Kinetics of in-situ Combijstion

7/28/2019 Kinetics of in-situ Combijstion

1/323


2/323


3/323


4/323


5/323


6/323


7/323


8/323


9/323


10/323


11/323


12/323


13/323


14/323


15/323


16/323


17/323


18/323


19/323


20/323


21/323


22/323


23/323


24/323


25/323


26/323


27/323


28/323


29/323


30/323


31/323


32/323


33/323


34/323


35/323


36/323


37/323


38/323


39/323


40/323


41/323


42/323


43/323


44/323


45/323


46/323


47/323


48/323


49/323


50/323


51/323


52/323


53/323


54/323


55/323


56/323


57/323


58/323


59/323


60/323


61/323


62/323


63/323


64/323


65/323


66/323


67/323


68/323


69/323


70/323

4 . ANCILLARY EXPERIMENTAL RESlJLTS 53

54 , , ,I , , , , I I " ' I I " '

-+-- Nitrogen injected (Run H B 0 4 )- - +- Air injected (Run H B 0 3 )

53 - -

M 52 -d

-

3

Eati

$ 51:

ru0

M.3

-

50 - -

LTO HTO 1

49 I , , , / I / I 1 , , 1 , , 1 , 1 , , , 1 1 , I ,

100 200 300 400 500 6000Temperature, "C

Figure 4.2: Huntington Beach Oil - Sample Weight Versus Tempera ture (Ru ns HB03a nd H B 0 4 )


71/323

4. ANCILLARY EXPERIMENTAL RESlJLTS 54

I- Air injected (Run VEN13) Nitrogen injected (Run VEN 12)

M

3 6 1 -

a

5*

0

60 :

59 -

58 -

LTO HTO 75 7 1 l I l l I I , , , , , , , , , , , L-100 200 300 400 500 6000 Temperature, O C

Figure 4.3: Hamaca Crude Oil - Sample Weight Versus Temperature (Runs VEN12and VEN13)


72/323

55

51

50

49

OI)

ai' 48

2

3

8

5 47;s

cc0

Y

46

45

440

, , , I , , I , l , I ( I I I ,f I , I I , , ,

Nitrogen injected (Run VEN17)--+- Air injected (Run VENl8)

--

--

--

-

I , , I , , , I , / , , , ,I , , / , I

600

\ :-LTO HTo II

100 200 300 400 500Temperature, O C

Figure 4.4: Hamaca Crude Oil - Sample Weight Versus Temperature (Runs VEN17arid VENl8)


73/323

4 . A NCILLAR Y E XPE RIMENTAL RESULTS 56

4.3 DSC and TGA Results

Differential scanning calorimetry (DSC) and therrnogravimetricanalysis (TGA)

were performed on samples of Huntington Beach an d Harnaca crude oils. These

analyses were made at The Middle East Technical TJiiiversityin Ankara, Turkey. In

addition, analyses were conducted 011 mixtures of crude oil and sand, and mixtures of

crude oil, sand and clay. In these analyses, the sample mixtur e size was abou t 9 mg.

Air was injected at a constant rat e of 53 in1/11iii, while the tem per atu re was increased

at a fixed rate of 2"C/r1in. The results indicate that iiiclusion of sand or clay in the

crude oil samples had little effect on the main c11aracterisi;icsof the thermograiiis.

Consequently, only DSC and TGA therniograins for sainples containing crude oil and

sand are presented in Figs. 4.5 - 4.8.

Th e rat e of decrease in weight with temper atu re on the TGA thermogram indicate

three main temperature regions. In the range, 27"-28O"C, a steep rate of dec,rease

in weight was observed. This is mainly due to vaporization of the light hydrocarbon

fractions, as evident from the slightly endothermic reaction observed on the DSC

thermograms. A gentle rate of decrease in weight was measured in the tempera -

ture range, 280"-400"C, corresponding to the first exothermic reaction on the DSC

thermogram. In the range, 400"-500"C, the weight decreases greatly due to rapid

oxidation of the hydrocarbons. This corresponds to the second, highly -exothermic

reaction seen 011 the DSC thermograms, which have peaks at about 450C. The two

temperature peaks on the DSC thermograms are in line with the LTO and HTO

peaks observed in kinetic experiments on these crude oils (Chapte r 6).

The DTA thermogram obtained by Tadema (1959), for a typical crude oil with

air flow (Fig. 2 . l ) , also shows two main oxidation reaction peaks at abou t 270C an d

400C:.


74/323

4 . ANCILL ARY EXPERIMENTAL RESULTS 57

9 #

I

I

I - - - - - -.# .

c c -- ---- - --

-:.p.,

/ i/ I

F

00\o

0

0v,

00m

0

2

0


75/323

4 . ANCILLARY EXP ERIMENTAL RESIJLTS 5 8


76/323

4. ANCILLARY EXP ERIMENTAL RESULTS 59

I I I qO2 0 0 0 89 -.

I I m

\\

II

I II .

,I II II

I/

,I,

I

/ I

//

III

II

II


77/323

4 . ANCIL LARY EXPERIMENTAL RESULTS 60

I I

'II


78/323

4 . ANCILLARY EXPERIMENTAL RESULTS 61

4.4 Grain Sieve Analysis Results

In the inain, 20-30 mesh Otta wa sand was used in the experiments. However, to

investigate th e effect of surface area on oxidation reactions, 45-75 mesh and 170-270mesh sands were also used.

A 45- 75 mesh sand was used only in Run C4 (kinetic tube run with c arbon ). This

sand, retained 011 sieve No. 75 after passing through sieve No. 45, has a mean grain

size diameter of 0.0284 cni. Results of sieve analysis of 20-301 mesh and 170-270 meshsands are shown in Table 4.3 and Table 4.4 respectively.

4.5 X -Ray Diffraction Results

A sample of mo rta r clay used in the experiments was subjected t o X -ray diffraction

analysis to determine t he main types of clay present. The analysis was carried out a t

the Geology Department, Stanford University. The results (Fig. 4.9) indicate morta r

clay to consist of kaolinite, quartz and some illite.


79/323


80/323

4. ANCILLARY EXPERIMENTAL RESlJLTS 63

0

z0

$


81/323

5. Combustion Tube Experimen -

tal Results

The main objective of conducting conibustion tube experiments i n this study is

to compare the H/C ratios from combustion tube experiment,swith those obtained

from kinetic, experiments. Nonetheless, a complete analysis of combustion tube data

has been made and the results are presented in this chapter.

Six coinbustion tub e runs were performed. Th e crude oils used were Cold Lake

bi tumen (1 1.5"API),Huntington Beach oil (20.8"API)and Hamaca (Venezuela) cru'de

oil (10.2"API).Properties of th e sand packs for the tu be runs are shown in Table 5.1.

To allow direct comparison of results from different runs , the following conditions were

kept const ant: air injection rate and pressure were 3 L/min and 100 psig respectively;

concentration by weight of oil and water in a sample were 4.5- 4.9% and 4.1-4.3%respectively, and for sand was 86.8 -91 O%; and for runs using clay, the concentration by weight was 4.6%.

Runs H B 0 5 and CL15 were aborted because cornbustio-ticould not be sustained

after obtain ing initial ignition. Th e probable causes or this a re as follows. In RunCL15, clay wa s no t included in the sample. It appe ared th e fuel concentration was

insufficient to sustain combustion. Th e sample in Run HB05 contained clay. How -

ever, since th e crude gravity was 20.8"API,th e amount of fuel deposited was probably

insufficient for a self -sustaining combustion.

64


82/323

5. COMBUSTION TlJBE EXPERIMENTAL RESULTS 65

Table 5.1: Properties of Sand Packs for Combustion Tube Runs

Run No.

CL13 CL15 H B 0 5~

VEN14~

VEN21

Chide

Oil gravity (OAPI)Length (cm)Weight (g)Oil (wt. %)Water (wt. %)Sand (wt. %)Clay (wt. %)Sand fines (wt. %)+So (frac. pore vol.)Sw (frac. pore vol.)Sg (frac. pore vol.)q5 (frac. bulk vol.)

Cold Lake bi tumen

11.586.577954.64.186.84.60

0.290.260.450.31

Cold Lake bitumen

11.586.875794.84.391.0

00

0.270.240.4903 4

HuntingtonBeach oil

20.886.878 104.54.186.84.60

0.290.260.450.31

Hamacacrude oil

10.285.477394.64.186.84.60

0.290.270.440.31

Hamacacrude oil

10.287.1746 14.94.390.8

00

0.260.230.510.35

Hamacacrude oil

10.285.474474.64.1

86.80

4.60.260.240.500.33

t 170-270 mesh s and .

Da ta for the four successful runs are stored in a computer diskette and is available

on request from SUPRI-A. Results of the combustion tube runs are discussed in the

following sect ions.

5.1 Cold Lake Bitumen Run CL121

The sample c,onsisted of a mixture of Cold Lake bitu men , water, 20 -30 mesh

sand and clay. When the sand pack temperature across the electric igniter reached

330C, air iiijection was initiated. The electric igniter was turned off 25 minutes after

commencing air inject ion.

A fairly stable burn was obtained within an hour of ignition as evident from the


83/323

5 . COMBUSTION TlJBE EXPERIMENTAL RESULTS 66

fairly stable produced gas readings (Fig.5.1). In the period 1 -7 hours, the averagemolar concentrations of the produced gases were: CO2, 11.1%; CO , 5.0%; 0 2 , 1.8%and N2, 81.1%.

Produced gas readings (except for nitrogen) were oscillatory. Produced oxygen

readings varied between 1 to 5% during the period 1-4.5 hours, but therafter oscilla -

tions diminished to within 1 to 2%. Carbon dioxide and carbon monoxide readings

varied in step, mirror imaging those of oxygen. Combustion t ub e experiments using

identical apparatus with 11.2"APISail Ardo crude oil (Fassihi 1991) also yielded un -

steady produced gas readings. IJsing the same apparatus, Holt (1992) also obtained oscillatory produced gas readings for 12"API Cymric heavy oil. Produced oxygen

readings varied between 1 to 5%. However, in runs where i:ron nitrate was added to

the sample, produced gas readings were steady, and the produced oxygen readings

varied only slightly in the range 1 to 2%. Kinetic tube ex:periments performed on

18.5"API Huntington Beach oil (De 10s Rios 1987) showed a significant reduction

in the activation energy for high -temperature oxidation (13TO) from 128 K J/mo l

for runs with no metallic additives to 109 KJ/mo l for samples containing zinc or chromium. The results of De 10s Rios and Holt indicate that certain metals increase

fuel reactivity and thus coiiibustion stability.

