STRAIN-RESPONSE CHARACTERIZATION FOR UNBONDED CONCRETE OVERLAYS SUBJECTED TO HEAVY AIRCRAFT GEAR WITH MULTIPLE AXLES

ABSTRACT

Concrete pavement performance, with regard to fatigue life, has previously been studied using strain gages installed in-situ in the pavement, with the primary focus on peak strain response. This study focuses on identifying and characterizing additional components of strain responses from unbonded concrete overlays for concrete pavement on airfields, and relating those strain responses to observed performance. Strain data collected through a series of full-scale tests at the Federal Aviation Administration (FAA) National Airfield Pavement Test Facility (NAPTF) are analyzed to observe effects of repeated loading with heavy aircraft gear with multiple axles on three different structural cross-sections of unbonded concrete overlays. Components of strain response, such as peak strain, cumulative area, percent recovery, pre-stress area, post-stress area and duration are defined; these additional strain response components are analyzed and correlated to peak strain response. The peak strain is found to be strongly correlated to the cumulative strain area component of the strain response, but not to other components of the strain response. Preliminary regression relationships with performance, defined in terms of structural condition index (SCI), could be established only with peak strain, percent recovery and number of axles for the top of unbonded overlay, or alternately with area components. This study provides the foundation for future study of airfield concrete pavement overlay performance, especially for the subsequent experiment, performed at the same location with same pavement cross-section but with weakened support conditions. List of Acronyms

TABLE OF CONTENTS

  List of Tables………………………………………………………………………….. vii List of Figures………………………………………………………………………….. xiii Acknowledgements…………………………………………………………………….. xvi   CHAPTER 1. INTRODUCTION……………………………………………………… 1 1.1. Background…………………………………………………………………. 1 1.2. Objectives………………………………………………………………… 7 1.3. Scope……………………………………………………………………… 8   CHAPTER 2. LITERATURE REVIEW……………………………………………… 10 2.1. Background……………………………………………………………….. 10 2.2. Concrete Properties……………………………………………………….. 10 2.3. Concrete Pavement Fatigue ………………………………………………. 12 2.3.1. Stress Ratio Models……………………………………………….. 12 2.3.2. Slab Fatigue……………………………………………………….. 14 2.3.3. Variable Amplitude Loading……………………………………… 15 2.3.4. Mechanistic Approach……………………………………………. 15 2.3.5. Numerical Approach (Finite-Element Methods)…………………. 16 2.3.6. Fuzzy Logic……………………………………………………….. 16 2.4. Airbus A-380 and Boeing 777……………………………………………. 16 2.5. Current FAA Design Procedure…………………………………………… 18 2.6. Structural Condition Index……………………………………………….. 19 2.7. Strain Gage Characterization……………………………………………..    19 2.8. Summary of Important Findings from Literature Review……………….     20   CHAPTER 3. FULL-SCALE TESTING OF UNBONDED OVERLAYS ..………….. 22 3.1. NAPTF …………………………………………………………………….. 22 3.2. Pavement Cross-Sections…………………………………………………. 23 3.3. Types of Loading and Aircraft Gears…………………………………….           25 3.4. Instrumentation…………………………………………………………… 29 3.4.1. Strain Gages……………………………………………………….. 29 3.4.2. KM-100B…………………………………………………………….. 30 3.5. Performance……………………………………………………………….. 41   CHAPTER 4. METHODOLOGY …………………………………………………… 42 4.1. Overview of Proposed Methodology……………………………………… 42 4.2. Detailed Processing of Data………………………………………………. 43 4.2.1 Data Extraction/Filtering………………………………………….. 44 4.2.2. Responses from Track 0…………………………………………… 44 4.2.3 Matlab Programming……………………………………………… 45 4.2.3.1 Type 1 for North Test Items………………………………. 50 4.2.3.2 Type 2 for North Test Items ………………………………. 53 4.2.3.3 Type 3 for North Test Items ………………………………. 55 4.2.3.4 Type 4 for North Test Items ……………………………… 56 4.2.3.5 Type 1, 2, 3 and 4 for South Test Items…………………… 58 4.2.4 Statistical Analysis…………………………………………………. 60   CHAPTER 5. ANALYSIS ……………………………………………………………….. 62 5.1. Relationship of Components of Strain Gage Response to Peak Strain…..     62 5.2. Relationships between Peak Strain and other Components of Strain Gage Response…………………………………………………………………… 72 5.3. Strain Gage Response with Change in Loading and Cross-Sections………. 80 5.4. Performance………………………………………………………………    86 5.4.1. At the Top of Overlay……………………………………………… 86 5.4.2. At the Bottom of Overlay…………………………………………. 88 5.4.3. At the Top of Underlay……………………………………………. 89 5.4.4. At the Bottom of Underlay……………………………………….. 90 5.5. Summary…………………………………………………………………… 91   CHAPTER 6. CONCLUSIONS AND RECOMMENDATIONS ……………………. 93 6.1. Findings…………………………………………………………………… 93 6.2. Conclusions………………………………………………………………… 96 6.3. Recommendations……………………………………………………………. 99     REFERENCES………………………………………………………………………… 101   APPENDIX A. Strain gage  coordinates and calibration factor……………………..     111 APPENDIX B. Matlab Codes ……………………………………………………….. 114 APPENDIX C. Mean and Standard Deviation for Selected Gages …………………. 127 APPENDIX D. Performance Prediction………………………………………………. 174

