
small (250x250 max)
medium (500x500 max)
large ( > 500x500)
Full Resolution


April 2006 Research Report: UCPRC RR 2005 06 Calibration of Incremental Recursive Flexible Damage Models in CalME Using HVS Experiments Authors: Per Ullidtz, Dynatest Consulting Inc John Harvey, UC Davis Bor Wen Tsai, UC Berkeley Carl Monismith, UC Berkeley This work was completed as part of Partnered Pavement Research Program Strategic Plan Element 4.1: “ Development of the First Version of a Mechanistic Empirical Pavement Rehabilitation, Reconstruction and New Pavement Design Procedure for Rigid and Flexible Pavements ( pre Calibration of AASHTO 2002)” PREPARED FOR: California Department of Transportation Division of Research and Innovation Office of Roadway Research PREPARED BY: University of California Pavement Research Center Berkeley and Davis Stage 5 Distribution ii UCPRC RR 2005 06 DOCUMENT RETRIEVAL PAGE Report No: UCPRC RR 2005 06 Title: Calibration of Incremental Recursive Flexible Damage Models in CalME Using HVS Experiments Authors: Per Ullidtz, Dynatest Consulting Inc.; John Harvey, UC Davis; Bor Wen Tsai, UC Berkeley; and Carl Monismith, UC Berkeley Prepared for: California Department of Transportation Division of Research and Innovation Office of Roadway Research FHWA No.: F/ CA/ RR/ 2006/ 49 Date: Stage 5, May 2007 Contract number: UCPRC RR 2005 06 Client Reference No: UCPRC RR 2005 06 Status: Final, Caltrans approved Abstract: Caltrans is in the process of implementing Mechanistic Empirical design procedures. All mechanisticempirical methods must be validated/ calibrated against the behavior of real pavements. This should be done before implementing models in design methods to ensure that designs will be reasonable. The Heavy Vehicle Simulator ( HVS) provides a first step in this validation/ calibration process. The short test section can be carefully constructed with well characterized materials and instrumented to measure the pavement response. The climatic conditions may be controlled or monitored closely and all load applications are known exactly. The pavement may also be tested until it fails. The HVS may be seen as “ large scale” laboratory equipment, between the “ small scale” laboratory equipment ( triaxial tests, bending tests etc.) and the reality of real pavements, which have uncertainties regarding materials, loads and climatic conditions. The two HVSs owned by Caltrans have been used on 27 flexible pavement test sections, with varying combinations of asphalt and granular layers. Temperature control was used during the tests. Most sections have been instrumented with Multi depth Deflectometers ( MDDs) to compare the measured pavement deflections ( at several depths) to the deflections predicted by mechanistic methods, during the full duration of tests carried to “ failure” ( in terms of rutting or cracking). Results from mechanistic models have been compared with the deflection measurements and performance as a first step prior to empirical calibrations with field results. The complete time history of each test has been compared rather than just the beginning and end measurements. This report presents the validation of the mechanistic models for asphalt fatigue and for permanent deformation with the HVS test results. Keywords: Mechanistic empirical, full scale testing, calibration, response, performance, flexible pavement . Proposals for implementation: None Related documents: Kannekanti, V., and Harvey, J. June 2005. Sensitivity Analysis of 2002 Design Guide Rigid Pavement Distress Prediction Models. Draft report prepared for the California Department of Transportation. Pavement Research Center, Institute of Transportation Studies, University of California Berkeley, University of California Davis. UCPRC RR 2005 01 Signatures: P. Ullidtz Principal Author J. Harvey Co principal Investigator C. L. Monismith Co principal Investigator D. Spinner Editor M. Samadian Caltrans Contract Mgr. Stage 5 Distribution iii UCPRC RR 2005 06 DISCLAIMER The contents of this report reflect the views of the authors who are responsible for the facts and accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the State of California or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation. Stage 5 Distribution iv UCPRC RR 2005 06 EXECUTIVE SUMMARY The first step in a mechanistic empirical ( ME) pavement design or evaluation is to calculate pavement response — in terms of stresses, strains, and/ or displacements — using a mathematical ( or mechanistic) model. In the second step, the calculated response is used as a variable in empirical relationships to predict structural damage ( decrease in moduli or cracking) and functional damage ( rutting and roughness) to the pavement. Both of these steps must be reasonably correct. If the calculated response bears little resemblance to the pavement’s actual response, there is no point in trying to use the calculation to predict future damage to the pavement. In other words, only if the calculated response is reasonably correct does it make sense to try to relate the damage to the pavement response. This study’s purpose was to evaluate the overall trends of the damage models in the draft software package called CalME against those of Heavy Vehicle Simulator ( HVS) tests for which data was available. The report presents simulations of HVS tests using the set of distress models included in CalME. These models are for the typical flexible pavement distresses observed in California: asphalt fatigue, asphalt rutting, unbound layers rutting, and reflection cracking. An Incremental Recursive approach ( see item 4 below) was used for the simulations included in this report because this approach can accurately indicate pavement condition at different points during a pavement’s life. Approaches Included in CalME CalME software provides the user with four approaches to evaluating or designing a flexible pavement structure: 1. Caltrans’ current methods: the R value method for new flexible structures and the deflection reduction method used by Caltrans for overlay thickness design for existing flexible pavements. 2. “ Classical” Mechanistic Empirical ( ME) Design, which is based largely on the Asphalt Institute Method which uses very simple methods to characterize materials, climate, and traffic inputs. 3. An Incremental approach, which is a standard Miner’s Law approach that permits damage calculation for the axle load spectrum and expected temperature regimes, but without updating of the material’s properties through the life of the project. This is an approach similar to the one for cracking of asphalt included in the NCHRP 1 37A Pavement Design Guide, also referred to as the Mechanistic Empirical Design Guide ( MEPDG). This type of approach is calibrated against an end failure state ( such as, 25 percent cracking of the wheelpath) and it assumes a linear accumulation of damage to get to that state. 4. An Incremental Recursive approach in which the materials properties of the pavement — in terms of damage and aging — are updated as the pavement life simulation progresses. The current Caltrans methods and the Classical method are very fast in terms of computational time, and user input is highly simplified. In CalME both of these options perform a “ design” function, calculating and presenting pavement structures that meet the design requirements for the design traffic, materials, and climate. For design practice the Classical and Caltrans methods should be used to produce a set of potential pavement sections. The Incremental Recursive method should then be run to check the lowest cost alternative designs in the set to be certain that they meet design requirements. Once the final design has been selected, its Incremental Recursive output provides a prediction of the pavement condition across its entire life. The prediction of the pavement’s condition through its life from the Incremental Recursive output can be used as the first prediction for use in a pavement management system. Stage 5 Distribution v UCPRC RR 2005 06 Use of Heavy Vehicle Simulator Data to Evaluate Models The Incremental Recursive models included in CalME were used to predict performance for all twenty seven of the flexible pavement HVS tests performed so far as part of the Accelerated Pavement Testing ( APT) program operated for the California Department of Transportation ( Caltrans) by the University of California Pavement Research Center ( UCPRC). The HVS test data in this report come from tests performed between the years 1995 and 2004. The HVS response data and corresponding laboratory test data were extracted from the UCPRC HVS database. During HVS testing, pavement response  in terms of deflections at the surface and/ or at multiple depths  may be measured. A Road Surface Deflectometer ( RSD) measures deflections at the surface and is similar to the Benkelman Beam used to develop the current Caltrans overlay design method in the 1950s. A Multi depth Deflectometer ( MDD) measures deflections at multiple depths. In order to accurately predict the gradual degradation of a pavement, the response model must predict measured deflections with reasonable accuracy. Although a model might predict deflections correctly, this ability does not guarantee that the model can also accurately predict the stresses and strains in all the pavement layers. However, the opposite is true: if a model predicts deflections incorrectly it will also produce incorrect stress and strain predictions. Therefore when attempting to calibrate ME models from HVS tests, the research team’s first concern was to make sure that resilient deflections were predicted reasonably well for the duration of the test and for all load levels. This prediction depended on the moduli of all the pavement layers and on the changes to the moduli caused by fatigue damage, slip between asphalt layers, non linear elastic characteristics of unbound layers, and the effect of confinement on granular layers. Once reasonably good agreement was achieved between the measured and the calculated deflections then the permanent deformation models could be calibrated with confidence. Differences in boundary conditions, strain levels, and loading times, all of which can produce varied effects in materials, result in differing moduli values. In this study, methods used for determining moduli ( also referred to as “ stiffness”) values included backcalculation from Falling Weight Deflectometer ( FWD) and MDD data, and direct measurement — employing laboratory triaxial testing for unbound materials and flexural frequency sweep testing for asphaltic materials. Stiffnesses for the study’s asphalt materials were taken primarily from flexural frequency sweep data. Stiffnesses for the unbound layers came primarily from MDD data backcalculation. In practice the FWD is seen by the research team as the primary tool for stiffness measurement of all layers already constructed because it is used in the field on the full pavement system; this is thought to be appropriate because the boundary conditions are those of real pavement, and most Caltrans’ work will be rehabilitation and reconstruction with at least some layers already in place. The research team saw the flexural beam test as the primary means for measuring the stiffness and fatigue characteristics of asphalt overlay materials for new layers. For new pavement construction, a combination of FWD testing on existing pavements and triaxial testing can be used to develop a database of stiffnesses of unbound granular layers and subgrades based on different characteristics, such as Unified Soil Classification System ( USCS) classification. The purpose of this study was to evaluate the overall trends of the CalME damage models against those of the HVS test results. This was accomplished by comparing deflections calculated using moduli determined from initial measurements and CalME damage calculations with measured deflections under HVS loading. The results presented in this report verify that, overall, the CalME damage trends for deflection and permanent deformation under loading are correct. During HVS testing, deflections often increase markedly, sometimes becoming more than twice as high at the end of the test as they were at the beginning because of damage to the asphalt concrete caused by the repeated wheel loads. However, the flexible pavement design model of the NCHRP 1 37A Design Guide does not consider any decrease in the asphalt modulus as a result of fatigue damage ( except for rehabilitation designs). In fact, the NCHRP 1 37A Design Guide includes a model for aging that predicts a continuous increase in the stiffness of the asphalt concrete layers across the life of the pavement, which results in increased stiffness and smaller predicted deflections as the pavement is subjected to trafficking. While the Stage 5 Distribution vi UCPRC RR 2005 06 aging is potentially important, the effect of updating stiffness for aging and not updating it for fatigue damage results in calculation of very unrealistic elastic responses in the pavement during its life. This makes it impossible to use the model to simulate an HVS test and, inversely, to use HVS tests to calibrate the model, except for pavements with extremely thick asphalt concrete layers where little fatigue should develop. Results of HVS Test Simulation Using CalME The series of HVS tests in this report are grouped here by goals, which are defined as follows: • Goal 1, a comparison of new pavement structures with and without asphalt treated permeable base ( ATPB) layer under dry conditions, moderate temperatures, 20° C ( HVS Sections 500RF, 501RF, 502CT, 503RF) • Goal 3 Cracking, a comparison of reflection cracking performance of ARHM GG ( the acronym ARHM, asphalt rubber hot mix gap graded, refers to the material specification at the time of construction in April 1997.) and dense graded asphalt concrete ( DGAC) overlays placed on the cracked Goal 1 sections, dry conditions, 20° C ( HVS Sections 514RF, 515RF, 517RF, 518RF) • Goal 3 Rutting, a comparison of rutting performance of ARHM GG and DGAC overlays of previously untrafficked areas of Goal 1 pavements, dry conditions, 40° C or 50° C at 50 mm depth, four different tire/ wheel types ( HVS Sections 504RF, 505RF, 506RF, 507RF, 508RF, 509RF, 510RF, 511RF, 512RF, 513RF) • Goal 5, a comparison of new pavement structures with and without ATPB layer under wet conditions ( water introduced into base layers), moderate temperatures, 20° C ( HVS Sections 543RF, 544RF, 545RF) • Goal 9, initial cracking of asphalt pavement with six replicate sections in preparation for later overlay, new pavement, ambient rainfall, 20° C ( HVS Sections 567RF, 568RF, 569RF, 571RF, 572RF, 573RF) CalME models that the simulations evaluated included: • A stiffness model for asphalt concrete modulus as a function of reduced time based on the model used in NCHRP 1 37A Design Guide, with some adjustments based on field observations; • An asphalt concrete fatigue model that predicts damage, in terms of decrease in modulus, as a function of load repetitions, tensile strain, and stiffness, using parameters from flexural beam testing; • An ability to model partial bonding between asphalt concrete layers; • A model that adjusts the stiffness of unbound layers as a function of the combined bending resistance ( a function of their stiffness and thickness) of the layers above them; • A model that adjusts the stiffness of unbound layers as a function of load level, with an increased load level increasing the moduli for the granular layers and decreasing modulus for the subgrade ( clay); • A permanent deformation model for asphalt concrete as a function of permanent shear strain near the pavement surface beneath the edge of a tire, with permanent shear strain predicted by the calculated elastic shear strain and elastic shear stress; • A permanent deformation model for unbound layers as a function of the vertical strain at the top of each layer; and • A reflection cracking model based on tensile strain calculated using a regression equation developed from a large number of Finite Element analyses and the same damage parameters developed for asphalt concrete fatigue. Stage 5 Distribution vii UCPRC RR 2005 06 Response Models During most of the HVS tests, resilient deflections were measured using the RSD and the MDD. The following figure summarizes the measured deflections with those calculated using CalME damage models for all of the sections in terms of the ratio of the initial deflections before HVS loading to the final deflections at the end of the loading. Assumptions made regarding differences between moduli from different measurement methods, shift factors, slip between layers, and non linear elasticity of unbound layers to obtain reasonably good agreement between measured resilient deflections and those calculated with CalME are discussed in the report. The observed behavior of the aggregate base ( AB) and subbase layers under HVS loading contradicts the commonly accepted wisdom for granular materials, which is based primarily on triaxial testing. The observed behavior is discussed in the report and is modeled in CalME. Using these assumptions, it was possible to model resilient deflections reasonably well for the full history of all of HVS test sections using the layered elastic analysis program ( LEAP) response model. Final/ initial deflection 0.00 1.00 2.00 3.00 4.00 5.00 0.00 1.00 2.00 3.00 4.00 5.00 Measured Calculated RSD MDD Equality Figure ES 1. Ratio of initial to final deflection. Damage of Asphalt