29 Attrition and Breakage of a Western Reference Oil Shale at Process Temperatures Ulrich Grimm and Glenn Swaney U.S. Department of Energy Morgantown Energy Technology Center Morgantown, West Virginia ABSTRACT A high-temperature, rotating-drum attrition tester has been used to study the friability of a Mahogany zone (Parachute Creek member, Green River Formation) oil shale. This DOE reference shale has a Fischer assay of 27.5 gal/ ton and was obtained from the Exxon Colony mine located near Parachute, Colorado. This feed material was screened to 9.5 to 12.5 mm, heated in the drum to the temperature of interest, and tumbled a desired amount. Testing conditions included nitrogen, air, and carbon di oxide atmospheres; temperatures from ambient through 900C; and tumbling amounts of 0 to 2,400 revolutions. The breakage function of this shale is fairly similar to that of a western shale studied previously by the authors, yield ing a product size distribution that is typically trimodal or quadrimodal. Attrition levels for air atmospheres were somewhat higher than for nitrogen and produced more multimodal distributions under otherwise equivalent con ditions. The primary loss in particle strength occurs in the 300C to 500C range, and is attributed to kerogen decom position. Maxima in the particle-size distribution at 10 and 80 mm were evident. The breakage kinetics for the feed size of this shale are nonlinear. INTRODUCTION Breakdown of kerogen during the retorting of oil shale and combustion of the spent material generally weakens the remaining mineral matrix, leaving it substantially more friable. Processing of this material tends to generate fines, which can have detrimental environmental and economic consequences due to contamination and separation prob lems. Although in situ processes would seem to circumvent most of these attrition-related difficulties, they still must deal with the fines that contaminate the product oil. The aim of this work is to quantify the most significant factors affecting this attrition process, including (but not limited to) shale material properties, processing conditions, and the attriting environment. The ultimate objective is to develop a model capable of predicting the extent and nature of attrition in a proposed process that requires little or no experimental attrition data. In this study, attrition is defined as an unwanted erosion process in which daughter particles are substantially smaller than the parent, leading to a bimodal or multimodal par ticle-size distribution. The subject of attrition frequently is ignored because it is a complex, poorly understood process. However, research in this area has been increasing recently, particularly as more and more fluidized-bed applications are considered and developed (Ray and others, 1987). With oil shale, the difficulty of the general problem of attrition is considerably compounded by the complex morphology of the material. Attrition is quantified in practical applications using arbitrarily defined indices, such as grindability and hardness, but these permit only comparison of closely re lated materials and conditions and provide no insight into the attrition process itself. Attrition tests are classifiable as either single-particle or multiparticle, depending on the nature of the sample. Multiparticle tests readily provide statistically significant data but reveal little about the details of the attrition/- breakage process. While single-particle tests can provide such details, they generally require a prohibitively extensive and labor-intensive experimental program. Because much of the interest in attrition of oil shale centers on fluidized- bed processing, we have used a low impact-velocity, multiparticle tester, theRotating Attrition Test facility (RAT), to study attrition and breakage of four shales to date (Grimm and Swaney, 1989, 1990a, 1990b) (Figure 1). Based on the ASTM D4058 drum test for measuring catalyst friability, the RAT is a 25.4-cm-diameter Inconel 617 drum suspended and rotating in an electric furnace. Depending on the particle size, the impact energy afforded by this drum size is somewhat greater than or equivalent to that found in a typical fluidized bed (Grimm and Swaney, 1990). Hollow shafts allow the drum interior to be purged with a selected process sweep gas, and a ceramic felt filter on the exit shaft prevents loss of fines. A shelf mounted inside the drum prevents particle segregation of the charge
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
29
Attrition and Breakage of aWestern Reference Oil Shale at
Process Temperatures
Ulrich Grimm and Glenn SwaneyU.S. Department of Energy
Morgantown Energy Technology Center
Morgantown,West Virginia
ABSTRACT
A high-temperature, rotating-drum attrition tester has
been used to study the friability of a Mahogany zone
(Parachute Creek member, Green River Formation) oil
shale. This DOE reference shale has a Fischer assay of
27.5 gal/ ton and was obtained from the Exxon Colonymine located nearParachute,Colorado. This feedmaterial
was screened to 9.5 to 12.5 mm, heated in the drum to the
temperature of interest, and tumbled a desired amount.
