrels per 1000 square feet),= clingage factor (barrels per 1000 square feet),

WWL L = average liquid stock density (pounds/gallon),= average liquid stock density (pounds/gallon),DD = tank diameter (feet),= tank diameter (feet),

NNFC FC = no. of fixed= no. of fixed--roof support columns (dimensionless),roof support columns (dimensionless),FFC C = effective column diameter (feet).= effective column diameter (feet).

2222

DeckDeckseamsseams

(if bolted)(if bolted)

Rim sealRim seal(closure device(closure devicebetween thebetween thedeck and thedeck and thetank shell)tank shell)

DeckDeckfittingsfittings

(only if they(only if theyopen throughopen through

the deck tothe deck tothe liquid)the liquid)

FloatingFloating--Roof TankRoof Tank

StandingStandingStorageStorageLossLossEmissions areEmissions arebased on vaporsbased on vaporsthat get past thethat get past thefloating roof (afloating roof (afixed roof blocks thefixed roof blocks thewind, but it is vented &wind, but it is vented &does not contain the vapors).does not contain the vapors).

2323

FloatingFloating--Roof TankRoof TankStanding Storage LossStanding Storage Loss

Equation for estimating standing storage loss, Equation for estimating standing storage loss, LLSS::LLSS = [= [FFRR + + FFFF + + FFDD] ] P*P* MMVV KKCCWhere:Where:

FFR R = total rim= total rim--seal loss factor (pound moles/year),seal loss factor (pound moles/year),FFF F = total deck= total deck--fitting loss factor (pound moles/year),fitting loss factor (pound moles/year),FFD D = total deck= total deck--seam loss factor (pound moles/year),seam loss factor (pound moles/year),P*P* = vapor pressure function (dimensionless),= vapor pressure function (dimensionless),MMV V = stock vapor molecular weight (dimensionless),= stock vapor molecular weight (dimensionless),KKC C = product factor (dimensionless).= product factor (dimensionless).

EquipmentEquipment--relatedrelated StockStock--relatedrelated

2424

FloatingFloating--Roof TankRoof TankRim Seal Loss FactorRim Seal Loss Factor

Equation for total rimEquation for total rim--seal loss, seal loss, FFRR::FFRR = [= [KKRARA + + KKRBRB VV nn] ] D D Where:Where:

KKRA RA = zero= zero--windwind--speed loss factor (lb moles)/(ft yr),speed loss factor (lb moles)/(ft yr),KKRB RB = wind= wind--dependent loss factor (lb moles)/(mph)dependent loss factor (lb moles)/(mph)nn(ft yr),(ft yr),V V = average ambient wind speed (miles per hour)= average ambient wind speed (miles per hour)

[use [use VV = 0 for IFRTs and Domed EFRTs],= 0 for IFRTs and Domed EFRTs],n n = wind= wind--dependent loss exponent (dimensionless),dependent loss exponent (dimensionless),D D = tank diameter (feet).= tank diameter (feet).

2525

FloatingFloating--Roof TankRoof TankDeck Fitting Loss FactorDeck Fitting Loss Factor

Equation for individual deckEquation for individual deck--fitting loss, fitting loss, FFFiFi::FFFiFi = = KKFAiFAi + + KKFBiFBi ((KKVV V V ))mimi

Where, for deck fitting Where, for deck fitting ii::KKFAi FAi = zero= zero--windwind--speed loss factor (lb moles/yr),speed loss factor (lb moles/yr),KKFBi FBi = wind= wind--dependent loss factor (lb moles)/(mph)dependent loss factor (lb moles)/(mph)m m (yr),(yr),KKV V = fitting wind= fitting wind--speed correction factor (dimensionless),speed correction factor (dimensionless),V V = avg. ambient wind speed (miles per hour)= avg. ambient wind speed (miles per hour)

[use [use VV = 0 for IFRTs and Domed EFRTs],= 0 for IFRTs and Domed EFRTs],mi mi = wind= wind--dependent loss exponent (dimensionless).dependent loss exponent (dimensionless).