Materials Controlled strain fatigue tests conducted on beams were used to derive model parameters for the decrease in modulus for all the asphalt materials — except for the ATPB, where laboratory tests were not available. Working under the assumptions used in the modeling and using a shift factor with these damage models produced the correct changes in resilient deflections during all the HVS tests. Stage 5 Distribution viii UCPRC RR 2005 06 For reflection cracking, a simple model was used to calculate the strains in an overlay caused by existing cracking in the original top layer. Using this model and the laboratory fatigue model, reasonably correct resilient deflections were also predicted. Relating visual cracking to the calculated asphalt damage proved to be difficult, and no single relationship could be derived. Goal 1 and Goal 5 showed differences between the drained and the undrained sections; for Goal 3 Cracking, the increase in visual cracking was quicker than for Goal 1 and Goal 5; and for Goal 9, visual cracking occurred at much less calculated damage than for the other experiments. It is possible that the relationship between visual cracking and calculated damage depends on the thickness of the asphalt layers, and that reflection cracks and cracks in new pavements develop differently. It is also possible that the development of visual cracking depends on factors that the simulations did not consider. No single relationship could be established between the relative increase in deflection and the amount of surface cracking ( shown in the following figure), but it may be noted that visible cracking was not observed until deflection had increased by 50 percent or more. Cracking versus relative deflection 0 2 4 6 8 10 12 1 1.5 2 2.5 3 3.5 4 4.5 Deflection/ initial deflection Cracking m/ m2 500RF 501RF 502CT 503RF 514RF 515RF 517RF 518RF 543RF 544RF 545RF 567RF 568RF 569RF 571RF 572RF 573RF Figure ES 2. Cracking versus increase in deflection. Permanent Deformation of Asphalt The following figure shows the measured and predicted final permanent deformation of the asphalt layers from Goal 1, Goal 3, and Goal 5, where data were available. Permanent deformation was calculated for the upper 100 mm of the asphalt layer( s). Stage 5 Distribution ix UCPRC RR 2005 06 Permanent deformation in AC ( pro rated) 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 Measured, mm Calculated, mm Equality 20 º C at surface 45 or 50 º C at surface Figure ES 3. Measured and predicted final permanent deformation of asphalt. In Figure ES 3, the 20° C outlier point with a measured deformation greater than the calculated one is Section 543RF, which was a wet, drained test where the ATPB stripped and collapsed. The two relatively low values at high temperatures are from the test with a bias ply dual tire and from the test with an aircraft tire. The correlation coefficient between measured and calculated values is 0.82 and the standard error of estimate is 2.2 mm. The parameters for predicting permanent shear strain were based on Repeated Simple Shear Tests at Constant Height ( RSST CH). Permanent Deformation of Granular Layers The permanent deformation of the granular layers for Goal 1, Goal 3, and Goal 5 are shown with average measured values in Figure ES 4. It should be noted that the permanent deformations are rather small except for Section 543RF, the wet drained section, where the permanent deformation includes part of the ATPB. Stage 5 Distribution x UCPRC RR 2005 06 Permanent deformation of granular layers 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 Measured, mm Calculated, mm Goal 1 Goal 3 Goal 5 Equality Figure ES 4. Final permanent deformation of granular layers. Permanent deformation was calculated both at the top of the AB and at the top of the aggregate subbase ( ASB). The two materials are rather similar and might have been treated as a single layer. Permanent Deformation of Subgrade The final permanent deformation of the subgrade is even smaller than that of the granular layers, with a maximum measured value of less than 2 mm. In addition, the data scatter is as large as that of the granular layers, with an average coefficient of variation of 70 percent. This is far from ideal for the calibration of a subgrade permanent deformation model. The mean measured and predicted final deformations are shown in Figure ES 5. The subgrade and granular results indicate that rutting of the unbound layers is probably not a major concern for existing Caltrans pavements that need rehabilitation unless there is poor drainage or significant amounts of water are entering cracks in the asphalt layers. Stage 5 Distribution xi UCPRC RR 2005 06 Permanent deformation of subgrade 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Measured, mm Calculated, mm Figure ES 5. Final permanent deformation of the subgrade. Total Permanent Deformation at Pavement Surface Figure ES 6 shows the final calculated permanent deformation at the pavement surface versus the measured final deformation averaged from profilometer measurements along the HVS test area. Calculated final permanent deformations that underestimated the measured final permanent deformations were worst for the Goal 5 sections in which water was dripped into the base layers, especially for the drained section, 543RF in which the ATPB stripped. The correlation coefficient between measured and calculated deformations is 0.61 and the standard error of estimate is 2.6 mm. Conclusions and Recommendations The overall results from this study indicate that Incremental Recursive models provide reasonable results when predicting the response and performance of pavement under HVS loading. However, now that the models have been shown to match the mechanics of the flexible pavements under HVS loading, additional work remains to be done before these models can be used for pavement design and performance prediction. There are significant differences between HVS testing and field results, and the approach used in this study has limitations because of those differences. These include the effects of age and of seasonal variation that have not been quantified in the simulations because HVS tests are of relatively short duration and are performed, to varying degrees, in controlled environments. Field calibration is required to evaluate the response difference between the field pavement and the Incremental Recursive simulation that should be attributed to aging and seasonal effects. It is likely that the effects of aging can be dealt with using shift factors. Stage 5 Distribution xii UCPRC RR 2005 06 Permanent deformation at pavement surface 0.0 5.0 10.0 15.0 20.0 0.0 5.0 10.0 15.0 20.0 Measured, mm Calculated, mm Goal 1 Goal 3 45/ 55° C Goal 3 20° C Goal 5 Goal 9 Equality Figure ES 6. Final permanent deformation at the pavement surface. The effects of rest periods between loadings and of faster traffic have also not been included in the calibration. It is expected that different shift factors will result because of rest periods and different trafficking patterns. Lastly, moduli from frequency sweep data, triaxial tests, FWD tests, and MDD deflections used in this study are similar but they are not identical. The NCHRP 1 37A Design Guide study proposes relying primarily on triaxial testing to characterize the stiffness of flexible pavement layers and the permanent deformation parameters of asphaltic materials. Recommendations are made in this report for the most practical and economical methods for characterizing materials based on the understanding that the majority of Caltrans’ work over the next several decades will be rehabilitation and reconstruction, with some addition of lane capacity. Recommendations are also made regarding the next steps to develop the CalME models. These include: 1. Perform a sensitivity analysis using “ typical” values for properties and climate in the database established to date, and compare the results from the Classical, Incremental, and Incremental Recursive methods included in CalME to evaluate reasonableness of sensitivity across the three methods. 2. Simulate mainline highway case studies and test track data ( such as WesTrack and NCAT track) using the recommended methods for characterizing flexible pavement materials in conjunction with the Incremental Recursive models in CalME, and compare the simulated and measured results, as was done for the HVS results presented in this report. This step will provide validation for the models Stage 5 Distribution xiii UCPRC RR 2005 06 3. Address the variability of the input parameters ( moduli, thicknesses, traffic loading, etc.) and uncertainty on the damage models. Several approaches should be considered, including the approach used in the NCHRP 1 37A method. 4. Make final decisions regarding use of cemented layers in the flexible pavement structure, then calibrate. It is generally recommended that “ semi rigid” pavements, in which asphalt concrete is placed directly on cement treated base ( CTB) or lean concrete base ( LCB), not be used because of the relatively quick reflection of shrinkage cracks. However, because Caltrans has used semi rigid pavements in the past and they remain in the current design method, it is therefore important to have models for the response and performance of these layers. The models in the NCHRP 1 37A Report should be the starting point for such a validation and calibration exercise. Stage 5 Distribution xiv UCPRC RR 2005 06 TABLE OF CONTENTS Executive Summary........................................................................................................................ .................... iv Approaches Included in CalME ...................................................................................................................... iv Use of Heavy Vehicle Simulator Data to Evaluate Models ............................................................................. v Results of HVS Test Simulation Using CalME .............................................................................................. vi Response Models......................................................................................................................... .............. vii Damage of Asphalt Materials ..................................................................................................................... vii Permanent Deformation of Asphalt ........................................................................................................... viii Permanent Deformation of Granular Layers................................................................................................ ix Permanent Deformation of Subgrade ........................................................................................................... x Total Permanent Deformation at Pavement Surface.................................................................................... xi Conclusions and Recommendations................................................................................................................ xi List of Figures........................................................................................................................ .......................... xvii List of Tables ............................................................................................................................... .................... xxv 1.0 Introduction................................................................................................................... ......................... 1 1.1 Models and Approaches Included in CalME .......................................................................................... 1 1.1.1 Validation Using Heavy Vehicle Simulator Data .......................................................................... 2 1.2 HVS tests.......................................................................................................................... ...................... 4 1.2.1 Goal 1 and Goal 3 Tests................................................................................................................. 4 1.2.2 Goal 5 Tests ............................................................................................................................... ... 6 1.2.3 Goal 9 Tests ............................................................................................................................... ... 7 1.3 Response and Damage Models ............................................................................................................... 8 1.3.1 Asphalt Modulus ............................................................................................................................ 8 1.3.2 Fatigue........................................................................................................................ ................. 21 1.4 Weak Bonding........................................................................................................................ .............. 24 1.5 Unbound Layers ............................................................................................................................... .... 25 1.5.1 Triaxial Tests.......................................................................................................................... ..... 25 1.5.2 Influence of Stiffness of Layers above an Unbound Layer.......................................................... 26 1.5.3 Influence of Load Level ............................................................................................................... 39 1.6 Permanent Deformation ........................................................................................................................ 40 1.6.1 Asphalt ............................................................................................................................... ......... 40 1.6.2 Unbound Materials...................................................................................................................... 47 1.7 Reflection Cracking .............................................................................................................................. 48 2.0 Goal 1 Cracking Test Simulations ........................................................................................................ 50 2.1 Goal 1 Resilient Deformations.............................................................................................................. 50 Stage 5 Distribution xv UCPRC RR 2005 06 2.1.1 Section 501RF Resilient Deflections ( Undrained)....................................................................... 50 2.1.2 Section 503RF Resilient Deflections ( Undrained)....................................................................... 56 2.1.3 Section 500RF Resilient Deflections ( Drained)........................................................................... 60 2.1.4 Section 502CT Resilient Deflections ( Drained)........................................................................... 65 2.2 Visual Cracking Versus Damage of the Top Asphalt Layer, Goal 1 .................................................... 69 2.3 Goal 1 Permanent Deformation............................................................................................................. 72 2.3.1 Section 501RF Permanent Deformations..................................................................................... 73 2.3.2 Section 503RF Permanent Deformations..................................................................................... 75 2.3.3 Section 500RF Permanent Deformations..................................................................................... 77 2.3.4 Section 502CT Permanent Deformations..................................................................................... 79 3.0 Goal 3 Reflection cracking tests............................................................................................................ 81 3.1 Resilient Deflections ............................................................................................................................. 81 3.1.1 Section 517RF DGAC on Section 501RF Resilient Deflections ................................................. 81 3.1.2 Section 518RF ARHM on Section 503RF Resilient Deflections................................................. 86 3.1.3 Section 514RF DGAC on Section 500RF Resilient Deflections ................................................. 91 3.1.4 Section 515RF ARHM on Section 502CT Resilient Deflections ................................................ 96 3.2 Visual Cracking Versus Damage of the Overlay, Goal 3, 20 º C.......................................................... 101 3.3 Permanent Deformation Goal 3, 20 º C................................................................................................. 104 3.3.1 Section 517RF 75 mm DGAC Permanent Deformations .......................................................... 105 3.3.2 Section 518RF 38 mm ARHM Permanent Deformation ........................................................... 107 3.3.3 Section 514RF 75 mm DGAC Permanent Deformations .......................................................... 109 3.3.4 Section 515RF 38 mm ARHM Permanent Deformations ......................................................... 112 4.0 Goal 3 Rutting experiments................................................................................................................. 