In comparison, produced gas readings were less oscillatory for Run VEN5 (Hamaca

crude oil with clay) as described in the next section. Based 011 kinetic experiments

(Table 7.2), HTO activation energy for samples identical to those in Runs CL13

and V E N 5 are 219 KJ/mo l and 150 K J / m o l respectively. T h e results indicate that

co~iibust ionis more stable for a crude with a higher fuel reactivity. The decrease in gas

reading oscillations in the late r half of the experiment was probably due t o increase

in temperature of the insulation jacket thereby permitting increased fuel oxidation.

Apparent H/C and m -ratios based on produced gas analysis data are presented

in Fig. 5.2. In the period 1-7 hours, the average apparent H,IC and rn-ratios are 1.60


84/323


85/323

5. C OMBlJSTlONTlJBE EXPERIMENTAL RESULTS 68

and 0.311 respectively. By comparison, th e atomic H/ C ratio for the original crude

is 1.53 based on elemental analysis (Ch apter 4). The s imih r i t y in H /C ratios of thefuel and original crude oil indicates that there was practic:ally 110 low-temperature

oxidation in Run CL13, and that distillation is chiefly responsible for fuel deposition.

Temperature profiles are presented in Fig. 5.3. The avera,gecombustion zone tem -

peratu re was 500 "C. This relatively high combustion zone temper atu re is also indica -

tive of the absence of low -temperature oxidation. Visual inspection of t he sand pack

at the end of the run indicated a coriibustion zone thickness of 1.5 cm. Penberthy

and Ramey (1966) inferred from their experiinents a combustion zone thickness of about 2 cni. From the slope of th e straight line drawn through th e combustion front

location data (Fig. 5 .4 ) ) th e combustion front velocity is calculated t o be 11.2 c m / h r

(0.37 f t / h r ) . From Eq. 2.10, the coriibustion front velocity is directly proportional

to the air flux and inversely proportional to the product olf fuel concentration and

air -fuel ratio. S i m e air flux was ~ o n s t a i i t ,the constant combustion front velocity

obtained implies a constant fuel concentratioii-air/fuel ratio product.

Cumulative oil and water voluiiies and produced oil gravity versus time are pre -sented in Fig. 5.4. Produced oil gravity increased to as high as 16.6"APIcompared

to 11.5"API for the original crude. Similar increases in procluced oil gravity were

observed in the South Belridge in -situ combustion project: 18"API compared to

12.9"API for the original crude. Produced oil gravity increases due to increasing

light hydrocarbon content of the crude as a result of distillation ahead of the com-

bustion zone. The heavy frac tions left as residue consitmutethe fuel burned at the

conibust ion zone.

Viscosity of the produced oil decreased significantly with time as a result of the

increasing light hydrocarbon content of the oil (Fig. rj.5). At the end of the run, the

produced oil viscosity was 70 cP a t 35C compared to 10,000 c P for the original crude.

Results of this tube run and those of Haiiiaca oil are sumiiiarized in Table 5.2.


86/323

5. C O M B l J S T l O NTlJBE EXPERIMENTAL RESULTS 69

0ye.I - UI

c'!x 0 09 09 \9I I I I I0c'!3

P

I I I l o o


87/323

5. COMBUSTION TUB E EXPERIMENTAL RESULTS 70

I I I I

P

I I I I I0 3 00 00 0 00 0\D v, 2m m


88/323

5. COMBIJSTION TU BE EXPE RIMEN TAL RESULTS 71

2 2 2 0 0 W * N 000 \D- 30NI I I I I I I I T W

III

0\\

- \

r;: \\1

\\

\\

\\\

\\\\\

\\

\\

\\\

\

\\

\\\

\\

\\\

\\

\\

\\

b\

h\

\a

\o\

\\

\

Q\

\I I I I & I I I I \I

o\

\-3

2 0 x 2 2 0 80\90r.09M(pampold i e i o i30 u o p m q ) u o g m p oi d a A p q n u n 3

1 I I 1 I

I8

u13 aqni30 doi u o y iuo q uoyisnquo330 a x m i s r a

0 0 0 0 W * Nw


89/323

5. COMBIJSTION TIJBE EXPERIMEN TAL RESULTS

a0c,r(

v10

b>r(

Temperature, O C

Figure 5.5: Oil Viscosity Versus Temperature (Run CL13)


90/323

5 . COMBUSTION TU BE EXPERIMENTAL RESULTS

Table 5.2: Summary of Combustion Tube Experimental Results

Initial oil gravity ( "A 1Initial oil viscosity ( c P )

Injection pressure (psig)Air injection rate (L /min)Air flux (scf /hr-f t2)

Comb. front temp. ( " C)Comb. front velocity (cm/hr )

Produced liquids:Oil gravity ("API)Oil vicosity (cP)

Produced water vol. (1111)Produced oil vol. (ml)Oil recovery (wt . %)

Produced gas:Ave. prod. rate (L /miu )COa (mole %)CO (mole %)0 2 (mole 9%N2 (1110le %)

ni -ratio

Apparent H/C ratioResidue after burn (g)Fuel c011c. (lb/ft3 bulk vol.)O2 utilization efficiency (%IAir -fuel ratio (scf/lb u e l )

Run No.

CL13 VEN5

11.510,000

at 35C1003.00130.9

50011.2

16.670

at 35C39226371

3.0911.15.01.8

81.1

0.311

1.6033.01.14991.7165

Air requirements (scf/ft3 bulk vol.)Heat of combustion (Btu/lbj,,l) 16,253

208

10.214,000

at 50C1003.00130.9

50010.5

14.0260

at 50C35328778

2 8011.64.91.1

81.4

0.298

1.6328.70.78495.0167139

16,458

VEN14 VEN21

10.214,000

at 50C1003.00130.9

3507.7

9.817,000

at 50C28534694

2.554.32.010.182.8

0.312

4.35143.9

54.5

-

-

-

-

10.214,000

at 50C1003.00130.9

50011.1

11.22600

at 50C27129888

3.059.74.33.8

81.2

0.308

1.778.6

0.53982.6169112

16,773

73


91/323

5. COMBUSTION TUB E EXPERIMENTAL RESULTS 74

5.2 Harnaca Crude Oil Run VENli

Th e sample consisted of a mixture of Hainaca crude oil, water, 20-30 mesh sand

and clay. Air injection was initiated when the temperature of the sand pack across

the electric igniter reached 330C. The electric igniter was switched off 30 minutes

after air injection started.

Stable combustion was observed almost from the start of the run as indicated by

th e stable produced gas composition readings (Fig . 5.6). In. the period 1-7.5 hours,

the average c,onceiitrationsof the produced gases were: CO;?,11.6%; CO, 4.9%; 0 2 ,

1.1% an d Nz, 81.4 %. Oscillatory produced gas readings observed in Run CL13 wereabsent in this run. As explained in Section 5.1, this result is probably due to Hamaca

crude oil being more reactive th an Cold Lake bitumen.

Apparent H / C and ~ n - r a t i o s based on gas analysis are presented in Fig. 5.7. Th e

average apparent H/ C and 77.2-ratiosin the period 1-7.5 hours are 1.63 and 0.298

respectively. Based on elemental analysis (C hapter 4) , the atomic H/C ratio of the

original Hainaca crud e is 1.65. Th e fact tha t H/C ratios of th e fuel and original crude

are alinost the same in this run indicates the absence of lo w-temperature oxidation

and the predominance of distillation as the fuel deposition mechanism.

Th e average c,ombustioiitemperature was 500C (Fig. 5.8). From cornbustion front

versus time data (Fig. 5.8))combustion front velocity was 10.5 cm /hr (0 .35 f t /h r ) .

Figure5.9 shows cumulative volumes of produced wa ter and oil and oil grav ity

versus time. Produced oil gravity at th e end of th e run was 3.8"API higher tha n tha t

of the original crude. Viscosity of the produced oil dropped to 260 cP a t 50C from

its original value of 14,000 cP (Fig. 5.10).


92/323

5. COMBUSTION TlJBE EXPERl MENTAL RESULTS 75

00 052 * P4 0

I I I I

..........................

z :


93/323


94/323

5. COMBUSTION T UBE EXPERIMENTAL RESULTS 77


95/323

5. C O M B I J S T I O NTIJBE EXPER IMENT AL RESULTS 78

00 W * N 0$ 1 20 Wz d 3I I I I I I I 1 - m

h

5

IIIII1I1I1I\11III1

I1

I\I

II\

1

I \1

\

\\

\\

\b

\

\ \

I I I I II

0 0 0s * N0u13 'aqni 30 do1 ~ 0 . 1 31.10.13uopsnquro3 30 a m e i s r a


96/323

5. C O M B I J S T I O NTlJBE E XPER IME NTA L RESULTS 79

Figure 5.10: Oil Viscosity Versus Temperature (Run VEN5)


97/323

5. COMBUSTION Tl JBE EXPERIMENTAL RESULTS 80

5.3 Harnaca Crude Oil Run VENlL4

In this run no clay was added to th e sample. During the first two hours of the

run , produced oxygen concentration varied, increasing t o as high as 13%. The electric

igniter was left 01 1 for two hours into the run t o aid ignition. However no ignition was

achieved. This run was a low -temperature burn as indicated by the temperature and

gas analysis data.

Produced gas composition readings stabilized after about three hours (Fig.5.1 1) .

In the period 3-9 hours, produced gas molar concentrations were: CO2, 4.3%; CO,

2.0%; 0 2 , 10.1% and N J , 82.8%, Th e average apparent H / C and rn-ratios were 4.35

an d 0.312 (Fig.5.12). Given the atomic H/C rati o of 1.65 for th e original crude,

the high apparent H /C ratio observed during the run indicates tha t low -temperature

oxidation was the main oxidation reaction.

