|
Document Type |
: |
Article In Conference |
|
Document Title |
: |
EARLY PERFORMANCE RESULTS OF SULPHUR EXTENDED ASPHALT (SEA) نتائج استخدام خليط الكبريت والأسفلت أعمال رصف الطرق بالمنطقة الشرقية بالمملكة العربية السعودية |
|
Document Language |
: |
Arabic |
|
Abstract |
: |
EARLY PERFORMANCE RESULTS OF SULPHUR EXTENDED ASPHALT (SEA)
PAVEMENTS IN EASTERN SAUDI ARABIA
W. AKILI
Professor of Civil Engineering
and Senior Research Engineer
University of Petroleum and Minerals
Dhahran, Saudi Arabia
~l.G. ARORA
Senior Research Engineer
Research Institute
Universi ty of Petroleum and ~Iinerals
Dhahran, Saudi Arabia
ABSTRACT
For the past four years the Research Institute of the University
of Petroleum and Minerals, Dhahran, Saudi Arabia, has been carrying out
a research program on sulphur based paving systems for potential application
in the Kingdom. To test sulphur extended asphalt (SEA) mixes under
trafficked conditions, three experimental test sections have been constructed
with SEA as part of the trafficked network near Dhahran. Test Roads
One and Two, constructed in 1979, have 30/70 and 45/55 sulphur-asphalt
weight ratio respectively. Test Road Three, constructed in 1982, has 30/70
sulphur-asphalt weight ratio and one percent cement to augment the strength
of the mix under soaked conditions.
This paper briefly describes the three pavements and presents the
inservice data generated from Test Road Two which has experienced extremely
heavy axle loads that have caused premature cracking along one section.
The paper additionally, comments on the potential application of SEA pavement
systems in Saudi Arabia.
met~ods of: production, transportation. placement and compaction. Although
many investigators have reported that SEA properties are equivalent or
superIor to those of conventional asphaltic concrete pavements, AI-Otaishan
and Terrel (5) have reported a faster inservice loss of strength on certain
SEA pavements in the USA, and earlier cracking tendencies of SEA were noted
by Shields et al (6).
The Research Institute. University of Petroleum and ~linerals, Dhahran,
Saudi Arabia, has been conducting research on the utilization of sulphur in
pavement construction with emphasis on developing SEA pavement materials
compatible with the environmental and traffic conditions prevailing in the
Kingdom. The tasks pursued and data acquired have been reported on by Akili
and Dabbagh (7), Akili and Uddin (8), Akili (9) and byCourval and Akili (10).
As part of the Research Institutes sulphur extended asphalt (SEA) pavement
research, three full scale SEA pavement projects were constructed in cooperation
with Saudi Ar:lbian ~linistry of Communications/Roads Department, to
check construction procedures, derive Ilseful data and evaluate inservice
performance of SEA materials.
447
transportation, placement, compaction and quality control te5ts, was almost
the same as with conventional asphalt concrete pavements. Various steps
involved are shown schematically in Figure 6. A trial patch may be necessary
to arrive at the correct percentage of SEA binder to be fed which may vary
from inter-volume to volume equivalent of optimum asphalt percentage.
Quality control tests during construction were carried out on loose
aslaid samples taken from behind the paver at regular intervals. Tests have
included: binder extraction, specific gravity, Marshall stability and resilient
modulus (MR). Field cores were also taken, immediately after compaction
to check layer thickness and percent compaction. The test results
derived from all three projects were within a close range of the mix design
data, indicating adequate batching, mixing, laying and field compaction
conditions. Typical results of loose samples are sho\ffi in Table IV.
Test Road One
The SEA mix on this test project was laid in t\,O Ii fts. The initial 60
mm thi.ck lift (layer II 2) was placed in June 1979. The mix was placed and
compacted in the normal manner. SEA mix arrived on site a temperature of
approximately1600C, a significantly higher temperature than the desi.rable
135 - 1400C range. Inspite of the relatively high mix temperature and the
hot windy and dusty weather conditions that prevailed, no emission problems
were encountered except for some minor eye irritation experienced by the
workers around the paver. The cOr.1paction scheme was left entirely up to the
contractor. Breakdown rolling ,,as accomplished by vib~3ting roller. The
final lift, 60 ~~ thick, was placed in ~ovember 1979 under extremely favorable
weather conditions. Pbcement and compact:.cn ",as accomplishcd as in
initial lift.
