The work reported on is an extension of work done in Washington State to examine the systematic occurrence and variability in pavement mat density. The focus will be on the extent of the pavement that is affected by variable density due to temperature differentials and the fact that randomly based testing does not identify the occurrence of cyclic density differentials. Temperature differentials 14 deg C (25 deg F) or greater generally cannot be compacted to the same level of density as the surrounding mat and therefore lead to significant density differentials (increase in air voids of 2% or more). A cyclic pattern typically occurs, matching each delivered truckload of mix, although temperature differentials can occur randomly or not at all, depending on the remixing device. The largest extent of pavement is affected when no remixing occurs and temperature differentials develop for every truckload of hot mix. These temperature differentials can cover the entire width and affect up to 50% of the mat. If the delivered hot mix is thoroughly remixed before placement, temperature differentials are minimal. Although density will vary in any paving operation, it was found that a uniform temperature mat greatly increases the ability to achieve a uniform density. Because of this cyclic pattern of variable density, random sampling for in-place density does not properly identify or quantify this problem. It is recommended that temperature differential areas be determined during construction and excluded from the random sampling used for acceptance testing. The issue of variable densities due to temperature differences or aggregate segregation should be identified and eliminated at the start of the project.
Abstract