1053
`
`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`Flammable Gas imaging System Using Infrared Absorption
`
`Syslémo Imagour do Go: lnflammablo Udiisani i'Absorption Inlrarougo
`
`wa
`
`Toshihido ka
`Hirolumi U
`Kohiohi Sumida
`Tokoohl lehlo
`
`Osaka Gas 00.. Ltd. JAPAN
`
`ABSTRACT
`
`MMMWWWWHMMM leaks
`mam.moolfidotflmouuolnuiiorknwchiooksovorawidouu
`'
`Wmmal—fimimoahgofflammflomnoinormoflymmmmomkod
`oyologolhorwmrod-fimoimocosofhommndinglandocapo.Thogoalism
`Moodovioobouovwflodloon-oflonppilcaflomwoughmomoimuon
`pioumlakmdmiyofflammmaumdmoumndinglmdsmpo.
`ralhotilmwhgoormlcalodnm mmmanplyoxtmcigao.
`Woooloclodaviou-lizaion
`“mum-mad
`deom-bmdgubbmbmm.mn
`simulafimohdinioddulnmmommpmomdbulanoooudboviouaizod
`oimultanoouolymoboomdomonstration oxpodmoms on visualization using
`Mmdcmmmmmmbmwhmloddmwoouldboimaged
`undueontrollodomirutmflalomdifiono.Thonoxtstopinlhiolinoolrosoarohwiii
`be to dovolop o dovioo suitable bl" industrialization.
`
`RESUME
`
`Deslodwnbgioodomafiondoohmdogazhflmmwm
`Wenoomdoddvoioppomonwodoiounirunnouvooumyondo
`donationphsoflimoodooolypodohiiuwrunonslomflo. Coll-mm
`ufiliaolaropréoonmfimviouoloonbmrbddoogazinflammfilaquinomM
`normolomommlbloormllmaindwodoohnogosoniomporooldooonoo
`onvironnams. L'objooiil as: do woof un TIMI mioux adaplo aux applications
`locales on utliloam dos Images du go: in nmmablo oi dos silos onvironnama fllmoos
`olmuilanémom. pluifil quo d'ufllisot dos instruments do moouro complexes qul no
`fool
`'oxtrairo lo oaz.
`r oommm loo roohordioo “memoir”, nous ovens «locum uno
`momododovbudioafionufilisamhowacionofimoodabsocpflondoohfimos
`dos gaz a boss d'hydrocarburos. Dos simulation ayanl mo qu'il flail pooohlo
`dovisuofisorolmminnomomlomédmnofipropmooiiowmmousavono
`prooodo a dos main do dimonolraiion oonoomanl lo faioabliilo d'un.
`roprooonlafion viouoilo dos lultoa do on: a l'oide do oamflao lnlrarougoo oi do
`sources lumlnouooo done un onvironnomonl oonirliiéio dovoloppornonl d'un
`dlopooitil adapio a I'induolrlo son: in proohaino étopo do calla llgno do rochorchos.
`
`FLIR Systems, Inc.
`Exhibit 1010-00001
`
`