Tes t Road ThO
This te~t road was lai.d in \iovC~!!~eT 1979. DU:in::; SEA paving. the
binder content in the base course ~as i.ncreased from an initi31 value of
5.~·,. dctermi;1()d by laborator:-- design, to 5.6, and later on to 5.86. OUTing
laYing and CO:;:p;lC t.lon the 5. ~: mi:x has jud);cd to be s tiff and on the tlry
side. Thi:> incc:lsC in bi ntlcr 1LVc~l improved the tc:xtur-c of the paved Sll1f:
lce and aicled cOllpact ion.
DiffLc~ltics in compact ion \,ere ini tially enc0untered that were attributablc
to :E.\ mix densi ficar.ion at either too high or- too Iowa temperature:.
Rolling was later adjusted by keeping the roller a set distance from paver,
whic~ pcr~itted b1eakdu~n r-olling to be carried out between 130 to 1100C.
In this projcct, SE~ mi:x wns used in the base course as well as in the
lcrlIin,; cours,. 11le bas, course 1,:15 Llid in two 70 I1lf.1 lifts in ~Iay 1982.
Tih ~(r1fing course was bid iiI .lun" J~lS~ in :1 single lift 60 1TU1l thick. Pn1blems
one! clelays Iwre encountcl(c! during wearing caUlse Llying that wcrc
attributable to breakdown in the connection between the heated sulphur
tanker and the sUlphur-asphalt module (SAM) responsible for blending sulphur
and asphalt.
On this project, the SEA mix arrived at the paving site at a temperature
ranging between 1300 to 1450 C. Breakdown rolling was accomplished with
2 passes of 10-ton steel tandem roller at 1200 C. This was followed by intermediate
rolling with 2 passes of 8-ton steel roller. Finish rolling was
with a Pneumatic tired roller at a temperature of about 950 C. Generally six
passes of the Pneumatic roller were found adequate to produce a satisfactory
smooth and homogeneous finish.
PERFO~~~CE EVALUATION
Performance observations carried out periodically on all three SEA
projects have included the following:
(a) Traffic counts and number of ~xle applications.
(b) Benkelman beam rebounds on the outer lane (slow lane).
(c) Degree of cracking and rut depth measurements.
In addition, a variety of tests including resilient modulus (MR) and
split tensile tests were conducted on pavement cores extracted from selected
as well as random locations. See Figure 6 for more details.
Latest traffic counts on the test projects have revealed that traffic
volume has decreased over the last two years on the Test Road One which
carries about 300 to 400 ADT with 25 percent t:ruck traffic in both directions.
The decrease in the traffic on this road is mainly due to the construction
of a parallel artery. Traffic on Test Road T\~o has been very heavy
ranging from 5000 to 60GLJ ADT \~ith 50, in the category of truck trailers.
Test Road Three mainly serves the University community and carries an ADT
of 1000 to 1500 vehic les wi th about 5~ truck s .
Since Test Road Two carries the heaviest traffic, its performance has
been closely monitored and the findings of the evaluation survey last conducted
in April 1982, are described belo~.
Traffic Counts and Axle Loads
Econol i te tr:1ffic counters \~hich count the traffic in graphical and
digital forms were used to count the numLer of axle load applications roundthe-
clock over a period of one week. Latest counts have shown that ~le
average daily traffic on all three SL\ north bound lanes is 5000 to 6000
vehicles. ~l:mual counts were also used to establish ratio of he:1Vy to light
vehicles and the nLunber of loaded trucks to total :J:de applications. \~ith
respect to the SIOl lane. 92 to 97"0 of all :1X Ie applic:1tiolls were due to
loaded trucks with an average of about 2Sl10 trucks/day.