Testing conditions included nitrogen, air, and carbon di
oxide atmospheres; temperatures from ambient through
900C; and tumbling amounts of 0 to 2,400 revolutions.
Thebreakage functionof this shale is fairly similar to thatofawestern shale studied previouslyby theauthors, yield
ing a product size distribution that is typically trimodal or
quadrimodal. Attrition levels for air atmospheres were
somewhat higher than for nitrogen and produced more
multimodal distributionsunderotherwiseequivalent con
ditions. The primary loss in particle strength occurs in the
300C to 500C range, and is attributed to kerogen decom
position.Maxima in the particle-size distribution at 10 and
80mmwereevident.Thebreakagekinetics for the feed size
of this shale are nonlinear.
INTRODUCTION
Breakdown of kerogen during the retorting of oil shale
and combustion of the spent material generally weakens
the remainingmineralmatrix, leaving it substantiallymore
friable. Processing of thismaterial tends to generate fines,which can have detrimental environmental and economic
consequences due to contamination and separation prob
lems.Although in situ processeswould seem to circumvent
most of these attrition-related difficulties, they still must
deal with the fines that contaminate the product oil. The
aim of this work is to quantify the most significant factors
affecting this attrition process, including (but not limited
to) shale material properties, processing conditions, and
the attritingenvironment. The ultimate objective is to
develop a model capable of predicting the extent and
nature of attrition in a proposed process that requires little
or no experimental attrition data.
In this study, attrition is defined as an unwanted erosion
process inwhichdaughterparticlesare substantiallysmaller
than the parent, leading to a bimodal or multimodal par
ticle-size distribution. The subject of attrition frequently is
ignoredbecause it isa complex,poorlyunderstood process.
However, research in thisareahasbeenincreasingrecently,
particularly as more and more fluidized-bed applications
areconsidered and developed (Rayandothers, 1987).With
oil shale, thedifficultyof thegeneral problemofattrition is
considerably compounded by the complexmorphologyof
and labor-intensive experimental program. Becausemuch
of the interest in attrition of oil shale centers on fluidized-
bed processing, we have used a low impact-velocity,multiparticle tester, theRotatingAttritionTest facility (RAT),to study attrition and breakage of four shales to date
(Grimm and Swaney, 1989, 1990a, 1990b) (Figure 1). Based
on the ASTM D4058 drum test for measuring catalyst
friability, theRAT is a 25.4-cm-diameter Inconel 617drum
suspended and rotating in an electric furnace. Dependingon the particle size, the impact energy afforded by this
drum size is somewhat greater than or equivalent to that
found in a typical fluidized bed (Grimm and Swaney,1990). Hollow shafts allow the drum interior to be purged
with a selected process sweep gas, and a ceramic felt filter
on the exit shaft prevents loss of fines. A shelf mounted
Figure 1. Rolling Attrition Tester (RAT) schematic.
MODELING
Two size-reduction theories have been extant formanyyears. In the first, Rittinger's surface theory, energy input
is proportional to the surface area formed. The second,
Kick's law, states that energy isproportional to thevolume
or weight of the comminuted product. It is generally con
sidered that Kick's law applies to impact pulverizing and
Rittinger's law to finegrinding (BritishMaterialsHandlingBoard, 1987). Raw shale consists of variousmineral grains
bound by kerogen; retorting the shale and subjecting it toan attriting environment removes most of this binder and
tends to resolve the material into its component, char
acteristic mineral grains, which are mostly <10 mm (L.J.
Shadle and D.C. Galloway, pers. comm.). It is an interest
ing observation that for attrition such as this,which forms
a characteristic fines distribution, the fines account for
nearly all the surface area formed,which isproportional to
theirvolume ormass.Thusboth Rittinger's law andKick's
law essentially are equivalent in this case (Ray and others,1987). These both imply that an attriting process that con
sumesenergyat a constant ratewill form fines at a constant
rate. In experimental studieson the attrition coal in a fluid
ized bed, Knowlton (1986) observed that the rate of fines
generation initially is rapid but then levels to a steady rate
as Kick's or Rittinger's law would indicate. He attributed
this nonlinear induction period to the breakage of sharpedges on the surfacesof the particles. Studies ofother size-
reductionprocesseshave shown that feedparticlespossess
a"memory"
of their previous treatment,whichmust fade
away before materials with different histories can be sub
jected toa faircomparison. Inotherwords, simplyknowingthe composition and particle-size distribution of a sample
is not enough to predict how it will attrite; details such as
the presence of microcracks and surface roughness can
play a crucial role. As samples are subjected to a common
processing environment, these differences tend to vanish.
