Experiments in Fluids 14, 409-415 (1993)

Experhnenmmnu 9 Springer-Vertag 1993

Infrared imagery of an air [ CO2 axisymmetric jet D.N. Gordge*, R.H. Page Mechanical Engineering Department, Texas A & M University, College Station, TX 77843-3123, USA Received: 5 March 1992/Accepted: 18 January 1993

Abstract. This experiment uses an infrared imaging system to investigate a subsonic, non-isoenergetic, air/CO~ axisymmetric jet. The classical limitations of using IR imagery with hot gases are presented and a novel approach to overcome these limitations is proposed. The results suggest that radial and axial irradiant profiles measured with the IR imager, when non-dimensionalized, collapse onto curves of similarity. This behavior could allow temperature, velocity, and concentration profiles to be deduced from the IR image.

List of symbols I

directional, spectral irradiance (Watt/cm 2) dimensionless irradiance d nozzle diameter x, r coordinate system x0 virtual origin displacement B constant C concentration T object temperature (~ U velocity (m/s) hemispherical, total emissivity PCOz partial pressure carbon dioxide Subscripts

A R b m o t tr amb atm meas

axial profile property radial profile property property measured from the background behind the jet property measured on the jet's centertine axis property measured at the jet outlet property measured from the target transition region property measured in the ambient surroundings atmospheric property measured property

1 Introduction The freely expanding axisymmetric jet has many applications and has been the source of engineering investigations * Present address: SETD, ACS, SY73, Naval Air Warfare Center, Patuxent River, MD 20670, USA Correspondence to: D. N. Gordge

for many years. The characteristics of the jet's temperature, velocity, and constituent distribution are well documented in many sources. Early works reported investigations with results obtained using probes that were traversed through the flow field. More recent endeavors have taken advantage of modern techniques such as laser velocimetry, marker nephelometry, and laser-Rayleigh scattering to make non-intrusive measurements of the jet's characteristics. In this paper, we consider the use of another non-intrusive technique, infrared (IR) imaging, to study the characteristics of the axisymmetric jet. Infrared imaging systems were initially developed for military applications during World War II. Much of the early development was classified until the 1950's. Military and industrial application continued the advancement throughout the 1960's with analog scanners and imagers. Digital technology of the 1970's and 1980's allowed real-time images and computer manipulation of the data. Unlike IR radiometry that quantifies discrete radiant emission at individual frequencies over a wide portion of the IR spectrum, IR imagery quantifies the emissions detected within a much narrower bandwith. IR imaging is relatively inexpensive and has been used successfully for qualitative flow visualization for certain flow fields (Flaig 1976; Merzkirch 1987). An IR imaging system is an energy detector. It creates a visual image by detecting radiant heat emitted from an object and converting this energy to an electronic signal. The object's temperature may then be calculated using basic laws of physics if the object's emissive characteristics are known. (Most commercially available IR imaging systems are therefore marketed as "thermographic" instruments.) The prospect of using IR imaging to visualize and analyze fluid dynamic characteristics of a complex gas flow has numerous possibilities. There are, however, technical difficulties in using IR imagery to analyze gas flow fields. For instance, the emissivity of a gas is dependent on gas composition, temperature, pressure, and the optical thickness (i.e., mass path length) through the gas. A gas will typically exhibit spectral radiating properties with bands of intense absorption and emission. Assuming blackbody

410 (e = 1.0) or even gray body (e < 1.0 but constant) characteristics in the analysis of a turbulent gas flow field will result in substantial computational errors. Users of IR imagery for thermal analysis must also account for other sources of radiation that may have a significant impact on their measurements. These other sources, including background, atmospheric, and reflected ambient emissions, must be excluded from the analysis. For these reasons, quantitative results describing the characteristics of a hot gas flow field obtained using IR imagery are difficult to obtain. The purpose of this investigation was to determine if there was quantitative information that could be derived from IR images of a hot gas flow field. We used a commercially available IR imaging system to examine a hot, subsonic, air/CO 2 axisymmetric free turbulent jet exhausting into a quiescent reservoir. We chose this classical flow field due to its simple stucture, its universal application, and the technical base already established concerning the characteristics of the jet.

