Correlation of Nuclear Density Readings with Cores Cut from Compacted Roadways

Abstract

A field procedure for use of the nuclear density gauge was developed that resulted in nuclear gauge density data closely resembling in-place density obtained from cores. Procedural recommendations intended to improve its accuracy were made. The nuclear gauge data were collected on seven sites during Connecticut Department of Transportation pavement construction projects in 2003 and 2004. The findings indicate that the six individual nuclear gauge density units used for this study do not produce similar results and do not consistently correlate with core densities. The differences between the core density values obtained by the three laboratories and the nuclear gauge readings were significantly higher than the 0.1% maximum theoretical density (MTD) reported accuracy currently used by Connecticut Department of Transportation (ConnDOT) for acceptance on projects. Significant variation was also found between the results of core density samples obtained by three different laboratories on the same samples. In addition, the variability of the nuclear density gauge error differed not only from gauge to gauge, but also from location to location and is present for both nuclear gauge density testing modes (backscatter versus thin lift). The effect of the nuclear gauge orientation during testing with respect to the new mat was statistically different but very small; the mean difference was 0.05% of MTD. When the nuclear gauge was in the longitudinal direction, the density reading was slightly higher than those taken in the transverse direction. The time recording interval for the nuclear gauge was found to be relatively significant with respect to minimizing the difference between the nuclear gauge densities and core densities. The longer the recording time interval for the nuclear density gauge, the smaller the difference between the core density values and nuclear gauge densities. If the nuclear gauge continues to be used for project Quality Assurance, it is recommended that the recording time interval be 1-minute and that the acceptance reported accuracy be set to 1% of MTD. A new project-by-project nuclear gauge density correction procedure that requires 10 cut cores to improve accuracy is presented in this report.