Top down cracking appears to be a common mode of HMA pavement distress in at least several states and countries. Traditionally, pavement cracking is thought to initiate at the bottom of the HMA layer where the tensile bending stresses are the greatest and then progress up to the surface (a bottom-up crack). Most traditional transfer functions used in mechanistic-empirical structural design are based on this concept including the one discussed in this Guide. However, the late 1990s saw a substantial focus on a second mode of crack initiation and propagation: top-down cracking.
Figure 1: Likely Top-Down Cracking on I-37 South of San Antonio, TX
Although not fully understood at this time (mid 2003), there are three basic views on the of top-down cracking mechanism:
Likely,
the mechanism is some combination of the above. The bottom line is that
HMA top-down cracking is not thoroughly understood and, at this time, is
generally not considered as a causative factor for pavement cracking although it
probably should be. Further, for two states that recently studied cracking
origins (Florida and Washington State), both reported that top-down cracking is
far more common than assumed. In fact, the Florida DOT reports that
top-down cracking is dominant for their HMA pavements due for rehabilitation.
Currently, the National Cooperative Highway Research Program (NCHRP) is
addressing the issue with Project 01-42: Identification of the Design
Conditions and Critical Factors That Are Related to the Top Down Fatigue
Cracking Mechanism.
Two simple suggestions may help in the identification of top-down cracking. First, in thick HMA pavements, consider top-down cracking as a possible cracking mechanism. Generally, previous research has found that in pavements thicker than about 160 mm (6.3 inches) top-down cracks can be and often are the dominant form of cracking. Do not assume pavement cracks are bottom-up. Second, before deciding on a maintenance and/or rehabilitation strategy, take a pavement core on a suspect crack (see Figure 2). Usually, a pavement core will show whether a crack is top-down or bottom-up. It will also show the extend to which the crack has propagated, thus defining the extend of needed milling prior to overlay.
In general, world-wide literature can summarize the problem in the following list:
Gerritsen et al. (1987) reported that pavements in the Netherlands were
experiencing premature cracking in the
wearing courses. Further, the cracks did
not extend into the
intermediate/binder course. These surface cracks occurred both
inside and outside the wheelpath areas, and, in some cases, soon after paving. This caused Gerritsen et al. to conclude that there was likely more than one
causative effect. The surface cracking outside of the wheelpaths had low mix
strength characteristics at low temperatures. Further, they noted low binder
penetration values could be related to higher thermal stresses. The surface
cracks in the wheelpath areas were largely attributed to radial shear forces
under truck tires near the tire edges. Their conclusion was that both thermal
and load related effects caused the observed surface cracking. They recommended
that the binder film thicknesses be increased to reduce early age hardening of
the mixes.
Dauzats et al. (1987) also published results that described surface initiated
cracking on pavements in France. They noted that the cracks could be either
longitudinal or
transverse and occurred typically three to five years following
construction. They estimated that these types of cracks were initially caused by
thermal stresses and then further propagated by traffic loads. It was noted that
a rapid hardening of the asphalt binder likely contributed to this type of pavement
distress.
Work reported by Matsuno and Nishizawa (1992) noted that longitudinal surface
initiated cracking of the HMA
wearing course was commonly observed in Japan about
one to five years following construction. Their observations and analyses are of
special interest. First, they observed that the
longitudinal cracks did not
extend under overpasses (shaded areas). Second, analysis of
FEM results showed
that very high tensile strains occur at the edge of truck tires at or near the
surface of the HMA wearing course. These high strains occur when the upper
portion of the HMA is at a low stiffness due to high surface temperatures. They
also noted that if the HMA is not hardened due to
aging effects, the small cracks
that form are eliminated by the kneading action of tires. This changes as the
HMA
ages. They analyzed two thicknesses of HMA: 200 mm (8 inches) on heavy traffic routes and l00
mm (4 inches) on light traffic routes. For both thickness cases using a peak surface temperature of
60°C (140°F) (decreasing with depth) and associated stiffness of about 200 MPa
(29,000 psi) at 60ºC (140°F),
they reported similar tensile strains of over 1400x10-6 mm/mm
(inch/inch) near the pavement
surface. Thus, they concluded that HMA thickness is not a major factor with this
type of cracking.