114 4.1 Section 504RF No Overlay, Wide Base Single Tire........................................................................... 115 4.2 Section 505RF DAGC Overlay, Bias Ply Dual Tire........................................................................... 117 4.3 Section 506RF DGAC Overlay, Radial Dual Tire .............................................................................. 119 4.4 Section 507RF DGAC Overlay, Wide Base Single Tire .................................................................... 121 4.5 Section 508RF ARHM Overlay, Wide Base Single Tire.................................................................... 123 4.6 Section 509RF ARHM Overlay, Radial Dual Tire ............................................................................. 125 4.7 Section 510RF ARHM Overlay, Radial Dual Tire ............................................................................. 127 4.8 Section 511RF ARHM Overlay, Wide Base Single Tire.................................................................... 129 4.9 Section 512RF DGAC Overlay, Wide Base Single Tire .................................................................... 131 4.10 Section 513RF DGAC Overlay, Aircraft Tire..................................................................................... 133 5.0 Goal 5 Wet conditions..................................................................................................................... ... 135 5.1 Section 543RF ARHM Overlay, Drained ........................................................................................... 136 5.2 Section 544RF ARHM Overlay, Undrained ....................................................................................... 145 Stage 5 Distribution xvi UCPRC RR 2005 06 5.3 Section 545RF DGAC Overlay, Undrained ........................................................................................ 153 5.4 Visual Cracking versus Damage of the Top Asphalt Layer, Goal 5 ................................................... 160 6.0 Goal 9 Modified Binder ( MB) road, initial tests ................................................................................. 162 6.1 Materials Characterization .................................................................................................................. 162 6.2 Section 567RF MB Road .................................................................................................................... 169 6.3 Section 568RF MB Road .................................................................................................................... 173 6.4 Section 573RF MB Road .................................................................................................................... 176 6.5 Section 571RF MB Road .................................................................................................................... 179 6.6 Section 572RF MB Road .................................................................................................................... 183 6.7 Section 569RF MB Road .................................................................................................................... 186 6.8 Visual Cracking Versus Damage of the Top Asphalt Layer, Goal 9 .................................................. 190 7.0 Summary and recommendations ......................................................................................................... 191 7.1 Shift Factors and Damage Equations Used in Simulations ................................................................. 191 7.2 Response Model.......................................................................................................................... ....... 192 7.3 Damage of Asphalt Materials.............................................................................................................. 197 7.4 Permanent Deformation of Asphalt..................................................................................................... 199 7.5 Permanent Deformation of Granular Layers ....................................................................................... 201 7.6 Permanent Deformation of Subgrade.................................................................................................. 203 7.7 Total Permanent Deformation at Pavement Surface ........................................................................... 205 7.8 Recommendations ............................................................................................................................... 207 8.0 References..................................................................................................................... ..................... 209 9.0 Appendix....................................................................................................................... ..................... 212 9.1 Glossary ............................................................................................................................... .............. 212 9.2 List of Units ............................................................................................................................... ........ 214 9.3 List of Parameters in Equations .......................................................................................................... 214 9.4 Parameter Values Used in Simulations ............................................................................................... 216 9.5 Section 569RF Simulated with a CTB Model from an HVS Nordic Experiment............................... 217 Stage 5 Distribution xvii UCPRC RR 2005 06 LIST OF FIGURES Figure ES 1. Ration of final to initial deflection .........................……………………………………………... vii Figure ES 2. Cracking versus increase in deflection ..........................……………………………………...… viii Figure ES 3. Measured and predicted final permanent deflection of asphalt ........................... .....…………...... ix Figure ES 4. Final permanent deformation of granular layers .............................……………………………… x Figure ES 5. Final permanent deformation of the subgrade .............................……………………………...… xi Figure ES 6. Final permanent deformation at the pavement surface............................……………………..… xii Figure 1. Layout of 20° C test sections. Goal 3 rutting sections are distributed in the area between the 20° C test sections. ............................................................................................................................... ....................... 4 Figure 2. Drip watering system for Goal 5 tests. ................................................................................................. 7 Figure 3. Layout of Goal 9 test sections. ............................................................................................................. 9 Figure 4. Example of modulus versus reduced time relationship. ..................................................................... 10 Figure 5. Modulus versus temperature for different viscosity versus temperature relationships....................... 12 Figure 6. Frequency sweep data for Goal 1 and Goal 3 materials compared to models. ................................... 13 Figure 7. Example of input parameters for the modulus versus reduced time relationship for the AC bottom layer. 21 Figure 8. Example of damage versus number of load applications. .................................................................. 23 Figure 9. Example of Equation 6 damage parameters of AC bottom layer ( Goal 1)......................................... 24 Figure 10. Simple Drucker Prager failure condition.......................................................................................... 28 Figure 11. EAC 10,000 MPa, no slip, 40 kN load. .............................................................................................. 29 Figure 12. EAC 10,000 MPa, slip, 40 kN............................................................................................................ 30 Figure 13. EAC 2,000 MPa, slip, 40 kN.............................................................................................................. 31 Figure 14. EAC 2,000 MPa, slip, 100 kN load. ................................................................................................... 33 Figure 15. Displacement field in particulate sample........................................................................................... 34 Figure 16. Displacement field in elastic solid ( FEM). ........................................................................................ 35 Figure 17. Modulus of Layer 2 as a function of the stiffness of the asphalt layers, for the undrained sections. 36 Figure 18. Modulus of Layer 2 as a function of the stiffness of the asphalt layers, for the drained sections. ... 37 Figure 19. Modulus of subgrade as a function of the stiffness of the pavement layers. .................................... 37 Figure 20. Results of RSST CH tests. [ Note: FMFC indicates field mixed field compacted specimen taken by coring the pavement. AV5.5 indicates cores with approximately 5.5 percent air voids. Each title in the legend indicates the RSST CH test temperature ( 40, 50, or 60 ° C) and average shear stress ( MPa).]...... 42 Figure 21. Normalized plastic strain versus number of load repetitions. ( Note: legend is the same as in Figure 20). 44 Figure 22. Average values for Figure 21 curves................................................................................................ 45 Figure 23. Best fitting Gamma function. ( Note: legend is the same as in Figure 20 and Figure 21. ................. 46 Stage 5 Distribution xviii UCPRC RR 2005 06 Figure 24. Input parameters for permanent deformation ( rutting) of subgrade in second column. ................... 48 Figure 25. Comparison of fitted vs. calculated strain for AC on AC overlay, 2D. ........................................... 49 Figure 26. Section 501RF pavement structure................................................................................................... 51 Figure 27. Section 501RF temperatures during testing...................................................................................... 52 Figure 28. Section 501RF 40 kN top modules deflection.................................................................................. 52 Figure 29. Section 501 RF 40 kN resilient compression of pavement layers. ................................................... 53 Figure 30. Section 501RF 40 kN deflection of subgrade................................................................................... 53 Figure 31. Section 501RF 100 kN deflection of top modules............................................................................ 54 Figure 32. Section 501RF 100 kN resilient compression of pavement layers. .................................................. 54 Figure 33. Section 501RF 100 kN deflection of subgrade................................................................................. 55 Figure 34. Section 501RF calculated moduli at 40 kN and actual temperature. ( Note: in this and all other figures showing change in elastic moduli ( E) under loading the lines are plotted for the modulus of each layer, i. e., E1 is the modulus of the first layer, E2 is the modulus of the second layer, etc.)..................... 55 Figure 35. Section 503RF pavement structure................................................................................................... 56 Figure 36. Section 503RF temperatures during testing...................................................................................... 56 Figure 37. Section 503RF 40 kN deflection of top modules.............................................................................. 57 Figure 38. Section 503RF 40 kN compression of pavement layers. .................................................................. 57 Figure 39. Section 503RF 40 kN deflection of subgrade................................................................................... 58 Figure 40. Section 503RF 100 kN deflection of top modules............................................................................ 58 Figure 41. Section 503RF 100 kN resilient compression of pavement layers. .................................................. 59 Figure 42. Section 503RF 100 kN deflection of subgrade................................................................................. 59 Figure 43. Section 503RF calculated layer moduli at 40 kN and actual temperature........................................ 60 Figure 44. Section 500RF pavement structure................................................................................................... 60 Figure 45. Section 500RF temperatures during testing...................................................................................... 61 Figure 46. Section 500RF 40 kN deflection of top modules.............................................................................. 61 Figure 47. Section 500RF 40 kN resilient compression of pavement layers. .................................................... 62 Figure 48. Section 500RF 40 kN deflection of subgrade................................................................................... 62 Figure 49. Section 500RF 100 kN deflection of top modules............................................................................ 63 Figure 50. Section 500RF 100 kN compression of pavement layers. ................................................................ 63 Figure 51. Section 500RF 100 kN deflection of subgrade................................................................................. 64 Figure 52. Section 500RF calculated moduli at 40 kN and actual temperature................................................. 64 Figure 53. Section 502CT pavement structure................................................................................................... 65 Figure 54. Section 502CT 40 kN deflection on top of AC. ............................................................................... 65 Figure 55. Section 502CT 40 kN compression of pavement layers................................................................... 66 Figure 56. Section 502CT 40 kN deflection of subgrade................................................................................... 66 Figure 57. Section 502CT 100 kN deflection at top of AC. .............................................................................. 67 Stage 5 Distribution xix UCPRC RR 2005 06 Figure 58. Section 502CT 100 kN resilient compression of pavement layers. .................................................. 67 Figure 59. Section 502CT 100 kN deflection of subgrade................................................................................. 68 Figure 60. Section 502CT calculated moduli at 40 kN and 20 º C. ..................................................................... 68 Figure 61. Cracking versus relative decrease in modulus of top AC layer for Goal 1....................................... 69 Figure 62. Goal 1, cracking versus increase in deflection. ................................................................................ 70 Figure 63. Permanent compression of AC layers. ............................................................................................. 73 Figure 64. Permanent compression of granular layers....................................................................................... 73 Figure 65. Permanent deformation of subgrade................................................................................................. 74 Figure 66. Permanent deformation at pavement surface.................................................................................... 74 Figure 67. Permanent compression of AC layers. ............................................................................................. 75 Figure 68. Permanent compression of granular layers....................................................................................... 75 Figure 69. Permanent deformation of subgrade................................................................................................. 76 Figure 70. Permanent deformation at pavement surface.................................................................................... 76 Figure 71. Permanent deformation of the AC layers. ........................................................................................ 77 Figure 72. Permanent compression of the granular layers................................................................................. 77 Figure 73. Permanent deformation of the subgrade. .......................................................................................... 78 Figure 74. Permanent deformation at pavement surface.................................................................................... 78 Figure 75. Permanent compression of the AC layers......................................................................................... 79 Figure 76. Permanent deformation of granular layers. ...................................................................................... 79 Figure 77. Permanent deformation of subgrade................................................................................................. 80 Figure 78. Permanent deformation at surface of pavement. .............................................................................. 80 Figure 79. Section 517RF pavement structure................................................................................................... 