2 Methods

The structure of a turbulent, subsonic, axisymmetric jet exhausting into a quiescent reservoir is well documented in many references (Pal 1954; Schetz 1980; Blevins 1984, or Dowling and Dimotakis 1990 and the references therein). The jet is initially composed of a core region in which the jet's exit temperature and velocity remain constant for uniform exit conditions (Fig. 1). Because of the difference in energy and momentum transfer, these two core regions may not necessarily be identical. The core is nearly free of shear, although a shear layer mixing region exists between the core region and the reservoir. Growth of the shear layer erodes the core regions until the centerline velocity and temperature on the axis of the jet begin to decay. The jet then enters a transition region where the details of the initial conditions at the outlet of the jet are lost as the turbulent eddies in the shear layer obliterate the details of the nozzle core flow. The resultant eddy-dominated flow is called self-preserving or fully developed (Blevins 1984 and the numerous sources cited therein). IR images of the initial and fully developed regions were visualized using an AGEMA-780 short wavelength IR thermal imaging system. This particular scanning system uses a cryogenically cooled indium antimonide (InSb) photodetector which is sensitive to IR emissions in the 3-5/am bandwidth. Commercially available computer hardware (Thermal Imaging Computer (TIC-8000)) and software (CATS version 1.04) were used for IR image acquisition. The TIC8000 computer board performs an 8 bit analog to digital conversion of the energy irradiated onto the detector. The 140 by 140 pixel image is stored by the computer in a file with processing information from the IR imager. Energy irradiated onto an IR imager is composed of photons emitted from the target (I,), the atmosphere (latin) , the

I

Fully developed region ' ~ ; i

Ij

~

J

Transition region " ~ ' ~ ' ~ " ,

,,ii!,c_~- Shear layer

m~x~ngregion

Initial region

Jet outlet i

i '~

Potentialcore flow Fig. l. Structure of the axisymmetric jet background (Ib), and ambient energy (lamb) that is reflected off the target itself (Prengle et al. (1985)). These photon energies are converted to an electronic signal via the imager's linear photon counting detector. The irradiant energy arriving at the imager can be expressed as

I . . . . = It + lb + lamb +/arm'

(1)

it should be noted that the units of measure for irradiance are W / c m 2. The irradiance data obtained with the IR imaging system is subject to geometric and calibration constants peculiar to the make and model of the system. Thus, we will use a proportionality constant "B" where necessary to denote actual irradiance data. This does not adversely affect the results of this experiment since the relationships are eventually normalized and presented in dimensionless form. An air/CO2 jet was used due to its visibility in the 3-5/am bandwidth of the IR scanner. Air, being composed predominantly of nitrogen and oxygen, has poor radiating characteristics in the 3 5/am band. Carbon dioxide has six bands of absorption/emission. These bands are centered at 15, 10.4, 9.4, 4.3, 2.7, and 2.0/am. Bands at 15, 4.3, and 2.7/am are strong bands and are characterized by intense absorption (and emission) within the band (Edwards and Balakrishnan 1973). The width of the band is a function of the CO 2 temperature, partial pressure, and optical path length through the gas. The emissions from the 4.3/am CO 2 band are within the 3 5/am sensitivity window of the IR scanner. Water vapor has an absorption/emission band that crosses the 3-5/am band. For this experimental work, however, the jet was composed of dry compresssed air and dry bottled CO2. The experiments were conducted in a well ventilated, environmentally controlled laboratory. Some ambient water vapor may have been entrained by the jet from the ambient surroundings, but this amount is assumed negligible compared to the mass of the jet. A simple arrangement was constructed to provide steady flow outlet conditions throughout the test envelope. The exhaust nozzle (25.4 mm nominal diameter) was constructed

411 Table 1. Nozzle outlet conditions

Partial pressure

CO z

Series

To (~

Uo (m/s)

A (atm)

B (atm)

C (atm)

D (atm)