`

`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`1059
`
`IITRDWL'I‘IOI
`
`Strict attention is paid to protective and disaster prevention
`equip-ant at LIB terninals because of the large quantities of fie-able
`gas handled there. Within this context. gas detectors play an
`important role in discovering gas leaks at an early stage in the rare
`event of a leak. and in gathering infornaticn for appropriate measures.
`storage and production equip-ant like we / LPG tanks and vaporizers
`are currently equipped prinarily with point-type gas detectors. but
`extensive monitoring.
`like wide-arse gas leak mitoring. requires many
`sensors and much labor for maintenance control and other work.
`To
`overcome this problem. Osaka use has focused on developing even larger
`area detection and monitoring devices. such as systems for measuring
`nethans gas concentration on the optical path using lie-Ne laser.
`The present research is related to developing s new eethod for
`lonitoring flammable gas leaks in which the surrounding landscape as
`well as nethane. propane and butane leek conditions are displayed in
`reel—tine on a TV monitor. Designated lee): action that nor-ally
`requires gas detection frat Iultiple locations using conventional
`point-type gas detectors can be taken i-edietely after a leak is
`vie-ed on screen. and operators can accurately amraise the rapidly
`changing dispersion and holding behavior of leaking gas using this
`eethod.
`(gee rig. 1.)
`The present report focuses on the operating principle. si-uletion
`evaluations on visualisation performance as well as results fr:-
`danonstration experisents for a visualisation method based on the
`infrared absorption characteristics of flammable gas.
`OPERATING PRINCIPLE OP VISUAL-12313101 MODE
`
`since flammable gas has certain absorption characteristics at
`specific wavelengths as shown in Ho. 2. we can tell that flammable gee
`is present by measuring infrared ray propagation at that wavelength.
`usesuring methods are classified as either active or passive. depending
`on whether the infrared rays eeasured are supplied by a special
`external light source. or from infrared rays radiating fro. the
`surrounding landscape.
`laee rig. 3.}
`By the sale token. evaluation
`eethods for visualisation performance vary even within the active
`eethod because evaluation are based on the type of light source. The
`operating principles for active and passive visualisation Isthoda are
`outlined below.
`Neither nethod is suitable facing skyward free the ground because
`a background that reflects or radiates infrared rays is an essential
`condition for operation.
`
`Active lethod Being An Ee-le Laser
`
`the
`when laser light is absorbed by gas on its optical path.
`relationship between 10,
`the intensity of light incidental to the gas.
`and I.
`the intensity of light that has passed through the gas. is
`generally expressed with the Lambert-Beer equation shown below.
`I - Io expt-lq 1 c)
`
`(1!
`
`Here. 1:1 is the absorption coefficient at wavelength 3. of the
`subject gas. 1 is the gas and light working length. and c is the gas
`concentration.
`We can evaluate the amount of light absorbed if we use the
`equation to find the absorption coefficient k}, for flan-able gas
`targeted for detection.
`Il'he lie-Is laser has been used for sole tine
`nor as a light source for eeesuring nethane gas by the infrared
`absorption method because one of the oscillating wavelengths of the
`laser conical to the V3 line of the lethane absorption spectrum.
`In
`the present study, white cells were also used on propane and butane -
`targeted for visualisation — in order to asasure their respective
`absorption coefficients.
`II'he results. shown in Table l. confir-sd that
`
`FLIR Systems, Inc.
`Exhibit 1010-00002
`
`

`

`1060
`
`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`a LED-pa ae-Na laeer can detect methane. propane and butane
`eimultanecuely.
`A method that expande laser henna via an optical eyetaue wee uaad
`with the laaer deacrih-ed above to irradiate an area targeted for
`aonitoring.
`I'l'he method immee the reliability and lowere the coat of
`device- hecauee it lecke lovable parte like thoee need in the bean acan
`eathod.
`In order to viaualiee newbie gae with thie device
`configuration.
`the aacunt of light ahecrhed by geeae auet exceed the
`detectable level of the infrared ilege eeneor.
`The following equation
`1! need to ante-e vieualieation conditione when the area irradiated by
`leeer ie eufficiently large,
`the area ie irradiated at a constant In
`inteneitr. end thereel radiation effecte free the background are
`ninieieed by a cold filter on the front of the image eenaor.
`(2}
`11:0 (1 - Brod-2 k; l 0)} > Iad
`lad ia a
`Here, 7 ie the reflectance ratio for the background.
`light inteneity differential that in need to dietinguieh differences in
`concentration when viewing a viaualired inege. and that ie determined
`emerieentelly according to the type of infrared inage eeneor need.
`The equation for vindication conditione indicatea that even a
`low-concentration gee cloud can be vieualiaed when identical infrared
`inaga eaneore are need ae long In the reflectance ratio of the
`background aa well ae the inteneity of laear irradience are both
`relatively large.
`
`Paeeive lathod Uaing Background Thermal Radiation
`
`3 of an infrared
`The peeeive method ia comrieed ae ehovn in Pig.
`caaera and a background that radiatee infrared raye. Here the
`background and flamehle gee preeent between the background and the
`can-re are iaaged eienltaneouely.
`In ten. of the ebeorption wavelength for fie-able gaeea given
`in Pig. 2. gee cannonente in the ateoephere, euch ea E20 and CO). have
`euch a “all 'atacepheric eindoe' for eheorption that there ie little
`eheorption and dilpereion within a range of aeverel ten- of Ietere.
`Here the trenaeieeion of infrared reya propagated through the
`ataoenhere ie entree-ed by the equation halal.
`(111!_=_
`v-
`dx
`k},p(| B)
`IV in the radiation intenaity. x 1! the length of the
`Here,
`medium it}, ie the eheorption coefficient of file-aehle gee at wavelength
`A, p ie the concentration of the aheorhed nediua (fle-eable gee) and 3
`ie the Planck'e equation of theraal radiation.
`hauling that the background is a gray body with e enieeivity of E,
`and that flamehle gaa ie uniformly dietrihuted from the background to
`the infrared canara,
`then if. we define Tbaak the background temperature
`and T0" the gee temperature. we can uee the following equation to
`expreae radiation intenaity in wavelength range 1.1 - M detected by the
`infrared eeneor.
`
`{3)
`
`Iv = EjrmLkaxm-h c an).
`+ I: B(LT"){l—exp(-k‘ c and).
`The tiret tern on the right aide of the equation repreeenta the
`tramieeion of background radiation. while the aecond tare rapreeente
`the radiation and ebeorption of the ehaorbed eediua.
`If the background
`ie a black body and the background and Elemeahle gee teenaraturee are
`the acne.
`than that-eel radiation iron the background will not be
`aheorbed by the gee. However,
`the propagation of background theraal
`radiation will be altered by the preeence of flan—able gee becauee the
`ataoepheric teaperatura will alwaya he alightly different fro- the
`
`(4)
`
`FLIR Systems, Inc.
`Exhibit 1010-00003
`
`