458
Rebound Deflections
Benkelman beam rebound deflections were measured along the outer
wheel path of the slow lane using a standard axle load of 82 KN. Test
points were located at 0.9 m from pavement edge and at 20 m intervals along
the length of SEA Sections A, Band C, and 40 m intervals in the control
Section D. Pavement temperature was also measured near such deflection
point at a standard depth of 40 mm. Temperature was taken at each test
point immediately following the deflection measurement. The measured deflections
were normalized for a standard reference temperature so that valid
comparisons can be drawn. The Asphalt Institute in its publication MS-17(11)
recommends a standard temperature of 21 0 C. Since higher temperatures are
often encountered in this region, deflections were normalized for a higher
reference temperature of 350 C, representative of average pavement temperatures
in Eastern Saudi Arabia. The corrected deflections were further
subjected to variability checks as recommended by Smith and Jones (12) and
the Asphalt Institute (11). Typical values of deflection rebounds are given
in Figure 7 for SEA Section B (the thinner section). The 85th percentile
deflection was computed for each section which may be assumed to represent
the characteristic deflection of the section. The results are shown in
Table V.
Rebound deflections can be related to pavement life in terms of standard
axle applications. TRRLs full scale pavement design experiments,
particularly from Alconbury Hill experiment constructed in 1957, support a
relation between early deflection and performance of the form given below as
described by Lister (13).
Life a 1
(deflection) 3
If the above relationship applies to the road experiment under consideration,
the life expectancy of Sections A, B and C would be 87, 25 and 69
percent respectively of the control Section D. Tne results derived are at
best approximate and their validity will have to be substantiated by a more
relevant correlation based on data that relates SEA pavement rebounds to
axle load applications under local conditions.
Rutting and Cracking
Latest rutting and cracking measurements in the wheel paths of Test
Road Two were carried out at the same time and locations where deflections
were measured. Rut depth was measured with reference to a 2 m long
straight edge. Cracking was determined in linear dimension with one meter
square aluminium frame follOl;ing the procedure reported by Smith and Jones
(12). A simple rating system recommended by Bulman and Smith (14) of TRRL
as shOlm in Table VI has been adopted for the analysis. Typical results
are shOlm in Figure 7 for Section B (Thinner Section) .
TIle deformation and cr:Jcking indices were assigned nWllerical v:Jlues
ranging from 0 to 4. In this manner, an overall rating for each section
can be obtained by adding rut depth and cracking indLx for each test point
within the section. OVLrall pavement condition rating for each section is
summarized in Table V where rebound deflections are presented. It. is noted
from above table that the Section B which has exhibited the highe5t JeboWld
deflections has cracked extensively, as the observed cumulative cracking is
69.3% of the critical value. This is not surprising since the base course
thickness in the abovesection was r~dllced by 20%. Therefore, early failure
of this section in comparison to full depth SEA section, has been anticipated
Sections A and D are virtually free from any visible crack~. Section C
has developed some cracks, mostly within the slow lane near pavement edge.
The crack patterns within Section B are multiple pavement edge cracks, alligator
cracks and mid-lane longitudinal cracks. Crack width varies from
"just visible" to a maximum of 30 mm. Figure 8 shows photogr"ph of typical
cracking in Section B within a one meter square aluminium fra~e.Core Analysis
(vlindrical cores 100 mm in diameter. were txtracted from the pavement
at ceTt~in loc;:tions of Benkelman lJeaC, dpflectiOT measunments in adcli.tion
to cores extracted randomly from v3ri~Js locations at different time intervals
to study the effect ot .Jging and tI3.fficking on properties like bulk
densi ty. resilient modulus (~lR) a:,:: s;;,li t te!lsi Ie strength.
Th," MR tes-::s ....·ere conducted us ing a repi ti tive loading device developed
by Sch imi.Jt (15). Figure9 shows photograph of the MR test in progress.