Mechanisms of particle breakage/attrition can be
grouped into three categories according to decreasingimpact energy (1) shatter or fracture,where high energyproduces a wide distribution of fragment sizes; (2) chip
ping or cleavage, where the energy is just enough to load
a fewregionsof theparticle to the fracturepoint,generatingseveral large fragments; and (3) abrasion,where low ener
gy causes localized stressing, shattering a small region on
the periphery of the particle.
Over the years a large volume of literature on
comminution and grinding has evolved,much of which is
aimed at thedesign ofballmill circuits.Recently, however,
fully and semi-autogenous grinding have become im
portant comminutionmethods (Menacho, 1986). If a large
amount of material is charged, the RAT can behave
effectively as a tumbling mill, but the expense of sample
preparationdictates thatweusea small charge in thedrum
that is completely lifted by the shelfon each pass, so it actssomewhat like an impact crusher. In either case, the grind
ing (attriting/breaking) action is basically autogenous,
which is intrinsically difficult to analyze in that all three
thesedata represent thebreakagedistribution function. As
a size fraction approaches that of the feed, the percentage
undersizemust become 100%; hence, the lines descendingfrom the topof thegraphareexpected. It ismore interestingthat the production of fines seems to level off or exhibit a
maximum for the 6.3- to 9.5-mm feed size. The decrease in
fines for smaller feed sizes results from reduced impact
energy. For the largest size, impact energy exceeds the
strengthof theasperities,and the rateofattrition is roughlyproportional to the specific surface area,which is less than
for the smaller particles.
N 60-
- 40 fjn
40 - 150 tan
160 - 840 fjm
0.84 - 1.18 mm
1.18 - 3.36 mm
2.36 - 6.3 mm
6.3 - 9.5 mm
9.5 - 12.6 mm
400 500 600
Temperature (deg C)
Figure 4. Attrition vs temperature under nitrogen.
n 60
WOS-86
Air
9.5 - 12.5 mm Feed
400 Revolutions
B888888|- 40 ym
40 - 150 ijm
150 - 840 jUll
0.84 - 1.18 fjm
1.18 - 3.36 mm
2.36 - 6.3 mm
6.3 - 9.5 mm
9.5 - 12.5 mm
400 500 600
Temperature (deg C)
Figure 5. Attrition vs temperature under air.
WOS-86
CO,9.5 - 12.5 mm Feed
400 Revolutions
400 500 600
Temperature (deg C)
Figure 6. Attrition vs temperature under carbon dioxide.
The estimation of the process matrix, or breakage rate
constants, are depicted in Figure 10 for the largest feed
sizes under all atmospheres at 500C. Because the
ATTRITION AND BREAKAGE OF WESTERN REFERENCE OIL SHALE 33
WOS-86 COj 9.5 - 12.5 mm Feed 600t:
160 - 840 /urn
2.36 - 6.3
1000 1500
RAT Revolutions
Figure 7.Attrition vs time under carbon dioxide at 500C.
WOS-86 Nfc 500V: 400 Rev
40
0.15 0.2 0.3 0.5
Feed Size (cm)
Figure 8. Attrition vs feed size under nitrogen at 500C.
disappearance of the feed is given by exp(-fr,f), /c, is given
by the slope of the line on this semilogplot. Clearly the
breakage is not first order, since the plots are strongly
nonlinear. Austin and others (1973) attributed this type of
behavior in laboratoryballmills to the fact that theparticles
have a distribution of strengths and are subjected to a
distribution of forces.Thismeans thatweaker particles are
destroyedmuch faster;as theyaredepleted, the rateof feed
disappearance reduces.
FRACTION OF FEED REMAINING
0.1
0.01 -
0.001
fj
WOS86
500C
:
o
6
D
O
Atmosphere:
o N2? CO2
A Air
-
AO
:
a
OD
-
A
A
i i i i i i i
A
i i i i
500 1000 1500
RAT REVOLUTIONS
2000 2500
Figure 10. Feed decomposition at 500C.
Figure 11 is a plot of the estimated initial breakage rate
constants (k{, s) as a function of feed size for all atmospheres
Figure 9.Attrition vs feed size under nitrogen at 750C.
