Non-nuclear density testing devices and systems to evaluate in-place asphalt pavement density

A field evaluation of portable non-nuclear density gauges was conducted to determine their effectiveness and practicality for quality control and acceptance of asphalt pavement construction. Three portable non-nuclear gauge models were evaluated, including the TransTech PQI Models 300 and 301, and Troxler PaveTracker 270 lb. All non-nuclear models consistently read lower than the nuclear density gauge. PQI Model 301 read 11.2 to 27.2 pcf lower than the nuclear gauge: PQI Model 300 ranged from 4.2 to 26.6 pcf lower; PaveTracker varied from 1.8 to 17.7 pcf lower. An analysis of variance determined that several factors affected the difference between the nuclear and non-nuclear readings, and it was recommended that a calibration be conducted uniquely for each project to block the effect of the factors. A daily calibration to the nuclear density gauge was recommended using a 10-point calibration slope functions, since it has less error and more simplistic approach for field purposes. The current nuclear density specification was reviewed and analyzed, and it was determined that the current n=7 sample size yielded a confidence interval of ± 1.5 pcf, and ± 0.9% density. It was recommended that adjustments be made to the current specification if risk levels are to be reduced. Sample size for non-nuclear gauge testing for a given lot on project was determined to be n=30 test sites, based on a 95% confidence level, mat and slope-function error, and confidence intervals of ± 1.0 pcf and ±0.6% density. A statistically-based procedure for determining the allowable difference between density gauges was detailed. When independent sites are used for non-nuclear test comparisons, 30 test sites are necessary to achieve a true difference of 1.0 pcf, based on the pooled variance, alpha risk of 5%, and beta risk of 20%. When the same test sites are used for comparison (split sample), 10 comparison test sites are necessary at the same risk levels. Finally, issues to consider for implementing the non-nuclear test specification were detailed, including nuclear density gauge requirement, operator familiarity with the devices, battery charging, adhering to manufacturer recommendations, computing the slope function, test site layout, and training.