; Premature Pavement Distress
Learning Center
Plans & pricing Sign in
Sign Out
Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

Premature Pavement Distress


  • pg 1
									Premature distress of a pavement on expansive black cotton soil in the
Horn of Africa
M. B Mgangira
Senior Researcher, CSIR Built Environment, PO Box 395 Pretoria, South Africa
P Paige-Green
Chief Researcher, CSIR Built Environment, PO Box 395 Pretoria, South Africa

ABSTRACT: This paper discusses a typical example of the distress experienced by a flexible pavement con-
structed over black clay soil, also commonly known as black cotton soil, where minimal precautionary meas-
ures were implemented. Black cotton soils are generally recognized as expansive soils and change signifi-
cantly in volume with changes in the moisture condition due to wetting-up and drying-out during the wet and
dry seasons respectively. The volume changes lead to differential deformation which then induces distress
and leads to damage of the pavement constructed on such soils. By way of example the paper describes an in-
vestigation of a road in the Horn of Africa in an area with extensive deposits of the black clays, demonstrating
the importance of correctly characterizing these soils prior to the pavement design and taking appropriate
counter measures against expected movement of the roadbed that will cause damage to the pavement..

1 INTRODUCTION                                            sequently no appropriate countermeasures were con-
                                                          sidered, led to this investigation.
Black clay or cotton soils, defined by Morin (1971)          The investigation was conducted in two main
as dark grey to black soils with high clay content,       phases. The first phase comprised a review of the
usually over 50%, are commonly expansive. Expan-          design traffic data, project specifications, engineer-
sive soils are susceptible to volume changes, swell-      ing design analysis and a visual assessment of the
ing as water is absorbed and shrinking during the         pavement condition along the project route with a
drying process as a result of moisture adsorption by      final report submission to the client. The second
the smectite clays. Thus a road constructed over          phase involved a field investigation programme in
such soils will be subjected to cyclic volume change      conjunction with a sampling programme to obtain
with changes in the moisture condition between the        material for laboratory testing of the roadbed mate-
wet and dry season. The swell-shrink phenomenon           rial characterization (Atterberg limits, particle size
                                                          distributions, etc). Detailed profiling was conducted
will consequently have an effect on the performance
                                                          in each of the six test pits. The test pits were exca-
of a pavement due to the induced stresses and strains     vated at locations that were severely distressed as
following the movements caused by the volume              well as at less distressed locations. In addition to
changes in the subgrade material. This paper dis-         field and laboratory data collection, a further review
cusses the investigation of a section of a rehabili-      of previous investigation reports was conducted. The
tated road (involving widening) in the Horn of Af-        results of the roadbed characterisation together with
rica region, which was constructed over black cotton      the recorded observations of the profile exposed in
soil and developed excessive longitudinal cracks one      each of the six test pits and information from the
year after construction, requiring remedial measures.     previous investigation reports were used to establish
It should be appreciated that the region’s climate has    the cause of the longitudinal cracking.
a marked wet and dry season which means the sub-
grade will be expected to swell and shrink with the       2 ROADBED MATERIAL CHARACTERISTICS
changing seasons.
    The need by the contractor to confirm whether         In addition to carrying out routine geotechnical en-
the observed distress was related to the roadbed          gineering tests on the roadbed material, X-Ray Dif-
characteristics which may not have been adequately        fraction (XRD) test results, conducted previously
taken into account during the design phase and con-       and during this investigation were also referred to.
                                                          The results showed the presence of montmorillonite
and both kaolinite and halloysite. The principal ex-      dex properties such as grain size distribution, clay
changeable cation was calcium. Magnesium was              content and plasticity are the most widely used in
present in significant quantities with potassium oc-      practice for identifying and classifying expansive
curring in minor quantities. The parent material of       soils (Nelson & Miller 1992). Consequently, the de-
the black cotton soil (BCS) was a Cenozoic basalt.        termination of the swell potential for the subgrade in
Both residual and transported clays from the weath-       this investigation was based on a number of methods
ered basalt occurred on the project.                      that establish a relationship between expansion po-
   Table 1 gives a summary of the roadbed proper-         tential and Plasticity Index and clay contents. Table