`

`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`105“
`
`background teeperature and black body backgrounds are rarely found in
`outdoor environmnts where the device is actually used.
`The equation below is used to express visualisation conditions
`when to is the intensity of radiation detected by the infrared caeera
`with no Elm-able gas present.
`IIv-Icl>lsd
`Just as with the active eethod.
`lad is a light intensity
`differential that is used to distinguish differences in concentration
`when viewing a visualised image. and. that is determined experimentally
`according to the type of infrared image sensor used.
`Infrared camera sensitivity and the transnit wavelength of the
`infrared bandpass filter mounted on the canera deter-ins detection
`sensitivity with this lethod. As the transmit hand of the filter
`narrows to latch the absorption peak within the transmit band as shall:
`in Fig. 2.
`the as ratio increases. nor-ever the snount of transmitted
`light tends to decrease at the sane time, and imaging is more
`difficult. There exists preferable band width of the filter that
`provides well balance of the an ratio and the asount of transmitted
`light.
`ror this reason. a bandpass filter that can be varied for the
`background conditions of the desired visualisation area as well as for
`gas absorption characteristics is highly desirable.
`
`(5!
`
`Active lethod Using A white Light Source
`
`An He-Its laser that oscillates at a wavelength of 3.39 pm is large
`in site relative to its output. so the device must be enlarged to
`monitor gas leaks over a wide area.
`lith that in hind, we devised a
`visualisation device that uses a relatively ssall and inexpensive white
`light source.
`Ithe light source used here is a globar leap of sintered
`silicon carbide (sit).
`aecsuse the globe: lain is a conductor, it can
`be heated directly by current flow. hissivity for the globar leap
`ranges from 0.7 to 0.8 at wavelengths ranging from 3 to 4 pm. and its
`radiated infrared spectrum conforms to Planck's law of thermal
`radiation. Globe: lamps are simply structured and readily provide a
`large output. so they feature relatively easy device operation, and
`they eliainate the need for detecting faint infrared rays in the border
`area of sensor sensitivity as is often the case sith a passive eethcd
`and an active one with s laaer.
`Like the passive method. visualisation with a globe: lamp requires
`gas absorption and radiation evaluations using wavelength integration
`rather than visualisation performance evaluations based on the Lambert-
`Beer equation used for lasers. Here. performance evaluations are
`extra-sly ccsplex because simlations are performed sith new para-stars
`like light source output and convergence performance as well as
`background reflectance that are appended to the passive method.
`VISUALIHT‘ION PERFUME! BINULNI'IO‘N USING THE PASSIVE METHOD
`
`Because so many parameters affect visualisation performance with
`the passive method.
`the lost efficient loans of evaluating perforlsnce
`is through simulation calculations. The following results were
`calculated using propane as an exanple.
`The graph in Fig. i-a represents variations in infrared ray
`intensity. Iv-Io in equation (5), caused by both propane'e absorption
`and radiation when the concentration of the gas was varied.
`In case
`the temperature of propane is lower enough than that of the background.
`the quantity of the absorbed infrared ray exceeds that of the
`radiation. In this situation the variation.
`the value of vertical axis.
`beet-cs negative. and the gas is visualised darker on a TV monitor.
`Here the ‘telsperature difference between the background and the gas
`was varied from 0°C to 8°C. and background emissivity was fixed at 0.9.
`Here infrared intensity corresponding to a detection level of the
`infrared camera whose noise equivalent temperature differential (mm
`is 0.1“C was set at 1. Generally the greater the temperature
`differential.
`the greater the differential between fla-able gas
`
`FLIR Systems, Inc.
`Exhibit 1010-00004
`
`