82 Figure 80. Section 517RF AC temperature during testing................................................................................. 82 Figure 81. Section 517RF 40 kN deflection of top modules.............................................................................. 83 Figure 82. Section 517RF 40 kN compression of pavement layers. .................................................................. 83 Figure 83. Section 517RF deflection of subgrade.............................................................................................. 84 Figure 84. Section 517RF 100 kN top modules deflection................................................................................ 84 Figure 85. Section 517RF 100 kN compression of pavement layers. ................................................................ 85 Figure 86. Section 517RF 100 kN deflection of subgrade................................................................................. 85 Figure 87. Section 517RF calculated moduli at 40 kN and actual temperature................................................. 86 Figure 88. Section 518RF pavement structure................................................................................................... 86 Figure 89. Section 518RF AC temperature during testing................................................................................. 87 Figure 90. Section 518RF 40 kN top modules deflection.................................................................................. 87 Figure 91. Section 518RF 40 kN resilient compression of pavement layers. .................................................... 88 Figure 92. Section 518RF 40 kN deflection of subgrade................................................................................... 88 Figure 93. Section 518RF 100 kN top modules deflection................................................................................ 89 Stage 5 Distribution xx UCPRC RR 2005 06 Figure 94. Section 518RF 100 kN compression of pavement layers. ................................................................ 89 Figure 95. Section 518RF 100 kN deflection of subgrade................................................................................. 90 Figure 96. Section 518RF calculated moduli at 40 kN and actual temperature................................................. 90 Figure 97. Section 514RF pavement structure................................................................................................... 91 Figure 98. Section 514RF AC temperature during testing................................................................................. 91 Figure 99. Section 514RF 40 kN top modules deflection.................................................................................. 92 Figure 100. Section 514RF 40 kN compression of pavement layers, MDD1 and MDD2................................. 92 Figure 101. Section 514RF 40 kN compression of pavement layers, MDD3 and MDD4................................. 93 Figure 102. Section 514RF 40 kN deflection of subgrade................................................................................. 93 Figure 103. Section 514RF 100 kN top modules deflection.............................................................................. 94 Figure 104. Section 514RF 100 kN compression of pavement layers, MDD1 and MDD2............................... 94 Figure 105. Section 514RF 100 kN compression of pavement layers, MDD3 and MDD4............................... 95 Figure 106. Section 514RF 100 kN deflection of subgrade............................................................................... 95 Figure 107. Section 514RF calculated moduli at 40 kN and actual temperature............................................... 96 Figure 108. Section 515RF pavement structure................................................................................................. 96 Figure 109. Section 515RF AC temperature during testing............................................................................... 97 Figure 110. Section 515RF 40 kN top modules deflection................................................................................ 97 Figure 111. Section 515RF 40 kN resilient compression of pavement layers. .................................................. 98 Figure 112. Section 515RF 40 kN deflection of subgrade.................................................................................. 98 Figure 113. Section 515RF 100 kN top modules deflection.............................................................................. 99 Figure 114. Section 515RF 100 kN compression of pavement layers. .............................................................. 99 Figure 115. Section 515RF 100 kN deflection of subgrade............................................................................. 100 Figure 116. Section 515RF calculated moduli at 40 kN and actual temperatures. .......................................... 100 Figure 117. Cracking in overlay versus relative decrease in modulus of overlay, Goal 3............................... 101 Figure 118. Goal 3, 20 º C, cracking versus increase in deflection.................................................................... 102 Figure 119. Section 517RF permanent deformation of AC layers................................................................... 105 Figure 120. Section 517RF permanent deformation of granular layers........................................................... 105 Figure 121. Section 517RF permanent deformation of subgrade. ................................................................... 106 Figure 122. Section 517RF permanent deformation at pavement surface. ...................................................... 106 Figure 123. Section 518RF permanent deformation of AC layers................................................................... 107 Figure 124. Section 518RF permanent deformation of granular layers........................................................... 107 Figure 125. Section 518RF permanent deformation of subgrade. ................................................................... 108 Figure 126. Section 518RF permanent deformation at pavement surface. ...................................................... 108 Figure 127. Section 514RF permanent deformation of AC layers................................................................... 109 Figure 128. Section 514RF permanent deformation of granular layers, MDD1 and MDD2........................... 109 Figure 129. Section 514RF permanent deformation of granular layers, MDD3 and MDD4........................... 110 Stage 5 Distribution xxi UCPRC RR 2005 06 Figure 130. Section 514RF permanent deformation of subgrade. ................................................................... 110 Figure 131. Section 514RF permanent deformation at pavement surface. ...................................................... 111 Figure 132. Section 515RF permanent deformation of AC layers................................................................... 112 Figure 133. Section 515RF permanent deformation of granular layers........................................................... 112 Figure 134. Section 515RF permanent deformation of subgrade. ................................................................... 113 Figure 135. Section 515RF permanent deformation of the pavement surface................................................. 113 Figure 136. Section 504RF pavement structure............................................................................................... 115 Figure 137. Section 504RF permanent deformation at pavement surface from profilometer.......................... 115 Figure 138. Section 504RF calculated permanent deformation of pavement layers........................................ 116 Figure 139. Section 505RF pavement structure............................................................................................... 117 Figure 140. Section 505RF temperatures during testing.................................................................................. 117 Figure 141. Section 505RF permanent deformation at pavement surface from profilometer.......................... 118 Figure 142. Section 505RF calculated permanent deformation of pavement layers........................................ 118 Figure 143. Section 506RF pavement structure............................................................................................... 119 Figure 144. Section 506RF temperatures during testing.................................................................................. 119 Figure 145. Section 506RF permanent deformation at pavement surface from profilometer.......................... 120 Figure 146. Section 506RF calculated permanent deformation of pavement layers........................................ 120 Figure 147. Section 507RF pavement structure............................................................................................... 121 Figure 148. Section 507RF temperatures during testing.................................................................................. 121 Figure 149. Section 507RF permanent deformation at pavement surface from profilometer.......................... 122 Figure 150. Section 507RF calculated permanent deformation of pavement layers........................................ 122 Figure 151. Section 508RF pavement structure............................................................................................... 123 Figure 152. Section 508RF temperatures during testing.................................................................................. 123 Figure 153. Section 508RF permanent deformation at pavement surface from profilometer.......................... 124 Figure 154. Section 508RF calculated permanent deformation of pavement layers........................................ 124 Figure 155. Section 509RF pavement structure............................................................................................... 125 Figure 156. Section 509RF temperatures during testing.................................................................................. 125 Figure 157. Section 509RF permanent deformation at pavement surface from profilometer.......................... 126 Figure 158. Section 509RF calculated permanent deformation of pavement layers........................................ 126 Figure 159. Section 510RF pavement structure............................................................................................... 127 Figure 160. Section 510RF temperatures during testing.................................................................................. 127 Figure 161. Section 510RF permanent deformation at pavement surface from profilometer.......................... 128 Figure 162. Section 510RF calculated permanent deformation of pavement layers........................................ 128 Figure 163. Section 511RF pavement structure............................................................................................... 129 Figure 164. Section 511RF temperatures during testing.................................................................................. 129 Figure 165. Section 511RF permanent deformation at the pavement surface. ................................................ 130 Stage 5 Distribution xxii UCPRC RR 2005 06 Figure 166. Section 511RF calculated permanent deformation of the pavement layers.................................. 130 Figure 167. Section 512RF pavement structure............................................................................................... 131 Figure 168. Section 512RF temperatures during testing.................................................................................. 131 Figure 169. Section 512RF permanent deformation at pavement surface from profilometer.......................... 132 Figure 170. Section 512RF calculated permanent deformation of the pavement layers.................................. 132 Figure 171. Section 513RF pavement structure............................................................................................... 133 Figure 172. Section 513RF permanent deformation at pavement surface from profilometer.......................... 134 Figure 173. Section 513RF calculated permanent deformation of pavement layers........................................ 134 Figure 174. Cores from trafficked area of Section 543RF after HVS loading show stripping and disintegration of ATPB, as well as signs of moisture damage between the the three lifts of asphalt concrete ( Bejarano et al. 2003). ............................................................................................................................... .................. 135 Figure 175. Section 543RF pavement structure............................................................................................... 136 Figure 176. Section 543RF temperatures during testing.................................................................................. 137 Figure 177. Section 543RF Road Surface Deflectometer, at 40 kN. ............................................................... 137 Figure 178. Section 543RF Road Surface Deflectometer, at 100 kN. ............................................................. 138 Figure 179. Section 543RF 40 kN top module. ............................................................................................... 138 Figure 180. Section 543RF 40 kN top of aggregate base. ............................................................................... 139 Figure 181. Section 543RF 40 kN top of aggregate subbase........................................................................... 139 Figure 182. Section 543RF 40 kN deflection of subgrade ( 850 mm depth). .................................................... 140 Figure 183. Section 543RF 100 kN top module. ............................................................................................. 140 Figure 184. Section 543RF 100 kN top of aggregate base. ............................................................................. 141 Figure 185. Section 543RF 100 kN top of aggregate subbase......................................................................... 141 Figure 186. Section 543RF 100 kN deflection of subgrade ( 850 mm depth). .................................................. 142 Figure 187. Section 543RF Permanent deformation of asphalt layers............................................................. 142 Figure 188. Section 543RF permanent deformation of granular layers plus top of subgrade.......................... 143 Figure 189. Section 543RF permanent deformation in subgrade ( 850 mm depth)........................................... 143 Figure 190. Section 543RF permanent deformation at pavement surface. ...................................................... 144 Figure 191. Section 543RF calculated layer moduli, at 40 kN and actual temperatures. ................................ 144 Figure 192. Section 544RF pavement structure............................................................................................... 145 Figure 193. Section 544RF temperatures during testing.................................................................................. 145 Figure 194. Section 544RF Road Surface Deflectometer, at 40 kN. ............................................................... 146 Figure 195. Section 544RF Road Surface Deflectometer, at 100 kN. ............................................................. 146 Figure 196. Section 544RF 40 kN top module. ............................................................................................... 147 Figure 197. Section 544RF 40 kN top of aggregate base. ............................................................................... 147 Figure 198. Section 544RF 40 kN top of aggregate subbase........................................................................... 148 Figure 199. Section 544RF 100 kN top module. ............................................................................................. 148 Stage 5 Distribution xxiii UCPRC RR 2005 06 Figure 200. Section 544RF 100 kN top of aggregate base. ............................................................................. 149 Figure 201. Section 544RF 100 kN top of aggregate subbase......................................................................... 149 Figure 202. Section 544RF Permanent deformation of asphalt layers............................................................. 150 Figure 203. Section 544RF permanent deformation of aggregate base. .......................................................... 150 Figure 204. Section 544RF permanent deformation on top of basecourse. ...................................................... 151 Figure 205. Section 544RF permanent deformation at pavement surface. ...................................................... 151 Figure 206. Section 544RF calculated moduli of pavement layers................................................................... 152 Figure 207. Section 545RF pavement structure............................................................................................... 153 Figure 208. Section 545RF temperatures during testing.................................................................................. 153 Figure 209. Section 545RF Road Surface Deflectometer, at 40 kN. ............................................................... 