AASHTO makes progress on new design guide

In one of the largest movements in the history of U.S. pavement design, a group of 19 Lead States has begun implementing the Mechanistic-Empirical Pavement Design Guide (MEPDG). The national effort began in the early 1990s, when an invitational workshop on mechanistic pavement design was held in Irvine, California. At that meeting, a vision was created to develop a national mechanistic-empirical pavement design procedure.

Work started under the auspices of the AASHTO Joint Task Force on Pavements in 1996 with the largest National Cooperative Highway Research Program project funded in the 40-year history of NCHRP, says David Newcomb, P.E., PhD, vice president of research and technology, National Asphalt Pavement Association. The development effort concluded earlier this year with the release of MEPDG version 1.0. And AASHTO, the state highway transportation officials group, has convened a task force to develop Version 2.0, a faster, more user-friendly version of the software.

AASHTO has not yet approved the new design guide. It is currently under review and will be balloted by the Subcommittee on Materials, the Subcommittee on Design, and the Standing Committee on Highways within the next year, says Newcomb.

“One could say that the flexible pavement portion of MEPDG is descended from the Asphalt Institute’s ME design methods that have been around for more than 25 years,” says Gary L. Fitts, P.E., a senior regional engineer with the Asphalt Institute. “Dr. Matthew Witczak led the development of both procedures and they use the same fundamental concepts.”

More Accurate
A primary benefit of the new design guide is improved accuracy, says Linda Pierce, P.E., Washington state DOT and chair of the Lead States Group that is aiding the implementation of the MEPDG. Program inputs of material properties, climate, and traffic loading are much more detailed than in the past. For example, the new design guide does not use equivalent single axle loadings. Instead, it uses “axle load spectra,” which call for actual loads per truck axle over a range of trucks. Most states have this data in terms of weigh-in-motion studies, Pierce says.

“If you do the design and you use more accurate input values, you have more reliability that the pavement is going to last for the period that it was designed for,” says Pierce. “But if you use average inputs as we have in the past, and you’re shooting for an average 20-year life, it’s going to last somewhere between 15 and 25 years.”

Strictly speaking, the MEPDG is not yet a pavement design tool; it is a pavement analysis tool. Pavement thickness is an input, not an output of the program. “One of the biggest benefits will be the ability to predict pavement distress over the design life of the pavement, giving the designer a tool to refine the structural section to give the maximum benefit for the least cost,” says a spokesperson for the Arizona transportation department, a Lead State.

Longer term, Pierce says the MEPDG will give pavement designers a way to quantify the benefits of special asphalt mixes, such as stone matrix asphalt or mixtures made with modified binders. But none of the asphalt pavements currently being used to establish performance—most of them from the Long Term Pavement Performance (LTPP) program—used modified binders or were SMAs. “So unless somebody does an accelerated testing program, it’s going to take us a number of years to establish performance records for those pavements,” says Pierce. “NCHRP now has a study on how to bring modified binders into the new design guide.”

How It Works

Many states currently use design systems that evolved from the AASHO road tests done on a test track near Ottawa, Illinois, in the 1950s. (AASHO was the American Association of State Highway Officials.) The tests were done over two years, in one climate, using one subgrade material type and one dense-graded hot mix asphalt. Traffic loadings of less than two million equivalent single axle loads (ESALs) were considered. The results were all empirical, or observed. The researchers would run trucks on the test track, collect distress data, and repeat the cycle. Since then, many states have extrapolated those loadings to some 100 million ESALs or more – and derived pavement thicknesses accordingly. But in reality, there is no way to tell if the relationship between traffic loadings and pavement thickness should be linear, or what that relationship is.

Enter the new MEPDG, which uses both mechanistic and empirical design procedures. With the mechanistic portion, a computer model is given conditions of climate, traffic, materials and layer thicknesses. The material is characterized based on fundamental engineering properties. In the case of asphalt, the dynamic modulus is used.

“You define your materials based on fundamental engineering properties and then you use a computer program to calculate tensile strain at the bottom of the asphalt layer,” says William Barstis, P.E., the Lead States representative from the Mississippi DOT. “Then you take your pavement responses—strains and deflections—and use equations known as transfer functions to relate those responses empirically to the magnitudes of distresses that will occur, such as cracking and rutting.”

For asphalt pavements, the MEPDG predicts five types of distresses: thermal cracking, fatigue cracking, rutting, longitudinal cracking and ride quality as measured by the International Roughness Index (IRI). So designers, then, select trial layer thicknesses of asphalt (and base structure), and the design guide produces levels of distresses that are compared to the designer’s criteria. If the distresses are too severe over the design life, the designer can go back into the program and change an input—bump up the grade of binder, or make the stabilized base thicker, or whatever is desired—to strengthen the pavement.

