Recovery Room

Gravel Roads Article August 16, 2003
Printer-friendly version

Low-volume roads constructed in regions susceptible to freezing and thawing periods are often at risk of load-related damage during the spring thaw period. The reduced support capacity during the thawing period is a result of excess melt water that becomes trapped above the underlying frozen layers. Many agencies place spring load restrictions (SLR) during the thaw period to reduce unnecessary damage to the roadways.

Minnesota has utilized SLR since around 1937. Prior to 1999, load restrictions were placed and removed based on quantitative information (including, but not limited to, falling weight deflectometer [FWD], frost tubes and air temperatures) and physical observations made on roadways. In 1999, the Minnesota Department of Transportation (MnDOT) modified the criteria used for the placement of load restrictions and set the duration of load restrictions, for both flexible and aggregate-surfaced roads, to eight weeks for all frost zones. A weather-based model is used to determine the starting date, where the cumulative thawing index for a frost zone must exceed 14°C days (25°F days) based on the three-day forecast, with predicted increases well in excess of 14°C days (25°F days).

In addition to the above modifications to the placement and removal of load restrictions, the Minnesota Legislature mandated that Minnesota counties and cities follow MnDOT’s period of spring load restrictions unless signs are posted otherwise. The period of SLR set forth by MnDOT is effective for all flexible pavements; however, experience suggests that many aggregate-surfaced roads require additional time relative to flexible pavements to recover strength sufficient to carry unrestricted loads.

Testing for typical

The principal phase of the current research program involved seasonal measurements of in situ shear strengths, measured using the DCP, on various Minnesota aggregate-surfaced county routes. Strong efforts were made to select roadway sections representing different frost zones and performance characteristics. The following evaluation sequence (totaling 18 test dates per year), for the weakening and recovery periods of 2000, 2001 and 2002, was followed to verify whether these types of roadways take more than the typical eight weeks to gain adequate strength for supporting legal load limits:

  • Testing one time per week during the typical eight weeks of SLR;
  • Testing one time per week during an additional four weeks after SLR are removed; and
  • Baseline testing on three separate dates during both the summer and fall seasons (totaling six testing dates) for use as “strength recovered” data points.

The DCP seasonal data were reduced to characterize changes in bearing capacity over time for each test section. These trends (i.e., percent-strength recovery) were then adjusted to reflect the incremental changes in strength with respect to the time since the start of SLR. These resulting values were then used to determine the average length of time for aggregate-surfaced roads to recover to a baseline condition.

Dynamic cone penetrometer

The dynamic cone penetrometer (DCP) was chosen as the test device for this study because it can measure the in situ strengths of base and subgrade materials, is portable and is inexpensive.

An alternative would be the FWD. However, this device was not utilized due to the significant spacing between test sites, which consequently limited the ability to perform a large number of tests during the required testing period. Additionally, the FWD does not provide a direct measure of strength, as the DCP, and there are problems associated with the testing of unbound surfaces using the FWD which generally worsen during the spring weakening period (e.g., difficulty obtaining correct loading condition [combination of drop height and plate diameter], sensor contact issues).

The DCP test involves lifting the hammer vertically to the bottom of the handle bracket and then allowing it to freely fall, by gravity, to the stopping point at the anvil/coupler assembly. The impact of this mass forces the cone tip to penetrate into the testing material. The penetration depth is then measured using the graduated drive rod or vertical scale immediately after each blow. The magnitude of this depth will depend on the strength of the material being tested and also the moisture content present. For instance, penetration rates are greater for weaker materials or materials with larger moisture contents than for stronger materials or materials with lower moisture contents.

Data reduction

Two methods were used to characterize changes in bearing capacity over time for each test section:

  • Area under the DCP penetration index (DPI) profile (AUDP); and
  • Damage quantified using predicted rutting depths.

The following sections briefly describe these results and analyses.

The AUDP (i.e., area under [to left of] the penetration versus DCP penetration index [DPI] curves) was calculated for each county, route, site and test date. Fig. 1 illustrates typical curves generated, where the area was calculated over the shaded region.

Please note the DPI is a measure, in mm/blow, of the vertical movement (i.e., penetration) of the lower shaft into the test material produced by one hammer drop. Therefore, DPI values are greater for weaker materials or materials with larger moisture contents than for stronger materials or materials with lower moisture contents. Consequently, DPI values are good indicators of seasonal changes in bearing capacity.

