A combination of finite element method (FEM) analysis, traditional pavement design methods, and forensic analysis can reveal information about a pavement’s history and its future performance that can not be determined with traditional evaluation methods alone. Engineers and researchers recently applied this three-tiered approach to the rehabilitation evaluation and design of Runway 9-27 at Houston Airport System’s (HAS) George Bush Airport (IAH). The results of the investigation demonstrate the utility of a new generation of tools for site-specific pavement evaluation and rehabilitation design, and suggest a change in the way engineers approach these projects in the future.
Seeking a quick solution
More than 40 million passengers pass through the gates of IAH annually. Houston’s largest airport has five runways to accommodate 700 departures per day. So, when a 2005 pavement condition assessment identified potential failure of a runway, HAS had to act quickly.
HAS hired PBS&J — nationally recognized for its runway pavement expertise — which, assisted by Applied Research Associates (ARA) and CMS Engineering Group, undertook a rigorous investigation of Runway 9-27 to determine the potential causes for the evident distress and to develop a rehabilitation design.
“Typically, airports demolish airfield pavements at the end of the 20-year design lives and build new ones,” said HAS Assistant Director of Aviation Adil Godiwalla, P.E. “In contrast, HAS generally performs major rehabilitations of airfield pavements sufficient to extend the lives to 40 years or more. This practice enables us to save approximately half the cost, perform the construction in half the time, and mitigate environmental impact associated with pavement replacement.”
Traditional pavement rehabilitation evaluations typically focus on identifying the material properties of the existing pavement to determine the best method for reconstruction. Investigative methods generally include a topographic survey of the pavement surface; a subsurface geotechnical investigation and core sampling; non-destructive testing such as heavy-weight deflection (HWD); back-calculating HWD and core sample data (thicknesses of various layers) to derive elastic properties (moduli) of the individual pavement layers and sub-grade; and calculating elastic properties using a layered elastic design pavement program (LEDFAA) developed for design of pavements that can handle very large aircraft, or other layered-elastic analysis modeling tools. LEDFAA was recently superseded by the Federal Aviation Administration Rigid and Flexible Iterative Elastic Layered Design (FAARFIELD) program.
The traditional approach has proven value, but it fails to provide important information that’s key to developing a site-specific, cost-effective rehabilitation plan: the causes for the pavement distress. When performed in parallel with a traditional investigation, forensic investigation and FEM can provide the missing elements needed for effective pavement rehabilitation design.
“In order to have the information we need to develop a truly ‘site-specific’ pavement design, we have to know the cause for the evident distress,” said Jim Hall, Ph.D., P.E., of ARA. ”It could be a problem in the mix or the way the pavement layers were constructed, or it could be that the pavement wasn’t designed for certain site characteristics, like the weights it has had to handle or the weather, or any combination of reasons.”
This is where forensic investigation comes in. Researchers use what they learn about the pavement’s history to ensure that the new design doesn’t experience the same distress.
“In addition,” Godiwalla said, “the data enables the team to perform a life-cycle cost analysis on the alternate pavement rehabilitation techniques to determine the most cost-beneficial and high-performance pavement design.”
In the case of Runway 9-27, the team had to discover the root causes for crazing (closely-spaced, top-down cracking), tearing, and shoving (corrugation or ripples) of the asphalt at the landing areas and high-speed exits.
Tradition and forensics in the computer age
On-going advances in computer technology open the door for development of increasingly sophisticated programs to analyze pavement conditions, allowing for more cost-effective, yet more thorough, investigations. Deficiencies in the traditional pavement analysis approach can be mitigated not only by supplementing the investigation with forensic analyses to determine causes for apparent distress, but also by using 3D FEMs to analyze relationships between weak interlayers (identified on the basis of visual examination), and shear stresses and permanent deformation.
The Federal Aviation Administration (FAA) recently incorporated the use of FEM into its pavement design criteria, effectively changing the way civil airports approach runway and taxiway maintenance. FAARFIELD, the FAA’s new airport pavement thickness design computer program, is based on a 3D FEM that can incorporate more detailed and complex characteristics of construction materials than the LEDFAA it supersedes. Other FEMs for pavement thickness design are available, such as ARA’s ISLAB 2000.
Perhaps FEM’s greatest strength is its ability to model complex configurations and incorporate a variety of materials. The method’s ability to incorporate specialized models that accurately reflect material behavior enables development of site-specific designs capable of accommodating anticipated aircraft for the next 20 years.
Runway 9-27 investigation
The Runway 9-27 investigative team adopted a non-traditional approach for its pavement assessment and design, supplementing HWD and field and laboratory efforts with FEM and forensics. The initial design phase focused on the following two crucial factors:
- aircraft weight and the number of repetitions; and
- material properties of the existing pavement.
Forming the foundation of the investigation was a detailed analysis of the pavement’s discrete layers — polymodified asphalt concrete (AC) underlain by a lime/cement/flyash (LCF) base course and a stabilized subbase course. A 9,400-foot-long section in the middle of the 100,000-foot-long runway included a half-inch-thick, stress-absorbing membrane interlayer (SAMI) within the 7-inch-thick AC surface course.