Temperature profiles ar e shown in Fig. 5.13. The average combustion zone tem -

perature was 350C.A kinetic t ube experiment (Ru n VEN15,described in Chapter 6)

was performed on a sample identical to tha t used in this tube run: 350C corresponded

to the temperature at the saddle between low- and high -temperature oxidation peaks.

Therefore it is concluded that only low -temperature oxidation occurred during this

tub e run. Combustion front location d ata are shown in Fig. 5.14 and indicate a low

burning f ront velocity of 7.7 c m / h r (0.25 f t / h r ) compared to about 11 cm/hr for Runs

CL13 and VEN5.

Produced oil gravity and viscosity data are presented i n Figs.5.14 and 5.15 re-

spectively. Produced oil gravity (9.8" API) was slightly lower than that of the orig-

inal crude (10.2" API). Produced oil viscosity was 17,000 C P at 50C compared to

14,000 cP for the original crude. The decrease in oil gravity and increase in oil vis-

cosity were due to low -temperature oxidation. Alexander e t aZ. (1962) also observed

an increase in viscosity of oil subjected to low -temperature oxidation. Bae (1977)


98/323

5. COMBUSTION TlJBE EXPERIMENTAL RESULTS 81

000

0\o 0

0

(o/o a~0U.I) a%Lxo SapIxo u o q n 3


99/323

5 . COMBIJSTION T lJBE EXPERIMENTAL RESULTS 82

Y

Ea

2'i-i2

..2


100/323

5. C O M B I J S T I O NTlJBE EXPERIMENTALRESrJLTS 83

I I I I I0 0 0 30 0\D vr z09 m c\1


101/323

5. COMBUSTION TrJBE EXPERIMENTAL RESULTS 84

I \I I I I AI I I- I \I

w

0 8 ? vs 0 x 0 0$2 * N

c!9 r-s 0 09 +(pampoid p o i 0 u o p x y ) u o p m p oi d aAy q nu rn 3

I I I I Id

ur3 aqni 30 doi uroi3 iuoi3 uogsnquro330 a 3 u q s r a

0 0 0


102/323

855. COMBUSTION TrJBE EXPERIMENTAL RESULTS

1o6

1o5

%s

til

YM

wl0

T

1o4

1o3

- Original crude- Sample at 5.4 hr -+ Sample at 7.5 hr - Sample at 9.8 hr

20 30 40 50 60 70 80Temperature, O C

Figure 5.15: Oil Viscosity Versus Temperature (Itun VEN14)


103/323

5 . COMBUSTION TUBE EXPERIMENTAL RESULTS 86

also found a decrease in API gravity of oil which had undergone low -temperature

oxidation. Combustion did not occur in this tube run because fuel concentration waslow, for reasons described in Section 5.4.

After th e tube r un, the sand pack was examined. The b~urnedzone was black in

color as opposed t o dark grey in the othe r three runs. Samples of this burned zone

were fired in an oven. An average sample weight loss of 2.170was observed compared

to 0.1-0.5% for the other three tube runs (Table 5.2). These results indicate thatsome heavy hydrocarbon residue was left unburned in this run .

The percentage of oxygen used in LTO is given by Eq. 5.1 (Rarney et al . 1992).

Percent oxygen in LTO = 100 ( Z a p p a r e n t - .e) - CO,+C O )(5.1)4 (0.2682N2- O z p )

Based on Eq. 5.1, 35% of oxygen injected went into LTO and 46% of oxygen injected

was produced and did not generate heat. Thus only 19% went into HTO reactions.

T h e result of this operation was a slight decrease in API oil gravity and a large

increase in viscosity of the produced oil. A material balance on oil indicated 4:3% wasdeposited as an immobile residue and 57% was displaced with no or little improvement

in quality.

The oxygen-carbon ratio, y, has been estimated, assuming all produced carbon

oxides are the products of HTO reactions. Th e mass of fuel consumed per second

in HTO reactions is qo(CO+ C 0 2 ) ( 1 2+x ) / ( 6 0 x 22.4138),as derived for Eq. 7.2 inChapter 7. The duration of th e tube ru n was 593 minutes. Thus:

qo(CO +CO,) (12 + X ) x 593 x 6060 x 2!2.4138Mass of fuel in HTO reactions == 57.8 g

using dat a on Table 5.2 and 1c = 1.6. Based 011 mass conservation of t he hydrocarbon

fuel, the initial mass of oil in the sand pack is equal to the sum of the mass of oil


104/323

5 . COMBUSTION TlJBE EXPERIMENTAL RESULTS 87

recovered plus the unburned hydrocarbon residue plus the fuel in HTO reactions, as

follows:

367.7 = (346.9 + 1 4 3 . 9 ) ~t 57.8

Tha t is:

y = 0.63

This estiiiiatedvalue of y , 0.63, is considerably larger th an that obtained by elemental

analysis, 0.25, of a produced oil sample from this tube run (Section 4.1). However,

based 0 1 1 gas analysis in kinetic tube Run VEN15 (Table 6.7), the oxygen-carbon

ratio is 0.5; a closer match.

Elemental analysis of a produced oil samp le from Run VEN14 confirmed t he oxy-

genation of oil as a result of LTO. Clearly LTO should be avoided in field operations

by regular monitoring of the app arent H/C ratio. Combustion tu be Run VEN14 is

one of the most important results of this study.

5.4 Hamaca Crude Oil Run VEN21

Th e sample consisted of oil, wate r, 20-30 mesh sand a nd 4.6% by weight of 170 -270

mesh sand. One hour after commencing air injection, the igniter was turned off.

A fairly stable burn was observed. Th e produced gas compostion readings were

however oscillatory (Fig. 5.16). In the period 1-7 hours, th e average values were: COZ,

9.7%; CO, 4.3%; 0 2 , 3.8% and N2, 81.2%. Apparent H/C and rn-ratios averaged 1.77and 0.308 respectively (Fig. 5.17). Th e average cornbustioti zone temperature was

500C (Fig. 5.18). The combustion front velocity averaged 11.1 cm/h r (0.36 f t / h r ) .

Produced oil gravity increased by 1 " API while oil viscosity decreased to 2,600 CP a tc ~ Oo 1 from 14,000 CP for the original crude (Figs.5.19 and 5.20).


105/323

5. C O M BUSTZON TlJBE EXPERIMENTAL RESULTS 88

Except for tlie oscillatory gas composition readings, results of this tube run were

similar to that of Run VEN5 in which clay was included in t he sample. Voussoughie t al . (1982) inferred from combustion tube experiments t81iat clay did not have a

catalytic effect on combustion. A possible effect of clay and fine sand on combustion

is as follows. Both clay and 170-270 mesh sand particles axe smaller than those of

20-30 ~ i i e shsand. It is conceivable therefore that these smaller particles increase

oil entrapment and thereby increase fuel concentration. Oil entrapment may be the

result of permeability reduction and t he greater surface area by these smaller particles.

Similarly, with no clay or sand fines, fuel concentration is decreased. This may lead to low -temperature burns as observed in Run VEN14. Future research is suggested

to investigate the relative effect of clay and sand grain size 1011 fuel concentration.

Typical products of thermal cracking of crude oil are hydrogen and short chain

paraffins, e.g. methane (Burger e t al. 1985). In all tube runs, hydrogen an d methane

were not present in the produced gas, based on gas chromatograph ineasurenients.

The absence of these gases, a n d thus thermal cracking, indicate that distillation is

the main mechanisinfor fuel deposition.In th e next chapter, results of kinetic experiments will be presented. In part icular

we will see whether a comparison can be made between kinetic and combustion t ub e

experiments for identical samples.


106/323

5. C O M BCISTION TU BE EXPE RIME NTAL RESULTS

0

3 0CI 0

89


107/323

5. C O MBlJSTIONTUBE EXPERIMENTAL RESULTS

I +

m 3

90


108/323

5 . COMBUSTION TUBE EXPERIMENTA L RESULTS

I I I I

0 0 0 0 3

w Ti- m m

L\D 2

91


109/323

5. C,OMBIJSTIONT IJBE EXPERIME NTAL RESULTS

2 0 0 W * m 000 W m3 3 s +I I I I I I I I T o o

\\

I

I I

\ I I

I

I

\ I

\ I P

I

\ I

\ F ) ;

II

\o\

\ \

92


110/323

5. C ~ O M B l J S T l O NTUB E EXPERIMENTAL RESULTS

1o6

1 5

1o4

io3

93

1Original crude- Sample at 5.1 hr - -+- Sample at 6.1 hr- nmnle a t f ; f ; hr

I I I I I I I I I I l a

20 30 40 50 60 70 80Temperature, "C

Figure 5.20: Oil Viscosity Versus Temperature (Run V E N 2 l )


111/323

6. Kinetic Experimental Results

The original objective of performing kinetic tube experiments was to study the

rate, order and riiechanisrns of reactions during cornbustion of crude oil. However

based 011 data from initial kinetic tube runs, two puzzling results were obtained.

1. During high -temperature oxidation (HTO), apparent H /C ratios were typically

in the range 0 to 1 compared to values of 1 to 2 based on combustion tube

experiiiient s (Chapter 5).

2. An Arrhenius graph of HTO data did not yield a stra.ightline as pred icted by

a previous oxidation reaction model (Fassihi 1981).

Using the same experimental technique, Burger e t al . (1985) and Fassihi (1981)

also found apparent H/C ratios between 0 and I for crude oil in the HTO range.Fassihi (1981) obtained the following apparent H/C ratios at' the HTO peaks: 0.3

(Huntington Beach oil), 0.2 (Venezuela J o b 0 crude oil) and 0 . 1 (San Ardo crude oil).

Burger et al. (1985) attribu ted the low ap parent H / C ratios t'o th e fuel being made u p

of heavy oil fractions. Since heavy crude oil is a mixture of aromatic and saturated

hydrocarbons, whose limiting atomic H/C ratios are between 1 and 2, a different

explanation must be sought.