Barstis says the transfer functions that relate the mechanistic strains to the empirical portion of MEPDG all have coefficients. And when researchers wrote the MEPDG, for the most part they used LTPP data to establish coefficients that reflect pavement performance on a national scale. The data are not specific to any state or region.

Therefore, it falls to the states to “calibrate” the MEPDG to their given conditions, traffic levels, and material characteristics. “You have to take your local data from your state and revise the coefficients for your local conditions,” says Barstis.

“That is the essence of our State Study 170,” he says. Indeed, Mississippi has hired Applied Research Associates Inc. (ARA) to do that study, which is being conducted concurrently with several of 12 related studies for MEPDG by Mississippi. The objective of Study 170 is to implement the MEPDG for the state. It will address the following:

  • Provide for training of Design Guide users and other stakeholders
  • Develop and execute a plan for securing the appropriate design input data on material and traffic characterization, and other design inputs
  • Conduct sensitivity analyses and make comparisons of MEPDG designs with current procedure
  • Develop and execute a plan for calibration of MEPDG performance and distress models.

The 12 related studies being undertaken by Mississippi include two in-house support studies, two traffic studies, four soils studies, two HMA studies, and two concrete studies. The two asphalt studies include one that will characterize Mississippi HMA mixes using dynamic modulus testing in preparation for future implementation of the MEPDG. Selected mixes will also be evaluated using the Asphalt Pavement Analyzer. The other study concerns structural characterization of asphalt drainage course layers.

Mississippi currently uses AASHTO’s 1972 Interim Guide to design flexible pavements. The Interim Guide is based upon the AASHO Road Tests. Now, Barstis is excited about implementing the new Design Guide, but says his state will wait until Version 2.0 of the software comes out to use it for design.

In the meantime, Mississippi will focus on calibration and validation. “We have a lot of work to access our pavement management data and construction data and materials data on pavement sections that have been built,” says Barstis. “All that data is needed by our consultant, ARA, to calibrate and validate these models for Mississippi conditions.”

Barstis says the new design guide offers the following advantages:

  • Design input of 50 million or more axle loadings, compared to fewer than 2 million with the AASHO Road Tests
  • Design input of a wide range of structural sections, compared to a limited range in the former system
  • Design input of all climates over 20 to 40 years, compared to one climate in the former system
  • The potential to include new materials, such as polymer asphalts and crumb rubber mixes, in the new Design Guide.

“Now with the new methods for characterizing asphalt mixes with modulus, we can actually run the dynamic modulus on a mix with crumb rubber in it, and see what the difference is in the engineering property versus a neat asphalt,” says Barstis. “We have a greater spectrum of materials that we can look at and analyze.”

What’s Next?
In addition to her role as chair of the MEPDG Lead States Group, Linda Pierce is chair of AASHTO’s DARWin task force, which is in charge of the new Design Guide software and its evolution. (DARWin stands for Design, Analysis and Rehabilitation for Windows.)

The group has begun to consider modifications and enhancements to the MEPDG software. “We’re going to drop the MEPDG and just call it DARWin,” says Pierce. States involved thus far include Washington, Virginia, South Carolina, Vermont, New York and Wisconsin.

The DARWin task force intends to sunset the existing AASHTO design software, Pierce says. The second intent is to develop a request for proposals package that will address enhancements or modifications to the software. Version 2.0 will be an AASHTO document, Pierce points out; version 1.0 is an NCHRP product.

Pierce says that preliminarily, the ability to do thickness design is one goal of the DARWin task force. So is the ability to maintain a database of aggregate sources or binder suppliers in the software. And the software can be modified to run faster.

In summary, the new Design Guide offers the following abilities:

  • To relate pavement material properties to the structural behavior of the pavement system
  • To analyze pavement responses to a wide variety of loading and climatic conditions
  • To investigate the behavior of pavements in the presence of new loads or with new materials
  • To understand the influence of the variability of materials on pavement behavior.

Concrete and Asphalt in Missouri
The new Design Guide can be used to design pavements with both asphalt and concrete. In Missouri, the DOT has used the guide to design 50-plus projects with both asphalt and concrete. The state does alternate bidding, so in many cases a contractor can bid a project with either an asphalt design or a concrete design. The award goes to the low bidder, and both pavements were designed for 45-year design lives.

“Nobody is ever completely happy with something like this, but for the full-depth pavements, 20 projects went to asphalt and 25 to concrete,” says John Donahue, P.E., pavement engineer in the Missouri DOT. “So we think it’s reasonably competitive between the two designs. We’re not artificially handicapping either industry.”

Donahue says of the awards for 45 projects, $300 million went for asphalt pavements and $360 million was spent on concrete. Some maintenance was included in the 45-year design lives. For asphalt, it was assumed that the pavement would need a mill-and-fill resurfacing at 20 years and another one at 33 years. For the concrete pavement, the state assumed that it would need diamond grinding and 1.5 percent full-depth replacement at 25 years of service.

Dan Brown is the Principal of Technicomm