New FAA pavement design software

After a 10-year research and development effort, the Federal Aviation Administration (FAA) is set to debut a new software package for airport pavement thickness design. The new program is called FAA Rigid and Flexible Iterative Elastic Layered Design, but is known by its acronym, FAArfield. In addition to putting the finishing touches on the FAArfield software, the FAA is also substantially rewriting the Advisory Circular (AC) covering Airport Pavement Design and Evaluation (AC 150/5320-6D). The revised AC will make FAArfield the FAA’s standard thickness design procedure for both rigid and flexible pavements, including overlays, and will retire the FAA nomograph-based design procedures.

Background

The FAA’s transition to computer-based design procedures has been gradual, and was originally driven by the introduction of 6-wheel aircraft gears in the 1990’s, as represented by the Boeing B777. At the time, it was recognised that design nomographs for the unique 6-wheel gear did not exist, and moreover, that the extension of the underlying design method (in particular the CBR equation with alpha factor) to 6-wheel configurations and higher might not be correct. Fortunately, an alternative procedure was available. LEDNEW, a computer-based design procedure involving layered elastic computation of stresses and strains, was originally developed by the US Army Corps of Engineers (USACOE). LEDNEW was adapted to meet FAA requirements and released as LEDFAA version 1.2 in 1995.

Even though LEDFAA 1.2 provided a short-term solution for pavement designs involving 6-wheel gears, it was recognised that a significant R&D effort was needed to develop a computer-based method that would be able to handle complex gear configurations for new large aircraft, and would fully replace the design nomographs for the existing fleet. The major problems with LEDFAA 1.2 fell into two areas:

Failure models

LEDFAA is a mechanistic-empirical design procedure in which computed responses are related to pavement life predictions through empirical failure models. However, the failure models embedded in LEDFAA 1.2 were significantly out of date, being based on full-scale traffic tests conducted by the USACOE that ended in the early 1970’s. Therefore, a new set of full-scale tests, involving 4- and 6-wheel gears and pavements constructed to current FAA standards, was needed.

Rigid pavement analysis

LEDFAA’s structural model makes use of layered elastic analysis (LEA) for all types of pavement structures. However, the basic assumptions of LEA make it incapable of computing stresses in jointed structures. Since the critical stresses in rigid pavements generally arise at joints and slab edges, this is a significant drawback in rigid pavement design. In LEDFAA 1.2, workarounds were devised to approximate the critical edge stresses in rigid pavements, but these were unsatisfactory for arbitrary gear loads. After considering a number of alternatives, it was determined that a three-dimensional finite element model (3D-FEM) would provide the best solution for design of rigid pavements and overlays.

FAArfield development

The FAA’s Airport Technology R&D Branch, located at the William J. Hughes Technical Center in Atlantic City, New Jersey, began the task of upgrading LEDFAA 1.2 and developing the FAArfield program. Full-scale testing of both rigid and flexible pavements was conducted at the National Airport Pavement Test Facility (NAPTF), which opened in 1999. Flexible pavement full-scale tests with 4- and 6-wheel gears were conducted in 1999 and 2002, and the results of these tests were then used to update the flexible pavement failure models in LEDFAA version 1.3, which was released in 2004. (Change 3 to AC 150/5320-6D made LEDFAA 1.3 an alternative to the design nomographs for all traffic mixes, not just those involving 6-wheel gears.) However, upgrading of the rigid pavement models had to wait for completion of the rigid pavement full-scale tests at the NAPTF. The rigid pavement tests were completed in December 2004, but post-traffic investigations and analysis of the data took another 18 months.

In the meantime, the FAA was developing a 3D-FEM based analytical model that would be suitable for routine pavement design applications. Most 3D-FEM programs are sophisticated, general-purpose programs intended for use in a research or academic setting. Many are not user-friendly, but require a large amount of detailed preparation on the part of the user, which can provide opportunities for error. It was therefore a challenge to integrate the 3D-FEM subprogram into the FAA standard design procedure, which intentionally limits user input to certain key variables. In FAArfield, the specifically finite element functions, such as 3D mesh generation, material model selection, boundary conditions, etc., are all standardised and performed internally by the program. The typical FAArfield user would not see the finite element mesh (although 3D mesh generation information is available to advanced users).