Figure 6.1 presen ts a correlation of a ton i c ,H/C ratio as a function of oil gravity

an d Universal Oil Products (UOP) K -factor. This graph is based on the Hougen

and Watson correlation charts for petroleum hydrocarbon properties (Hougen e t al .

94


112/323

6. KINETIC EXPERIM ENTAL RESlJLTS

9I $1

94

4

I

I

\

95


113/323

6. KINETIC EXPERIMENTAL RESULTS 96

1954). T he Hougen and Watson c-orrelationcharts represent i t wide range of pe troleum

hydrocarbon types including crude oil distillation fractions with atinosplieric boiling po ints of 1000F. Figure6.1 is a variation of a correlation presented by Lirn (1991).

T h e IJOP K -factor characterizes a crude oil based 011 the cube root of the mean

boiling po in t. A curve fit of UOP K -factor wa s produced as a function of specific

gravity and kinematic viscosity:

7.78 8.24- - ] SG8.24 +- (6.1)12.758 - (2.17 + 1% ( I . / P >(2.87A - SG)(2.87A - )S GIJOP K -factor =

where:S G = oil specific gravity (6O0F/60"F),

A = 1 +8.69 log [(T +460)/560] ,8 = 1 +0.55410g[(T +460)/560],p

p = viscosity ( c P ) a t T ,

T = temperature, O F .

I := density (g/cc) a t T = S G / [I +0.000321(T - 60) x 10.00462AP'

Equation 6.1 was then used to produce the atomic H,/C ratio from a Hougeri

and Watson correlation of weight fraction of hydrogen. Th e Hougen and Watson

correlation wa s tested against measured atomic H/C ratios for 23 crude oils whose

oil gravities ranged between 10" and 36"API.Th e results are presen ted in Table 6.1.

Measured atomic H/ C ratios a n d those obtained from the IJOP K -factor correlation

are plotted on Fig. 6.2 for comparison. Th e straight line in Fig. 6.2 represen ts a perfect

iiiatch between the tiieasured da ta a nd H/C ratios from the correlation. A t o n i c H / C

ratio data for crude oils agree reasonably well with H /C ratios obtained from the

Hougen arid Watson correlation.

Based 011 correlation of residuum at equivalent iiorinal boiling po int of 335C

(635F) versus original oil gravity (Lirn 1991): residuum gravity is 5"API for an


114/323

6. h ' INETIC EXPERIMENTAL RESULTS

Table 6.1: Atomic H / C Ratios From Hougen and Watson Correlation

Oil gravitySource (OAPI)

Alexander et al . 10.3(1962) 12.6

16.317.021.321.824.024.425.829.535.636.0

Wilson e t al .(1962)

12.922.0

24.028.730.6

Bousaid and 13.9and Ramey (1968) 22.1

Hvizdos e t a l . 10.0(1982)

Greaves e t al. 22.8(1989)

This study 10.211.5

Viscosity( 4

5,129625

358406

1243279416733

1,92225 1

34I2 7

9

3,480158

1,002

40

14,00010,000

Temp.(OF)

130130

303030303030

130130130130

100100

100100100

7777

185

100

12295

UOPK -factor

11.6211.58

11.8311.9111.4611.8911.9512.3011.9311.9512.0412.09

11.5912.13

11.8612.5911.97

11.6311.94

11.68

11.80

11.6611.75

A1;omic H / C Atomic H / C(correlation) (measured)

I .461.50

1.601.621.621.721.761.811.791.841.921.94

1.521.57

1.661.551.641.631.701.621.661.761.871.79

1.50 1.571.75 1.68

1.74 1.651.94 1.871.85 1.73

1.53 1.561.73 1.65

1.46 1.48

1.71 1.54

I .46 1.651.48 1.53

97


115/323

6. KINETIC EXPERIMENTALR E S t J LT S 98

- e Alexander et al. (1962)rn Wilson et al. (1962)

0 Hvizdos et al. (1982)- 0 Greaves et al. (1989)

-2.2 -

A Bousaid and Ramey (1968)

-2.0 - A This study -

-5 -; 1.8 -cd -

-W -

-.3- 1.6 -2 -

---

---

-

-

-

---

0.8 1 o 1.2 1.4 1.6 1.8 2.0 2.2 2.4Atomic H/C ratio (Hougen and Watson correlation)

Figure 6.2: Measured Atomic H/ C Ratios Versus Values Derived From Hougen andWatson Correlation


116/323

6. KINET IC EX PER IME NTA L RESULTS' 99

original oil gravity of 10"API. Figure 6.1 indicates H/C ratios greater than 1 even for

an unrealistically low fuel gravity of 0"API.Fassihi (1981) obtained the atomic H / C ratios of distilltationcuts of Huntington

Beach oil based on elemental analysis. The atomic H/C ratio decreased from 1.95 for

a distillation cut at 150C (80 psig) to 1.5 for a distillation cu t a t 5*50"C,compared to

1.64 for the original oil. Fassihi's results indicate th a t fuel b-urnedduring combustion

would have atomic H/C ratios slightly lower than those of the original crude, as

typically observed in combustion tube experiments.

The apparent H/C ratios observed during HTO from kinetic tube experimentsare much lower than the lowest H/C ratios on Fig. 6.1. This fact constituted a main

reason for performing kinetic tube experiments in this study. If atomic H/C ratios

and thus fuels oxidized in kinetic and combustion t ub e experiments are different, then

kinetic experimental results do not reflect oxidation reactions in the combustion zone.

To resolve this puzzle an d other problems related t o th e previous kinetic oxidation

experiments, eleven kinetic experiments were carried out using three crude oils: Cold

Lake bitumen, Huntington Beach oil and Hamaca crude oil. In addition, three runswere made using carbon (Reagent grade, 60 -mesh) an d a mixture of carbon and Cold

Lake bi tumen. The conditions for these runs are listed in 'Table 6.2. Unless stated

otlierwise, the samples consisted of crude oil and 20-30 mesh Ottawa sand. For some

runs (as indicated) the sample matrix contained clay or 170--270 mesh sand. Air was

the usual injection medium. However, in three kinetic runs, nitfrogenwas followed by

air, or vice versa, as indicated on the table.

6.1 Kinetic Tube Experimental Results

Seven kinetic tube experiments were conducted in wliich air was flowed through

an oil-sand mixture. Composition of the produced gas and tempera ture da ta are


117/323

6. KINETIC EXPERIM ENTAL RESIJLTS 100

Ru n No.

Table 6.2: ExperiinentalConditions for Kiiiletic Runs

Crudeoil or fuel*

a**c 4 * * *

c:Lc:1

CL2CL5

CL14

H B 0 2

VEN6VEN7;

VEN 15

VEN23

CLlOt

VE N lo t

V E N ~ S ~ ~

c(:

C: LB+ :

C L BCL BCLBCLB

HBO

HCOHCOHCOHCOHCOHCO

(:;asAow-rate,

L/min

0.700.70

1.08

1.100.950 .300.70

1.09

0.700.700.700.700.700.70

Pres-sure,psig

-

100100

85

808595

100

85

100100100100100100

Temp.rate,

OC:/hr

5050

50

5050

150 50

50

505050 50 5050

Samplelength,

cm

5.66.0

6.2

5.35.55.45.0

5.4

5.55.51.84.94.34.3

Sample Initial weight %weigh 1,,

g Sand

52.3635 94.352.3505 93.9

53.8340 91.6

53.2740 92.952.8474 92.353.0834 91.652.6357 86.8

53.6500 92.3

57.7968 86.857.5726 86.817.2630 86.849.8876 90.844.4586 90.835.4507 86.8

* C = Carbon, (:LB = Cold Lake bitumen, HBO = Huntington Beach oil,HCO = Hamaca crude oil.** 20-25 mesh sand.*** 45-70 mesh sand.t Nitrogen injected un ti l en d of LTO, then air injected.t t Temperature decreased on reaching 35OOC.

Air injected until end of LTO, then nitrogen injected.O Includes 4.6% by weight of 170-270 mesh sand.

Clay-

00

0

000

4.6

0

4.64.64.6

00

O 0

Oil/ Fuel Water

1.9 3.81.9 4.3

2.9 4.0

3.0 4.13.8 3.94.1 4.34.6 4.1

4.1 3.7

4.6 4.14.6 4.14.6 4.14.9 4.34.9 4.34.6 4.1


118/323

6. KINETIC EXPERIMENTALREStJLTS 101

shown in Figs. 6 .3 - 6.9. The following observations were mi%dt:.

0 The oxygen consumption da ta indicates two oxidatioin reactions. Th e first re -

action (LTO) starts at about 170C and has a peak at about 250C. This is

followed by a second oxidation reaction (HTO) which has a peak at about

400C.

0 Temperature of the kinetic tube was programmed to increase a t 50"C/hr. How-

ever, temperatures at the LTO and HTO peaks were higher than programmed

temperatures, indicating both oxidation reactions to be exothermic.

0 Oxygen consumption during LTO increased significantly if the sample matrix

contained clay (Table 6.3, page 109). For example, consider Hamaca oil Runs

VENG (Fig. 6.7), VEN15 (Fig. 6.8) and VEN23 (Fig. 6.9). In Run VEN15 (no

clay), the ratio of th e oxygen consumption peak during LTO to that during HTO

was 0.11. However in Run VENG (with clay), th e rat io increased t o 0.90, while

in Run VEN23 (with 170-270 mesh sand ), th e ratio was 0.18. Clay increased

oxygen consumption during LTO.

0 Th e initial amoun ts of oil in th e saniples were: 2.66 g (Run VENG), 2.44 g (Run

VEN15) and 1.63 g (R un VEN23) . Given th e same air injection rate , 0.7 L/min ,

the oxygen consumption peak during HTO was higher in Run VEN15 (14.0 %)

- with less oil - than in Run VEN23 (6.8 %). The sample in Run VEN15 ha d

50 % more oil than in Run VEN6. However, the HT O peak in Run VEN15

was about three times higher than that in Run VENG (4.6 %). An increase in

oxygen consu~npt ionduring LTO resulted in a decrease in oxygen consumption

during HTO.

These perplexing observations appear to be a result of changes in t he fuel during

LTO, and are explained in Section 6.7.