ties. In total, 116 moisture content determinations       2 summarizes the classification of potential swell of
and 36 plasticity index tests were performed. It          the subgrade using various techniques.
should be noted that the field investigation was car-
ried out at the end of the wet season.                    Table 2. General classification of potential swell of
                                                          the roadbed
Table 1. Physical properties of the subgrade                    Researcher           Classification             Remark
             Property          Unit       Value Range      Carter & Bentley     High to very high     Potential swell based
 Specific gravity               -          2.43 – 2.45     (1991)                                     on plasticity index
 Percent clay                   %            18 - 73       Chen (1988)          High to very high     Expansive soil classi-
 Percent silt                   %            25 -70                                                   fication based on plas-
 Percent sand                   %            3 - 12                                                   ticity index
 Liquid limit                   %           43 - 103       Holtz & Gibb         High to very high     Potential swell based
 Plasticity index               %            26 - 54       (1956)                                     on plasticity index
 Activity                       --         0.60 – 1.11     Nelson & Miller      Critical degree of    Expansive soil classi-
 Linear shrinkage               %            18 - 39       (1992)               expansion             fication based on lin-
 Optimum moisture content       %            22 - 49                                                  ear shrinkage
 Natural moisture content       %            24 - 53       Seed et. al (1960)   High to very high     Evaluation of potential
 Maximum dry density          Kg/m3        1108 - 1278                                                expansiveness based
                                                                                                      on clay fraction and
                                                                                                      plasticity index
    From the results in Table 1 and visual assessment      Van der Merwe        High to very high     Potential severity of
of the soil, the general description of the roadbed is     (1964)                                     volume change based
dark grey silty clay with high to very high plasticity.                                               on clay content and
The CBR of the roadbed material was generally less                                                    weighted plasticity in-
than 4 %.                                                                                             dex
    The objective of carrying out the tests was to
fully characterize the subgrade material to determine         The classification methods show that a soil hav-
its swell potential. The following section discusses      ing clay content in excess of 30 % and a Plasticity
the approach used in determining the swell potential      Index of greater than 35 % will be classified as a soil
of the subgrade                                           with high swelling potential. The average Plasticity
                                                          Index of the subgrade was 41 % with a clay content
3 EVALUATION OF THE BLACK COTTON                          average of 48 %. The results in Table 2 show that,
  SOIL AS CONSTRUCTION ROADBED                            irrespective of the method used, the subgrade has a
                                                          high to very high predicted swell potential. The ac-
3.1 Expansiveness potential                               tivity of the samples ranged between 0.68 and 1.28
                                                          based on the definition suggested by Seed et al
   Several investigators have proposed methods for
                                                          (1962). The average value for the activity was 0.85.
determining the swelling properties of soils (Holtz &
                                                              From the above analysis, it can conclusively be
Gibb 1956; Seed et al 1962; Van der Merwe 1964;
                                                          stated that the subgrade is expansive and will there-
Chen 1988; Carter & Bentley 1991; Kariuki & van
                                                          fore be susceptible to volume change with variations
der Meer 2003; Rao et al 2004; Nwaiwu & Nuhu
                                                          in the moisture condition.
2006 and others in Nelson & Miller 1992, to men-
                                                              Figures 1 and 2 illustrate the moisture content
tion a few). The basis for the determination of the       distribution within the subgrade at two test pit loca-
swelling potential differ, but generally range from       tions. The distribution of the in-situ moisture con-
methods based on the Plasticity Index, to those that      tent shows a variation longitudinally as well as ver-
include placement conditions (Rao et al 2004) or          tically.
those based solely on properties such as electrical
conductivity and specific gravity (Nwaiwu & Nuhu
2006). However, classification tests for the soil in-
                                                                                                                      Test Pit 1


          In-situ Moisture Content (%)




                                                                                                                                                                                                             Under carriageway
                                                                                                                                                         Middle of shoulder

                                                                                                                                                                                          Edge of pavement
                                                                                                 6.65 m from CL

                                                                                                                                    4.95 m from CL
                                                  9.90 m from CL

                                                                                                                                                                                                             0.9 m from CL
                                                                    8.2 m from CL


                                                                                                                  Depth below BCS surface level
                                                                                    0-20 cm    20 - 40 cm                          40 - 60 cm                60 - 80 cm

                                                 Figure 1: In-situ moisture content distribution within the roadbed: Test Pit 1

                                                                                                                      Test Pit 3



             In-situ moisture Content (%)




                                                                                                                                                           Middle of Shoulder

                                                                                                                                                                                                             Under carriageway
                                                                                                                                                                                         Edge of Pavement
                                                   10.2 m from CL

                                                                                                                               6.0 m from CL

                                                                                              7.5 m from CL

                                                                                                                                                                                                             1.1 m from CL
                                                                    8.9 from CL




                                                                                                                  Depth below BCS surface level
                                                                                    0-20 cm       20-40 cm                     40-60                 60-80                      80-100

                                                 Figure 2: In-situ moisture content distribution within the roadbed: test Pit 3

   It is clear from the distribution that the area be-                                                                                                  The results of the Liquid Limit and Plasticity
low the centre-line or near the centre-line of the                                                                                                   Index tests together with previous consolidometer
road is drier than in the shoulder areas as would be                                                                                                 test results were used to estimate the expected
expected under equilibrium conditions (Emery                                                                                                         swell pressures under the road. Swell pressure val-
1992). Higher moisture content in the areas below                                                                                                    ues were used to estimate the depth of the active
the embankment slopes is also evident.                                                                                                               zone of expansive soils (Rao et al. 1988). A more
                                                                                                                                                     reliable method to locate the active zone is by car-
3.2 Estimated heave                                                                                                                                  rying out soil suction measurements (Chen 1988)
                                                                                                                                                     but such testing was not possible during this inves-
   It is important that the expansiveness of the soil
                                                                                                                                                     tigation. Identification of the active zone which
is adequately identified for appropriate design so-
                                                                                                                                                     gives the depth of material that will contribute to
lutions (Nelson & Miller 1992). In this regard it
                                                                                                                                                     the total heave at the top of the expansive clay
becomes essential that the magnitude of expected
                                                                                                                                                     layer is necessary in order to compute the esti-
heave is determined.
mated heave using the proposed closed solution in                    The method presented in Rao et al (1988) was used
Rao et al (1988). It is this depth of material that                  to calculate the values in Table 3. Estimated heave
should optimally be removed and replaced with an                     values using the method of van der Merwe (1964;
inert material in order to achieve a stable moisture                 1976) are also given
state, which typically ranges between 1 m and 4 m,                      Only roadbed material characteristics within the
but greater values are not uncommon (Nelson &                        shoulder and edge of the pavement were used for
Miller 1992).                                                        the computation as this area was deemed critical in
   Table 3 gives the estimated swell pressure val-                   the observed pavement distress. The final pressure
ues and estimated heave values in the event that                     on the surface of the roadbed was calculated by
unstable moisture conditions exist in the roadbed                    considering the layer thicknesses and the corre-
and assuming that 1.0 m of the expansive roadbed                     sponding unit masses at the shoulder position for
material had been removed and replaced by a suit-                    the different test pit locations.
able non-expansive material. This was the sug-
gested solution after the first cracking appeared.