0.07 0.1 0.15 0.2 0.3 0.5
RELATIVE FEED SIZE
Figure 11. Rate constants vs feed size.
0.7
34 COLORADO SCHOOL OFMINES QUARTERLY
commonly observed in batch grinding systems. Here x,and x,+i are the upper and lower sieve sizes, respectively,for size fraction i, and a is a filled constant. Because thismaterial is so easilybroken and attrited, it is not possible to
accuratelymeasure these rate constants. However, all the
data show a similar trend, leveling off toward the smallerfeed sizes. This is consistentwith some results on rod mill
breakage of ore minerals, including quartz (Kelly and
Spottiswood, 1982).
CONCLUSIONS
Much more work is needed to develop a predictive
model, including the incorporation of a fines-distributionexpression into the breakagematrix, and finding relevantparameters to characterize the effect of temperature and
shale composition. In spite of this, several conclusionsmaybe drawn at this point:
1 . WOS-86 is very friable after kerogen pyrolysis. This
would make it difficult to use in processes with high face
velocities of combustion/gasification gases for residual
carbon burnout.
2. Temperatures above 900C and a carbon dioxide at
mosphere may be advantageous in promoting mineral
fusion to reduce the formation of fines.
3. Agglomerating, sintering, or other strengtheningmineral reactions in the range 600C to 900C are favored
by oxidizing atmospheres.
4. Product distributions are multimodal, with
characteristic grain sizes of 10 and 80mm evident.
5. Breakage kinetics for the feed size of this shale are
nonlinear,due to thewidevariability in individual particle
strengths.
REFERENCES
Adams, D.C, and Mahajan, O.P., 1987, Morphology of retorted
oil shale particles: Energy & Fuels, v. 1, p. 23-28.
Austin, L.G., Shoji, K., and Everett,M.D., 1973, Anexplanation of
abnormal breakage of large particle sizes in laboratory mills:
PowderTechnology, v. 7, no. 1, p. 3-7.British Materials Handling Board, 1987, Particle attrition:
Clausthal-Zellerfeld, BRD, Trans Tech Publications.
Grimm, U., and Swaney, G.,1989, Eastern oil shale attrition and
breakage: testing and modeling, in Proceedings of the 1989
EasternOilShaleSymposium:UniversityofKentucky, Institute
forMining andMinerals Research.
Grimm, U., and Swaney, G., 1990a, Attrition and breakage of a
New Albany shale at process temperatures, in Lazar, D.J., ed.,Proceedings of the 1990 Eastern Oil Shale Symposium:
University of Kentucky, Institute for Mining and Minerals
Research.
Grimm, U., and Swaney, G., 1990b, Attrition and breakage of
western oil shale, in Gary, J.H., ed., Twenty-Third Oil Shale
Symposium Proceedings: Colorado School of Mines Press,p. 162-168.
Herbst, J.A., and Fuerstenau, D.W., 1980, Scale-up procedure forcontinuous grinding mill design using population balance
models: International Journal of Mineral Processing, v. 7,
p. 1-31.
Kelly, E.G., and Spottiswood, D.J., 1982, Introduction to mineral
processing: NewYork, Wiley-Interscience, 491 p.
Knowlton, T.M., 1986, Attrition and entrainment studies related
to fluidized-bed gasifiers: Institute ofGas Technology, Project
61077 Quarterly Rept. (April 15-July 14).
Menacho, J.M., 1986, Some solutions for the kinetics of combined
fracture and abrasion breakage: Powder Technology, v. 49,
no. 1, p. 87-95.
Miknis, F.P., and Robertson, R.E., 1987, Characterization ofDOE
referenceoil shales:Mahoganyzone,ParachuteCreekmember,Green River formation oil shale, and Clegg Creek member,
New Albany shale: U.S. Dept. of Energy Rept. DE-FC21-
86MC11076.
Miknis, F.P., Turner, T.F., Ennen, L.W., Chong, S.-L., and Glaser,
R., 1988, Thermal decomposition of Colorado and Kentuckyreference oil shales: U.S. Dept. of Energy Rept. DOE/MC/
11076-2625.
Ray, Y.-C, Jiang, T.-S., and Wen, C.Y., 1987, Particle attrition
phenomena in a fluidized bed: Powder Technology, v. 49,no. 3, p. 193-206.