`

`1062
`
`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`absorption and radiation. no tho ditforontial intonoity of infrarod
`rays roaching tho oonsor varios trundously.
`Ito know, howovar,
`that
`gas talporaturo'. which is “tr-oily difficult to dotoct is not tho oano
`as tho background tonporaturo bocauso of background onissivity.
`fig.
`4—!) shown tho variations in infrared ray intensity as “11 as
`tig. l—a than tho tuporaturo difforontial botwoon tho background and
`propano gas is fixed at 2°C. In this caso background emissivity is
`variod from 0.8 to 1.0. Pro. this In know that tho highor tho
`ooissivity.
`tho graatar tho attonuotion duo to tho cos. and that a
`ouissivity of #OJ causos variations vary closo to 2°C in tho
`tumoraturo dittorantial hotmn tho background and gas.
`Thoso variations in tho tooporaturo diffsrantial hetooon
`background and gas as won an in background emissivity causo infrarod
`ray intonoity - an indicator of visualization porfonanco - to vary
`drantically.
`lthon visualiration performanoo otudios ooro oonductod by
`varying paramotors for nothano and hutsno gas so roll. no found that
`butano oxhihitod tho om absorption tondancioo as propano.
`Visualisation by tho pasoivo oothod using can-arcially availahlo
`canons is linitod at thio tino to nothano. propano and hut-no.
`Considorino tochnical tronds indicating rogular ioprovenonts in tho
`sensitivity of infrarod cannot-as. hmvor, to anticipate that this
`nothod in the future at flu-ashlo gas visualization tochnology because
`of foaturos lika the siao and sinlicity of dovioo structurol. Donor
`emulation, so wall as r-oto and Vida-ans. nonitoring capability.
`
`VIEWIIATIOII mums UBIIIG ”DEL-3D GAB mus
`
`Visualisation oxporinonts using nodolod filo-able gas leaks woro
`conductod in ordor to confirm just 1mmr an actual Ell-bio on look in
`ingod haood on our undorstanding at tho infrarod absorption
`charactoristics of gasoa. ovaluations on intrarod canora poriomnco
`and tho rosults ot studios. such as porfornsnco evaluation simulations.
`
`maxi-atoll Condition
`
`Tho bohavior of. outdoor gas looksgo from cylindsrs and tubes of
`{la-tabla gas at a maximum rato of 3 litors per minute was videotapod
`with a visualisation dovico installod sovoral notors away from tho
`roloaso oito. Fin. 5 shooa tho dovice structure.
`Il'ho aoount of
`looking gas was adjustod by a mans not controllor that oaintainod a
`constant nass (loo rato.
`Tho visualised inagos ooro gray—scale inagos
`shot in mac format. and voro storsd on a video taps rocordor (WE).
`with tho activo mothod, a scroon nado of styrono was inotallod in
`the background to nintain consistent background rofloctanco
`conditions. both to lointain an adoqunto snount of rofloctod light and
`to confirm rofloction conditions for tho looking gao.
`Tho background
`screen.
`light souroo and cmra woro installod in positions which
`prmntod tho airror rotloction.
`Tho light sourooo vor- an Bo-na lasor
`and aloha: laws. and visualilod isagos for both sources oars rocordod.
`Iith the passive oothod.
`tho background consistod of. ground
`surfaco and surrounding landscapo lado of concroto.
`Tho charactsristics of tho bandpass filter mounted on tho infrared
`colors £or tho passive oothod wora diff-Ironic from those for tho activo
`method. Bocauso tho intonaity of infrarod rayo in tho activo nothod is
`ouch aroator thon that of tho passivo sothod.
`tho bandwidth of tho
`filtor for tho activo ono can ho narrator to obtain better all ratio.
`
`hporilontal nosulta
`
`Active and posoivo oothod ioagoo rocordod on m aro ohm in
`!'igs. 6 and '7. rospactivoly.
`I'ig. 6 shows propane looking at a £10!
`rats of 3 litors por ninuto frat a 9—H:
`tdiauotor) ruhhor tuba
`positionod in tho cantor of tho monitor screen.
`Tho styrono scroon in
`positionod in tho background. and tho distancos hotooon tho background
`andthoruhbartuhoasooliasbotooonthoruhbortuboandtho
`visualisation dovico uoro 0.5 n and 3 n, rospoctivoly.
`fig. 1 shows a
`
`FLIR Systems, Inc.
`Exhibit 1010-00005
`
`