119/323


120/323

6. KINETIC EXPER IMENT AL RESULTS

0 0 0 (300 0 0 0rr, * m NI I I I

2 08W

W rr, d- m P 4

(% qow) ua8Axo 'SapIxo u 0 q . q

W

vr

d-

.2

.r(

I+

m

N

103


121/323

6. IrlNETIC EXPERIMENTALRESrJLTS

0 0 (3

8 8 13 0800 0

W v, *I I I I I

/s?

I

I1

+

+.I fI I 7..v, m N 3

(% qouI) ua%dxo sapyxo u o q n 30 -

I I I I I d0 0 0 0z W * N02

104


122/323

6. KINETIC EXPERIM ENTAL RESULTS l0rj

(% a1ouI) ua8Axo sapyxo u o q n 3

I I I I I I0 0 0 0W * c\I0 0 002


123/323

6. KINETIC E XPERIME NTAL RESlJLTS

0 0 0

* z N z 08I I I

000m

I

106


124/323

0. KINETIC EX PERI MENT AL RESIJLTS 107

0

t I I1 1 1 I2 0

8% m r-40 0

08rr,

0P4

++ + *

+ \+++ +

B+

+++++

/t=

+++*%a

++

I I 30

I I 1 I

2

I I

0 03 N0\D0


125/323

0'. KINETIC EX PERI MEN TAL RESIJLTS

0 0 0 (30 0

2 00 0 0 0vl e m rn

I I I I

++

;?

(% qow) ua%xo 'saprxo u0q.q)

I I I I I I0 0 0 0\o e N

0

2

10 8


126/323


Table 6.3: Oxygen CJonsumptioi~Peaks at LTO a n d HTO

Run No.

Sample m atrix(i n addition to

20-30 mesh sand)

Oxygen consum p tion peak (mole %:I

LTO (a ) HTO (b:)

CL2C L 5CL14HB02VEN6VEN15V E N 2 3

-

clay

clay

170-270 mesh sand

-

-

0.790.913.230.944.091.501.22

2.9ti5 .323.352.094-57

14.026.81

0.270.170.960.450.900.110.18

The apparent atomic H/C and rn-ratios were calculated from gas composition

data using Eqs. 2.7 and 2.6. Th e results ar e presented in Figs.6.10 - 6.16. Based 011

the apparent atomic H/C an d m-ratio graphs , th e following observations were made.

0 The apparent H/ C ratio increased to values ranging from about 15 to 40 at the

LTO peak temperature, indica ting that a large amount of oxygen entered into

LTO reactions which did not produce carbon oxides.

0 Fairly coilstant apparent H/C ratios were observed following the first oxygen

co~isumpt ion peak . The HTO reaction may be considered to be the oxidation

of a fuel consisting of a hydrocarbon with a particular H/C ratio.

0 The apparent H/C trends of these runs suppor t the conclusion th at there are two

inain oxidation reaction ~ ~ i e c h a n i s m s :oxygen addition to the hydrocarbon with

little carbon oxide generation at low temperatures, follclwed by high tempera ture

oxidation of this fuel.


127/323


0 The ni -ratio, fraction of carbon converted to carbon rtionoxide, decreased from

about 0.4 for the LTO temperature range to about 0.3 for HTO. The m-ratiowas fairly c,onstant tliroughout the teii iperature range of these experiments.

This observation also indicated the existence of two oxidation reactions.

0 Apparent H/C ratios in the HTO temperature range were typically between 0

and 1 (Table 6.4). As discussed at the beginning of this chapt er, th is observation

posed a major problem because fuel atomic H/C ratios of 1 to 2 were indicated

by combustion tube results, elemental analysis and the Hougen and Watson

correlation.

Table 6.4: Average Apparent H/C and vi-Ratios a t HTO Temperature Range

Run No. Period (hrs) m-rat io Apparent H/ C ratio

CL2 5.33 - 6.33 0.260 0.71CL5 5.25 - 6.25 0.268 0.49CL14 4.20 - 5.30 0.266 0.00HB02 4.83 - 6.25 0.247 0.37VEN6 5.10 - 6.30 0.246 0.21

VEN15 5.30 - 6.80 0.288 1.02VEN23 4.80 - 5.70 0.276 1.04


128/323

6. KINETIC EXPERIMENTAL RESIJLTS 111


129/323

6. Ir ' INETlC EX PE RIM ENTA L RESULTS' 112

3

2 0 2 2 0 2 0 0 0I I

T-:I I I I I

61


130/323

0'. KINETIC EXPERIME NTAL RESlJLTS 113


131/323

6. KINETIC EXPERIMENTA L RESIJLTS 114

O *o @?9

ospoooO

I +

h

I II

I O n W -4

031

- 0 0 *8

a0 0 2

I I no' 0 0 v, 0


132/323


0 p . I - UI3

2 2 0 2 2 x 2 2 0 0t';3 I I I I I I I I T

II I


133/323

6. KINETIC EXPERI MENTAL RESULTS 116

O0%J 0 O n --a - w 8

t O @ @ *


134/323



135/323

6. KINETIC EXPERIMENTAL RESlJLTS 118

6.2 An Oxygenated Fuel for HTO

During LTO, oxygenated hydrocarbon products such as ketones, aldehydes, al-

cohols an d peroxides are formed (Burger an d Sahuquet 1972). Therefore these oxy-

genated products form part of the fuel that is oxidized during HT O in kinetic tube

experiments. The oxygen contained in these reaction produlcts should be included in

the calculation of H/C arid m-ratios.

The following analyses a n d experiments indicate that the fuel during HTO is a11

oxygenated hydrocarbon. This is an i i i iportant finding of this study.

1. Taking into consideration oxygen in the LTO products consumed during HTO

reactions, th e appa rent hydrogen -carbon ratio, s , was coiiiputed on a total-gas-

volume basis. That is, following Eq. 2.7:

where the gas efflux rate, qo, is a function of time, an d the integration limits

are froin the start of LTO to the end of HTO.

Apparent H / C ratios based 011 Eq. 6.2 were compared with those obtained

from combustioiitube r u m . For direct compariso~i ,identical samples were used

in each set of kinetic arid coriibustion tube expe r i~nen ts(Table 6.5). Th' S was

achieved by preparing samples sufficient for both experimeiits. Kinetic exper-

iriieiit samples were placed in a deep-freezer for subsequent runs. In addition,the same air injection pressure (100 psig) was used in both types of experiments.

The co~ribustionzone temperature in Run VEN14 was 350C. This tem -

peratu re condition was approximated in the corresponding kinetic tube exper -

iment (Run VEN19) by increasing the te mperature of the kinekic tube to a


136/323

6. Ir ' lNETIC EXPERIMENTAL RESULTS

Table 6.5: Kinetic an d Combustion Tube Experimental Se ts

Run No.

119

Crude oil Mixture Combusliion Kinetic-

Cold Lake crude + 20-30 mesh sand CL13 CL14 bi tumen + clay

Hamaca crude + 20-30 niesh sand VENFi VEN6

Hamata crude + 20-30 mesh sand VEN14 VEN19+clay

Haiiiaca crude + 20-30 mesh sand VEN21 VEN23+ 170-270 mesh sand

~iiaxiriiumof 35O"C, then decreasing temperature to 180C while air was in-

jected (Figs. 6.17 and 6.18).

Figures 6.19 - 6.26 present ap paren t H/C ratios based on total-gas-volume

calculations. Results for the kinetic and combustion t ub e experimental sets are

compared in Table 6.6. Apparent H /C values for each set of kinetic and com-

bustion tube experiments were in good agreement. Further, appar ent H/ C

values were similar to those of the original crudes as determined by elemen -tal analysis. Th e exception was the set of runs, Nos. VEN14 (Fig.5.12) and

VEN19 (Fig.6.25), where the apparent H / C ratios were high as a result of

low-temperature oxidation.


137/323

6. KINETIC EXPERIMENTAL RESIJLTS 120

0 0 0 0

I I I

0 0 0d m N 2 0

I

m N 3

(% a1our) ua8llxo sapyxo u o q n a

30

0

20v)

0d

0N

0


138/323

6. KINETIC: EXPER IME NTA L RESULTS 121

3

2 0 2 2 x 2 0 d 0": I I I I I I I

I I I

20 09 m c\l


139/323

6. KINETIC EXPERIMENTALRESIJLTS 122

m


140/323


141/323

6. KINETIC EXPERIMENTALRESrJLTS 124

I I I

2 5 In 0c\(


142/323

0'. KINETIC EXPERIMENTALRESrJLTS 12.5


143/323

6. KINETIC EXPERIMENTAL RES [JLTS 126


144/323

6. KINETIC EXPERIMENTA LRESlJLTS

------%


145/323


iI I I I I 0

u) 02 2vr 0m c\1 c\I


146/323



147/323


Table 6.6: Apparent H / C Ratios for Kinetic and Combustion Tube Experiments

Kinetic experiments

Run No. Sample matr ixt Old method* New method**

Cold Lake bitumen:Original crude - - -ClL2 - 0.71 1.41CL5 - 0.49 1.32CL13, CL14 clay 0.00 1.74

Hunting ton Beach oil:

HR02 - 0.37 1.77Original crude - - -

Hamaca crude oil:Original crude - - -

VEN5, VEN6 clay 0.21 1.81VEN 15 - 1.02 1.65VEN14, VEN19 - 1.51 4.47VEN21, VEN23 170-270 mesh 1.04 1.75

sand

Combustion Elementaltube expts. analysis

- 1.501.63 -

4.35 -1.77 -

- -

* At HTO using Eq. 2.7.** At HTO using E q. 6.2.t In addition to 20 -30 mesh sand.


148/323

6. KINETIC EXPERI MENTAL RESULTS 131

2. The measured aniount of oxygen consumed is:

02 , (7nensured)= 0.2682N2- - O Z p (6.3)

If oxygen is consumed to form C 0 2 , CO and water, then from stochiometry

(Eq. 2.4) , tlie apparent amount of oxygen consumed irx

(6.4)X co4 2

2,(nppwerzt) = - co2+co)+cc)~+-Let:

AO2, = 0 2 , ( n L e a s u r e d )- O2,(appc!rent) (6.5)

Th e cumulative difference, J A 0 2 ,d t , m ay be calculated as shown in Figs. 6.27 -

6.34. The cumulative differe~iceincreases in the LTO period due to oxygenation

of t he fuel, but is essentially zero a t t he end of HTO. This indicates that most

of the oxygen from the oxygenated products takes part in the HTO reactions.