154 Figure 210. Section 545RF Road Surface Deflectometer, 100 kN. ................................................................. 154 Figure 211. Section 545RF 40 kN top module. ............................................................................................... 155 Figure 212. Section 545RF 40 kN top of aggregate base. ............................................................................... 155 Figure 213. Section 545RF 40 kN top of aggregate subbase........................................................................... 156 Figure 214. Section 545RF 100 kN top module. ............................................................................................. 156 Figure 215. Section 545RF 100 kN top of aggregate base. ............................................................................. 157 Figure 216. Section 545RF 100 kN top of aggregate subbase......................................................................... 157 Figure 217. Section 545RF permanent deformation of asphalt layers............................................................. 158 Figure 218. Section 545RF permanent deformation of aggregate base. .......................................................... 158 Figure 219. Section 545RF permanent deformation at pavement surface. ...................................................... 159 Figure 220. Section 545RF calculated moduli of pavement layers.................................................................. 159 Figure 221. Visual cracking versus relative decrease in modulus of layer 1, Goal 5 Wet conditions. ............. 160 Figure 222. Goal 5, cracking versus increase in deflection. ............................................................................ 161 Figure 223. MB road, AC modulus versus reduced time parameters from frequency sweep. ......................... 162 Figure 224. Moduli from FWD compared to frequency sweep tests, Goal 9 ( MB road). ................................ 163 Figure 225. MB road, damage parameters for AC in first column. .................................................................. 165 Figure 226. MB road, backcalculated modulus of AB versus time. ................................................................ 166 Figure 227. MB road, modulus of AB versus stiffness of AC......................................................................... 166 Figure 228. MB road, subgrade modulus versus stiffness of pavement layers. ................................................ 167 Figure 229. Section 567RF pavement structure................................................................................................ 169 Figure 230. Section 567RF load levels. ........................................................................................................... 169 Figure 231. Section 567RF temperatures during testing................................................................................... 170 Figure 232. Section 567RF Road Surface Deflectometer................................................................................. 170 Figure 233. Section 567RF MDDs at 90mm and 330 mm. ............................................................................. 171 Figure 234. Section 567RF permanent deformation of MDDs......................................................................... 171 Figure 235. Section 567RF permanent deformation at pavement surface from profilometer........................... 172 Stage 5 Distribution xxiv UCPRC RR 2005 06 Figure 236. Section 567RF calculated moduli at 40 kN and actual temperature.............................................. 172 Figure 237. Section 568RF pavement structure................................................................................................ 173 Figure 238. Section 568RF temperatures during testing................................................................................... 173 Figure 239. Section 568RF Road Surface Deflectometer................................................................................. 174 Figure 240. Section 568RF permanent deformation at pavement surface from profilometer........................... 174 Figure 241. Section 568RF calculated layer moduli at 40 kN and actual temperature..................................... 175 Figure 242. Section 573RF pavement structure................................................................................................ 176 Figure 243. Section 573RF temperatures during testing................................................................................... 176 Figure 244. Section 573RF Road Surface Deflectometer................................................................................. 177 Figure 245. Section 573RF permanent deformation at pavement surface from profilometer........................... 177 Figure 246. Section 573RF calculated layer moduli at 40 kN and actual temperature..................................... 178 Figure 247. Section 571RF pavement structure................................................................................................ 179 Figure 248. Section 571RF temperatures during testing................................................................................... 179 Figure 249. Section 571RF Road Surface Deflectometer................................................................................. 180 Figure 250. Section 571RF MDDs at 90 mm, 300 mm, and 525 mm. ............................................................. 180 Figure 251. Section 571RF permanent deformation of MDDs......................................................................... 181 Figure 252. Section 571RF permanent deformation at the pavement surface. ................................................. 181 Figure 253. Section 571RF calculated layer moduli at 40 kN and actual temperature..................................... 182 Figure 254. Section 572RF pavement structure................................................................................................ 183 Figure 255. Section 572RF temperatures during testing................................................................................... 183 Figure 256. Section 572RF Road Surface Deflectometer................................................................................. 184 Figure 257. Section 572RF permanent deformation at the pavement surface. ................................................. 184 Figure 258. Section 572RF calculated layer moduli at 40 kN and actual temperature..................................... 185 Figure 259. Section 569RF pavement structure................................................................................................ 186 Figure 260. Section 569RF temperatures during testing................................................................................... 186 Figure 261. Section 569RF Road Surface Deflectometer................................................................................. 187 Figure 262. Section 569RF MDDs at 90 mm, 300 mm, and 525 mm.............................................................. 187 Figure 263. Section 569RF permanent deformation of MDDs......................................................................... 188 Figure 264. Section 569RF permanent deformation at pavement surface from profilometer........................... 188 Figure 265. Section 569RF calculated layer moduli at 40 kN and actual temperature..................................... 189 Figure 266. Cracking versus relative decrease in modulus of AC layer for Goal 9 ( MB road)........................ 190 Figure 267. Goal 9 ( MB road), Cracking versus increase in deflection ( RSD) ................................................ 190 Figure 268. Ratio of final over initial deflection. ............................................................................................. 195 Figure 269. Cracking versus calculated decrease in modulus of top layer. ...................................................... 197 Figure 270. Cracking versus increase in deflection. ......................................................................................... 198 Figure 271. Measured and predicted final permanent deformation of asphalt................................................. 199 Stage 5 Distribution xxv UCPRC RR 2005 06 Figure 272. Final permanent deformation of granular layers. ......................................................................... 201 Figure 273. Final permanent deformation of the subgrade. .............................................................................. 203 Figure 274. Final permanent deformation at the pavement surface ( profile data). ........................................... 205 Figure 275. Modulus of AB layer backcalculated from FWD tests in center line. ........................................... 217 Figure 276. Pavement structure for Section 569RF.......................................................................................... 219 Figure 277. Damage parameters used for DGAC of Section 569RF................................................................ 219 Figure 278. Damage parameters used for CTB of Section 569RF.................................................................... 220 Figure 279. Section 569RF Road Surface Deflectometer................................................................................. 220 Figure 280. Section 569RF MDD resilient deflections..................................................................................... 221 Figure 281. Section 569RF permanent MDD deformations. ............................................................................ 221 Figure 282. Section 569RF average permanent deformation from pavement profile....................................... 222 LIST OF TABLES Table 1. Summary List of HVS Tests ................................................................................................................... 3 Table 2. Design Thicknesses for Goal 1 Sections ................................................................................................. 5 Table 3. As built Thicknesses for Goal 5 Sections ............................................................................................... 6 Table 4. As built Thicknesses of Goal 9 Sections ................................................................................................ 8 Table 5 Influence of Slip Value in LEAP on Calculated Vertical Deflections and Horizontal Strains ............. 25 Table 6. Triaxial Tests on Subgrade.................................................................................................................... 26 Table 7. Triaxial Tests on Aggregate Base ( AB) ................................................................................................ 26 Table 8. Layer Thicknesses Used for FWD Backcalculation.............................................................................. 36 Table 9. Moduli Parameters from FWD ............................................................................................................. 38 Table 10. Moduli Parameters from Calibration to MDD Deflections................................................................. 38 Table 11. Initial Moduli Used in HVS Simulations ( MPa)................................................................................. 39 Table 12. Summary of Moduli ( MPa)................................................................................................................. 39 Table 13. Parameters Used in Equation 20......................................................................................................... 41 Table 14. AC on AC, 2D Structural Parameter Combinations ........................................................................... 49 Table 15. Initial Moduli MPa ( 1.8 km/ h, 40 kN, Actual Temperature) for Each Section................................... 71 Table 16. Final Moduli MPa ( 1.8 km/ h, 40 kN, Actual Temperature) for Each Section.................................... 71 Table 17. Percentage Decrease in Layer Moduli for Each Section..................................................................... 72 Table 18. Damage Parameter for Asphalt Layers at End of Test for Each Section ............................................ 72 Table 19. Initial Damage Parameters for “ Old” Asphalt Overlay Sections ........................................................ 81 Table 20. Layered Moduli at Start of Test, MPa .............................................................................................. 103 Table 21. Layer Moduli at End of Test, MPa ................................................................................................... 103 Table 22. Percentage Decrease in Moduli ........................................................................................................ 104 Table 23. Initial Permanent Deformations, Goal 3, in mm............................................................................... 104 Stage 5 Distribution xxvi UCPRC RR 2005 06 Table 24. Tire Types and Pressure, MPa .......................................................................................................... 114 Table 25. Estimated Temperatures for Section 513RF ..................................................................................... 133 Table 26. Parameters for FWD Moduli versus Reduced Time, Goal 9 ............................................................ 164 Table 27. MB Road, Stiffness Parameters for AB............................................................................................ 167 Table 28. Summary of Damage Equations and Shift Factors Used in All Simulations.................................... 192 Table 29. Measured and Calculated Road Surface Deflectometer Deflections ( RSD), in mm......................... 193 Table 30. Measured and Calculated Deflections of the Top Multi depth Deflectometer ( MDD), in mm........ 194 Table 31. Final Permanent Deformation of Asphalt, in mm............................................................................. 200 Table 32. Final Permanent Deformation of Granular Layer, in mm................................................................. 202 Table 33. Parameters used for Granular Materials in Equation ( 29) ................................................................ 202 Table 34. Final Permanent Deformation of Subgrade, in mm.......................................................................... 204 Table 35. Parameters Used for Subgrade in Equation ( 29)............................................................................... 204 Table 36. Final Permanent Deformation at the Pavement Surface, in mm....................................................... 206 Stage 5 Distribution 1 UCPRC RR 2005 06 1.0 INTRODUCTION The first step in creating a Mechanistic Empirical ( ME) pavement design or evaluation is to calculate pavement response — in terms of stresses, strains, and/ or displacements — using a mathematical ( or mechanistic) model. In the second step, the calculated response is used as a variable in empirical relationships to predict structural damage ( decrease in moduli or cracking) and functional damage ( rutting and roughness) to the pavement. Both of these steps must be reasonably correct. If the calculated response bears little resemblance to the pavement’s actual response, there is no point in trying to use the calculation to predict future damage to the pavement with the empirical relationship. In other words, only if the calculated response is reasonably correct does it make sense to try to relate the damage to the pavement response. 