130 140 150 160 170 180 190

414 451 376 394 389 375 351

20.8 38.5 10.3 30.8 13.2 31.5 17.6

0.500 0.270 1.000 0.306 0.630 0.288 0.500

0.227 0.126 0.467 0.216 0.430 0.204 0.227

0.080 0.063 0.225 0.139 0.250 0.131 0.080

0.043 0.065 0.080 0.062 0.043

of a 457 mm section of smooth wall copper conduit. The jet exit temperature was stabilized between 351 to 451 ~ The partial pressure CO 2 was varied from 0.043 to 1.000 atmospheres (atm). Outlet velocity varied from 10.3 to 38.5 m/s. The Reynolds number, based on the jet's centerline exit velocity, kinematic viscosity, and nozzle diameter, ranged from 12,000 and 38,000. The detailed test matrix is shown in Table 1. The nozzle flow characteristics were initially assessed using air only (i.e., no CO2 in the flow). Limited quantities of dry, bottled CO 2 precluded extensive, long duration measurements of the air/CO 2 jet's axial and radial temperature and velocity profiles. The desire to maintain homogeneous test conditions in the laboratory also prevented prolonged CO2 discharge into the environment. Detailed measurements of the air only jet were made using a traversing mechanism (Gordge 1989). These results showed that the nozzle configuration used for this experiment produced temperature and velocity profiles (air only jet) that collapsed onto curves of similarity in the fully developed region. Dry air is a poor radiator and is virtually "invisible" in the 3 - 5 lam bandwidth. Baseline "background" images were acquired of the nozzle with hot air only, Fig. 2 a. These images are composed of background (lb) and atmospheric (Iatm) irradiance terms shown in Eq. (1). CO 2 was then added to the jet and time averaged (20-40 s) "foreground images" were acquired, Fig. 2 b. The foreground images included itradiant energy from the heated a i r / C O : jet (It), as well as background emissions (Ib) transmitted through the jet, atmospheric emissions (Iatm), and ambient (lamb) radiation reflected off the jet. To isolate the CO 2 emissions of the jet, the background image was digitally subtracted (pixel by pixel) from the foreground image. An example of this result is shown in Fig. 2 c. In isolating the CO 2 emissions, the following assumptions were made: 1) atmospheric emissions are identical in the foreground and background images, 2) an insignificant amount of 3 - 5 gm background energy will be absorbed by the narrow 4.3 lam CO 2 absorption band as it transitions through the jet, and 3) low temperature ambient radiation reflected off the a i r / C O / j e t is negligible due to the narrow bandwidth associated with CO 2 emissions corn-

pared to the sensitivity bandwidth of the imager and the low reflectivity of the CO: gas itself. The value of the subtraction method is more apparent further downstream. Figure 3 provides a series of images at x =18.31 diameters from the jet's outlet. In the baseline image (Fig. 3 a), details of the thermal gradient in the wall behind the jet are evident. The jet is barely recognizable in the foreground image (Fig. 3 b), but becomes readily visible following the subtraction process. Both processed images, Figs. 2c and 3c, are composed of the jet's isolated CO2 irradiance (It) only. It should be noted that the subtraction process available with the CATS computer program was not used for the quantitative analysis. The CATS subtraction process is performed in terms of perceived temperature, not irradiance. A separate subtraction routine was developed to perform this process in terms of irradiance and is explained in detail in appendix (B) of the original work (Gordge 1989). The process also allowed for automation for handling large quantities of experimental data. Once the CO z emissions were isolated, the irradiance data was non-dimensionalized using Eqs. (2) and (3). For simplicity, references to irradiance (I) in the remainder of the paper will refer to isolated CO 2 irradiance only. I ~R = - -

Im I

~A = - - .

Itr

(2)

(3)

Other researchers have shown similarity in the energy, momentum, and constituent distributions in the fully developed region (Schetz 1980, Blevins 1984, or Dowling and Dimotakis (1990) and the references therein). A virtual origin shift (Hill and Page 1969) is often necessary for the data to collapse neatly. The virtual origin shift, x o, is used in the dimensionless similarity coordinate: r / ( x - Xo).

3 Typical results and discussion

3.1 Radial profile Our test apparatus allowed us to vary the proportion of CO 2 in the air/CO: jet while maintaining constant outlet temperature and velocity conditions. When we considered the effect of initial CO 2 concentration, such as with the Series 150 test where the P C O 2 varied from 0.225 to 1.000 atm, we found that the irradiance data collapsed onto a common curve (Fig. 4). This result was consistent for each series we tested. Thus, the dimensionless irradiance profiles suggest the development of the CO2 jet (PCO 2 = 1.000) is identical to the development of the CO2 portion of the air/CO 2 jet (PCO2 < 1.000). Looking further downstream, we found the shape of the dimensionless irradiance profile in the fully developed region

Fig. 2. a Background, b foreground, and e resultant IR image from series 1 4 0 A (x = 1.81 diameters from jet outlet, To = 451 ~ U o = 38.5 m/s, P C O z = 0.270 atm)