3. To verify tha t t he fuel in kinetic tube experimeiits is oxygenated, an experiment

(Run VEN7) was performed in which air was injected until approximately the

end of LTO (310C).Thereafte r, nitrogen was injected. Th e results are shown

in Fig. 6.35. Carbon oxides were produced during H'TO even when no oxygen

was injected. Th e average C02 and CO molar concentrations during that period

were 0.1% and 0.05% respectively. This experiment clearly demonstrated that

oxygen from t he oxygenated fuel took pa rt in the HTO reac,tions. Th e cumula-

tive difference between measured and apparent amount of oxygen consumed is

shown in Fig. 6.36. At the end of t he run, the cumulative difference is not zero.

This result indicates tha t tlie oxygenated fuel was not, totally consumed. Car -

bon oxides were still being produced at the end of the run. After the run, the

sand n i x was examined and found to be black in color, indicating that residual

hydrocarbons 011 the sand mix had not been burned completely.


149/323

6. KINETIC EXPERI MENTAL RESlJLTS 132

0v,

I I I I\


150/323

6. KINETIC EXPERIMENTALR E X J LT S 133

010

I I I I

v I


151/323

6. KINETIC EXPERIM ENTAL RESlJLTS 134


152/323

6. KINETIC EXPERIMENTAL RESULTS 13 5

00 0* m 005 00

I I I I

,i


153/323

0'. KINETIC EXP ERIME NTAL RESIJLTS 136

0 0 0 003 c\1 00 02 2 2 2 00 \o

I I I I I I I


154/323

0'. KINETIC' EXPERIMENTAL RESIJLTS 137

0 0 0 003 hl 00 03 2 2 E: 00 \o

I I I I I I

0hl

a

1

%2\


155/323


0

2I I I I


156/323

6. KINETIC EXPERIMENTALR E S l J LT S 139

0 0

0 0 0

U

9

rcr0


157/323


00rr,

0

s0

20 0 0 0 C>0g 0 0 0 C>52

0 0g 0rI - 7

(% a1oUI) u a x x o sap;rxo u o q n aI I I I

zI I

0 0 0g d N0 0

00


158/323

6. KINETIC EXPERIMENTALRESrJLTS 141

0 0 0 00

3 2 2 z 3 c\1 0I

0zI I I I I I

Yz5


159/323

6. K I N E T l C E X PE R lM E NTA L RESIJ LTS 142

4. As described in Section 4.1, elemental analysis was performed 011 an oil sample

from coiiibustion tube Run VEN14 which was a low-temperature oxidation run.Th e atomic oxygen -carbon ratio of the oil sample was 0.25. This shows conclu -

sively that oxygenated hydrocarbons are formed when crude oil is subjected to

low- temperature oxidation.

6.3 Apparent O/ C Ratio From Gas Analysis

Following the stochiometric equation for fuel oxidation (13q. 2.4), the equation for combustion of an oxygenated fuel is:

2

(6.611L 2 y2 4 2 2

:H,O, + (1 - 7 + - - ) O , -+ (1 - m ) C O ,+m(C7O)+ -H,O

where y is the atomic O/ C ratio. Based 011 the carbon balance, molar concentration

of oxygen froni th e oxygenated fuel m ay be expressed as (CC$+CO)y/2 . Taking intoconsideration this additional amount of oxygen, the expresljion for the atomic H / C

ratio, 2 , as given in Eq. 2.7, becomes:

Equation 6.7 may be rear ranged to give:

2(1 - m ) C O , +co (12 +)4m - Ok]y = GO,

where OZc= 0.2682N2 - O z p . The atomic O/C ratio may be determiIied from gas

analysis da ta if the atomic H / C ratio is known.IJsing Eq. 6.8 and atomic H / C ratios calculated on a total-gas-volume basis (as

described in Section 6.2), atomic O/ C ratios were determined. The results are shown

in Figs. 6.37 - 6.43 and surmiiarized in Table 6.7. Average a,tomic O/C ratios of the

fuel ranged between 0.4 (Runs CL 2 and CL5) to 0.9 (Run VEN6) . Th e results agreed


160/323

6. K l N E T I C EXPERIMENTAL RESlJLTS

Table 6.7: Averaged Apparent H/C, O/C and na-Ratios

Run No. H/ C ratio O/C ratio

C L 2 1.41 0.4CL5 1.32 0.4CL14 1.74 0.7HBO2 1.77 0.8VEN6 1.81 0.9VEN15 1.66 0.TjVEN23 1.75 0.4

m- atio

0.2600.2680.2660.2470.2460.2880.5!76

---

143

with the atomic O /C ratio, 0.25, determine( by elementa l analysis 0 , the oxygenated

sample in combustion Run VEN14. Barta et al . (1989) conducted elemental analysis

of the products of wet combustion of Athabasca bitumen. The products were found

to have atomic O /C ratios of about 1.

Experiments were performed in which low -temperature oxjdation was minimized

to reduce oxidation of the fuel (Runs CLlO and VEN10). Nitrogen was injected from

the start until the approximate end of LTO. Thereafter air was injected. In Run

CL10, a short exothermic LTO period was observed irniiiecliatelyafter switching to

air iiijectlon (Fig. 6.44). A sharp rise in temperature and oxygen consumption was

observed at the start of air injection. Th e results indicated an average appa rent

H /C ratio of about 1.5 in the HTO period (F ig .6 .45 ). Thi s result suggests tha t the

fuel contained little or no oxygen. In Ru n VENlO, the LTO period was also short

and exothermic (Fig. 6.46). However the HT O peak was grea tly diminished due to

pr ematu re ignition as evident by the sharp increase in temper ature. Erroneous H/C

ratio results were obtained due to low values of oxygen consumption and carbon

oxides.


161/323

6. KINETIC EXPERIME NTAL RESIJLTS 144


162/323

0'. KINETIC EXPERIMENT ALRESlJLTS 145


163/323

6. KINETIC EXPERIMENTAL RESIJLT,Y 146


164/323

0'. KINETIC EXPERIMENTAL RES(JLTS' 147


165/323

6. KIN E T IC E X PERlM E N T A L RES JLTS 148

I


166/323

6. K l N E T I C EXPERlMENTAL RESULTS 149


167/323

0'. KINETIC EXPERIMENTALRESlJLTS 150

0

\


168/323


0 0 0 0 C>

20 0 0- N0

g v) e m+ I I I I II++ I

\;=

\BE

23

0

++

++

++

++

f++*+++

+++

++++++++++++++++++++t4.*b

bI I I I

.iv10

@8


169/323

0'. KINETIC EXPERIMENTAL RESIJLTS 152

G

+0

$uIIE

' 8o oo 8

8,Po

! o


170/323


171/323


6.4 Heat of Combustion of 0xyge:nated Fuel

For a hydrocarbon fuel (CH,) which undergoes c o m b u ~ t ~ i o naccording to th e sto-

cliiometry given by Eq. 2.4, th e heat of reaction, AH, (Btu/lb,) , may be estimated

using Eq. 6.9 (Burger and Sahuquet 1972). It is assumed in IEq. 6.9 tha t the products

of combustion consist of gaseous carbon dioxide and carbon monoxide and condensed

water.

(94.0 - G7.9rr~+ 31.22) (6.9)1800AH -- 12 * x )

The heat of coiiibustion for an oxygenated hydrocarbon fuel (CH,O,) is consid -erably less than that for a hydrocarbon fuel, 011 a per unit, mass basis, because an

oxygenated fuel is part ia lly oxidized and also it s mass includes oxygen. The heat

of conibustion of an oxygenated hydrocarbon was estimated by considering tlie fuel

oxidation pa ths as shown in Fig. 6.47. In Fig. 6.47, path A is the. oxidation of a hy-

drocarbon fuel to form carbon oxides and water. Path B relpresents the oxygenation

of the fuel to form CH,O,, while Path C is the oxidation of the oxygenated fuel to

form carbon oxides and water.

Figure 6.47: Fuel Oxidation Paths

Let A HA be tlie heat of reaction per mole of oxygen consumed for Pa th A, and A H B

and AHc the heats of reaction for Paths B and C for the same mass of fuel as in


172/323

6. KINETIC EXPERlMENTA LRESlJLTS 155

Path A. AHA ( K J / m o l 0,) may be calculated from Eq. 6.1.0 (Burger a n d Sahuquet

1972).786.4 - 6 7 . 6 ~ ~+ 260.92

A H A = (6.10)

The heats of reaction (KJ /mo l 0,) for t h e main products of hydrocarbon oxygenation

(Burger and Sahuquet 1972) are as follows: carboxylic acid (430.8),aldehyde (363.1),

ketone (375.7) and alcohol or phenol (306.6). The averaged heat of reaction for these

oxygenated products is 369.0 KJ/mol 0 2 . For one mole of oxygen consurned for Path

A, the number of moles of oxygen c o n s u ~ i ~ e dfor Path B is, (y /2) / (1 - n / 2 +2 /4 ) .Therefore:

2 - ri +2 / 2

(6.11)

From conservatioriof energy:

AHc = A H A - H B (6.12)

IJsiiig Eqs. 6.10, 6.11 and 6.12:

(6.13)6 9 . 0 ~

786.4 - 67.6m +260.9.;= I -Hc

A H ALet R17LflSSdenote the molar ~ n a s sratio of C H , to CH,O,:

12 + 212 + 2 +16y? I L ( L S S = (6.14)

The heat of reaction for an oxygenated fuel, AH,, (Btu/lb.,,, of oxygenated fuel), is

propor tional to AH, but reduced by the heat of fuel oxygenation an d by the addition

of oxygen to the fuel mass. That is:

(6.15)

Substituting Eqs. 6.9, 6.13 and 6.14 into Eq. 6.15:

16 9 . 0 ~(786.4 -- 5 6 7 . 6 m t 260.92)(94.0 - 7.9772+31.22)1800(12 +2 +16y) AH,, =(6.16)


173/323

0'. Ir ' lNETIC EXPERIMENTAL RESULTS 156

IJsing Eq. 6.16, the heat of combustion for an oxygenated hydrocarbon fuel was

computed as a function of atomic H / C , O/C and m-ratios. 'The results are presented in Fig. 6.48. The heat of c~ombustiondecreases significantly with increasing atomic

O /C ratio, arid also decreases slightly with increasing m -ratio for typical m -ratio

ranges. For example, for fuel with an atomic H / C ratio of 1.5 and na-ratio of 0.3, the

heat of coinbustion dec,reasesfrom 16,000 Btu/lbj,,,, or an atomic, O/C ratio of 0.0,

down to 8,000 Btu/lbj,,l for an atomic O /C ratio of 0.5. If the fuel is oxygenated

as a result of low -temperature oxidation, an inefficient combustion will occur. Low-

t e~ i i p e r a tu r eoxidation should be avoided in field operat iors. For the same reason,oxygenated gasoline used in automobiles will give considerably lower mileage per

gallon t h an non-oxygenated gasolines.