Split tensile tests were carried out on the Marshall loading device at a
r.tte of loading of SOmm/min.. Load was applied along the two opposite generatOL
;.}n:Gugh ·;t2.inless steel curved strips. Failure observed in all cases
o.~cured ins tan taneously at the maximwn load sustained by the specimen. The
mode of loading and failure of sample are shown in Figure 10. Split tensile
s _rength was computed from the following equati on:
a
(Sin 2a - 2R)
where P = maximum load sustained, a = width of loading strip, h " height of
specimen. a " angle subtemled at the center by one half the width of the
loadhlg : trip in radians and R = radius of the specimen.
"anr1.or.l coring was dQne in December 79, Apri 1. 80, July SO, December 80
,n,! :!:;y 82. Since the road was officially opened to traffic in October 80,
on]}" :h" last two sets of cores have been subjected to full scale tTafficking
over a period of 3 and 20 months, respectively.
C:Jr-es extrClcted from base .;:ourses exhii,ited same range of values as
t~,0.><; from the I,earing courses. Therefore, test results of all SEA cores
were combined and a1alysed statisticz.lly. This way it was possible to
o0;ain large sample size required for meaningful statistical analysis. Simi;.;
J: analysis \o"l.: vJJ1(h:cted separate:~y on asphaltic concrete cores. The te~t
(~·su:t.;; aloni,x·,tft thp- nw:lber of .;ores tested are shown in Figure 11 and dis~
·.:"";"cd below.
LC varied :..Y(;!TI 2.261 to 2.270 glee for SEA pavpment layers in compari~:!
1 -::.~ 2.27& tG 2 . .>3, g/..:.c ;:01 aspr.alti<: c.oncrete pavement courses. Higher
·j(r:S~1:if~~ ":C1~: ,)~J:::~cr;"L~J \iit~\ the cores ext!""ar::ted in May 198~, obviously due
:.C: ;;.dcii t~ 01;:,.1 tTaftic dens,ific::llioll a.S comJ):lred l;i th the- cores extracted
Res i llent ~1od"lus
it vn ~fil .·~·r-,m 2.84 to S. 37 ;; lOoKPa fOl SEA pavement layers versu~
~.31 to 4.83 x 10 0 KPa for asphaltic concrete naterial. The coefficient of
·J:,,·i:?~.io1 in the nil) cases varied froJT1 !9.1 to 32.6 and 27.1 to 37.7% res[
wet i vely. The hi ghest value:> in bot h c "ses were again associated with th~
l,)SZ cores as they iHld been subjected t.o densification under traffic for
the longest period. Greater rise in MR value of SEA cdncrete relative to
asphal t j C Cc)!lcn·t;·" T!lay be attributed tl) changes in the internal structure
in~luding sulphur recrystali::ation kncMn to occur in the SEA pavement layers.
~lit .Tensile Strength
Tllis test was conducted on the 1982 cores, selecting those taken from
463
certain location along the Benkelman beam deflection points. There were
22 such cores out of which 12 were of SEA concrete and 10 of asphal tic
concrete. Split tensile strength of SEA concrete was found to be 1193.9 KPa
versus 1577.9 KPa for asphaltic concrete, the coefficient of variation in
the two cases being 7.9 and 10. sg" respectively. Although split tens ile
strength of SEA concrete was 24.3% lower than asphaltic concrete, the MR
value was found to be 11.1% higher than asphaltic concrete.
DISCUSSION
The general appearance and inservice data obtained so far from the
three SEA test projects described earlier ascertain the applicability of
SEA technology to conditions in Saudi Arabia. There have been, however,
undesirable and unexpected consequences that should be dealt with and are
conunented below.
Earlier Cracking of Section B - Test Road Two
The fact that Section B (the 20% thinner base course section) has
cracked and is approaching failure earlier than expected, is of concern.
Although it is true that SEA layers are inherently stiffer than conventional
pavements particularly at a high S/A ratio such as 45/55; the cause
of cracking in this case may be attributed partly to factors that are not
related to the presence of sulphur in the pavement. In this instance, analysis
of inservice results plus earlier construction information gathered
on the subbase have revealed that lime stabilized subbase in Section B,
unlike the rest of the road, was prepared in a hurry and it is doubtful
whether it had attained sufficient strength prior to placement of SEA
layers. This fact coupled with the extremely heavy truck traffic that the
road has been receiving since its opening, are ample reasons for initiation
of cracks which were detected early and have been increasing ever since.
