11. Quality Assurance

11.5.1 Representative Sampling

A primary goal is to obtain samples that are representative of the products being evaluated. Since test results are used to adjust mixture production, it would be counterproductive to make plant adjustments based on a non-representative sample.

Representative samples are critical for effective QC and acceptance programs. Inspection is more appropriate for identifying nonuniform portions of the work. Once inspection has identified a portion of the work that is nonuniform (different from the rest of the work) or defective materials are suspected, additional samples representing that specific element of the work should be taken. Test results from such samples should never be used as part of the lot averages—and certainly not for lot acceptance. They only represent the specific problem areas identified by inspection. Standard specifications should offer guidance on how to handle suspect areas.

It is important to collect enough material when sampling. The sample size is important not only to obtain enough material to complete the required tests but also to achieve a representative sample. Most standards will indicate a minimum sample size requirement.

The quantity of asphalt mix produced on a given project is usually divided into predefined portions called “lots.” QA and QC plans usually require samples to be taken at intervals within these lots. The intervals are generally measured in one of the following three ways:

1. Time intervals—Typical lot sizes when sampling by time interval would include either a half day’s production or a full day’s production.
2. Quantity intervals—Typical lot sizes when sampling by quantity interval would include 5,000 or 10,000 tons (tonnes), depending on anticipated production rates.
3. Distance intervals—Typical lot sizes based on linear feet (meters) of paving. For example, cores are commonly taken based on a linear distance relative to stationing (a length of 100 ft [or 100 or 1,000 m] along a survey route) and an offset from a fixed longitudinal line.

Lot sizes are normally established by standard specifications, as defined well in advance of construction. The size and method of measurement of a lot will be used to determine locations and quantities of QC and acceptance tests. The lot will also be used to define what portion of the work is represented by a given set of test results. The limits of the lot are especially significant if corrective action or removal is required.

11.5.2 Random Sampling

The use of random sampling techniques helps to obtain unbiased representative samples, where every element in the population has an equal chance of being selected for sampling. As shown in Figure 162, it is theoretically possible for the purely random numbers to require that all samples for the entire project be taken over a very small area. The best and most practical method of ensuring that samples include the full range of construction is a process called stratified random sampling. This practice separates lots into equally sized sublots, from which random samples are then obtained. ASTM D3665 provides the standard practice for random sampling of construction materials.

Source: Asphalt Institute
Figure 162. Purely Random Sampling vs. Stratified Random Sampling

Stratified random sampling can be applied to different types of intervals by taking the number of samples needed and distributing them over times, quantities, or locations.

Watch Video

Coring

Taking pavement cores is one of the most important ways to verify that asphalt construction meets specifications. Cores provide direct evidence of thickness, density, and overall quality—information that can’t be fully confirmed by surface measurements alone. This video walks through the complete process step by step: how to determine the right number of samples, how to select random locations to avoid bias, and how to properly operate diamond bit coring equipment. It also includes best practices for extracting, labeling, and handling cores to ensure accurate testing results. Together, these steps help agencies and contractors confirm performance, document compliance, and build confidence that the finished pavement will stand the test of time.

Table 11 illustrates how to calculate random sampling tonnage for loose asphalt samples from truck beds. The example is for a project with 4,000-ton (tonne) lots and 1,000-ton sublots. Random numbers can easily be generated by computer. In the example, the random number is applied to the sublot amount only, then the cumulative amount is calculated. For sublot 3, remember 2,000 tons have already been produced in sublots 1 and 2. Therefore, the random tonnage within sublot 3 (263.7) is added to the 2,000 tons, and the truck containing ton 2263.7 is sampled.

Table 11. Example Random Sampling by Tonnage Table

Source: Asphalt Institute

The above example presents sampling based on quantity increments. They are useful for planning random samples of loose mix during plant production. These same methods and spreadsheet solutions could be applied if samples were to be obtained based on location. For example, sampling increments for roadway coring would be planned at both a given station and a distance from the centerline. The random numbers could be generated by hour if lots were assessed by half-day or one-day intervals.

11.5.3 Sampling Asphalt Mixtures

Watch Video

Sampling Procedures

Sampling requirements are typically set by the agency and should be identified in the contractor’s written QC plan. Acceptance sampling is often performed by the contractor but is required to be witnessed by agency personnel. Agency personnel typically take possession of the samples immediately after they are taken. It is important that all samples be properly documented with, at minimum, the project number, the date sampled, the material type, and sampling location (lot and/or sublot number). Other identifying information may also be required by the agency.

Sampling methods are available for sampling plant mix at many points along the construction process, including at the plant, from delivery trucks, from windrows at the jobsite, and from behind the paving machine. Some agencies may require one sampling point for QC and a different sampling point for acceptance.

For these sampling methods, the procedures described in AASHTO R 97 or ASTM D979 should be followed. Specifiers should carefully consider the sampling location. Each allowable sampling location in these standards has pros and cons related to the probability of obtaining a representative sample, the probability of causing permanent defects in the new mat, and the relative safety of the technician while sampling.

A common thread throughout the procedures is that they all require at least three randomly selected increments, combined to make a field sample meeting the minimum size requirement.

Samples of the compacted plant mix can also be taken by cutting cores from the cooled pavement following methods described in ASTM D5361 or AASHTO R 67. A brief discussion of each preferred sampling location is presented below.