ACPA provides training programs in which designs are generated using the AASHTO 1993 guide, M-E PDG and StreetPave software, and are optimized for performance. Ayers says slab thickness is often used as a basis for comparison between design elements. However, says Ayers, it is preferable to compare estimated costs for the overall pavement structure including the subbase, load transfer, slab configuration, etc.

To establish a baseline design using AASHTO 1993, Ayers established the following inputs for an example:

Traffic: 20 million 18-kip rigid Equivalent Single Axle Loadings

Reliability: 90 percent

Concrete Modulus of Rupture: 600 psi

Concrete Modulus of Elasticity: 4.050 million psi

Load transfer coefficient: 3.2

Drainage coefficient: 1.0

Initial serviceability: 4.5

Terminal Serviceability: 2.5

Initial and final serviceability reflect the initial construction quality (primarily smoothness) and the end of the design life of the roadway (the point at which major rehabilitation or reconstruction is required).

With those inputs, AASHTO 1993 produces a calculated slab thickness of 11.5 inches. With AASHTO 1993, concrete strength and load transfer are important parameters. However, specific guidance as to the configuration of load transfer is not provided. Concrete durability and dimensional stability issues are also not addressed.

Many important design elements cannot be accounted for using AASHTO 1993. Simply altering traffic levels, reliability, support conditions, and levels of serviceability is not a true design optimization strategy, Ayers says.

Inside the M-E PDG

In 2008, the M-E PDG was adopted as an Interim Design Procedure by AASHTO. Full implementation by the states will take a number of years, and some states may not adopt it. The “final” version of the program, referred to as DARWin ME, is currently under development. However, the research grade software is available as a free download at (http://trb.org/mepdg/home.htm).

Ideally, optimization should be conducted based on a state or regional calibration of the M-E PDG. State or regional calibrations take into account such factors as climate and locally available materials.

The bases for comparison among the various design features are the three failure criteria for concrete pavements in the M-E PDG. The three criteria include transverse slab cracking, joint faulting and smoothness as determined by the International Roughness Index (IRI).

In his example of how M-E PDG works, Ayers includes the following variables:

Concrete strength;

Coefficient of thermal expansion of the concrete;

Subbase type and thickness;

Dowel bar size;

Edge support; and

Joint spacing.

30-year design

Failure criteria was set to the default values for cracking, faulting and IRI. Slab cracking was set at 15 percent; a faulting threshold of 0.12 inches was used; and terminal IRI was set at 172 inches per mile.

Those values, Ayers emphasizes, can have a significant effect on your final design. Establishing realistic failure values for a specific project is a key to successful use of the M-E PDG software.

Two primary climatic zones were analyzed in the example: Wet/freeze in Chicago, Illinois; and dry/no freeze in Phoenix, Arizona. Traffic was based on 5,000 AADT, and M-E PDG default values were used for traffic variables. Fully 100 percent of the design traffic was allocated to the design lane. The compound annual growth rate was fixed at 2 percent, and the design period was 30 years.

Next, Ayers chose suitable values for: soil type; granular subbases; concrete properties (modulus of rupture and coefficient of thermal expansion); dowel bar diameters and spacing according to ACPA guidelines; and edge support. A number of pavement configurations were evaluated, including the following: a 12-foot lane with no shoulder; a 13-foot widened lane; a 14-foot widened lane, and a tied concrete shoulder.

Graphs were developed illustrating the climatic effects on estimated transverse cracking, estimated faulting, and estimated IRI. In each case, the independent variable is shown as the slab thickness, and the failure mode is plotted as the dependent variable.

Similarly, the effects of base type on estimated transverse cracking, faulting, and IRI were plotted for various base types. Again, the independent variable is the slab thickness and the failure mode is shown on the vertical axis. For example, at a slab thickness of 10 inches, using a granular base in Chicago with 50-percent reliability, cracked slabs go to zero.

The effects of various modulus of rupture values, coefficients of thermal expansion, dowel diameters and spacing, and edge support were plotted to show what happens to the various failure modes for varying slab thicknesses.