Fig. 3. a Background, b foreground, and e resultant 1R image from series 1 4 0 A (x = 18.81 diameters from jet outlct, To = 451 ~ Uo = 38.5 m/s, P C O ~ = 0.270 atm)

4~

413 ,0

(x/d > 6) was self preserving. Figure 5 illustrates data from

0.9

Series 130 A compared to the characteristic profile from Series 150 A. The shape of the dimensionless radial profile was consistent for all of the data acquired in the fully developed region of the jet regardless of initial outlet temperature, velocity, or CO 2 concentration. For this experiment, the collapse of the irradiance data was optimal when x o -- 0. That is, when there was no correction for a shift in the virtual origin. Unfortunately, limitations in the test apparatus prevented completely independent control of the outlet temperature and velocity variables. We recognize that this limitation does not produce all inclusive evidence concerning the result. However, over the length and breadth of the test conditions the dimensionless irradiance radial profiles remained constant. Recall that the radiant emission of CO2 is a non-linear function of the gas' temperature, partial pressure, and optical thickness. The consistency of the irradiance profile in the fully developed region may be partially attributed to the similar behavior of the mean temperature, velocity, and concentration distributions. The similarity in the irradiance profile, however, is unexpected given the non-linear nature of CO 2 emissions with temperature. Although a detailed discussion of molecular gas radiation physics is beyond the scope of this paper, a short discussion may be beneficial. Increasing the temperature of a gas increases its emissions in two ways. First, the emitted intensity increases according to Planck's law. Secondly, gas molecules emit in bands whose individual lines originate in excited rotation-vibration states. As the temperature increases, more states are excited and the band broadens as more lines emit. As the band broadens, the intensity of individual lines near the band center drop, while the intensity of the lines in the wings increases (Wormhoudt et al. (1985)). The broadening of the absorption/emission band greatly influences the spectral emissivity of the gas. An exponential wide-band model is used to demonstrate the relative influence of temperature, path length, and partial pressure on CO 2 emissions. Edwards and Balakrishnan (1973) developed a model based on rotational correlations for band absorption/emission of CO2 (and other gases) which demonstrated good correlation with existing data. The model specifically addresses the 4.3 ~tm CO 2 absorption/emission band. Figure 6 provides calculated data for the energy emitted by a fixed volume of isothermal gas (comparable to a finite element within the gas flow field). Each set of data represents the variation of one parameter (temperature, path length, or PCO2) while holding the other factors constant. As seen in the figure, the temperature of the emitting volume has significantly more influence than varying the partial pressure or path length. Because of the dependence on mass path length, varying the gas partial pressure from 0.1 to 1.0 atm had the same influence as varying the optical path length tenfold (i.e., from 1 cm to 10 cm). Hence, the radiant emissions of the jet are mostly influenced by

O.8

e~

':~

07

06

eo

o~

~5 05 t~

o.~ 0.3 ~ 0.2 / 0.1

o PC02=1.000 ~ PCO 2 = 0.467 o PCO 2 = 0,225

~

0

~

I

- O. 2

J

- 0.1

I

~

q~ ~ ^ . I

0.1

0

0.2

r / ( x - x O)

Fig. 4. Dimensionless irradiance profiles in the downstream region of series, 150 A demonstrating insensitivity to outlet CO2 conditions (TO= 376 K, Uo= 10.3 m/s, PCO 2 = 1.000 to 0.225 atm, x = 6.31 diameters from jet outlet) 1.0

~-

0.9 0.8 0.7 0.6

0.5 0./. 0.3 0.2 0.1 - 0.2

- 0.1

0.1

0

0.2

r/(x-x 0)

Fig. 5. Dimensionless irradiance profiles in the downstream region of series, 130A demonstrating self-preserving nature of irradiance profile (To= 414 K, Uo=20.8 m/s, PCO2 =0.500 atm) /,50

~

1.1

~ iip a r a m e

400 ~"

"

S

~

f .'~

~ 350

T-- 400 ~ Optical dis

ters

1.0 vo 0,9 0.8 x

=5 m

0.7

.~

PC02= 05atm

,', E

0.5 ~. o

o.~ r

,;' 250

0.3 ~

,;

300

,

i 2

Temperature Distance PCO 2 [] --& . . . . . -~--

i , i , i , I 4. 6 8 10 Emitted energy x 10.4 (W/cm 2)

,

0.2 C~ 0.1 0 0 12

Fig. 6. Influence of temperature, C O 2 concentration, and optical path length on energy emitted in the 4.3 p.m band width by a fixed volume of CO 2

414 1.0' 0.9 0.8

and temperature (C/Crn) (T - Tamb)

0.7

g-

(T m - Tamb)

0.6

-o

.g

o.s

Velocity (UlUm)

E 0./4

o Z

0.3 0.2

~.