6.5 Error Analysis

Estimates were made of the errors in apparent H/ C a,nd nz-ratios determined

froin gas analysis. Error analyses were performed for a typical kinetic experiment(run 110. CL5) an d a combustion tu be experiment ( run no. VEN5).

Apparent H/ C ratios were calculated using Eq. 2.7 which is:

(0.2682N2 - 2 p - GO2 - C O / 2 )x =4 - (6.17)

The error in apparent H / C ratio, A x , may be approximated by the derivative of x in

Eq. 6.17 with r e s p e d to th e dependent variables:

(GO2*G O )

A N 2 ,nozp, GO2 and A C O represent the accuracies of th e gas analysis readings.The partial derivatives are:

1.0728- -

d X

8N2 (GO2+GO) (6.19)


174/323


175/323


The m -ratio is:CO

= (COZ +CO)The error in nh-ratio, A m , may be approximated by:

where:

co-1-3 r n- -dCO ( C O ,+CO) (COZ+ C o j "

CO - - -

d?Tl

d C O , ( C O , +CO),

(6.20)

(6.21)

(6.22)

(6.23)

(6.24)

(6.25)

(6.26)

For a conservative estimate, absolute values of the partial derivatives in Eqs. 6.20

- 6.22, 6.25 and 6.26 were used. Th e gas analyzers and gas chromatograph were

calibrated using standard gases whose compositions were known to f 0 . 0 1 % . Since

two calibrations were performed, one at the beginning and one at the end of an

experiment, th e accuracies of the gas analysis readings were taken to be *0.02%.

Error analysis results for Runs CL5 (at HTO) and VEN5 axe presen ted in Table 6.8.

Errors in apparent H /C and m-ratios are smaller in Run VEN5 than in Run CL5.

This is a consequence of the difference in magnitude of the gas concentrations between

combustion tube runs and kinetic t ub e experiments.


176/323

6. K l N E T l C EX PE RI ME NTAL RESULTS 159

Table 6.8: Estimated Errors in Apparent H /C a,nd nz-Ratios

Concentration (mole %)

Run No. C O S CO 0 z p NS Ax Ani x f A x ni f A n i

CL5 1.77 0.69 18.4 77.94 0.06 0.008 0.49 f 0.06 0.268 f 0.008

V E N 5 11.58 4.92 1.05 81.44 0.01 0.001 1.63 f 0.01 0.298 f 0.001

- - - _ _ - - - -

6.6 Previous Oxidation Reaction Model

An oxidation reaction model was presented by Fassihi (1981) and has been used

by subsequent workers (D e 10sRios 1987, Shallcross 1989). A brief description of the

model together with an example application (Figs. 6.49-6.50) follows.

In th e previous model, the following assumptionswere made.

1. The oxygen co~ i sumpt ioncurve can be separated into three component curves

corresponding to low, medium an d high tempe rature oxidation reactions (LTO,MTO and HTO, respectively). Each of these reactions is assumed to be the

oxidation of a particular fuel.

2. No carbon oxides are produced during LTO.

3 . The fuel during HTO is carbon.

For each reaction, th e rate (equated to th e oxygen corisuiiiptionrate) is:

(6.27)

where q is the volumetric flowrate, OaCis the molar concentration of oxygen con -

sumed, A and L are the cross-sec,tionalarea an d length of t he sand m ix in the kinetic

tube, Art is the pre-Arrhenius constant, Poz is the oxygen p(artia1 pressure, C f is the


177/323

6. h'ZNETIC EXPERIMENTAL RESCJLT5' 160

iiistaritaneoiisfuel concentration, E is the activation energy, R is the universal gas

constant, T is the absolute temperature, vi and n are the reaction orders in respectto Po, and C f respectively, arid i is the index for either LTO, MTO or HTO.

Further, each reaction ra te is proportional to t he decrease in th e fuel concentration,

so that:

(6.28)

where cy; is a proportionality coiistant. Integrating Eq. 6.28 from t = t to t = 00

(where Cfl = 0) yields:

From Eq. 6.27:( I 0 2 c

( :7L t fl'( = ALATIP;;~ezp(-E;/RT)

Substituting Eq. 6.30 into Eq. 6.29:

where:

(6.29)

(6.30)

(6.31)

(6.32)

A graph of the natural logarithm of the left hand side of Eq. 6.31, termed the

relative reaction rate, versus 1/T should yield a straight line with a slope of - E ; / R ,

and an intercept of ln,Lj;. The exponent, n;, may be obta ined by tr ia l -and -error or

through a non-linear regression riiethod. An example of such an Arrhenius graph is

showii in Fig. 6.49b.

Starting with E H and ,BHobtained for HTO, Eq. 6.31 m ay be used to ca lculatean approximate oxygen consumption curve, Curve I in Fig. 6.50a. Curve I may be

subtracted from the experimental data to yield Curve 11, the curve for MTO and

LTO oxygen cotisumptioii (Fig. 6.50a). Assuming fuel for H TO to be carbon, the

oxygen consumption may be represented by t he equivalenl,carbon dioxide and ca rbon


178/323

6. KIN E TIC EXPERIMENTAL RES I JLT,5' 161

0 Carbon dioxide produced 0 Carbon monoxide produced Oxygen consumed

- - - - - - - Temperature

(a) Gas Composition and Temperature versus Time

Inverse temperature (l/deg. K)

(b) Arrhenius Graph for HTO Data

500

400

util

300 %

Y i$5200 a

b

100

0

Figure 6.49: Huntingtoi iBeach Oil - Analysis Based on Previous Model ( A f t e rShall -cross, 1989)


179/323

6. KIN E TIC EXPERIME NTAL RES lJLTS 162

Curve I

Curve IICurve IIIEquivalent carbon oxides produced

.............

- - - - - - -a Oxygen consumed data

(a) Generation of MTO Curve

HTO (calculated)MTO (calculated)

Total (calculated)

- - - - - - -............. LTO (calculated)

h

@

(b) Total Oxygen Consumption Curve

Figure 6.50: Huntington Beach Oil - Analysis Based 011 Previous Model (After Shall -cross, 1989)


180/323

0'. KINETIC EXPERIMEN TAL RESULTS 163

inonoxideconceritration curve, S = COz + C O / 2 . Curve I m ay be subt rac ted from the

S curve to yield Curve 111, the curve representing LTO a nd M'I'O. The assumption tha tcarbon oxides are not produced for LTO implies tha t Cu rve 111 represents t he oxygen

coiisuniptioii for MTO only. An Arrhenius graph based 01 3 Curve 111 should yield

E M and , l j ~for MTO. Subtrac,tingCurve I11 from Curve I1 should yield the oxygen

coi isumptio~icurve for LTO, from which E L and [)L may be obtained . Superposition

of Curve I, Curve 111 and the LTO curve should yield th e total curve for the oxygen

consu~iipt ionof the three reactions (Fig. 6.50b).

Kinetic experirnental data were analyzed using the previous oxidation reactionmodel. Th e following problems were encounter ed.

1. T h e Arrhenius graph for the HTO reaction does not yield a satisfactory straight

line (Figs.6.51 and 6.52). After reaching a maxir-riumvalue, the slope of the

graph becomes positive. Previous researchers appear to ignore the data points

after th e maximum value in th e Arrhenius graph (for example, Fig. 6.4913). How-

ever these data points may represent some kinetic be11aviour not considered in

th e previous model.

2. The assumption that carbon is the fuel for HTO is only an approximation of

the combustion reaction. A kinetic tube run was performed for a sand rnix

containing a mixture of Cold Lake bitumen a nd carbon. Th e results (F ig. 6.53)

indicate a separate carbon peak at about 450 "C, compared to about 400 "C

for the HTO peak of the crude. As described in Section 6.2, fuel for HTO in

kinetic tube experiments is an oxygenated hydrocarbon.

3. Kinetic experiniental results indicate only two oxidation reactions, LTO followed

by HTO (Section 6.1).

4. Previous researcliers have found that a11 increase in the surface area of the

porous iiiedium increases the LTO reactions (Drici and Vossoughi 1985) which


181/323

6. KINETIC EXPERIMENTALRESlJLTS 164

-7.01 I I I I I I I I I I I , , I , I I I0.00140 0.00145 0.00150 0.00155 0.00160 0.00165 0.00170

Inverse temperature, l/deg. K

Figure 6.51: Cold Lake Bitumen (R un CL14) - Arrhenius Crribph for HTO Dat a Based 011 Previous Model


182/323

165. KINETIC EXPERIMENTALRESlJLTS

n = 0.2

- l o l l l l l r l l l l l l l l l l l l l l l l l , l r l I I

0.00125 0.00130 0.00135 0.00140 0.00145 0.00150 0.00155 0.00160 0.00165Inverse temperature (l/deg. K)

Figure 6.52: Carbon (Run C 3 ) - Arrhenius Graph Based on Previous Model


183/323


0 0 0 0 C>

I I I I I8 Fi i3 00 0 0\o v, *

m hl 3

(% qow) ua8Axo SapFxo u o q n 30


184/323

0'. KINETIC EXPERIMENTALR E S I I LT S 167

result in the forination of more fuel (Shahani and Hansel 1984, Vossoughi e t al .