The investigative team analyzed existing material properties by using laboratory testing and back-calculating the elastic properties of the individual pavement layers and sub-grade from existing HWD data. To enable the team to develop a confident assessment of the existing pavement section and a realistic model for use in development of rehabilitation alternatives, they reconciled the differences between the laboratory-measured compressive strengths, resilient moduli, and field back-calculated values using three techniques. First, they used a layered elastic computer model of the pavement, WinJulea, to simulate the pavement response under a 25,000-pound load. Secondly, FEM was used to assess typical HWD back-calculation results using different layer combinations. Using ISLAB 2000, the team analyzed the role of weak interlayers — including the SAMI layer — in the pavement structure with respect to shear stresses and their impact on shoving and crazing within the AC layer.
“On the basis of our previous experience, we know that the location and thickness of the SAMI layers can play an important role in the development of shear-induced shoving of the asphalt surface,” said Dallas Little, Ph.D., of CMS. “The presence of a well-designed and sufficiently deep SAMI layer may reduce the potential for both reflection cracking and the propagation of cracks from the bottom up by allowing energy to dissipate at the SAMI interface rather than through crack growth or shoving in the asphalt surface.”
In the final approach, ISLAB 2000 was used to assess the substantial difference in back-calculated moduli verses laboratory-measured values. Since the back calculations of LCF moduli are substantially lower than the laboratory-measured values, it was necessary to determine how to model the LCF behavior in the pavement structure. The team identified the following two alternative approaches:
- model the LCF with the thickness measured on field cores but with a lower effective modulus; or
- model the LCF with a modulus comparable with the laboratory-measured values, but with a reduced effective thickness.
The investigators also evaluated the relationship between air void and crack distributions in core samples and the distresses observed in different locations of the runway. The tool they used was X-ray computed tomography (CT), a non-destructive technique used to visualize the internal characteristics of pavement. In X-ray CT, an X-ray source emits a beam of known intensity through the pavement sample, producing an image that depicts relative densities of aggregate, air voids, et cetera. As the sample rotates 360 degrees with respect to its center and moves vertically, images are recorded for the entire sample volume.
Results of the X-ray CT scan indicated high air void content consistently at a depth of approximately 3 inches, indicating a point of separation or slippage layer. Investigators concluded that the pavement section was not acting as a homogenous layer to resist shear stresses imparted by aircraft.
Initially, investigators suspected that the SAMI layer was a major contributor to the pavement distress. However, the CT scan and analysis disproved their initial assumption, while providing another significant finding. Little explained, “Even though we were mistaken in our assumption that the SAMI layer was primarily responsible for the distress in Runway 9-27, our investigation revealed useful information about the relationship of the SAMI layer depth to shear stresses and pavement deformation.”
For the FEM analysis, the asphalt mix material properties were selected to represent the properties of a typical asphalt mix that was recently evaluated at Texas A&M University at a high temperature to simulate Runway 9-27 conditions. The applied load was that of a B-737-300 aircraft. A pavement section with the SAMI layer 2 inches below the surface experienced more permanent deformation than the section without the SAMI layer (about 11 percent increase in permanent deformation). The section with the SAMI layer located 4 inches below the surface experienced only a 4-percent increase in permanent deformation compared with the section without SAMI. As the SAMI layer was moved closer to the surface, the shear stresses within the hot mix asphalt surface increased. Little said that this could lead to near-surface distortion, and could possibly contribute to groove closure or distortion of the groove pattern. The system with a SAMI layer at 4 inches below the surface induced a favorable combination of low permanent deformation and low surface shear stresses. The system with no SAMI layer gave the least permanent deformation but the highest surface shear stresses.
Based on the forensic evaluation and FEM analysis, investigators concluded that the current distress in Runway 9-27 pavement is likely due to high shear stresses resulting from the braking and cornering action of aircraft, and to the geometry of the asphalt layers. They found that the asphalt binder used was very stiff. The synergistic effects of a stiff, polyethylene binder, oxidative aging in the top inch of the pavement, and the high, near-surface shear stresses resulted in shoving and crazing.
The analysis also identified the benefits of maintaining a section of the existing AC in the final overlay design. This layer, approximately 2 or 3 inches thick after milling, would provide cushioning for the portland cement concrete (PCC) overlay, reducing corner stresses imposed by loading and temperature-induced curling, and allow the PCC overlay and the LCF to function as two separate layers, reducing the bending stresses within the LCF.
The team performed an LEDFAA analysis to develop pavement alternatives using the material properties they identified for specific areas of the runway. To incorporate climatic conditions into the analysis, they used the ISLAB 2000 computer model to evaluate the PCC overlay thicknesses predicted by LEDFAA. On the basis of these analyses, they developed an optimum design thickness for the new pavement.
High-speed computers and advanced forensic technologies such as X-ray CT place a new generation of tools in the hands of pavement engineers. “Given the unique nature of the existing runway pavement section, the team identified the need to extend the study beyond the limits normally associated with pavement design,” said HAS Interim Director Eric Potts. “The study results supported this innovative approach to pavement rehab evaluation and resulted in a more cost-effective, constructible, site-specific rehab design package. As IAH celebrates its 40th anniversary in 2009, we’re looking forward to seeing how Runway 9-27 holds up in 2029, when the airport celebrates its 60th.”
Justin P. Jones, P.E., is an associate vice president in PBS&J’s Houston office. He can be reached at 713-252-2608 or firstname.lastname@example.org.