0.1 '

00.

0.05

0.10

~==~--2~

0.15

0.20

0.25

r/(x -x 0)

Fig. 7. Dimensionless profiles versus similarity coordinate

70 . . . . . .

N

65

\

60

-~--o---o\

N"

/450 N

3.2 Axial profile

Temperature _ _

"~ \\\ \\

nk.

"%

9 Series 140 " Series140

/400

""

g

E ss "~

"~'

~

350

~ Irradiance ---O SerJes140A * Series 160A C3 Series 140 B " Series 160 8 300 D, Series 140 C Series 160 C .

/45 ~ ~ "~ A~^ ~ "~ ~ ~c3~

~'~ /40

35

~ E

250

25

' 0

I /4

2

I 6

I 8

,

I , 10 12

1/4

16

~ 18

200 20

Distance from outlet x/d

Fig. 8. Centerline axial decay of irradiance (air/CO2 jet, prior to subtracting background) and temperature (air only jet) for series 140 and 160

10

, []

190A

190B

190C

190D

9

9

0

[]

,/" El,,

/t

8 .....................................

/

.s'"

~ -,;" . . . . . . . //t

S.~ . . . . . /sss

//

, r

8 6 / / / " []

.e_

,d Transition points ...............

2

\

/'"

(~/s s ss S

/'o

9 ...-'

.,~'"

//"

,A-'""

9 .--"

.........

,/ ,.m" "__,,'_ . . . . . . ~,~A"-"'"'" ___~,~ oA"-.... . . . . ..h.......

e..,a .... ~-]~-"" ..~"7[.-"6"'" 2

/4

6

temperature which may explain the self preserving nature of the dimensionless irradiance profile. Figure 7 compares our measured, normalized irradiance profile to standard similarity profiles for velocity, temperature, and concentration (multiple references summarized in Blevins 1984). The figure shows the dimensionless irradiance profile lies inside the normalized temperature, velocity, and concentration profiles. One possible explanation is that the irradiance detected by the IR imager is essentially a line integral through the jet. The cooler regions in the jet's fringe areas contribute disproportionately less energy to the integral measurement than the hotter regions within the jet. Additionally, the irradiance profile may be narrower than the temperature and concentration profiles due to scattering within the jet, atmospheric humidity entrainment, and absorption within the jet by the cooler gas along the fringe area. Thus, the profiles of Fig. 7 appear reasonable.

8

10

12

1/4

16

18

20

22

24

Distance from jet outlet x/d

Fig. 9. Linear irradiance decay for series 190 (To= 351 K, Uo= 17.6 m/s)

Another important characteristic of the axisymmetric jet is the behavior along the jet's centerline axis. Previous research has shown that the jet's potential core will vary with initial temperature and velocity conditions (Prengle et al. 1985). The rate of decay of centerline temperature and velocity increases with decreasing initial velocity and increasing initial temperature. We found that the irradiance profile begins to decay immediately upon exiting the nozzle as shown in Fig. 8. This characteristic is notably different from the axial decay of temperature and velocity profiles that exhibit a constant value along the jet's centerline within the jet's potential core (also shown in Fig. 8). Again, the irradiance's immediate decay is most likely attributed to the line-of-sight, line integral nature of the irradiance measurement. The immediate changes in temperature and CO2 concentration in the shear layer mixing regions that surround the potential core have a direct impact on the irradiance measurement. Conversely, changes in the mixing region surrounding the potential core do not influence the temperature or velocity on the jet's centerline axis within the core region. The rate of axial decay also changes substantially from the jet's initial region to the fully developed. This is demonstrated most clearly by plotting the inverse of the centerline irradiance (1/1 m) as shown in Fig. 9. Interestingly, both the initial and fully developed regions exhibit a linear decay, but the rate of decay changes dramatically from the initial region to the fully developed region. The difference was again attributed to the predominant changes in the initial region being limited to the mixing region surrounding the potential core. Previous work has demonstrated proportionate decay characteristics in the fully developed region for temperature, velocity, and concentration (Chen and Rodi 1980; Becker et al. 1967; Prengle et al. 1985). However, proportionate irradiance decay is again unexpected given the non-linear nature of CO z emissions with temperature.