`

`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`[053
`
`visualised image videotaped under direct sunlight 5 meters away from a
`leaking of butane gas container.
`The following summarizes the results.
`[1)
`lethane. propane and butane could all be visualized with the
`same device in the active method using an Ee-Ne laser and a
`glohar lamp as light sources.
`(2) Propane and butane could both be visualized with a commercially
`available infrared camera using the passive method as long as
`an adequate amount of radiated background light could he
`maintained.
`(Visualization performance was improved with
`strong sunlight, but the overall effect was essentially the
`same as with the active method using a white light source.)
`{3) Honitor screen concentration variations due to gas absorption
`as well as concentration variations due to background
`conditions were easily distinguishable if flammable gas was
`visualized together with the background in real time.
`{4) The frame rate of a TV monitor was found to play an important
`role like sensor sensitivity in order for observers to
`ascertain a gas leak from the visualized image.
`(5) Gas leaking from the surface was videotaped when a modeled
`underground flammable gas leak was observed at ground level by
`the visualization device.
`
`CONCLUSION
`
`(2)
`
`We focused on a flammable gas visualization method that uses the
`infrared absorption characteristics of hydrocarbon—based gas as a
`method for monitoring flammable gas leaks over a wide area. and we
`conducted studies on active and passive methods using He-fle lasers as
`well as an active method using a white light source.
`A visualization
`device was set up for performance evaluations, and videotaping
`experiments were conducted on methane. propane and butane.
`The results
`are given below.
`{1] Methane. propane and butane were simultaneously detected with
`the active method.
`Propane and butane were visualized by the passive method using
`a commercially available infrared camera when environmental
`conditions, such as background temperature. were controlled.
`[3} Visualizing flammable gas and the background in real time not
`only provides a simple mechanism. but also facilitates the
`ability to visualize the situation of gas leaks.
`(4) The visualization device set above ground has the possibilities
`to detect gas leaks from buried pipes.
`(5) As a result of the experiments.
`the flammable gas imaging
`together with surrounding landscape is confirmed a very
`effective measure to find leak locations and to observe the
`behavior of leaking gas.
`
`ACRNDHLEIEMENTS
`
`We would like to thank the Industrial Supplies s Equipment
`Division of Nikon Corporation for kindly providing an equipment package
`containing an infrared camera with cold filters and imaging equipment
`used in the present visualisation demonstration experiments.
`
`REFERENCES
`
`1.
`
`2.
`
`3.
`
`4.
`
`Infrared Engineering Society : Infrared Technology, can
`publishing/Tokyo. 1991
`Y. Yamasaki. et al.
`: Development of methene detecting system
`utilising Ee—Ne laser. Instrumentation Vol.34
`N07. 1991
`K. Gannbo. H. Tanaka, T. Tokioka : Atmospheric Environment
`(Atmospheric Sciences Lectures).Tokyo University Publishing. 1932
`T. Eennno et a1.
`1 Infrared sensorstII) IRCCD focal plane arrays
`sensitive in the 3 to 5 gm range, Technical report, Technical
`research and development institution. Japan defense agency, 1967
`
`FLIR Systems, Inc.
`Exhibit 1010-00006
`
`