                                    Table 3: Roadbed heave potential characteristics
                 Test Pit     Estimated      Estimated fi-     Estimated      Estimated heave   Estimated heave
                            swell pressure    nal pressure    depth of ac-          (mm)         (mm) Van der
                               (kN/m2)       on Black Cot-   tive zone (m)          Rao et al    Merwe (1976)
                                                ton Soil                           (1988)
                 1                 63                14            4.63               50                70
                 2                 49                9             3.60               33                43
                 3                 48               11             3.20               26                42
                 4                 44               38             2.82               17                25
                 5                 39               23             2.58               10                25
                 6                 39               39             2.58               10                18

   Rao et al (1988) have shown that partial exca-                    the material was also observed within the red clay
vation of expansive soil does not totally reduce the                 fill used for the shoulders.
total heave. This is confirmed in Table 3, which
shows that even after removing 1.0 m of unsuitable
material, movement could still take place. Note
that test pits 5 and 6 were in a road section show-
ing minimal distress.


The objective of estimating the expected heave
was to establish whether the design adequately
took the effect of anticipated swell pressures or
vertical movement into account. The results in Ta-                        Figure 3: Slickensided condition within the roadbed
ble 3 show that the condition could have been im-
proved by removal of the unsuitable material to a                       The moisture content distributions shown in
depth greater than 1.0 m and backfilling with inert                  Figures 1 and 2 reveal that the soil under the em-
material. The records showed that, on average, less                  bankment slopes and the shoulder area would be
than 1.0 m was removed from beneath part of the                      subject to wet and dry seasonal cycles, compared
fill, and this only after problems were first noted.                 with the soil under the carriageway. It is interesting
What this implies is that the pavement probably                      to note that, using one of the equations suggested
experienced greater movements than those shown                       by Emery (1992) for the prediction of equilibrium
in Table 3. Inspection of the roadbed exposed in all                 moisture content (EMC), which considers the per-
of the test pits showed the presence of severe                       centage passing 0.425 mm sieve and the linear
slickensides (Figure 3). This phenomenon which is                    shrinkage, the predicted EMC values under the car-
an indication of repeated movement/strain within                     riageway are 36.86 % and 29.72 %, compared to
                                                                     the measured values of 35.05 % and 30.6 %, as
presented in Figures 1 and 2. As the roadbed has                 There is a noticeable deformation, relative to
been confirmed to be expansive, it means that soil           the string, of the layers around the crack area. The
in the immediate vicinity at the edge of the pave-           convex-up shape of the upper surface of the se-
ment will expand more during the wet season and              lected layer near the crack is indicative of expan-
that the vertical movement will result in heave at           sion of the underlying clay material.
the edge of the pavement. During the dry season                  It is therefore essential that the design engineer
the soil shrinks and the edge of the road subsides.          addresses the influence of expansive soils both as
Roads are intolerant of this alternating heave and           naturally occurring undisturbed soils and as the
shrinkage and ultimately the movement produces               compacted, soils used in the road formation. Thus
longitudinal cracking (O’Connel & Gourley 1993)              it is necessary to identify the source of expected
as shown in Figures 4 and 5, resulting in weaken-            heave or shrinkage before selecting the final design
ing of the pavement along this location.                     solution (Nelson & Miller 1992).
                                                                 The obviously high expansive potential charac-
                                                             teristics of the roadbed certainly required special
                                                             design and construction strategies to counter the
                                                             anticipated damage due to volumetric movements
                                                             of the roadbed associated with seasonal moisture
                                                             changes. This was not adequately addressed in the
                                                             rehabilitation design or contract documents, espe-
                                                             cially as the rehabilitation included excavation of
                                                             part of the existing road structure (essentially the
                                                             damaged outer wheel track areas) and thus distur-