CHAPTER 1. INTRODUCTION

1.1 Background Aviation activity in the United States accounts for approximately forty percent of all commercial aviation and fifty percent of all general aviation activity in the world (FAA, 2008). As a key industry in the United States, air transportation has a significant impact on the economy. Air transportation faces the high cost of shutdowns due to rehabilitation of airfield pavements, which also results in unnecessary delays to the traveling public.  Military airfields face similar problems when operational efficiency is affected by poor pavement condition. Therefore, airfield pavements should be designed for better performance and constructed with a high degree of quality (Kohn and Tayabji, 2003). Concrete has been widely used in the construction of airport pavements because of its durability and capacity to sustain large loads (Portland Cement Association, 2010).   Crack generation and propagation are among the most serious problems in concrete pavements (Darestani, 2007; Hossain et al., 2003).  Therefore, safe operation of aircrafts requires airfield pavement condition assessment with timely performance of pavement maintenance and repair (Greene et al., 2004). Pavement distress, structural capacity, friction, and roughness are important factors for condition assessment (Greene et al., 2004). Once the pavement condition is properly assessed, appropriate maintenance and rehabilitation can be programmed and designed. With the onset of a new generation of aircraft gears, including large multi-axled gears, such as the triple dual tandem, airport pavement design assumptions need to be reexamined to determine the potential effects of gear spacing and load levels on the development of stresses and resulting fatigue life of concrete slabs (Roesler, Hiller, and Littleton, 2004). Extensive research on stress development and fatigue of airport concrete slabs has been performed, including the work from Westergaard (1926), the Lockbourne and Sharonville test sections (Parker et al. 1979), the PCA design for airfields (Packard 1973, 1974), and the work by Rollings (1981, 1986, 1990, 1998, 2001) for the U.S. Army Corps of Engineers (USACOE). Airfield concrete pavement fatigue has been extensively studied by Smith and Roesler (2003) and Littleton (2003) (Roesler, Hiller, and Littleton, 2004). Although portland cement concrete pavements have been used for the construction of airfield pavements for many decades, and typically perform well, eventually all pavements require rehabilitation or replacement. An unbonded concrete overlay offers an attractive alternative for its use as an airfield pavement rehabilitation technique for several reasons. Unbonded concrete refers to concrete pavement constructed over an existing concrete pavement. The concrete layers are separated by an interlayer, typically of hot-mix asphalt concrete, which acts as a shear zone, enabling the concrete layers to move independently of each other. This is why the term unbonded is used (Pavement Technology Advisory, 2005). One of the reasons unbonded concrete overlays are suitable for airfield pavements is that, by leaving the existing pavement in place, the in situ conditions of subgrade and base layers are essentially undisturbed, minimizing any opportunity for additional consolidation or settlement to take place. Another advantage is that the existing pavement can be taken into consideration in structural design, typically resulting in a thinner and less costly required pavement layer (Stoffels et al., 2008), which may be especially important for the heavy loads and thick pavement structures typically required for airfields. The Innovative Pavement Research Foundation (IPRF) has an objective of improving the current understanding of the influence of design parameters on unbonded concrete overlays of airfield pavements, thus enabling improvement of design methodologies and consideration of the new aircraft. In 2005, IPRF contracted for a study to prioritize and conduct the necessary research activities for the design of unbonded concrete overlays for airfields, including a series of full-scale tests to be built at the Federal Aviation Administration (FAA) National Airfield Pavement Test Facility (Stoffels et al., 2008). The construction of the first of the IPRF unbonded overlays took place between November 2005 and May 2006. Three 6000-ft² pavement cross-sections, with thicker overlay on thinner underlay, equal thicknesses of overlay and underlay, and thinner overlay on thicker underlay, were built, as shown in Figure 1. The full-scale loading was performed from July 2006 until November 2006 to study the effects of the relative thickness of the overlay and underlay, the correspondence between predicted and measured responses, the relative effects of two gear configurations, and the relationship of the failure mechanisms in the existing pavement and overlay. Figure 1. Test item layout with as-built thicknesses. Pavement instrumentation has recently become an important tool to monitor in-situ pavement material performance and quantitatively measure pavement system response to loading. Sensors, such as strain gages, have become available to monitor the health of the pavement and its performance (Weinmann et al., 2004). Embedded strain gages have been used to capture horizontal and vertical deformations of the pavement due to load on the surface of the pavement. Strain gage response has been used to determine the effect of static and fatigue cracks on a pavement (Roesler and Barenberg, 1999). The strain responses can be used to verify the assumptions in mechanistic response models, whether finite element or closed form solutions, and for development of relationships between load-induced responses and structural performance. In the IPRF unbonded overlay studies at the NAPTF, strain gages were installed in an attempt to capture critical pavement responses under both twin dual tandem and triple dual tandem gears. Strain responses from the strain gages were recorded. A typical response from one of the stain gages is shown in Figure 2.   Figure 2. Typical strain gage response as shown in FAA TenView.   Most of the previous studies on strain response in concrete pavements have focused on the peak measured response from the strain gages, as defined in Figure 2. Peak strain response has been used widely in different models to study the fatigue behavior and life for the pavement. For this thesis, different components of strain response were analyzed and compared to peak strain and to observed performance, to determine whether additional valuable information can be extracted from the strain gage data. For example, some of the components of strain gage response used are illustrated in Figure 3. Those components include the peak values of strain for each axle of the gear, the recovery between axles, the pre-stress and post-stress peaks, duration and the areas of the various portions of the strain responses. Details of the strain response components are developed and described in Chapter 4. Figure 3-b. Cumulative area of a strain gage response shown as shaded portion of the curve.   Although the responses from strain gages have been widely used, literature on the repeatability of these measurements is limited (Gokhale et al., 2009). Factors, such as inherent construction-related variability in pavement structure and materials, presence or absence of moisture, layer thicknesses, joint position, slab curling, and gage orientation, affect the variability of the responses from in-situ instrumentation. Therefore, for this study, strain responses are considered on a statistical basis, rather than by detailed comparison of selected typical responses. The complete strain response was considered by statistical characterization of the various peaks and component areas. By considering only the highest peak, other peaks and recovery areas which may have potential information relative to pavement performance, are neglected. Thus, different strain gage components, such as strain values for individual axles, cumulative area, pre-stress area, post-stress area, % recovery and duration, are characterized and correlated with the peak strain response for the strain gage. A preliminary analysis of pavement performance was established with various identified strain gage components. 1.2. Objectives. The overall objective of this research is to examine and characterize the strain gage responses from the full-scale testing of unbonded concrete overlays at the FAA NAPTF. The responses are characterized not just in terms of peak strain, but in terms of multiple peak values, recovery, and component areas of the responses. No such characterization for concrete pavements has been found in the literature review. Therefore, the development of the characterization of stain gage response is the primary objective and product of this research.   The following questions are addressed using statistical characterization of the interpreted strain data.