1.1 Models and Approaches Included in CalME This report presents the modeling of several series of flexible pavement Heavy Vehicle Simulator ( HVS) tests using the set of distress models included in the draft software package, CalME. These models are for the flexible pavement distresses typically observed in California: asphalt fatigue, asphalt rutting, unbound layers rutting and reflection cracking. CalME software provides the user with four approaches for evaluating or designing a flexible pavement structure: • Caltrans current methods, the R value method for new flexible structures, and the deflection reduction method for overlay thickness design for existing flexible structures. • “ Classical” Mechanistic Empirical ( ME) design, largely based on the Asphalt Institute method. This method uses a standard Equivalent Single Axle Load ( ESAL) for the traffic load, one temperature to characterize the entire range of temperatures the asphalt concrete ( AC) layer will experience, and the Asphalt Institute fatigue and unbound layers rutting equations, with an adjustment for air void content and binder content in the asphalt concrete. • An Incremental method, using the typical Miner’s Law approach, permitting damage calculation for the axle load spectrum and expected temperature regimes, but with no updating of materials properties through the life of the project. This is similar to the approach included in the NCHRP 1 37A Design Guide, also referred to as the Mechanistic Empirical Pavement Design Guide ( MEPDG). This type of approach is calibrated against an end failure state, such as 25 percent cracking of the wheelpath, and it assumes a linear accumulation of damage to get to that state. • An Incremental Recursive method in which the materials properties for the pavement are updated in terms of damage as the simulation of the pavement life progresses. The Incremental Recursive approach was used for the simulations included in this report, and is the only approach that can provide an accurate indication of pavement condition at different points during the pavement’s life. The research team proposes that pavement designers should begin their designs by applying either an existing Caltrans method or the Classical method. In CalME both of these options perform a “ design” function, calculating and presenting pavement structures that meet design requirements for a predetermined number of traffic loads. Then, the lowest cost alternatives in the set of candidate pavement structures meeting the design requirements with either of these methods should be checked by the designer with the more comprehensive and precise Incremental Recursive method to be certain that they meet the design requirements. Once a final design has been selected, its Incremental Recursive output can be used to provide a prediction of the pavement’s condition across its entire life. Some distresses and some materials are not considered in either the Caltrans or Classical methods, and can only be evaluated using the Incremental Miner’s Law approach or the Incremental Recursive approach. Stage 5 Distribution 2 UCPRC RR 2005 06 1.1.1 Validation Using Heavy Vehicle Simulator Data The Incremental Recursive models included in CalME were used to predict the performance of all the flexible pavement HVS tests performed to date as part of the Accelerated Pavement Testing ( APT) program operated for the California Department of Transportation ( Caltrans) by the University of California Pavement Research Center ( UCPRC). The HVS test data presented in this report come from tests performed between 1995 and 2004. The HVS response data and the corresponding laboratory test data were extracted from the UCPRC HVS database. HVS tests measure pavement response in terms of deflections, either at the pavement surface, using a Road Surface Deflectometer ( RSD), at multiple depths, using a Multi depth Deflectometer ( MDD), or both. The RSD is very similar to the Benkelman Beam used in the development of the Caltrans new flexible pavement and overlay design methods in the 1950s. In predicting the gradual degradation of the pavement it is important that the response model provides a reasonably accurate prediction of measured deflections. Although a correct prediction of deflections by the response model is no guarantee that it can also correctly predict the stresses and strains in all of the pavement layers, the opposite is true: if the model inaccurately predicts deflections, it will also provide inaccurate predictions of stresses and strains. Therefore, in trying to calibrate the ME models from HVS testing, the research team’s first concern was to make sure that the model predicted resilient deflections reasonably well for the duration of the test and for all load levels. This prediction depended on the moduli ( often referred to as “ stiffnesses” in this report and in the literature) of all of the pavement layers and on the changes to these moduli caused by fatigue damage, slip between asphalt layers, non linear elastic characteristics of the unbound layers, and the effect of confinement on granular layers. Once reasonable agreement was achieved between the measured resilient deflections and the calculated ones, then models of permanent deformation could be calibrated with some confidence. There are different methods for determining moduli and there are often differences in the results from each method ( which should be expected based on the literature.) The methods used in this study included backcalculation of moduli for all layers from Falling Weight Deflectometer ( FWD) and MDD deflection data , and direct measurement of moduli in the laboratory using triaxial tests for unbound materials and flexural frequency sweep tests for asphaltic materials. Differences in measured moduli across the different methods are due to variations in boundary conditions, strain levels, and loading times between the different measurement methods, the effects of which vary among materials. The FWD does not fit under the HVS, so there is no FWD data during an HVS test, there is only FWD data from before the HVS was placed on the pavement and from after the HVS was taken off the trafficked section. The simulations in this report primarily relied on stiffnesses for the asphalt materials taken from flexural frequency sweep data, and stiffnesses for the unbound layers taken from backcalculation of MDD deflection data. In practice, the research team views backcalculation using deflections from the FWD as the primary tool for obtaining the stiffnesses of layers in existing pavements, as opposed to laboratory testing of materials samples taken from the already constructed pavement. FWD deflections and backcalculation take into consideration the stiffness of the layers as they occur in the constructed pavement structure, including the effects of boundary conditions, water and temperature conditions, previous traffic and environmental conditioning, and interaction between layers acting as a system in the in place pavement structure. This is important because most of Caltrans future work will be rehabilitating and reconstructing pavements already in service. The research team sees the flexural beam test as the primary tool for measuring the stiffness and the fatigue characteristics of asphalt overlay materials for new layers. For new pavement construction, the team sees the use of databases of moduli for granular bases and subbases and for subgrades backcalculated from FWD tests on existing pavements, with the materials referenced by characteristics such as the Unified Soil Classification System ( USCS) classification and relative density. The databases should also include some laboratory triaxial tests for these materials, for comparison with any new base, subbase, and subgrade materials Stage 5 Distribution 3 UCPRC RR 2005 06 for which there is no previous FWD testing, and for which laboratory triaxial testing must be used to measure stiffness. The purpose of this study was to compare the overall trends shown by the damage models in simulations of the HVS tests against the actual trends measured in the same HVS tests. Asphalt concrete stiffnesses from flexural frequency sweep data were used in the simulations of the HVS tests, and the stiffnesses of the underlying moduli were adjusted from their initial values as the asphalt concrete stiffness changed with damage so they would match the measured and calculated deflections. The results presented herein show that, overall, the damage trends for deflection and permanent deformation under loading were verified. During HVS testing, deflections often increase markedly, with deflection sometimes rising more than twice as high at the end of the test than they were at the beginning. The flexible pavement design model of the NCHRP 1 37A Design Guide ( NCHRP 2004) does not consider any decrease in the asphalt concrete modulus as a result of fatigue damage ( except for rehabilitation designs). In fact, the NCHRP 1 37A Design Guide includes a model for aging that predicts a continuous increase in the stiffness of asphalt concrete layers across the life of the pavement, resulting in smaller predicted deflections as the pavement is subjected to trafficking. While aging is potentially important, the effect of updating stiffness for aging and not updating it for fatigue damage results in the calculation of unrealistic elastic responses during the pavement life. This may be acceptable for pavements with extremely thick asphalt concrete layers where little fatigue should occur, but it is impossible to use the model to simulate an HVS test and, inversely, to use HVS tests to calibrate the model. The HVS test series in this report were grouped by “ goals,” which are defined as follows: Table 1. Summary List of HVS Tests Goal General Conditions HVS Test Numbers Original Report References Goal 1: Comparison of structures with and without ATPB layer under dry conditions, moderate temperatures New pavement, dry conditions, 20° C 500RF, 501RF, 502CT, 503RF 14, 15, 16, 17, 18 Goal 3 Cracking: Comparison of reflection cracking performance of ARHM GG and DGAC overlays Overlays of cracked Goal 1 sections, dry conditions, 20° C 514RF, 515RF 517RF, 518RF 8, 11, 13 Goal 3 Rutting: Comparison of rutting performance of ARHM GG and DGAC overlays Overlays of previously untrafficked areas of Goal 1 pavements, dry conditions, 40° C or 50° C at 50 mm depth, four different tire/ wheel types 504RF, 505RF 506RF, 507RF 508RF, 509RF 510RF, 511RF 512RF, 513RF 7, 10 Goal 5: Comparison of structures with and without ATPB layer under wet conditions, moderate temperatures New pavement, wet conditions, 20° C 543RF, 544RF, 545RF 2, 3, 4, 5, 13 Goal 9: Initial cracking of asphalt pavement in preparation for later overlay New pavement, ambient rainfall, 20° C 567RF, 568RF, 569RF, 571RF, 572RF, 573RF 1 ATPB: Asphalt treated permeable base. ARHM GG: Asphalt rubber hot mix gap graded. DGAC: Dense graded asphalt concrete. Stage 5 Distribution 4 UCPRC RR 2005 06 The Goal 1, 3, and 5 tests were performed inside a metal shed built over native subgrade. The shed provided protection from sun, wind, and rain; however, changes in subgrade water content and the depth to the water table were recorded during HVS tests. Goal 9 tests were performed on a road with no cover other than what the HVS and its temperature cabinet provided. The remainder of this chapter presents the general descriptions of the HVS tests and the models used to simulate them. 1.2 HVS tests 1.2.1 Goal 1 and Goal 3 Tests Figure 1 shows the layout of the Goal 1 and Goal 3 cracking sections. Figure 1. Layout of 20° C test sections. Goal 3 rutting sections are distributed in the area between the 20° C test sections. All the sections in Goal 1 had two layers of asphalt concrete ( AC), an aggregate base ( AB), and an aggregate subbase ( ASB). In the two drained sections, part of the AB layer thickness was replaced by an asphalt treated permeable base ( ATPB) at a ratio of 1.4: 1.1. A constant temperature of about 20 º C was maintained during Goal 1. The design layer thicknesses are shown in Table 2. 518RF 517RF 515RF 514RF 503RF 501RF 502CT 500RF Undrained Drained ARHM GG DGAC Goal1 Goal3 Stage 5 Distribution 5 UCPRC RR 2005 06 Table 2. Design Thicknesses for Goal 1 Sections Layer Undrained ( mm) Drained ( mm) AC top lift 61 61 AC bottom lift 76 76 ATPB None 76 AB 274 182 ASB 229 229 The subgrade was clay, with varying plasticity across the pavements. Goal 3 was an overlay study with an aphalt rubber hot mix gap graded ( ARHM GG) ( the acronym ARHM, asphalt rubber hot mix gap graded, refers to the material specification at the time of construction in April 1997.) concrete and a dense graded asphalt concrete ( DGAC). The four Goal 3 Cracking tests were performed at the same temperature as Goal 1 ( 20 º C), with similar wheel types and loads. For these tests the design thickness of the DGAC overlay ( 75 mm) was approximately twice as thick as that of the ARHM ( 40 mm), and the overlays were placed over the four previously cracked Goal 1 sections. The remainder of Goal 3 was done at higher temperatures, with the two overlays placed on previously untested areas of the Goal 1 pavement, and several overlay thicknesses, tire types, and wheel loads used in the testing. ( Table 1 provides the reference numbers for the reports containing the details regarding thickness, tire types, and wheels loads.) All the HVS tests at 20 º C started with a wheel load of 40 kN, which was increased stepwise to 100 kN. Most of the load applications were at 100 kN. Bias ply tires on a dual wheel were used for the Goal 1 tests. The same radial tires on a dual wheel were used for the Goal 3 Cracking tests, and the Goal 5 and Goal 9 tests. Various tires and wheels were used for the Goal 3 Rutting tests, with the load for the entire duration of all but one test fixed at 40 kN. [ De Beer and Fisher ( 1997) describe the details of the tire contact stresses measured.] For all the tests except the Goal 3 Rutting tests, the wheel load was a dual wheel with a centerline distance of 305 mm and an assumed tire pressure of 690 kPa for all load levels. It was assumed that the wheels distributed the load over two circular areas; this assumption was reasonably correct for the low load level ( 40 kN) but not for the high load level ( 100 kN, where the actual load distribution was closer to two rectangles with one side twice the length of the other). All the 20° C sections were instrumented with MDDs. Each section in Goal 1 had two MDDs. Each section had two additional MDDs installed for Goal 3 testing. All MDD anchor depths were assumed to be 3,000 mm. Not all MDD modules functioned for the duration of the tests. The as constructed layer thicknesses given in Harvey et al. ( 1999) were used for the analyses of the Goal 1 and Goal 3 results presented in this report. As built layer thicknesses given in Bejarano et al. ( 2003) and Bejarano et al. ( 2005) were used to analyze the Goal 5 and Goal 9 results respectively. The remaining data was imported from the UC Pavement Research Center database, taken from a subset database named PRCHVS. mdb. Although actual wheel speeds varied, they were assumed to be 7.6 km/ h during HVS testing and 1.8 km/ h during deflection measurement on the MDDs. All tests other than the Goal 3 Rutting tests were performed with bidirectional loading. The Goal 3 Rutting tests were performed with unidirectional loading. The temperature of the test sections was controlled by a “ temperature control box.” The actual temperatures of the asphalt layers were recorded and used in the simulations. Poisson’s ratio was assumed to be 0.35 for all layers. Stage 5 Distribution 6 UCPRC RR 2005 06 1.2.2 Goal 5 Tests The Goal 5 tests were performed on the overlay structures of Goal 3 in locations where the overlay had not been placed on previously trafficked and cracked Goal 1 pavement. The designed structures were: • Section 543RF drained with ATPB and 40 mm ARHM GG wearing surface, • Section 544RF undrained ( no ATPB) and 40 mm ARHM GG wearing surface, and • Section 545RF undrained ( no ATPB) and 75 mm DGAC wearing surface. The structures’ as built thicknesses are shown in Table 3. Table 3. As built Thicknesses for Goal 5 Sections Layer Section 543RF ( mm) Section 544RF ( mm) Section 545RF ( mm) Wearing course 36 51 90 AC ( two lifts combined) 140 149 143 ATPB 64 none none AB 180 272 259 ASB 223 205– 310 206– 280 All of the test sections had a 2 percent transverse gradient and an approximate 0.5 percent longitudinal gradient in all the layers above the subbase. Holes with a diameter of 38 mm were drilled through the asphalt concrete layers on the uphill side of the three HVS test sections, and a drip watering system was installed to continuously put water into the pavement. Holes were drilled into the top of the ATPB layer of Section 543RF so that water entered into that layer. Sections 544RF and 545RF had holes drilled into the top of their aggregate bases, and water entered those layers. The water flow was greater into Section 543RF because of the initial high permeability of its ATPB layer. Considerably less water flowed into the other two sections because of the relative impermeability of their AB layer. Figure 2 shows the drip watering system. Each pavement section had water introduced into it for more than a month prior to HVS loading. This allowed the section to reach an approximate steady state moisture condition. Goal 5 testing used the same dual wheel, radial tire configuration as Goal 3 testing. Similar loading patterns and the same temperature control provided a basis for comparing the results from Goal 3 and Goal 5 tests to the result of Goal 1 ( dry condition) testing. In addition, two MDDs were installed in each pavement section, with depths similar to those used for Goal 1. Stage 5 Distribution 7 UCPRC RR 2005 06 Figure 2. Drip watering system for Goal 5 tests. 1.2.3 Goal 9 Tests Six HVS tests were performed on what were designed to be identical pavement structures. The primary purpose of the six tests was to provide fatigue cracked sections for subsequent placement of different kinds of overlay for HVS reflection cracking tests; these would be similar to the work performed in the Goal 3 cracking tests. The pavements were built so that they aligned with an existing access road. When the existing structure was removed, its subgrade was compacted to state standards. ( Figure 3 shows the layout of the six Stage 5 Distribution 8 UCPRC RR 2005 06 test sections.) The structures’ design thicknesses were: 410 mm of AB with 90 mm of dense graded asphalt concrete ( DGAC). ( Table 4 shows the as built thicknesses.) The dual wheel and radial tire configuration used for the Goal 3 tests was used for Goal 9. Loads were generally lower than those used during Goals 1, 3, and 5 to minimize rutting. Temperatures were maintained close to 20° C, as on the other cracking tests. Two MDDs were placed in each test section. However, most of the MDD modules below the surface either never functioned or failed during the test. Table 4. As built Thicknesses of Goal 9 Sections Layer 567RF 568RF 569RF 571RF 572RF 573RF AC 78 80 81 82 78 76 AB 352 349 337 352 349 337 1.3 Response and Damage Models Response and damage models are presented in this section, and they are discussed in terms of the materials properties common to the Goal 1, 3, and 5 tests. The materials properties of the Goal 9 tests are discussed with the description of the simulations for those tests later in this report. 