                                                             bance of the moisture equilibrated subgrade be-
Figure 4: Illustrative form of typical longitudinal cracks   neath the existing road.
                                                                 The overall performance of a pavement also de-
                                                             pends on the characteristics of the materials used
                                                             for constructing the other layers. Volcanic ash and
                                                             a tuffaceous material as well as weathered cinders
                                                             were used as fill for widening the road shoulders
                                                             and to raise the road. This was supposed to have
                                                             been relatively impermeable and inert to minimise
                                                             moisture and volumetric movements beneath the
                                                             shoulder area. Permeability test results previously
                                                             conducted on the materials showed that the materi-
                                                             als were not adequately impermeable. Red silty
                                                             clay was also used as backfill in sections where the
                                                             partial replacement technique was used as a correc-
Figure 5: Formation of longitudinal cracks near the edge
of the pavement
                                                             tive solution following the early development of
                                                             cracking. Based on tests on material from two bor-
   Figure 6 shows the location and the extent of a           row pits used on the project, the liquid limit values
typical crack at the edge of the pavement.                   for the replacement red silty clay ranged between
                                                             50 % and 63 % while the Plasticity Index values
                                                             varied between 13 % and 25 %. The clay content
                                                             was between 16 % and 68 %, with between 86 and
                                                             93 % of the material finer than 0.075 mm. Clearly
                                                             this shows that the fill materials themselves would
                                                             possibly require measures to be taken to minimize
                                                             potential expansiveness under fluctuating moisture
                                                             content conditions.
                                                                 Evaluation of the documentation showed that
                                                             there was a fundamental lack of appreciation of the
                                                             impact of the roadbed characteristics on the per-
Figure 6: Propagation of cracks into layers
                                                             formance of the pavement in the original design,
                                                             despite the evidence of considerable volumetric
movement of the existing road structure prior to        field investigations. Much of the cracking was
rehabilitation.                                         also associated with the presence of eucalyptus
   Correct interpretation of the soil characteriza-     (and other) trees alongside the road. No attempts
tion test results combined with any experience in       were made to remove these and ensure that they
similar materials should have revealed that the         were not allowed to re-establish.
roadbed was potentially highly expansive and that          It should be pointed out that during the course
there was therefore the need to predicting the po-      of the investigation, some aspects of the construc-
tential in situ soil movements likely to occur under    tion were found to not entirely conform to the
the pavement. Calculations could have been made         technical specifications. Some layer thicknesses
using several methods to estimate the potential         were less than specified, whilst several instances
heave along the project route as discussed. From        were seen where the constructed layer thickness
such an exercise, sections of the road on which         was significantly higher than the specified design.
countermeasures were required could have been           However, the variability of the layer thicknesses
                                                        found in the course of this study, which is not out
identified. The design should then have taken into
                                                        of the ordinary in construction practice, was not
account the impact of the estimated movement on         deemed to be the cause of the longitudinal cracks. I
the pavement structure. There was, however, no
evidence to suggest that such an exercise was per-      5 CONCLUSIONS
formed in the pavement rehabilitation design, and
thus no special solutions for dealing with the ex-      The paper has demonstrated by an example of a
pansive roadbed were considered necessary at that       typical pavement failure condition and the impor-
stage.                                                  tance of correctly characterizing problem soils so