2626

FloatingFloating--Roof TankRoof TankIFRT Bolted Deck Seam Loss FactorIFRT Bolted Deck Seam Loss Factor

Equation for total bolted deckEquation for total bolted deck--seam loss, seam loss, FFD D ::FFDD = = KKDD SSDD DD 22

Where:Where:KKD D = deck seam loss per unit seam length factor (lb mol/ft yr),= deck seam loss per unit seam length factor (lb mol/ft yr),

= 0.34 (API MPMS 19.2)= 0.34 (API MPMS 19.2)= 0.14 (EPA AP= 0.14 (EPA AP--42 and TANKS)42 and TANKS)

SSD D = deck seam length factor (feet per square feet),= deck seam length factor (feet per square feet),= 0.17 (noncontact deck with 6= 0.17 (noncontact deck with 6--ft wide deck sheets)ft wide deck sheets)= 0.28 (contact deck with 5’= 0.28 (contact deck with 5’ x 12’ rectangular panels)x 12’ rectangular panels)

DD = tank diameter (feet).= tank diameter (feet).

2727

Ambient Wind EffectsAmbient Wind EffectsAssumptions in the EquationsAssumptions in the Equations

Losses from Internal & Covered (domed external) Losses from Internal & Covered (domed external) Floating Roofs are not windFloating Roofs are not wind--dependent.dependent.

Vapor loss past the floating roof is independent of ambient Vapor loss past the floating roof is independent of ambient wind speed, (i.e., wind speed, (i.e., VV = 0), but air movement through the vents = 0), but air movement through the vents is adequate to evacuate those vapors that do get past the is adequate to evacuate those vapors that do get past the floating roof.floating roof.

Losses from External Floating Roofs are windLosses from External Floating Roofs are wind--dependent.dependent.

Full ambient wind speed, Full ambient wind speed, VV, is applied to rim seal losses., is applied to rim seal losses.Ambient wind speed, Ambient wind speed, VV, is modified by the Fitting Wind, is modified by the Fitting Wind--Speed Speed

Correction Factor, Correction Factor, KKVV (= 0.7), for deck fittings losses.(= 0.7), for deck fittings losses.

2828

Special Considerations forSpecial Considerations forCrude OilCrude Oil

•• Range of volatility of the components is Range of volatility of the components is much greater than for refined products.much greater than for refined products.

•• Heavier components may impede migration Heavier components may impede migration of lighter components.of lighter components.–– Thus retarding evaporation rates of stable oils,Thus retarding evaporation rates of stable oils,–– But also potentially retaining unstable But also potentially retaining unstable

components.components.

2929

Special Considerations forSpecial Considerations forLowLow--Volatility MixturesVolatility Mixtures

•• Low volatility means low level of emissions.Low volatility means low level of emissions.•• Thus these tanks are typically not regulated.Thus these tanks are typically not regulated.•• Properties have not been well defined.Properties have not been well defined.•• Large percent error would still be a small no.Large percent error would still be a small no.•• Nevertheless, better data is being pursued.Nevertheless, better data is being pursued.

3030

LowLow--VolatilityVolatilityStocksStocks

GasolineGasoline

DistillateDistillate0.1 psia0.1 psia

DistillateDistillate0.01 psia0.01 psia

3131

Floating RoofFloating RoofLanding LossesLanding Losses

•• Not a special event for fixedNot a special event for fixed--roof tanks.roof tanks.Emptying & refilling is accounted for in working loss.Emptying & refilling is accounted for in working loss.

•• Landing a floating roof:Landing a floating roof:–– Opens the vacuum breaker vent.Opens the vacuum breaker vent.–– Tank subsequently has losses while standing idle.Tank subsequently has losses while standing idle.–– Refilling expels vapors from under the deck.Refilling expels vapors from under the deck.

3232

Landing a Floating RoofLanding a Floating RoofEmptying draws Emptying draws

fresh air under fresh air under the deck.the deck.