• Do the various components of the strain gage response, including peak strain values for individual axles, cumulative area, pre-stress area, post-stress area, percent recovery and duration, directly relate to the peak strain?
• How do the components of strain gage response change with different gear configurations and different pavement cross-sections?
• If not directly related to peak strain, how do the components of the strain gage response relate to observed performance in terms of cracking or structural condition?

  1.3. Scope For this thesis, study is limited to analysis of the FAA NAPTF unbonded concrete overlay strain gage data. Data before the first visually-observed crack was considered, as these responses are most consistent and relate most directly to those that would be predicted using a closed-form solution or finite element program for design purposes. For the original study of the experiment, data was collected in different loading paths consistent with the wander patterns on an airfield. The total number of loading passes prior to the first observed crack for each test item is enumerated in Table 1. A test item is defined as a unique combination of structural cross-section and gear configuration.           Table 1. Vehicle Passes Utilized for this Study

  Peak strain responses from all tracks for all gages were analyzed for initial loading, which was performed with wheel loads of less than 40,000 pounds, as compared to the extended fatigue loading at 50,000-pound wheel loads. Code in Excel’s Visual Basic for Applications was developed to verify which strain gage responses were of the greatest magnitude. The model was developed with the help of Pennsylvania State University’s doctoral candidate Lin Yeh.  It was found that the greatest responses occurred when the outside wheel track was directly over the strain gages. Detailed characterization of the strain responses was limited to those passes, as indicated in the final column of Table 1.

STRAIN-RESPONSE CHARACTERIZATION FOR UNBONDED CONCRETE OVERLAYS SUBJECTED TO HEAVY AIRCRAFT GEAR WITH MULTIPLE AXLES