415 gion of the axisymmetric jet. The results of this experiment suggest that the irradiance detected using a narrow band infrared imaging system is also self-preserving in nature, although additional research is needed to consider higher temperature gases and binary mixtures. The method has the potential to overcome the classical limitations of thermal image/hot gas temperature correlation by use of dimensionless similarity relationships. Thus, dimensionless IR imagery may provide an alternate non-invasive method to monitor and detect changes in flow field temperature, velocity, or constituent distributions within a dynamic flow field.

< 3

._~

~2

O

............ ~

E 1

Downstream region

...................... ~

0 (15)

(10)

(5)

~

0

-

5

-

10

15

20

Distance from transition point (x-Xtr)/d

Fig. 10. Axial irradiance decay normalized using transition coordinates Most treatments of the jet's axial decay properties normalize the ordinate by dividing the centerline value by the outlet (or core region) value: U/Uo; (T-Tamb)/(To--T, mb); C/C o. Choosing the normalizing value for axial irradiance, however, is not immediately obvious since the centerline irradiance decay begins immediately upon exciting the nozzle. As shown in Fig. 9, a least squares curve fit (r 2 generally 0.96 or better) was determined for the initial and fully developed regions. The coordinates of the transition point from the initial to fully developed regions (Xtr, Itr) were used to normalize the data: ( x - X t r ) / d , and (I--lamb)/(Iu--lamb). Using the subtraction process resulted in l,mb = 0 and Eq. (3) was applied. This relatively simple method yielded data that collapsed neatly onto a c o m m o n characteristic curve as shown in Fig. 10. For comparison, data from several other series are also shown. Using this technique, one might expect the characteristic curve to pass through the coordinate (0,1). The slight offset seen in Fig. 10 is attributed to the transition from initial to fully developed regions being a finite region, vice an absolute point as we have shown in Figs. 9 and 10.

4 Summary and conclusion Previous research has shown similarity in momentum, energy, and constituent distributions in the fully developed re-

References Becker, H. A.; Hottel, H. C.; Williams, G. C. 1967: The nozzle-fluid concentration field of the round, turbulent, free jet. J. Fluid Mech. 30, 285-303 Blevins, R. 1984: Jets, plumes, wakes, and shear layers. In: Applied fluid dynamics handbook, pp. 229 278. New York: Van Nostrand Company Chen, C. J.; Rodi, W. 1980: Vertical turbulent buoyant jets: A review of experimental data. New York: Peragon Press Dowling, D.; Dimotakis, P. 1990: Similarity of the concentration field of gas phase turbulent jets. J. Fluid Mech. 218, 109-144 Edwards, D, K.; Balakrishnan, A. 1973: Thermal radiation by combustion gases. Int. J. of Heat Mass Transfer 16, 25 39 Flaig, J. 1976: IR flow visualization. Naval Air Systems Command, Department of the Navy, Washington, D.C. 20361, Final Report Volume No. 13, Report No. NAVAIR-13R-76 Gordge, D. 1989: Infrared thermography of a hot air/CO 2 axisymmetric jet (Master's Thesis). Texas A & M University, College Station, TX Hill, W. G.; Page, R. H. 1969: Initial developmet of turbulent, compressible free shear layers, ASME J. Basic Eng. 91, pp. 67-73 Merzkirch, W. 1987: Flow visualization, 2nd edition. New York: Academic Press Pai, S. I. 1954: Fluid dynamics of jets. New York: Van Nostrand Company Prengle, H. W.; Mahagaokar, U.; Shun-Kwok, T. 1985: Thermal and momentum structure of an emerging plume by remote sensing. In: Infrared methods for gaseous measurements (ed. Wormhoudt, J.), pp. 47-79. New York: Marcel Decker Schetz, J. A. 1980: Injection and mixing in turbulent flow. In: Summerfield (ed.), Progress in Astronautics and Aeronautics, Vol. 68, AIAA, pp.19 84 Wormhoudt, J.; Conant, J. A.; Herget, E H. 1985: High resolution infrared emission from gaseous sources. In: Infared methods for gaseous measurements (ed. Wormhoudt, J), pp. 1-46. New York: Marcel Decker