1982). As described in Section 6.1, kinetic experiments indicate clay increasesthe amount of oxygen consumed during LTO. These results are in qualitative

agreement with observations made by earlier researchers. Th e salient point is

that a proper model should incorporate the surface area of the porous medium.

6.7 Summary of Kinetic Experimental Results

eAs a result of excess oxygen during kinetic tub e experiments, the fuel th at reactsduring HTO is an oxygenated hydrocarbon, CH,O,.

e Knowing the atomic H / C ratio, the atomic O / C ratio m ay be de te rmined from

gas composition data.

e Clay increases LTO reactions to form oxygenated fuel with large atomic O/C

ratios, as indicated by results of kinetic tube Runs CL14 and VENG. Thus in

these two rims, the measured oxygen consumption during HTO is decreased

due to the high oxygen content of the fuel.

e In cornbustion tube experiments where the apparent H/C ratio is similar to

that of the original crude (e.g. Run VEN5), the atomic O/C ratio is essentially

zero. Thus no LTO has occurred and the fuel is not ox,ygenated.These efficient

burns have high combustion tempe ra tu re s and virtually no or little oxygen in

the produced gas stream.

e Combustion involving an oxygenated hydrocarbon fuel is an inefficient operation

from the standpoint of poor usage of oxygen injected, low heat of combustion

and reduced oil mobility.


185/323


0 Properly designed combustioiitube experimeIitsand field projec ts involve burn -

ing a hydrocarbon fuel which has not been oxygenated. Since kinetic tube ex - periments en tail HTO reactions for an oxygenated fuel, the results may not

per ta in to combustion that occurs in a combustion tube or in field operatioils.

Nevertheless, a new oxidation reaction model for :kinetic tube experiments is

presented in the next chapter.


186/323

7. New Oxidation Reaction Model

Based on experimental results described in Chapter 6 , it new oxidation reaction

niodel was formulated.

1. T h e oxidation of crude oil consists of two i n a i ~ istages: low -temperature oxida -

tion, followed by high - temperature oxidation. The oxygen coiisumption curve

may be sepa ra ted in to two, partially -overlapping curves.

2. In kinetic tube experiments, crude oil oxidation reactions occur in an oxygen

abundant environriient. The fuel that is oxidized during HTO is an oxygenated

hydrocarbon which is characterized by a particular se t of atomic H/C and O/Cratios.

3 . The surface area of the porous ~ n e d l u ~ i i ,represented by the size of the sand

reackion surfac,e airea,)affects the reactiongrains and /o r clay p a r t i d e s ( i s .

kinet ics.

To assess the effect of the surface area, two models were studied: a spherical -

fuel-geometry and a varying- fuel-geometry model. In the former (Fig. 7.1), the fuelis assumed to be deposited evenly around the sand grainis wliic~iare assumed to

be spherical in sh ape. As temperature increases, water and light hydrocarbons are

vaporized leaving behind a residue of heavy oil fractions. Due t,o excess oxygen being

presen t, the residue undergoes low -temperature oxidation to form an oxygenated fuel.

169


187/323

7. N E W OXIDATION REACTION MODEL 170

In the HTO range, t he fuel undergoes oxidation to form carbon oxides and water. Th e

inass, a n d therefore radius, of t he fuel decreases until th e fulnl is completely oxidized.As discussed in Section 7.1, this model failed to match kinet ic experimental data .

, oilwater

--*

t =O t = t t = te

Figure 7.1 : Schematic Diagram of Spherical-Fuel -Ckometry Model

, oil water

-

t =O t = t o t = t c t = t e

Figure 7.2: Schematic Diagram of Varying-Fuel -Geometry Model


188/323

7. N E W OXIDATION REACTION MODEL 171

In tlie varying-fuel-geometry model, the fuel has two geometries as shown in

Fig. 7.2. Initially the fuel covers the ent i re sand grain. Due to t he large surface area,fuel deposited directly on the grain surface is oxidized fas ter than that deposited at

grain contacts. At some later time, t, , fuel is present only a,t grain contacts. During

this time, until fuel oxidation is complete, grain contact geometry m ay be approxi-

mated by a shrinking toroid.

The following features are c o ~ i i ~ i i o ~ ito bo th tlie spherical-fuel-geometry arid varying -

fuel-geometry models.

1. The stochioiiietricequation for fuel oxidation is given by Eq. 6.6, which is:

r r i z y z

2 4 2 2CH,O, + (1 - - + - - )O + (1 - rz)CO2+rrzCO + - H 2 0 (7.1)

Let C O an d C O , be the mole percent of produced carbon monoxide and carbon

dioxide, respectively, and qo the effluent gas flow ra te ( l i ter /min) . One mol

of gas at standard conditions occupies 22.4138 liters. The atomic weights of

carbon, oxygen and hydrogen are 12.01 10, 15.9994 ;tiid 1 OOOO, respectively.

lJsirig the carbon balance, th e number of moles of fuel oxidized per second may

be expressed as q,(CO + G 0 , ) / ( 6 0 x 22.4138). Th e molecular weight of thefuel (CHzOy)is equal to 12.0110 + 2 + 1 5 . 9 9 9 4 ~ .Therefore, the mass of fuelconsumed per second, d r n j / d t , is:

drrij q,(CO +C0,)( 12.01 10 + J: +1 5 . 9 9 9 4 ~ )(7.2)- -at 60 x 22.4138

2. If I * , and ps are the the average radius and density of a sand grain (assumed

spherical), then tlie mass of a sand grain is given by 4n7-:ps/3. The total mass of

sand grains i n a sand mix of length, L , and cross-sectional a rea , A , and porosity,

4, is given by AL(1 - i)p,. Thus:

:3AL(1 - 4) No. of sand grains in the cell = --

4Tr :(7.3)


189/323

7. N E W OXIDATION REACTION MODEL

Table 7.1: Estimated Grain Sizes and Fuel Densities

AverageR U I ~ No. sand grain radius, r , (cm) Fuel density (g /cc)

c'3c 4CL2CL5CL14

HB02VENGVEN23

0.03900.01420.03 750.03750.0001

0.03750.00010.0004

2. 02.0

I .029:i .029:i.029

0.992I. .03711.037

172

3 . T h e average sand grain radius was based on sieve ana1,ysis(Tables 4.3 and 4.4).

Mixtures of clay and sand were also iised in some runs. In such cases, th e aver -

age sand and clay grain radius was estimated based on mass -weighted surface

area of each grain type. Th e average clay particle diameter was estimated to

be 0.004 c ~ i i(Burger e t al . 1985). The same method w,as used for mixtures con-

taining sands of two -size distributions ( e.g. 20-30 mesh and 45-70 mesh sand

mixtur e). The estimated grain sizes of the sand rtiixtures used are shown in

Table 7.1.

4. I n Table 7.1, the fuel densities for crude oils were assumed to be similar to t ha t

of their residua a t th e equivalent normal boiling point of 335C. The API versus

residuum correlation (Lim 1991) was used to estim ate t he fuel densities. For

kinetic tube r u m using carbon, the fuel density was that for carbon, 2 g/cc

(Chemical Rubber Co. 1982).


190/323

7. N E W OXIDATION R EACTION MODEL 173

7.1 Spherical -Fuel -Geometry Mod.el

In the spherical-fuel-geometry model, th e fuel is deposited evenly around the sand

grains whicli are assumed to be spherical in shape (Fig. 7.1). Let 7' be the radius of

the fuel surface and p j the fuel density at time, t .

4t 3

(7.4)Mass of fuel per sand grain = 7 7 r ( r 3 - ' t ) p j

IJsing Eqs. 7.3 and 7.4:

( T 3 - ' 3AL( l- ,)pj..3Total mass of fuel at t irne,t

=(7.5)'s

Differentiating Eq. 7.5 with respect to time and equating the result with Eq. 7.2

yields:

qo(.CO+CO2)(12.0110 f 2 + 15.9994~)AL(1 - , )p j = - (7 -6)dt 7'; 1 60 x 22.4138[r 3 - ';)

Integrating Eq. 7.6 from t t o t he end of fuel oxidation, t , (where 7' = rs ) gives:

(7.7) 3 3 t t' qo(CO+CO2)(12.0110+ 5 + 1 5 . 9 9 9 4 ~ ) d t

60 x 22.4138AL(l - ) p f7'3 = 1', - s

That is:

7' = r s ( l + y)+where:

Jtteqo(CO+C02)(12 .0110+ x +15.9994y)dt7 = 60 x 22.41:38AL(l - ) p j (7.9)

Th e rat e of decrease in fuel conce~i t ra t ions given by Eq. 6.28, which is:

(7.10)

where O J Cis th e mole percent of oxygen coIisuiiied, qt is the injected gas ra te ( L / m in ) ,

C H is the fuel co~icen t ra t ion , ndC X His a proportional ity consiiant. Since the molecular

weight of oxygen is 31.9988 and if n i j is the mass of fuel at time, t , then from Eq. 7.10:

(7.11)


191/323

7. N E W OXIDATION R E A C T l O N MODEL

The reaction rate for a part icular surface area (Burger e2 al . 1985) is:

d m- -k,*pirH exp(- H / RT) x surface area

dt

174

(7.12)

where k,* is an equilibrium c,oiistaiit and EH is the fuel activation energy at HTO.

Using Eq. 7.3 and r from Eq. 7.8:

Fuel surface area = 3AL(1 - ) (1 +y ) j i / r s (7.13)

Substituting Eq. 7.13 into Eq. 7.12 and equating the resulting equation with Eq. 7.11

yields:

(7.15)

A graph of the natural logarithm of the left hand side of Eq. 7.15 versus l /T for the carbon (Fig. 7.3) or crude (F ig. 7.4) HTO kinetic experimental data, however,

did not yield the expected straight line fit. In Fig.7.4, the Arrhenius graph for the

spherical-fuel- geometry model has been transposed by +7 y-a

Kinetics of in-situ Combijstion

Documents

Transcript of Kinetics of in-situ Combijstion