1.3.1 Asphalt Modulus Asphalt modulus was determined as a function of temperature and loading time, using the NCHRP 1 37A Design Guide model ( NCHRP 2004): ( ) ( ( tr)) E 1 exp log log β γ α δ + + = + ( 1) where E is the modulus in MPa, tr is reduced time in sec, α, β, γ, and δ are constants, and logarithms are to base 10. Reduced time is found from: aTg ref visc visc lt tr ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ = × ( 2) where lt is the loading time ( in sec), viscref is the binder viscosity at the reference temperature, visc is the binder viscosity at the present temperature, and aTg is a constant. Stage 5 Distribution 9 UCPRC RR 2005 06 Figure 3. Layout of Goal 9 test sections. Stage 5 Distribution 10 UCPRC RR 2005 06 100 1000 10000 100000 0.00001 0.0001 0.001 0.01 0.1 1 10 100 1000 10000 Reduced time, sec Modulus, MPa α/ 2 tr ( ) min δ= log E max min log E E ⎛⎜ ⎞⎟ ⎝ ⎠ α = Figure 4. Example of modulus versus reduced time relationship. Stage 5 Distribution 11 UCPRC RR 2005 06 Equation 1 may also be written: ( ) ( ) max min ' min ' / 2 / 2 log log log 1 1 E E E E tr tr tr tr γ γ α α α δ ⎛⎜ ⎞⎟ = + = + ⎝ ⎠ +⎛⎜ ⎞⎟ +⎛⎜ ⎞⎟ ⎝ ⎠ ⎝ ⎠ ( 3) where ( ) ( ( )) / 2 ' ln 10 and tr exp ln 10 α γ γ βγ = = − × trα/ 2 is the reduced time corresponding to log( E) = δ + α/ 2, as indicated in Figure 4. Equation 1 is normally used with frequency sweep data to characterize the master curve. The form of the master curve equation shown in Equation 2 provides some insight, and can be used if Emax and Emin were known. Emax is related to the limiting stiffness of asphalt binder at temperatures below the glass transition temperature. Emin appears to depend on the boundary conditions under which it was measured, with different values coming from backcalculation of in situ pavements and beam fatigue tests. From Equation 2 it can be seen that changing trα/ 2 will shift the curve left or right and changing γ’ will change the curvature. The loading time is determined from the speed of the wheel ( input on the incremental design screen in CalME) and from the depth at which the loading time is desired. Loading time is a rather uncertain notion, as it will vary for different types of responses. For example, the loading time for transverse strain will be much longer than it is for longitudinal strain because of the actual shape of the contact area of the tire on the pavement, which is longer in the longitudinal direction than the transverse direction. The reason for the longer loading time for transverse strain is that the transverse strain is tangential to the load, whereas the longitudinal is radial and therefore has a sign change. The loading time is calculated from ( 200 mm + 2×depth)/( wheel speed in mm/ sec). The reference loading time is 0.015 sec ( 15 msec, roughly corresponding to the loading time of a standard FWD, where loading time refers to a creep test), and the reference temperature is 20 º C. The NCHRP 1 37A Design Guide makes use of an “ effective depth” based on the equivalent thickness of the layers, which results in longer loading times. The guide, however, then converts loading time into frequency, using f = 1/ lt, rather than f = 1/( 2πlt), more than compensating for the longer loading time ( unless loading time is defined differently, i. e., it is not based on a creep test). Viscosity is found from: log( log( visc cPoise)) = A + VTS * log( tK ) ( 4) where tK is the temperature ( in ° K), and A and VTS are constants, and cPoise indicates units of centipoise For all of the asphalt materials in this report a value of A = 9.6307 [ 10.5254 with temperature in º R ( degrees Rankine)] and VTS =  3.5047 were used. These values correspond ( according to the NCHRP 1 37A Design Guide) to an asphalt with a penetration grade of 40– 50. If the minimum modulus, Emin, the maximum modulus, Emax, and the modulus at two different temperatures are known, the viscosity versus temperature relationship ( Equation 4) will have very little influence on the modulus versus temperature relationship. This can be seen in Figure 5, where the resulting modulus versus temperature relationship is shown for asphalts with penetration grades from 40– 50 to 200– 300 and for a PG64 22 grade asphalt. Stage 5 Distribution 12 UCPRC RR 2005 06 Influence of viscosity 0 2000 4000 6000 8000 10000 12000 14000  20 0 20 40 60 Temperature, C Modulus, MPa 40 50 60 70 85 100 120 150 200 300 PG64 22 Figure 5. Modulus versus temperature for different viscosity versus temperature relationships. The constants δ, β, γ, and aTg, and the modulus at the reference temperature ( 20 º C) were derived from flexural frequency sweep tests. The constant a is calculated by the program. The frequency sweep tests were available for the top and bottom asphalt layers of Goal 1 and for the overlays in Goal 3. The fit between frequency sweep data and model data is shown in Figure 6. In fitting the model to the frequency sweep data it was assumed that the minimum modulus ( 10δ) was 200 MPa ( δ = 2.3010). In the frequency sweep test the measured modulus was considerably lower. However, based on FWD testing, it was assumed that an asphalt layer’s modulus, even at very low frequencies and high temperatures, had a minimum value greater than the one measured in the frequency sweep test on a flexural beam. This variance is attributable to the differences in boundary conditions between a laboratory test, such as a flexural beam frequency sweep and the same material when it is part of a layered pavement structure in the field. Specifically, a flexural beam is suspended in space without confinement in a flexural frequency sweep test. In contrast, the same asphalt concrete material, confined below and on its sides when it is part of a pavement layer, has its modulus increased. ( In this confined condition, the aggregate in the asphalt concrete, which has its own relatively unchangeable high modulus, also has a large compressive stress component applied to it.) Figure 7 shows an example of the input parameters for the AC bottom layer. A modulus versusreduced time relationship was assumed for the ATPB, based largely on laboratory triaxial testing. The ATPB had a modulus of 1144 MPa at a temperature of 20 C and a loading time of 0.015 sec. A minimum modulus of 200 MPa was assumed. Stage 5 Distribution 13 UCPRC RR 2005 06 0 2000 4000 6000 8000 10000 12000 14000 16000  4  3  2  1 0 1 2 log reduced time E MPa 28 28 28 28 19 19 19 19 15 15 15 15 AC top 0 2000 4000 6000 8000 10000 12000 14000 16000  4  3  2  1 0 1 2 log reduced time E MPa 28 28 28 28 19 19 19 19 15 15 15 15 AC bottom 0 2000 4000 6000 8000 10000 12000 14000 16000  4  3  2  1 0 1 2 log reduced time E MPa 25 25 10 25 10 25 25 10 25 10 25 10 10 DGAC 0 2000 4000 6000 8000 10000 12000 14000 16000  4  3  2  1 0 1 2 log reduced time E MPa 25 25 25 25 25 25 10 10 10 ARHM Figure 6. Frequency sweep data for Goal 1 and Goal 3 materials compared to models. Stage 5 Distribution 21 UCPRC RR 2005 06 Figure 7. Example of input parameters for the modulus versus reduced time relationship for the AC bottom layer. 1.3.2 Fatigue The modulus of damaged asphalt was calculated as: ( ) ( ) ( ( tr)) E 1 exp log log 1 β γ α ω δ + + × − = + ( 5) where the damage, ω, was calculated from: ( t) MPa E MN A × × ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ × ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ = × × δ με με ω β γ α exp 200 3000 ( 6) where E is the modulus of damaged material, MN is the number of load repetitions in millions ( N/ 106), με is the strain in μstrain, t is the temperature in º C, and α, β, γ, and δ are constants ( these constants are the same as the constants in Equations 1 through 5, and different in Equation 6). Stage 5 Distribution 22 UCPRC RR 2005 06 The initial ( intact) modulus, Ei, corresponds to a damage, ω, of 0, and the minimum modulus, Emin= 10δ, to a damage of 1. Equation 5 leads to: ( ) ( ( ) ) ( ) ⎟⎠ ⎞ ⎜⎝ ⎛ ⎟⎠ ⎞ ⎜⎝ ⎛ = ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ = − = − × − min min log log , log log 1 , E E E E or E E E E E E or i i i i i ω δ δ ω ω ( 7) The NCHRP 1 37A Design Guide calculates the modulus of damaged asphalt, for rehabilitation purposes, using Equation ( 5), but at the same time the Guide defines ω as the relative decrease in modulus ( although this is mistakenly indicated as E/ Ei in the report for the Guide). This definition of ω included in the NCHRP 1 37A Design Guide is inconsistent with Equation 5, as shown in Equation ( 7) The one stage Weibull distribution, Equation ( 8), could be used instead of Equation 6: SR = exp(− α × Nβ ) ( 8) where SR is the stiffness reduction (= E/ Ei), N is the number of load applications, and α and β are constants, different from those used in previous equations. Combining Equations ( 7) and ( 8) one gets: ( ) β α β α ω N E E E E N E E SR E E E E i i i i i × ⎟⎠ ⎞ ⎜⎝ ⎛ = ⎟⎠ ⎞ ⎜⎝ ⎛ − × = ⎟⎠ ⎞ ⎜⎝ ⎛ = ⎟⎠ ⎞ ⎜⎝ ⎛ ⎟⎠ ⎞ ⎜⎝ ⎛ = min min ln min ln min ln ln ln ln ( 9) which has the same format as Equation 6. In the present version of CalME ( September 2006) it is assumed that α and β can be written in the format: ( ( ) ( )) A B t C ( w) A B t C w D t w ln exp ln ln = + × + × = + × + × + × × β β β β α α α α α ( 10) where t is the temperature in º C, w is the internal energy density ( ½ ×ε2×E), and αA, αB, αC, αD, βA, βB, and βC are constants fit from the beam fatigue data. Fatigue parameters were determined for the AC top and AC bottom of Goal 1 and for the DGAC and ARHM overlays of Goal 3, based on four point bending beam tests at 10 Hz under controlled strain. In determining the fatigue parameters it was assumed that for Equation 6 that β was equal to two times γ. This reduces the number of parameters to be determined by one, and it ensures that the damage is a function of the internal energy density ( ½ ×ε2×E). The parameters were determined by minimizing the root mean square ( RMS) difference between the calculated relative modulus ( E/ Ei) and the experimental data for values of E/ Ei > 0.3, the stiffness ratio to which most of the beam fatigue tests were carried out. Stage 5 Distribution 23 UCPRC RR 2005 06 An example of damage versus number of load applications is shown in Figure 8 which shows the damage at a temperature of 20 º C for a constant strain of 400 μstrain. The four materials are surprisingly similar with respect to damage. Damage of asphalt 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 500000 1000000 1500000 2000000 2500000 Number of load applications Damage Goal1Top Goal1Bottom Goal3DGAC Goal3ARHM Figure 8. Example of damage versus number of load applications. For ATPB a damage function ( Equation 6) was chosen based on the damage function for the bottom lift of Goal 1, but with a value of A about fifty times as high. Figure 9 shows an example of the damage parameters for Equation 6 for the AC bottom lift of Goal 1. The parameters of interest here are the values of the first column ( under the heading “ Fatigue, dE/ Ei”). The response type is the horizontal tensile strain at the bottom of the layer ( the minor principal strain) indicated by the Response type “ e”. The parameters A, α, Reference strain ( Respref), β, Reference modulus ( Eref), γ, and δ are given. The parameter Std A is a Standard Deviation factor on A that is used in the Monte Carlo simulations, which are not used in this report. The number of in situ load applications is divided by the Shift Factor given at the bottom of the column to allow for differences between laboratory testing and field conditions. A Shift Factor of 3 was used for all materials, i. e., three HVS loads were assumed to give the same damage as one laboratory load. Stage 5 Distribution 24 UCPRC RR 2005 06 Figure 9. Example of Equation 6 damage parameters of AC bottom layer ( Goal 1). 1.4 Weak Bonding Weak bonding between the top and bottom asphalt lifts of Goals 1, 3, and 5 was found for certain areas of the HVS test sections. It appeared that the top AC layer had moved horizontally with respect to the bottom AC layer. During forensic studies a brown discoloration and scratch marks were observed on the surfaces of the materials at the interface where slip had apparently occured. Cores showed no bonding at the interface. The layered elastic analyis program ( LEAP) option of CalME has a parameter that controls the degree of bonding between two layers, referred to as “ slip” in LEAP if less than full bonding and “ stick” if full bonding. Full bond corresponds to a high value ( 105 is used in CalME), and a value of 0 corresponds to full slip ( i. e., there is no bond between the layers). The logarithm of this parameter is decreased linearly until the final slip is reached, as a function of the number of loads in ESALs. The number of ESALs corresponding to final slip is not known, but the shape of the deflection versus number of loads curve can serve as a guide. The LEAP program treats the pavement structure as a continuum, which means that the two materials above and below the slip interface will be in contact for all points of the interface. This is not a completely realistic assumption, but with the present response models it cannot be changed. [ A three dimensional Finite Element Model ( FEM) would be required to change that assumption.] It is likely that this incorrect modeling of the slip interface will have different effects on deflections, on horizontal strains at the interface, and on the shear stress and strain at a depth of 50 mm. ( This depth is used for calculating the permanent deformation of the AC layer. This is described later). For deflection and shear Stage 5 Distribution 25 UCPRC RR 2005 06 stress, it is likely that friction between the two layers immediately under the wheel will result in only a partial slip occurring because of the high compressive normal force, whereas the maximum horizontal strains may well correspond to a condition closer to full slip. Table 5 Influence of Slip Value in LEAP on Calculated Vertical Deflections and Horizontal Strains Slip Value 0 10 4 10 3 10 2 10 1 1 10 102 103 104 105 d 0 mm* 0.526 0.526 0.525 0.517 0.473 0.396 0.370 0.367 0.367 0.367 d 625 mm* 0.276 0.276 0.276 0.272 0.253 0.228 0.220 0.220 0.219 0.219 Ex top* 230 230 229 222 182 83 14 0  2  2 Ex bottom* 232 232 232 232 226 196 168 162 161 161 * Note: d 0 mm is vertical deflection at surface, d 625 mm is vertical deflection at 625 mm depth, Ex top is horizontal strain at the bottom of the top asphalt layer, Ex bottom is the horizontal strain at the bottom of the bottom asphalt layer. Table 5 shows an example, taken from Section 501RF, of the influence of the slip value on the deflection at the surface ( d 0 mm) and the deflection of the subgrade ( d 625 mm), and the horizontal strain at the bottom of the top AC layer ( Ex top, tensile as positive) and at the bottom of the bottom AC layer ( Ex bottom). A slip value of 0.0001 was chosen for the main simulations. This corresponds to full slip between the layers. At the interface with full slip the shear stress will be zero. The shear stress used for calculating permanent deformation in the asphalt is at a depth of 50 mm, which is only slightly above the interface. When full slip develops, the shear stress at depth 50 mm will therefore decrease considerably. As was mentioned above, this may not be realistic, so a second simulation was carried out with full bonding to determine the permanent deformation of the asphalt. During this simulation the stiffness factors for the unbound layers and the shift factor for asphalt fatigue were adjusted to assure that the pavement deflection history was still correct. These simulations with no slip were only used for determining the permanent deformation of the asphalt. 1.5 Unbound Layers 1.5.1 Triaxial Tests Table 6 shows the results of triaxial tests on the subgrade material ( Harvey et al. 1996). Two specimens were tested, compacted at different density and moisture content ( MC) and either soaked or saturated. The tests were done at a confining stress of 7 kPa, which is close to the static confining stress at the top of the subgrade. The parameters C and n are defined by the equation: n d MPa C E ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ = × 0.1 σ ( 11) where E is the modulus, and σd is the deviator stress. The column “ E( 30 kPa)” indicates the modulus, at a deviator stress of 30 kPa, which is a typical stress at the top of the subgrade under a 40 kN dual wheel load. Stage 5 Distribution 26 UCPRC RR 2005 06 Table 6. Triaxial Tests on Subgrade Density MC% Condition C n E( 30 kPa) 2.06 g/ cm3 22.4 Soaked 36.2  0.34 55 MPa 2.12 g/ cm3 15.8 Soaked 66.5  0.32 98 MPa 2.12 g/ cm3 15.8 Saturated 41.5  0.27 57 MPa Triaxial tests were done on one AB specimen, with the results shown in Table 7. The sample was compacted to a density of 2.47 g/ cm3 at a MC of 5.5 percent. The top of the specimen was exposed for ten days to simulate the effects of exposure to air on the test section. The moisture content dropped from 5.5 percent to 2.9 percent during the ten days. After testing the specimen was saturated and tested again. Table 7. Triaxial Tests on Aggregate Base ( AB) Condition k1 k2 E( 50 kPa) Exposed 481 0.16 430 MPa Saturated 201 0.49 143 MPa The constants k1 and k2 are defined by the equation 2 1 0.1 k MPa k E ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ = × θ ( 12) where E is the modulus, and θ is the bulk stress. The column “ E( 50 kPa)” indicates the modulus at a bulk stress of 50 kPa. According to the theory of elasticity, the bulk stress in the AB a
Click tabs to swap between content that is broken into logical sections.