A study on large transport vehicles and their effects on pavements was reported
by Craus et al. (1994) in work done for the California Department of
Transportation. Their analyses showed that large tensile strains occur at the
top of the HMA wearing courses. Specifically, these strains are due to high tire
edge stresses for conditions where the upper HMA is at a low stiffness due to
high surface temperatures (stiffness ratios of less than 0.5 produced the largest
tensile strains). It is of special interest that the California and Japanese
studies drew the same conclusions concerning the cause of surface initiated
cracking.
More recently, Nunn (1998) reported that surface initiated cracking was common
on UK motorways. Typically, these surface cracks were observed about 10 years
after paving. Nunn noted that for pavements with HMA thicknesses exceeding
180 mm (7 inches), there was no evidence of fatigue cracking in the lower
intermediate/binder
course—only the wearing courses. Additionally, he showed that there was
a discontinuous relationship between the rate of rutting and the thickness of
the HMA layers. For combined HMA thicknesses greater than 170 mm (6.7
inches), the
rutting rates on about 50 pavement sections were about 200 times less than for
HMA layers with thicknesses less than 170 mm (6.7 inches). For sections with less than
170 mm (6.7 inches) of HMA the rutting rates were about 100 mm (1 inch) per million
ESALs and 0.4 mm
(0.016 inches) per million ESALs for greater than 170 mm (6.7 inches). Such
dramatic measurements suggest that a very different distress mechanism occurs at
the “breakpoint” thickness. Nunn also summarized recent work performed in the
Netherlands that showed for HMA thicknesses exceeding 160 mm (6.3 inches), cracks
initiated at the pavement surface and eventually penetrated to a depth of about
100 mm (1 inch). He also noted that the Netherlands work indicated for full depth cracks
in thinner pavements that the cracks propagated from the top of the pavement
surface downward. Nunn showed that the surface initiated cracking in the UK
could be either longitudinal or transverse. The transverse cracks were related
to low binder penetration values (typically about 15). He also stated that the
pavement sections with and without surface cracking had no significant
difference in measured deflections. He concluded the cause of the surface
initiated cracking was due to horizontal tensile stresses generated by truck
tires at the HMA surface. Wide based tires generated the highest tensile
stresses.
Myers et al. (1998) reported that surface initiated cracking in Florida was found
to represent 90 percent of the observed cracking in pavements scheduled for
rehabilitation. Thus, this type of cracking predominates in Florida. They noted
that this type of cracking is generally observed on pavements five to ten years
following construction. The HMA thicknesses in their study ranged
from 50 to 200 mm (2 to 8 inches). The cracks were most often longitudinal with surface crack
widths of about 3 to 4 mm (0.12 to 0.16 inches) decreasing with depth. The total crack depths ranged
from about 25 mm (1 inch) to the full depth of the HMA layer. Based on
computer modeling, it was concluded that tensile stresses under the treads of
the tire—not the tire edges—were the primarily cause of the cracks. Further,
wide base tires caused the highest tensile stresses. They noted that the tensile
stresses dissipate quickly with depth suggesting that this might be the reason
the cracks essentially stop growing; however, they felt this needed further
investigation. They concluded that surface initiated cracking is not a
structural design issue but more related to mixture composition. Specifically,
they concluded that more fracture resistant asphalt mixes are needed.
At the January 2000 TRB Annual Meeting, Uhlmeyer et al. reported that top-down
cracking is common to thicker Washington State DOT HMA surfaced pavements
(top-down cracking was typically observed when the average HMA thickness was
about 160 mm (6.3 inches) or greater). Such cracks were generally contained in the
wearing course and averaged 46 mm
(1.8 inches) in depth. The top-down cracks generally initiated
within three to eight years of paving. No hypothesis as to cause was made.
An initial review of the literature shows that the following references are
relevant to the topic: