By Gerald Huber
This is the second of two articles on the history of mix design. The first article covering the early design practices through the Marshall and Hveem systems was published in the spring 2013 issue of “Asphalt” magazine.
Superpave mix design was developed as part of the Strategic Highway Research Program (SHRP) from 1987 to 1993. The objective of the Asphalt Research Program was to develop a performance-based asphalt binder specification, a performance-based asphalt mixture specification and a mix design system. The Performance-Graded (PG) asphalt binder specification is the result of the research.
The performance-based mix specification was less successful. Although performance tests for asphalt mixture were developed and models were designed to predict mixture response (stress, strain, etc.) and to predict mixture performance (rutting, fatigue cracking, thermal cracking), the system ended up being too difficult to implement and was never used by state DOTs.
The mixture design system developed during SHRP became known as Superpave. Superpave mix design had three levels of increasing complexity. These levels were referred to as Level 1, Level 2 and Level 3 mix design. The performance-based mixture tests were to be used in the Level 2 and Level 3 designs. As the SHRP research progressed, it became apparent that testing and analysis for the performance predictions would be sufficiently complex to justify using a simple, empirical design method as the base or entry level mix design. When the performance-based tests and models were not implemented, the mix design method that remained was the base level (also known as Level 1). Superpave, as specified in AASHTO M323, is the baseline, entry-level mix design that was developed during SHRP.
So, Level 1 mix design was decided to be based on mixture properties. The problem facing SHRP was how to answer questions such as:
• “What is the proper level of air voids?”
• “How should minimum asphalt content be specified?”
• “How should gradation be specified?”
The Federal Highway Administration (FHWA) convened a Technical Working Group (TWG) to determine which empirical volumetric properties should be included in Superpave and how they should be defined. Participants included four members from state DOTs, four from industry associations, two from private companies, and one each from consulting and academia. The minutes of the first meeting state:
The purpose of the Technical Working Group is to determine key volumetric properties of asphalt mixes during the production process that affect performance and how the production process can affect those properties.
At the first meeting, it was agreed that the following properties would be used:
Air voids were to be calculated using the measured maximum theoretical gravity (Gmm) of the mixture and the measured bulk specific gravity of the compacted hot mix asphalt.
Voids in mineral aggregate (VMA)
VMA was to be calculated as proposed by the Asphalt Institute in their Manual of Asphalt Mix Design (MS-2). This calculation uses the bulk specific gravity of the aggregate (Gsb). A recommendation to calculate VMA using effective specific gravity (Gse) of the aggregate (back calculated from the Gmm) was discussed at length, but was not accepted.
Voids filled with asphalt (VFA) would be calculated using VMA and air voids.
At the time of the Technical Working Group, National Asphalt Pavement Association (NAPA) had been discussing volumetric properties in their Quality Improvement Committee. A significant question of the day dealt with VMA, specifically, “Why does VMA decrease when mix is produced in the hot mix plant as compared to the design laboratory?” A task group was assembled to develop a guidance document for contractors about how to control VMA.
Also in the same time era, the Colorado DOT had completed a research program of 24 mixes that looked at various properties such as mix gradation, gradation of the aggregate passing the No. 200 sieve (P200), amount of P200, fine aggregate angularity, and crushed particle count.
All of this input became part of the information that was considered by SHRP.
SHRP studied state DOT specifications and found that there was no consensus in properties specified or what criteria should be specified for the properties. The FHWA TWG made recommendations on what properties should be used but not on what criteria or how the criteria should change with mix type. Superpave needed volumetric properties but there were no clear answers to the questions of “Which properties should be used?” and “What criteria should be used for them?” So SHRP convened a group commonly called the Delphi group.
The Delphi process is a method of developing consensus among a group of experts. First, the experts had to have a working knowledge of the area of study. Experts do not debate who is right and who is wrong regarding some property. Instead, the experts are given a series of questionnaires designed to elicit their response to situations.
Fourteen experts were selected that represented state DOTs, industry representatives and university researchers. They received the questionnaire by mail (yes, snail mail was still in vogue).
The first questionnaire defined the areas to be investigated. On a scale of “very strongly disagree,” “strongly disagree,” “disagree,” “neutral,” “agree,” “strongly agree,” and “very strongly agree,” the participants indicated the relative importance of a property in the mix design system. For example, they were given the statement, “Design air voids should be part of the mix design”, and they would give their reaction to it.
There were seven aggregate properties and three mixture properties considered:
• gradation limits
• crushed faces
• natural sand content
• L.A. abrasion
• deleterious content
• sand equivalent
• air voids
Also on questionnaire 1, there was a set of follow-up questions for each property:
• What is the best way to measure the property (air voids, for example)?
• Are there any external influences that would change the level of air voids used?
• How does that factor affect air voids?
After the first questionnaire results were received, the panel members were brought together and the results were presented and discussed. There were several contentious points. One of these was whether minimum asphalt content should be controlled by film thickness or VMA. Both VMA and film thickness were carried forward to the next questionnaire.
By the end of the third questionnaire, it became clear which properties the group felt were important and VMA had displaced film thickness as the method for minimum asphalt content. In the fourth questionnaire, participants were asked to rank each property as to its importance for performance. The fifth and final questionnaire was used to estimate specification limits.
The key components of Superpave Mix Design were determined to be as follows:
• Compaction was to be done with a gyratory compactor.
• Air voids
Calculation of air voids using Gmm and Gmb.
VMA as the method of setting minimum asphalt content, and VMA should be calculated using aggregate bulk specific gravity (Gsb) versus effective specific gravity (Gse) back calculated from the Rice test (Gmm) results.
• Voids Filled with Asphalt (VFA)
There was much discussion about the need for this property.
The argument against was that it merely was the third of three inter-related properties. Only two of the three were necessary to be specified. In fact, the argument went, specifying all three (AV, VMA and VFA) could set up a situation where it is not possible to meet all three.
The argument for using VFA is that it was similar to degree of saturation, which had been shown to be related to rutting. Also, it was argued that minimum VMA was needed for durability and that maximum VMA would be indirectly controlled by having a maximum VFA. In the end, the Volumetric TWG decision was to recommend VFA.
• Aggregate gradation
This was debated by both the TWG and the Delphi group. The TWG debated the existence of a maximum density line. Research from the Asphalt Institute was convincing enough for the TWG to recommend the Superpave gradation specifications. The restricted zone was maintained in the specification at the time because European mixtures were “discovered” to have coarse aggregate skeletons, and the restricted zone was a method of ensuring coarse aggregate skeletons and preventing excess sand (especially natural sand) from being used.
Later research indicated that Fine Aggregate Angularity was adequate for controlling quality of the sand and so the restricted zone was a Primary Control Sieve (PCS) in the AASHTO specifications. The purpose of the PCS is to define coarse-graded and fine-graded mixtures.
• Crushed faces
The method of designating crushed faces was accepted by the Delphi group. The surveys defined the level of crushing for different applications. The main parameters that influence the demand for crushing was traffic level and depth from surface. Both of these parameters were incorporated into Superpave mix design.
• Natural sand
There was much concern about excessive use of natural sand. The epidemic of rutting that had occurred nationwide in the 1980s was partly related to excessive use of natural sand. The question became how to define or control the amount used. The National Sand and Gravel Association had recently begun promoting use of the Fine Aggregate Angularity (FAA) test and NCAT had completed two major studies to document FAA of different sands and to relate FAA to mixture rutting performance. This work became the basis of the decisions from the Delphi group.
Summary of the history
Today’s technology represents an evolution of ideas that have been evaluated through the years. Film thickness and voids in the mineral aggregate have been two ideas for controlling the minimum amount of asphalt binder in a mixture going back to the early days of asphalt mix design.
One of the limitations of current mix design systems (Marshall, Hveem or Superpave) is the inability to measure expected performance—specifically, the ability to measure rut resistance, fatigue cracking, low temperature cracking, asphalt binder aging or mixture resistance to moisture damage. Prediction of these properties for a given application was the goal of SHRP. But the SHRP products could not be implemented. In all three mix design methods, surrogate properties are used to control the performance properties of the asphalt mixture.
Rutting is controlled by
◊ Aggregate properties
• Crushed faces on coarse aggregate
• Fine aggregate angularity on fine aggregate
◊ Volumetric properties
• Air voids
• Voids filled with asphalt
Fatigue cracking is controlled by
◊ Asphalt content
• This is probably the most sensitive mixture property for fatigue
Low temperature cracking
◊ Low temperature grade of the asphalt binder
• Has the greatest influence
◊ Asphalt binder content
◊ Asphalt content
◊ Bond strength of asphalt-aggregate interface
• Enhanced by anti-strip agents
◊ Asphalt binder content
The future of mix design is being researched today. Mix properties will be measured and mixture performance will be predicted given the conditions to which the pavement will be subjected. Such research is occurring today. In the future, mixture design will be more intimately tied to structural design, much the way that the Mechanistic Empirical Pavement Design Guide (MEPDG) is currently designed. Mixture design incorporating performance prediction is at least 10 years in the future, more likely 15 to 20 years.
Gerry Huber is the Associate Director of Research for Heritage Research Group.
Read the first part of the History of Mix Design