Rather than create a 3D-FEM engine from the ground up, the decision was made to modify an existing general-purpose 3D-FEM software program (NIKE3D) to suit the requirements of pavement design. The decision to use NIKE3D, which was originally developed by the US Department of Energy’s Lawrence Livermore National Laboratory (LLNL), was driven partly by the need to have free access to the source code for program modifications, something not possible with commercial programs. Modifications made by the FAA included adding certain element types (“infinite elements”) not in the original formulation, substantially reducing the number of subroutines by eliminating those not specifically used in pavement analysis, and recompiling the Fortran program as a WindowsTM dynamic link library (DLL). The compiled program is distributed with FAArfield under terms of a software sharing agreement between the FAA and LLNL, the NIKE3D originators.

A second issue addressed in the program development was run time requirements. Because 3D-FEM programs are resource-intensive, there was originally concern that the required run times for most design procedures would be excessive. This was especially so since the iterative design procedure and use of a cumulative damage approach (summation of damage for all aircraft individually, as opposed to the “design aircraft” concept) means that multiple 3D-FEM computations are needed in the course of a typical design. However, rapid improvements in PC processors, as well as time-saving strategies implemented in the programming, have reduced typical run times to practical levels. Running FAArfield on the current generation of PCs, one can expect a new rigid pavement design involving a mix of 10-15 aircraft types to run to completion in about 5 minutes. Due to a more complicated overlay design algorithm, rigid overlay designs are likely to take significantly longer, though most will still run to completion in under 30 minutes. New flexible pavement designs, and flexible-on-flexible overlays, which do not involve 3D-FEM computations, will normally execute in a few seconds. Of course, as computer technology continues to improve, it is expected that run time requirements will continue to decrease accordingly.

Benefits and changes

The advanced structural and failure prediction models implemented in FAArfield have already resulted in significant cost savings through more accurate pavement life predictions, particularly for new large aircraft, thus reducing reliance on overdesign of structural layers. The FAA has estimated the overall cost avoidance due to the technological advances in FAArfield at over one billion U.S. dollars. This figure represents expenditures such as strengthening overlays for new large aircraft that would have been required under older design procedures, but which FAArfield can show to be redundant.

A variety of other improvements have been added to FAArfield in anticipation of a release date in early 2008. These include:

• Full coordination with design standards in AC 150/5320-6D. This includes automatic base layer thickness design for flexible pavements.
• Upgrade to the MicrosoftTM Visual Basic.NET 2005TM programming environment. This change will ensure optimum performance and compatibility with the latest PC operating systems.
• Complete renovation of the aircraft library using the latest data from aircraft manufacturers.
• Improved report generation capabilities.

One feature that will not change is the overall look and feel of the program compared with LEDFAA 1.3. With minor changes, such as the introduction of an options window, the design sequence and program structure as experienced by the user will be the same as LEDFAA 1.3. This was done to reduce the need for a steep learning curve. Also, every effort is being made to ensure backwards compatibility with LEDFAA; for example, job files created using LEDFAA will be readable by FAArfield.

Other FAA software programs

The introduction of FAArfield will be accompanied by a complete revision of AC 150/5320-6D, making FAArfield the standard design procedure. As part of this change, the existing design nomographs based on the CBR design equation and Westergaard’s plate bending theory will no longer be used for standard FAA thickness design. However, the FAA recognises that the CBR method (with alpha factor) and rigid plate theory will continue to be the basis of pavement evaluation by the ICAO (ACN/PCN) method for the foreseeable future. Therefore, the FAA has updated its published guidance for reporting airport pavement strength (AC 150/5335-5A), and included a new computer program called COMFAA. The COMFAA program has several capabilities:

• It computes Aircraft Classification Numbers (ACN) following the ICAO standard procedure for common aircraft types, as well as arbitrary and user-definable gear configurations. This is useful for computing the Pavement Classification Number (PCN) once a critical aircraft and maximum allowable load has been defined.
• In Pavement Design Mode, it computes pavement design thickness following the FAA conventional design methods (i.e., the CBR method for flexible pavements and plate theory for rigid pavements). Again, thicknesses are computed for library aircraft as well as for user-defined aircraft. In pavement design mode, the user enters an aircraft weight, traffic level and subgrade support condition; the program returns either the total pavement thickness (flexible) or the slab thickness (rigid).

In pavement design mode, COMFAA can be used to establish maximum gross weights for critical aircraft. In this sense, it is a replacement for the AC 150/5320-6D design nomographs when used for pavement structural evaluation.

In addition to the beta version of FAArfield (called FEDFAA 2.0) and COMFAA, the FAA makes available a number of other software programs. These are available on the FAA Airport Technology R&D Branch web site www.airporttech.tc.faa.gov/downloads and include:

• BAKFAA. BAKFAA (Backcalculation – FAA) is a program designed to be used in conjunction with falling-weight deflectometer (FWD) equipment as part of a pavement evaluation program. BAKFAA reads the data from a variety of FWD devices and returns backcalculated layer properties. The computational engine in BAKFAA is LEAF (see below), and BAKFAA can also be used in “forward” mode to perform LEA calculations.
• LEAF. The structural analysis program LEAF (Layered Elastic Analysis – FAA) is built into FAArfield, but can also be downloaded and run separately. BAKFAA can be used as a platform for running LEAF, but the FAA has made the Visual BasicTM source code available for programmers to run LEAF from their own applications.
• PROFAA. PROFAA (Profile – FAA) is not a pavement design program, but is used to compute roughness indices from elevation profile data. The program currently supports indices including: Straight Edge, Boeing Bump, International Roughness Index (IRI), California Profilograph (PI), and RMS Bandpass.

Workshops and training

As the FAA transitions to new pavement design procedures, we understand the need for education. To familiarise the global aviation community with the new procedures, the FAA is holding a series of software training workshops at various locations in the United States and worldwide. These workshops are taught by FAA R&D engineers and include practical design examples using the new software products. FAArfield workshops have been held in such cities as Atlanta, Georgia (USA), Washington, DC (USA) and New Delhi (India). A recent workshop in New Delhi sponsored by the Airports Authority of India (AAI) attracted 75 participants from 14 countries. Information on upcoming workshops will be publicised on the FAA Airport Technology R&D Branch website: www.airporttech.tc.faa.gov.

Acknowledgements/Disclaimer

The contributions of the following individuals to this research effort are gratefully acknowledged: Dr. Satish K. Agrawal, Manager, FAA Airport Technology R&D Branch; Dr. Gordon F. Hayhoe, FAA, Manager, National Airport Pavement Test Facility; Mr. Robert “Murphy” Flynn, FAA; Mr. Chuck Teubert, Dr. Izydor Kawa, Dr. Navneet Garg, Dr. Edward Guo, and Ms. Lia Ricalde of SRA International, Inc., Mr. Roy D. McQueen of R.D. McQueen Associates, and Mr. Richard Ahlvin.

The contents of this article reflect the views of the author, who is responsible for the facts and accuracy of the data presented within. The contents do not necessarily reflect the official views and policies of the Federal Aviation Administration. This article does not constitute a standard, specification, or regulation.

References

Federal Aviation Administration, Advisory Circular AC 150/5320-6D, “Airport Pavement Design and Evaluation,” Change 3, April 30, 2004, FAA Office of Airport Safety and Standards, Washington, DC.

Federal Aviation Administration, Advisory Circular AC 150/5335-5A, “Standardised Method of Reporting Airport Pavement Strength – PCN,” September 28, 2006, FAA Office of Airport Safety and Standards, Washington, DC.