Landing lossesLanding lossesdo not occur do not occur

during emptyingduring emptyingVacuum Vacuum breaker ventbreaker vent(see Section 2.1)

3333

Standing Idle LossesStanding Idle Lossesfor a Landed Floating Rooffor a Landed Floating Roof

•• Emissions occur while the tank stands idle Emissions occur while the tank stands idle after landing the floating roof.after landing the floating roof.

•• Modeling these emissions depends upon:Modeling these emissions depends upon:–– Liquid heel vs drain dry condition.Liquid heel vs drain dry condition.

If drained dry, available liquid is limited to clingage.If drained dry, available liquid is limited to clingage.

–– IFRT vs EFRT.IFRT vs EFRT.The vapor space under a landed IFR will behave in a The vapor space under a landed IFR will behave in a

manner similar to a fixed roof tank, but an EFR will manner similar to a fixed roof tank, but an EFR will additionally have wind effects.additionally have wind effects.

3434

Standing IdleStanding Idlewith a Liquid Heelwith a Liquid Heel

Depth of Depth of liquid heelliquid heel

is sufficientis sufficientto continueto continueto support to support

breathing lossbreathing loss

Daily Daily breathing breathing expels vapors expels vapors from under from under the deck.the deck.

3535

Standing IdleStanding Idledrained drydrained dry

Breathing Breathing loss ends loss ends when the when the

wetted areas wetted areas become dry.become dry.

The only The only remaining remaining liquid is liquid is clingage that clingage that didn’t drain.didn’t drain.

3636

WindWind EffectWind Effect

on EFRTon EFRT

Standing idle Standing idle loss from an loss from an EFRT with a EFRT with a liquid heel is liquid heel is

accelerated accelerated by wind.by wind.

3737

Filling LossFilling Lossfor a Landed Floating Rooffor a Landed Floating Roof

•• Refilling expels vapors from under the deck.Refilling expels vapors from under the deck.The quantity of vapors residing under the deck prior to The quantity of vapors residing under the deck prior to

refilling depends upon the standing idle condition.refilling depends upon the standing idle condition.

•• Incoming liquid generates additional vapors.Incoming liquid generates additional vapors.Thus even a completely vaporThus even a completely vapor--free tank will have free tank will have

vapors in the air displaced from under the deck.vapors in the air displaced from under the deck.

3838

Filling LossFilling Lossfor a Landedfor a Landed

FloatingFloatingRoofRoof

Additional Additional vapors are vapors are generated by generated by the incoming the incoming liquid.liquid.

Incoming Incoming liquid liquid

displaces displaces vapors from vapors from

under the under the deck.deck.

3939

Landing Loss SummaryLanding Loss SummaryEach scenario has a critical parameter that is not Each scenario has a critical parameter that is not

well quantified.well quantified.–– IFRT w/liquid heelIFRT w/liquid heel: the critical parameter is the : the critical parameter is the

saturation level (concentration of vapors) of the saturation level (concentration of vapors) of the airair--vapor mixture expelled from under the deck.vapor mixture expelled from under the deck.

–– EFRT w/liquid heelEFRT w/liquid heel: the effect of wind is an : the effect of wind is an additional critical parameter for this case.additional critical parameter for this case.

–– DrainDrain--dry tanks (IFRT or EFRT)dry tanks (IFRT or EFRT): the effective : the effective depth (thickness) of the layer of liquid left behind depth (thickness) of the layer of liquid left behind on the tank bottom is the critical parameter.on the tank bottom is the critical parameter.

4040

Status of Landing LossStatus of Landing LossEstimation MethodsEstimation Methods

•• API has developed a theoretical model.API has developed a theoretical model.Accounting for all the phenomena identified.Accounting for all the phenomena identified.

•• API has conducted limited field testing.API has conducted limited field testing.Which improved understanding of the theoretical model.Which improved understanding of the theoretical model.

•• Critical parameters remain elusive.Critical parameters remain elusive.If you have to come up with a number, the present state If you have to come up with a number, the present state

of the practice is to make a good faith effort using of the practice is to make a good faith effort using sound engineering judgment sound engineering judgment –– with lots of caveats.with lots of caveats.