Sunteți pe pagina 1din 7

Pavement investigation sh40

INVESTIGATION OF PAVEMENT
ON ROAD SH40 2012

PREFACE

The investigation was ordered by Roughton International Limited. The measurement was done
with an FWD KUAB 50 owned by a Georgian state agency and KUAB's ground penetration
radar of make Mal ProEx with 500 MHz and 1600 MHz antennas, with personnel from
Roughton International Limited, the Georgian agency and KUAB Konsult & Utveckling AB.
General information about the roads was supplied to KUAB by Roughton International
Limited. The calculations were made by KUAB.

EXECUTIVE SUMMARY

The following rehabilitation is suggested:

Sections 0 to 5100, 7350 to 8550, 9550 to 10850 and 13500 to end: patch the distressed areas of
the pavement and seal the longitudinal and transverse cracks.

Sections 5100 to 7350, 8550 to 9950 and 10850 to 13500: Remove the asphalt pavement, or
preferably grind and mix it with the existing base and add good quality crushed stone base and
asphalt concrete with the following thicknesses:

Section Asphalt, cm Base, cm


5100 to 7350 5 20 to 25
8550 to 9950 5 15 to 20
10850 to 13500 5 15 to 20

Page 1 of 7
Pavement investigation sh40

MEASUREMENT

The measurement was done in June 2012, with an FWD KUAB 50 owned by a Georgian state
agency and KUAB's ground penetration radar of make Mal ProEx with 500 MHz and 1600
MHz antennas, with personnel from Roughton International Limited, the Georgian agency and
KUAB Konsult & Utveckling AB.

The maximum peak load used was about 50 kN, and the rise time about 23 ms. The drop
sequence was in most points 50 kN and 50 kN, and in about 10% of the points 50 kN, 50 kN
and 50 kN, with measurement on all drops but the first. The diameter of the load plate was 300
mm, and deflection sensors were placed 0, 200, 300, 450, 600, 900 and 1200 mm from the load
plate center. Further details about the measurement can be found in the measurement data files.

The FWD measurement files show test point position expressed as distance from a reference
point, drop number, peak load, peak deflections, air temperature, E Mod, the position measured
with an EGNOS GPS, and time. EMod is the so called surface modulus, which is the modulus
of an assumed linearly elastic, homogenous and isotropic half-space with Poisson's ration 0.5,
that would get the same deflection in the center of the load plate as the actual pavement. The
measurement was made in the wheel path close to the pavement edge, approximately every 100
m in both directions. Slightly edited raw data files are attached, sh40H.fwd is the right side
when driving towards increasing distance, sh40V.fwd is the left side.

The GPR was recording 10 times per meter with the high frequency antenna and 5 times per
meter with the low frequency antenna.

Photos were taken every 10 m on most of the road, when driving one direction.

TRAFFIC

Traffic estimate for the design period was received from Roughton International Limited,
1 650 000 8.2 ton axles for the first 3.9 km, and 475 000 8.2 ton axles for the rest.

INITIAL INSPECTION OF THE DEFLECTIONS

Fig 1 and 2 show the deflections at 0, 300, 600 and 1200 mm from load plate center.

Fig 1. Deflections at 0 and 300 mm from load center.

Page 2 of 7
Pavement investigation sh40

Fig 2. Deflections at 600 and 1200 mm from load center.

There is much variation in pavement quality as well as subgrade modulus along the road.

The deflection bowls are typical for a flexible pavement, with much variation, from good in
relation to the traffic to very bad, as can be seen in the examples in Fig 3.

Fig 3. Examples of deflection bowls.

Page 3 of 7
Pavement investigation sh40

INITIAL INSPECTION OF THE LAYER THICKNESSES

Fig 4 shows the layer thicknesses from GPR measurement and test pits supplied by Roughton
International Limited in two depth scales. We did not obtain particularly good agreement
between the layers reported from the test pits. The GPR depths are therefore based on typical
propagation velocities, 0.13 m/ns for asphalt, 0.12 m/ns for base and 0.11 m/ns for subbase
rather than on detailed calibration with the actual materials. The layer thicknesses must be
considered as approximate estimates.