The experience described with respect to potential reduction in SEA
pavement thickness is inconclusive. It does point out, however, th~t
extreme care should be exercised when considering reduction in thickness
if sub-surface conditions are not sufficiently strong and traffic intensity
is very high,
Water Induced Damage of SEA Mixes - Test Road Three
Early mix design data using the Marshall test have revealed that the
fine fraction of some of the limestone aggregates available in Eastern
Saudi Arabia are prone to hater particularly when used in SEA mixes. ll1e
exp~rience described earlier with regard to the mix design of Test Road Three
attests to the fact that cement or hydrated lime. need to be added to the SEA
mix to increase the 2~-hour Marshall stability in order to meet specification.
For Test Road Three, 1% cement by weight of aggregate was later added
to the SEA mix at the pugmill.
This w:\ter induced d:lmage of SEA mixes evident when using certain limestones,
has been attributed to two general causes: (i) potential adverse
466
reactions between sulphur and certain limestones which could initiate microcr;
lcks in the SEA m;ltrix followed by macro-cracks and subsequent failure (ii)
the in;ldequacy of SEA bindeTs to hold some of the limestones aggregates together
and provide the water proofing properties required during early cure.
Econo~lic Considerations of SEA Pavements
The commercialization of SEA binders and SEA mixes is undoubtedly a
flIlction of their relative perfonnance and relative cost to asphaltic concrete.
Gene~ally since sulphur is twice as dense as asphalt cement, the cost
of sulrhuT must be equal or less than half the cost of asphalt cement for
positive economics when sulphur replaces asphalt on equal volume basis.
The price of sulphur and asphalt have risen sharply in the last few
years in response to world market conditions created by energy crisis. Undoubtedly,
their relative price will vary from one region to another depending
on a nu~ber of factors including availability and marketability.
In Saudi Arabia, both sulphur and asphalt are produced locally and are
available at relatively low price if compared with international market pricing.
AdJed advantages that Saudi Arabia may derive from SEA applications
arc: th,~ potential benefits of using sulphur-extended asphalts with lower
gr.ldc aggrcg;ltcs including desert sands. This work reported on by Akili (16)
is currentl~: in progress and appears to hold promise.
SU:-NARY AND CONCLUSIONS
Rlending liquid sulphur and asphalt produces a new binder termed sulphur
extended asphalt (SEA), where up to SO weight percent of asphalt
cement can be replaced by sulphur in the production of asphaltic concrete
mixes.
Three sulphur-extended asphalt (SEA) pavement projects have been
successfully constructed in Eastern Saudi Arabia as pa~t of an ongoing proI(
am that aims at utilizing the by-product sulphur, produced from the gas
~"therinC1 oroiect.: of Eastern Saudi Arabia. Two test roads (One and Three)
;~ere p:;,v~d wit!". 30/70 weight percent of sulphur/asphalt (S/A) binder. Test
Road Two, a scgmen"( of a heavi ly trafficked expressway, has S/.4. ratio of
45/55. All three roads are under surveillance. The field and laboratory
informJtions obtained to date support the following:
1. SEA pavement mixes can be produced, hauled, placed and compacted
with conventional methods and equipment.
,,) [l"oed on inscrvice infoI1TIation the engineering properties of SEA
mix(~ are c~~parJ.ble to conventional asphaltic concrete mixes
cxcq)t for the stiffening effect that sulphur imparts to SEA mixes
whell S/.\ catlos are higher than 30/70.
Tle ":1!11.,,··":tly stiffer SE..\ mix of T0st RO;lJ Two. where S/A ratio
i u,,;. r::." resulted in earlier cracking in a thinner section
467 |
|
Publishing Year |
: |
1983 AH
|
|
Added Date |
: |
Tuesday, January 13, 2009 |
|