Based on previous HVS experiments on new pavement and on simulation of the WesTrack experiment the following equations were found to be capable of estimating the severity, or density, of surface cracking, in m/m2, reasonably well:
Equation 6: Model for estimating damage at crack initiation.
| Where |
ωi is the damage at crack initiation, and
hAC is the combined thickness of the asphalt layers. |
Equation 7: Model for estimating crack density (severity)
| Where |
Cr is the crack density (severity in m/m2),
ω is the damage to the surface layer, and
ωo is a constant. |
ωo was determined based on the assumption that crack initiation corresponds to a severity of 0.5 m/m2.
Figure 5 shows the crack severity, in m/m2 of the wheel track, as a function of the fatigue damage determined from the strain at the bottom of the original asphalt layer. This damage is used to reduce the average modulus of both the original asphalt layer and of the overlay. In Figure 5 the original pavement sections, before overlay, are designated by their test numbers. The overlayed sections are given by the type and thickness of the overlay.
The heavy curves indicated by a thickness value are the crack severities calculated using Equation 6 and Equation 7. The approximate thickness of the asphalt layer before overlay was 80 mm, and the thicknesses of the combined asphalt layers after overlay were either about 125 mm or 170 mm.
Figure 5: Surface cracking as a function of fatigue damage
The original sections, before overlay, are seen to crack more rapidly than predicted from the equations. FWD tests showed that the modulus of the asphalt layer before HVS testing was considerably below the modulus from frequency sweep tests in the laboratory, whereas the moduli of the overlays from FWD tests were in good agreement with the frequency sweep data. It is possible that the low in situ modulus of the original asphalt layer was due to some initial damage to the material. If that were the case, this initial damage should be added to the fatigue damage in Figure 5. This would shift the curves to the right.
For the overlay sections the observed cracking does not correspond to the respective thickness curves. A better fit can be obtained if the reflection damage calculated from the strain over the existing cracks (Equation 1) is used with the following equations:
Equation 8: Model for reflection damage at initiation of reflection cracking
Equation 9: Model for estimating reflection crack density, as a function of reflection damage.
Reflection crack initiation was assumed to correspond to a density of 0.5 m/m2.
Figure 6 compares the observe reflection cracking on the overlay sections to the reflection damage predicted using Equation 8 and Equation 9, as a function of the reflection damage calculated from Equation 1 and Equation 3.
Figure 7 shows the predicted reflection cracking severity as a function of the observed severity.
Figure 6: Surface cracking as a function of reflection damage.
Figure 7: Predicted reflection cracking severity as a function of observed severity.
|