`

`1064
`
`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`Plant
`
` Portable imaging
`
`device
`
`Fixed Imagang devzce
`
`Mum:-
`
`1
`
`tho can lounge monitoring sylten
`Ina-.1: of
`based on an imaging
`
`‘
`I
`;
`
`.
`1
`
`T
`
`350 ‘
`‘
`
`300
`
`.1
`a
`g
`.1.
`:5
`8:200l
`:9:
`_
`2:. 150
`Ed
`0
`5
`
`V
`methane 1
`.———-r1~bur.ane
`:“ffloflane
`
`I
`
`1
`
`250
`
`100
`
`g; D
`
`Figure 2
`
`Infratod absorption spectra; of
`
`flu-unable gases
`
`WavelengL'n [113“:
`
`FLIR Systems, Inc.
`Exhibit 1010-00007
`
`

`

`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`1065
`
`Light. source
`
`\
`
`Infrared
`"-uu‘imqe sensor
`
`
`
`Infrared
`image sensor
`
`
`«2 Lxgh: radiated by the background
`+: Reflected Light by the light source
`
`figure 3
`
`Principlo 0! “ml. Ill inning
`
`'I'nhlo 1
`
`Ahlozption mutton-tinting at
`[la-lo 1.10:)
`
`subdue: all.-
`
`Absorption
`unafficiont
`"u“1cn‘1
`
`8.9 x 10
`
`
`Type of an : Propane
`Background Lunpontun:
`300 [It]
`Background «Mummy :
`0.9 [-1
`
`Gas tmplrature
`
`
`292 [K]
`
`
`
`
`
`Variationininfrared
`
`rayintensity[-1'1
`
`Lfi -
`
`"= blah-mm
`Gas concentration [mm]
`, ,
`
`
`in an ammonia-clan and infrared my
`Mauro l-n Variation:
`Anton-icy (Background on twat-tux. «nun-outta)
`
`FLIR Systems, Inc.
`Exhibit 1010-00008
`
`

`

`10“
`
`195 INTERNATIONAL GAS RESEAFCH CONFERENCE
`
`
`
`
`Tn: ofqns: hnpun
`Background (mature:
`JODIK]
`Gas twraturo 2
`298!“
`
`>u
`
`u Ehm
`
`'2
`_
`
`“*1 Ia“
`“ >
`c:
`“ m
`c“
`03
`*‘c
`1:...
`E
`>
`
`
`
`
`Gas emissivity
`
`r——-———~' AMWVAW
`I
`0.30 [-1
`\
`----0.85 1—1
`0.... 0.90 [-1
`-----o.95 [-1
`---—-1.0 H
`
`‘
`‘
`
`I
`
`‘
`‘
`
`
`
`Gas concentration [DWm]
`
`
`'2: lv-blnoqflbnzs)
`
`rtgntc C-b Varintionn in an. aoncontration and inlrnr-d rly
`lacunllty (luckutound uni-nlvity)
`
`Background screen
`(styrene)
`
`Globar
`
`Beam expander
`x//He-Na laser
`
`
`
`
`
`
` Mass flow
`
`controller Image
`
`.-rgcessing
`unlt
`
`
`'Hethane
`-Propane
`
`IR image sensor
`
`(Band-pass
`
`filter equipped!
`
`figurc 5
`
`Contiguration a!
`
`th- activu nothod talc fluvial
`
`FLIR Systems, Inc.
`Exhibit 1010-00009
`
`

`

`1995 INTERNATIONAL GAS RESEARCH CONFERENCE
`
`1067
`
`
`
`figure 6
`
`visualisod lug. uling tho active mthod (Prupnna)
`
`
`
`figure 7 Vinullisod lug. using the pal-iv. method
`
`(Butane)
`
`FLIR Systems, Inc.
`Exhibit 1010-00010
`
`