   Various countermeasures have been suggested          that appropriate measures can be taken at the de-
in the literature (Chen 1988; Basma & Tuncer            sign phase of both new and rehabilitation works. It
1991; Nelson & Miller 1992) and some of these           shows that:
should have been specified to minimize the antici-          • Ignoring the systematic assessment of such
pated movement of the expansive soil during and                 soils is most likely to lead to the premature
after construction while an equilibrium moisture                distress of road structures.
condition was being achieved. The design should             • The construction of roads over problem
have ensured that this equilibrium moisture content             soils such as black clay demands that addi-
be maintained as far as possible during and after               tional requirements should clearly be stated
construction to minimise the impact of movement                 in the contract documents to minimize the
on the pavement structure. The likelihood of avoid-             damage to road structures caused by vol-
ing all movement and consequential distress is un-              ume changes in these soils as a result of
realistic in most cases, but the severity of the dis-           moisture variation during construction and
tress can be controlled. The project documentation              in service.
review revealed that the need to control moisture          The conclusion drawn from the investigation
variation during construction was only identified       was that the cause of the premature distress of the
after problems occurred.                                road was that the potential expansiveness of the
   As problems were encountered on the project          subgrade materials was not timeously identified
various remedial measures were proposed by the          and inadequate precautionary measures were thus
engineers. These included selective removal and         taken to counter the effects of differential seasonal
replacement of the expansive subgrade soils, and a      moisture variation in the expansive roadbed.
number of cost-saving option such as benching of
the clays towards the edge of the pavement. It is       6 ACKNOWLEDGEMENTS
considered by the authors that this would have ex-
acerbated the problem as a result of the variable       This paper is published with the permission of the
thickness of the clay material and interrupted flow     acting Director, CSIR Built Environment.
paths of water within the road prism. Flattening of
the slopes with the clays removed from the sub-         7 REFERENCES
grade for replacement with inert materials was car-
ried out as the project progressed. Little attention    Basma, A.A & Tuncer, E.R., 1991. Effect of lime
was paid to the effective removal of water from ad-       in the volume change and compressibility of
jacent to the road prism and the drainage around          expansive clays. Transportation Research
culverts. This was clear from the regular presence        Board, Washington, D.C TRR 1296: 54-61.
of standing water along the roadside during the         Carter, M & Bentley, S.P 1991. Correlation of soil
                                                          properties, Pentech Press, London
Chen, F.H. 1988. Foundations on expansive soils.      Conf for Africa on Soil Mech and Found Engng,
  Developments in Geotechnical Engineering,           2: 166-167.
  Vol 54. New York: American Elsevier Science
Emery, S J 1992. The prediction of moisture con-
  tent in untreated pavement layers and an appli-
  cation to design in southern Africa. Bulletin 20,
  CSIR, Pretoria.
Holtz, W.G & Gibbs, H.J 1956. Engineering prop-
  erties of expansive clays. Transactions, Ameri-
  can Society of Civil Engineers, 121:641-677.
Kariuki, P.C & van der Meer, F 2003. Issues of ef-
  fectiveness in empirical methods for describing
  swelling soils. International Journal of Applied
  Earth Observation and Geoinformation, 4:231-
Morin, W.J 1971. Properties of African tropical
  black clays. Proceedings, 5th Regional Confer-
  ence for Africa on Soil Mechanics and Founda-
  tion. Luanda, Angola. 1: 51-59.
Nelson, J.D & Miller, D.J. 1992. Expansive soils:
  Problems and practice in foundation and pave-
  ment engineering. New York: Wiley.
Nwaiwu, C.M.O and Nuhu, I 2006. Evaluation and
  prediction of the swelling characteristics of Ni-
  gerian black clays. Geotechnical and Geologi-
  cal Engineering, 24:45-56.
O’Connell, M.J & Gourley, C.S 1993. Expansive
  clay road embankment in arid areas: moisture-
  suction conditions. TRL report PA1299/93,
  DFID, Berkshire, United Kingdom.
Rao, A.S, Phanikumar, B.R & Sharma, R.S 2004.
  Prediction of swelling characteristics of re-
  moulded and compacted expansive soils using
  free swell index. Quarterly Journal of Engi-
  neering Geology and Hydrogeology, 37: 217-
Rao, R.R., Rahardjo, H & Fredlund, D.G 1988.
  Closed-form heave solution for expansive soils.
  Journal of geotechnical engineering, 114
Seed, H.B., Mitchell, J.K & Chan, C.K 1960. The
  strength of compacted cohesive soils. Proceed-
  ings, ASCE Research Conference on Cohesive
  Soils, Boulder, pp 877-964.
Seed, H.B., Woodward, R.J & Lundgren, R 1962.
  Prediction for swelling potential for compacted
  clays. Journal of the Soil Mechanics & Founda-
  tion Engineering Division, American Society of
  Civil Engineers. Part 1, 88 (SM3):53-87.
Van der Merwe, D H 1964. The prediction of
  heave from the plasticity index and the percent-
  age clay fraction of soils. Trans. S Afr. Instn
  Civ. Engnrs, 6(6): 103-107.
Van der Merwe, D H 1976. Plasticity index and
  percentage clay fraction of soils. Proc 6th Reg

To top