Specifying agencies sometimes say that they limit the amount of reclaimed asphalt pavement (RAP) in asphalt mixes because these stockpiles are highly variable. Therefore, the agencies reason that allowing higher percentages of RAP could have a negative impact on the consistency of asphalt mixes and diminishes the performance of asphalt pavements. This perception of high variability may come from the notion that all RAP stockpiles are a heterogeneous agglomeration of reclaimed materials from a range of sources.

Photograph of an un-processed stockpile of pavement materials from multiple sources, also known as a "God Only Knows" a.k.a. GOK pile.

Many contractors maintain separate stockpiles of RAP obtained from different sources. When a large quantity of millings is obtained from a single project, for example 5000 tons or more, then it makes perfect sense to keep this RAP stockpiled separately from other RAP. Considering that the millings came from a given project that was originally built with under strict quality assurance specifications 15 to 20 years ago, it is fairly safe to assume the millings will be a consistent and high-quality material. In many areas of the country, this milled material is used directly in mix designs without further processing. An advantage of this practice is that the dust content of the RAP is kept as low as possible. High dust in RAP often limits how much RAP can be used and still meet mix design requirements.

On the other hand, for small quantities of millings, most contractors do not consider it worth the trouble of setting up a mix design for a small production run. Therefore, they typically combine small milling quantities with other sources of RAP to create the GOK stockpiles. It is common for the multiple-source RAP stockpiles to contain paving materials from small milling jobs, plant waste, rejected asphalt pavement material, and pavement rubble from complete demolition of roads or parking lots.

What contractors and agency personnel want to know is, what are the best ways to process multiple-source RAP stockpiles to achieve consistency in the material. The National Center for Asphalt Technology (NCAT) in Auburn, Ala., is working to find the best RAP management practices by surveying contractors across the U.S., examining quality control data, and identifying what techniques provide the most uniform RAP materials. NCAT has collected information on the current state of the practice for handling RAP and has gathered quality control (QC) data from contractors on their processed materials. This information has been used to draft preliminary best practices as outlined below.

Many contractors use independent companies to do their RAP crushing or have portable crushing units that move from one plant site to another. RAP crushing units differ by crusher type, screening capability, and other factors, so that there are numerous variations in processing operations.

However, the following general principles work well for achieving consistent RAP products that can be used successfully at higher percentages in new mixes:

NCAT tabulated standard deviations from more than 70 RAP stockpiles and from more than 60 crushed-stone products from quarries in six states. Analysis of the data has revealed some interesting results. First of all, contrary to the perception of many people, the analysis shows that RAP stockpiles are as consistent as virgin aggregates. Researchers compared the standard deviations of the percent passing the median sieve (the sieve with closest to 50 percent passing) for all RAP and all aggregate products. The mean of the standard deviations for RAP was statistically less than that of virgin aggregate for the median sieve. The variances of the standard deviations on the median sieve were the same for RAP and aggregate. The same analyses were conducted on the standard deviations for the percent passing the No. 200 sieve. In this case, the means and variances of the standard deviations for RAP and virgin fine aggregates were not statistically different.

A statistical analysis was also conducted to compare variability of fractionated RAP stockpiles to non-fractionated RAP. For this analysis, we compared standard deviations of asphalt content, percent passing the median sieve, and percent passing the No. 200 sieve. Surprisingly, the results do not show that fractionating RAP into different sizes will improve the consistency of the material. We expected that fractionating would further reduce variability in RAP stockpiles, but the data from contractors did not prove that out.

An important lesson from this analysis is that agencies should not use a method specification like requiring fractionation of RAP with the expectation that it will result in a more consistent RAP. A better approach is to set variability limits for RAP stockpiles and allow the contractor to do what is best and most economical to meet those limits. Based on the collection of RAP QC data from contractors we have so far, the following preliminary variability limits are suggested.

The above benchmark variability values may need to be revised in the future as more information is collected. However, using these kinds of limits for quality control of RAP stockpiles becomes more important when higher RAP contents are used in asphalt mixes.

After assessing their own RAP QC data, some contractors will find that they need to improve their processing operations. The eight basic principles above are a good place to start.