The first main reflex often shows a first pavement layer much thinner than reported from the
test pits. The pictures from the extracted cores shows that there are two separated asphalt
layers, and the GPR picks up the first. Judging from the pictures the quality of the lower layer
is such that it is probably can be considered as more like a good gravel base than an asphalt
layer. The GPR and the test pits picks up rather different base thicknesses. This is not
surprising when there is much variation in the pavement. Dielectric properties can be different
from visual properties.

The colors in the diagram have no specific significance, they are just to make the layers visible.

Fig 4. Layer thicknesses from GPR and test pits.

Page 4 of 7
Pavement investigation sh40

CALCULATIONS

Calculation principle

The calculations were made with the analysis program PVD.

Two different calculation methods were tested. One is the so called MET, Method of
Equivalent Thicknesses. A mathematical model of the pavement as a layered elastic structure
is created, based on the information from construction records, test pits and GPR. The essential
elements of the model are thickness and modulus of elasticity of each layer. In the
mathematical model the FWD load is applied, and the deflections calculated. The calculated
deflections are compared with the measured, and the properties of the layers adjusted until there
is agreement between the two sets of deflections. The above back-calculation is made at the
actual temperature of the measurement. When the final values are obtained, the modulus of the
asphalt layer is adjusted to a reference temperature with a temperature adjustment function.

In the thus obtained mathematical models the strains are calculated and compared to the criteria
for strain in the different layers. Different institutions suggest different strain criteria. One set
of criteria is 195 microstrain horizontal tensile strain for asphalt with 5000 MPa modulus at one
million load applications, and 885 microstrain vertical compressive strain for unbound layers at
one million load applications. The exponent for adjustment to actual number of load repetitions
was for asphalt 5.62 and for unbound materials 4. These criteria are taken from the Shell
pavement design manual, published by Shell research laboratory in London. Another example
is from the Swedish Transport Administration, which uses the more tolerant Kingham's criteria
for asphalt layers, and allows only about half the strain for subgrade.

When the criteria are not met, additional asphalt is added in the model until the strain is small
enough.

The other method is also based on equivalent thicknesses. Depending on traffic amount a
certain surface modulus is required If the modulus is too small, overlay is added and the new
surface modulus is calculated with the method of equivalent thicknesses. Two calculations are
made, one with an asphalt overlay on existing pavement, and one where the current asphalt is
removed and a base of calculated thickness and an asphalt layer of fix thickness are added.

The former method is more exact if the materials and thicknesses are well known, while the
second is more tolerant to variations and deviations between real pavement and theoretical
assumptions.

Calculation result

After testing a number of parameter sets for the two methods the latter method was selected.

Fig 5 shows the surface modulus at measuring conditions, the modulus after correction for
temperature and season and the requested modulus.

Fig 6 shows the strengthening needed to obtain the requested modulus. The new asphalt is
assumed to have the modulus 4000 MPa at design temperature, and the new base is assumed to
have 450 MPa, typical for a good base of crushed material.

Page 5 of 7
Pavement investigation sh40

Fig 5. Surface modulus at measuring conditions, surface modulus after correction for
temperature and season and the requested surface modulus modulus.

Fig 6. Calculated overlay.

The following rehabilitation is suggested:

Section Asphalt layer Crushed grave thickness


0 to 5100 0 0
5100 to 7350 5 20 to 25
7350 to 8550 0 0
8550 to 9950 5 15 to 20
9550 to 10850 0 0
10850 to 13500 5 15 to 20
13500 to end 0 0

With reference to previous estimates made by Roughton International Limited the sections with
0 overlay suggested would be treated with Treatment A, and the others with something very

Page 6 of 7
Pavement investigation sh40

similar to Treatment C. It would not be necessary to rip off exactly 20 cm, but to remove the
asphalt bound layers, or grind them and mix them with the base. Then the thickness in the table
with good base course crushed stone and 5 cm asphalt would be applied. Probably Treatment B
with 5 or 7 cm asphalt would be less costly than Treatment C, but there may be significant risk
for reflexion cracking.

Page 7 of 7

S-ar putea să vă placă și