In order to make direct comparisons between different test dates, the AUDP was normalized by taking the quotient of AUDP and maximum penetration depth. Plots of normalized AUDP (NAUDP) versus time were generated for each county, site, route and test year for determination of recovered (i.e., normal strength) values.

This roadway section begins to weaken around March 30, 2002, and achieves a peak strength loss on April 24, 2002, after which time the roadway begins gaining strength until 100% strength recovery is obtained around May 29, 2002. For this case, the data points collected on July 19, Aug. 14, Aug. 27, Sept. 18 and Oct. 24, 2002, were assumed to be reflective of this roadway’s recovered (normal) strength condition. The corresponding NAUDP values were averaged for use in determining percent recovery. For this case, the average recovered NAUDP value is 2.3 mm/blow.

Decimal percent recovery was then calculated using the ratio of NAUDP to NAUDPrecovered, where NAUDPrecovered is the average recovered NAUDP value for a given county, route, site and year. Plots of NAUDP/NAUDPrecovered versus test date were generated for each county, route, site and year for use in determining percent recovery over time. Fig. 2 presents an example of this type of graph.

Percent recovery, as a function of time since SLR start, was then determined between 50 and 100% (at 5% increments). For example, the SLR starting date and 100% strength recovery date for the case presented in Fig. 2 is Feb. 19, 2002, and May 29, 2002, respectively. Therefore, 100% strength recovery is achieved 14 weeks after the start of SLR. Fifty percent recovery occurs on May 22, 2002, resulting in 13 weeks to achieve this recovery level. Peak strength loss occurred nine weeks after the start of SLR (one week after spring load restrictions are typically removed).

The percent strength recovery data was further reduced to determine the average time, since SLR start, required to achieve given strength recovery levels. Fig. 3 illustrates the mean time, since SLR start, required to achieve the given strength recovery levels.

The results indicate that at the end of the typical eight-week spring load restriction period, aggregate-surfaced roads recovered only 50 to 75% of their recovered strength. Thereafter, time required increases with increasing recovery levels. Full-strength recovery (100%) is obtained on average between 9.5 and 10.5 weeks after the start of SLR, thereby exceeding the typical eight-week period of SLR by 1.5 to 2.5 weeks, respectively. As a result, aggregate-surfaced road strengths are typically only between 50 and 75% of their recovered strength values when SLR are removed from flexible pavements and unrestricted loads are allowed, and generally require between 1.5 and 2.5 additional weeks to reach recovered bearing capacity.

Damage, quantified using rutting depth, was calculated using the U.S. Forest Service surfacing thickness design model. Steps similar to those discussed for AUDP were followed using estimated rutting depth values. Fig. 4 illustrates the mean time, since SLR start, required to achieve the given strength recovery levels. The results indicate that at the end of the typical eight-week spring load restriction period, aggregate-surfaced roads have not yet achieved 50% of their normal recovered strength. On average, 8.5 weeks are required to achieve a 50 to 60% strength recovery level. Full strength recovery (100%) is obtained on average between 10.5 and 12 weeks after the start of SLR, thereby exceeding the typical eight-week period of SLR by 2.5 and 4 weeks, respectively. As a result, aggregate-surfaced road strengths typically have not yet achieved 50% of their recovered strength values when SLR are removed from flexible pavements and unrestricted loads are allowed, and generally require between 2.5 to 4 additional weeks to reach recovered bearing capacity.

Please note the values presented in Fig. 3 and 4 were obtained by averaging data from all test sites and, therefore, neglect varying locations, material type and layer thickness. Additionally, similar results were obtained using the two independent methods (i.e., AUDP and damage quantified using rutting depth).

The feeling is full

It is recommended that Minnesota Statute 169.87 “Seasonal Load Restriction; Route Designation,” subdivision 2 “Seasonal Load Restriction” be modified to reflect the additional time required by aggregate-surfaced roads to reach full bearing capacity. The resulting statute changes would allow SLR to remain effective on aggregate-surfaced roads for an additional two weeks beyond that imposed for flexible pavements (which is currently a floating ending date).

About the author: 
Embacher is a research project engineer in the Office of Materials at the Minnesota Department of Transportation. Van Deusen is a pavement design engineer at MnDOT. Funding for this research was provided by the Local Road Research Board.
Overlay Init