Rating  
Title  Calibration of incrementalrecursive flexible damage models in CalME using HVS experiments 
Subject  PavementsCaliforniaTesting.; PavementsCaliforniaLive loads.; PavementsCaliforniaDeteriorationMathematical modelsCalibration. 
Description  Title from PDF title page (viewed on September 5, 2007).; Performed for California Dept. of Transportation Division of Research and Innovation, Office of Roadway Research.; Authors: Per Ullidtz, John Harvey, BorWen Tsai, Carl Monismith.; "This work was completed as part of Partnered Pavement Research Program Strategic Plan Item 4.1 "Development of the First Version of a MechanisticEmpirical Pavement Rehabilitation, Reconstruction and New Pavement Design Procedure for Rigid and Flexible Pavements (preCalibration of AASHTO 2002)."; "April 2006."; "FHWA No.: F/CA/RR/2006/49"Document retrieval page.; Includes bibliographical references (p. 209211).; Final report.; Harvested from the web on 9/5/07 
Publisher  University of California Pavement Research Center 
Contributors  Ullidtz, Per, 1942; Harvey, John T.; Tsai, BorWen.; Monismith, Carl L.; California. Dept. of Transportation. Division of Research and Innovation. Office of Roadway Research.; University of California. Pavement Research Center.; Dynatest Consulting, Inc. 
Type  Text 
Identifier  http://www.its.berkeley.edu/pavementresearch/PDF/CalME%20Calib%20Rpt_RR200506.pdf 
Language  eng 
TitleAlternative  Calibration of incrementalrecursive flexible damage models in CalME using Heavy Vehicle Simulator experiments 
DateIssued  [2006] 
FormatExtent  xxvi, 222 p. : digital, PDF file with col. ill, col. charts. 
RelationRequires  Mode of access: World Wide Web. 
Transcript  April 2006 Research Report: UCPRC RR 2005 06 Calibration of Incremental Recursive Flexible Damage Models in CalME Using HVS Experiments Authors: Per Ullidtz, Dynatest Consulting Inc John Harvey, UC Davis Bor Wen Tsai, UC Berkeley Carl Monismith, UC Berkeley This work was completed as part of Partnered Pavement Research Program Strategic Plan Element 4.1: “ Development of the First Version of a Mechanistic Empirical Pavement Rehabilitation, Reconstruction and New Pavement Design Procedure for Rigid and Flexible Pavements ( pre Calibration of AASHTO 2002)” PREPARED FOR: California Department of Transportation Division of Research and Innovation Office of Roadway Research PREPARED BY: University of California Pavement Research Center Berkeley and Davis Stage 5 Distribution ii UCPRC RR 2005 06 DOCUMENT RETRIEVAL PAGE Report No: UCPRC RR 2005 06 Title: Calibration of Incremental Recursive Flexible Damage Models in CalME Using HVS Experiments Authors: Per Ullidtz, Dynatest Consulting Inc.; John Harvey, UC Davis; Bor Wen Tsai, UC Berkeley; and Carl Monismith, UC Berkeley Prepared for: California Department of Transportation Division of Research and Innovation Office of Roadway Research FHWA No.: F/ CA/ RR/ 2006/ 49 Date: Stage 5, May 2007 Contract number: UCPRC RR 2005 06 Client Reference No: UCPRC RR 2005 06 Status: Final, Caltrans approved Abstract: Caltrans is in the process of implementing Mechanistic Empirical design procedures. All mechanisticempirical methods must be validated/ calibrated against the behavior of real pavements. This should be done before implementing models in design methods to ensure that designs will be reasonable. The Heavy Vehicle Simulator ( HVS) provides a first step in this validation/ calibration process. The short test section can be carefully constructed with well characterized materials and instrumented to measure the pavement response. The climatic conditions may be controlled or monitored closely and all load applications are known exactly. The pavement may also be tested until it fails. The HVS may be seen as “ large scale” laboratory equipment, between the “ small scale” laboratory equipment ( triaxial tests, bending tests etc.) and the reality of real pavements, which have uncertainties regarding materials, loads and climatic conditions. The two HVSs owned by Caltrans have been used on 27 flexible pavement test sections, with varying combinations of asphalt and granular layers. Temperature control was used during the tests. Most sections have been instrumented with Multi depth Deflectometers ( MDDs) to compare the measured pavement deflections ( at several depths) to the deflections predicted by mechanistic methods, during the full duration of tests carried to “ failure” ( in terms of rutting or cracking). Results from mechanistic models have been compared with the deflection measurements and performance as a first step prior to empirical calibrations with field results. The complete time history of each test has been compared rather than just the beginning and end measurements. This report presents the validation of the mechanistic models for asphalt fatigue and for permanent deformation with the HVS test results. Keywords: Mechanistic empirical, full scale testing, calibration, response, performance, flexible pavement . Proposals for implementation: None Related documents: Kannekanti, V., and Harvey, J. June 2005. Sensitivity Analysis of 2002 Design Guide Rigid Pavement Distress Prediction Models. Draft report prepared for the California Department of Transportation. Pavement Research Center, Institute of Transportation Studies, University of California Berkeley, University of California Davis. UCPRC RR 2005 01 Signatures: P. Ullidtz Principal Author J. Harvey Co principal Investigator C. L. Monismith Co principal Investigator D. Spinner Editor M. Samadian Caltrans Contract Mgr. Stage 5 Distribution iii UCPRC RR 2005 06 DISCLAIMER The contents of this report reflect the views of the authors who are responsible for the facts and accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the State of California or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation. Stage 5 Distribution iv UCPRC RR 2005 06 EXECUTIVE SUMMARY The first step in a mechanistic empirical ( ME) pavement design or evaluation is to calculate pavement response — in terms of stresses, strains, and/ or displacements — using a mathematical ( or mechanistic) model. In the second step, the calculated response is used as a variable in empirical relationships to predict structural damage ( decrease in moduli or cracking) and functional damage ( rutting and roughness) to the pavement. Both of these steps must be reasonably correct. If the calculated response bears little resemblance to the pavement’s actual response, there is no point in trying to use the calculation to predict future damage to the pavement. In other words, only if the calculated response is reasonably correct does it make sense to try to relate the damage to the pavement response. This study’s purpose was to evaluate the overall trends of the damage models in the draft software package called CalME against those of Heavy Vehicle Simulator ( HVS) tests for which data was available. The report presents simulations of HVS tests using the set of distress models included in CalME. These models are for the typical flexible pavement distresses observed in California: asphalt fatigue, asphalt rutting, unbound layers rutting, and reflection cracking. An Incremental Recursive approach ( see item 4 below) was used for the simulations included in this report because this approach can accurately indicate pavement condition at different points during a pavement’s life. Approaches Included in CalME CalME software provides the user with four approaches to evaluating or designing a flexible pavement structure: 1. Caltrans’ current methods: the R value method for new flexible structures and the deflection reduction method used by Caltrans for overlay thickness design for existing flexible pavements. 2. “ Classical” Mechanistic Empirical ( ME) Design, which is based largely on the Asphalt Institute Method which uses very simple methods to characterize materials, climate, and traffic inputs. 3. An Incremental approach, which is a standard Miner’s Law approach that permits damage calculation for the axle load spectrum and expected temperature regimes, but without updating of the material’s properties through the life of the project. This is an approach similar to the one for cracking of asphalt included in the NCHRP 1 37A Pavement Design Guide, also referred to as the Mechanistic Empirical Design Guide ( MEPDG). This type of approach is calibrated against an end failure state ( such as, 25 percent cracking of the wheelpath) and it assumes a linear accumulation of damage to get to that state. 4. An Incremental Recursive approach in which the materials properties of the pavement — in terms of damage and aging — are updated as the pavement life simulation progresses. The current Caltrans methods and the Classical method are very fast in terms of computational time, and user input is highly simplified. In CalME both of these options perform a “ design” function, calculating and presenting pavement structures that meet the design requirements for the design traffic, materials, and climate. For design practice the Classical and Caltrans methods should be used to produce a set of potential pavement sections. The Incremental Recursive method should then be run to check the lowest cost alternative designs in the set to be certain that they meet design requirements. Once the final design has been selected, its Incremental Recursive output provides a prediction of the pavement condition across its entire life. The prediction of the pavement’s condition through its life from the Incremental Recursive output can be used as the first prediction for use in a pavement management system. Stage 5 Distribution v UCPRC RR 2005 06 Use of Heavy Vehicle Simulator Data to Evaluate Models The Incremental Recursive models included in CalME were used to predict performance for all twenty seven of the flexible pavement HVS tests performed so far as part of the Accelerated Pavement Testing ( APT) program operated for the California Department of Transportation ( Caltrans) by the University of California Pavement Research Center ( UCPRC). The HVS test data in this report come from tests performed between the years 1995 and 2004. The HVS response data and corresponding laboratory test data were extracted from the UCPRC HVS database. During HVS testing, pavement response  in terms of deflections at the surface and/ or at multiple depths  may be measured. A Road Surface Deflectometer ( RSD) measures deflections at the surface and is similar to the Benkelman Beam used to develop the current Caltrans overlay design method in the 1950s. A Multi depth Deflectometer ( MDD) measures deflections at multiple depths. In order to accurately predict the gradual degradation of a pavement, the response model must predict measured deflections with reasonable accuracy. Although a model might predict deflections correctly, this ability does not guarantee that the model can also accurately predict the stresses and strains in all the pavement layers. However, the opposite is true: if a model predicts deflections incorrectly it will also produce incorrect stress and strain predictions. Therefore when attempting to calibrate ME models from HVS tests, the research team’s first concern was to make sure that resilient deflections were predicted reasonably well for the duration of the test and for all load levels. This prediction depended on the moduli of all the pavement layers and on the changes to the moduli caused by fatigue damage, slip between asphalt layers, non linear elastic characteristics of unbound layers, and the effect of confinement on granular layers. Once reasonably good agreement was achieved between the measured and the calculated deflections then the permanent deformation models could be calibrated with confidence. Differences in boundary conditions, strain levels, and loading times, all of which can produce varied effects in materials, result in differing moduli values. In this study, methods used for determining moduli ( also referred to as “ stiffness”) values included backcalculation from Falling Weight Deflectometer ( FWD) and MDD data, and direct measurement — employing laboratory triaxial testing for unbound materials and flexural frequency sweep testing for asphaltic materials. Stiffnesses for the study’s asphalt materials were taken primarily from flexural frequency sweep data. Stiffnesses for the unbound layers came primarily from MDD data backcalculation. In practice the FWD is seen by the research team as the primary tool for stiffness measurement of all layers already constructed because it is used in the field on the full pavement system; this is thought to be appropriate because the boundary conditions are those of real pavement, and most Caltrans’ work will be rehabilitation and reconstruction with at least some layers already in place. The research team saw the flexural beam test as the primary means for measuring the stiffness and fatigue characteristics of asphalt overlay materials for new layers. For new pavement construction, a combination of FWD testing on existing pavements and triaxial testing can be used to develop a database of stiffnesses of unbound granular layers and subgrades based on different characteristics, such as Unified Soil Classification System ( USCS) classification. The purpose of this study was to evaluate the overall trends of the CalME damage models against those of the HVS test results. This was accomplished by comparing deflections calculated using moduli determined from initial measurements and CalME damage calculations with measured deflections under HVS loading. The results presented in this report verify that, overall, the CalME damage trends for deflection and permanent deformation under loading are correct. During HVS testing, deflections often increase markedly, sometimes becoming more than twice as high at the end of the test as they were at the beginning because of damage to the asphalt concrete caused by the repeated wheel loads. However, the flexible pavement design model of the NCHRP 1 37A Design Guide does not consider any decrease in the asphalt modulus as a result of fatigue damage ( except for rehabilitation designs). In fact, the NCHRP 1 37A Design Guide includes a model for aging that predicts a continuous increase in the stiffness of the asphalt concrete layers across the life of the pavement, which results in increased stiffness and smaller predicted deflections as the pavement is subjected to trafficking. While the Stage 5 Distribution vi UCPRC RR 2005 06 aging is potentially important, the effect of updating stiffness for aging and not updating it for fatigue damage results in calculation of very unrealistic elastic responses in the pavement during its life. This makes it impossible to use the model to simulate an HVS test and, inversely, to use HVS tests to calibrate the model, except for pavements with extremely thick asphalt concrete layers where little fatigue should develop. Results of HVS Test Simulation Using CalME The series of HVS tests in this report are grouped here by goals, which are defined as follows: • Goal 1, a comparison of new pavement structures with and without asphalt treated permeable base ( ATPB) layer under dry conditions, moderate temperatures, 20° C ( HVS Sections 500RF, 501RF, 502CT, 503RF) • Goal 3 Cracking, a comparison of reflection cracking performance of ARHM GG ( the acronym ARHM, asphalt rubber hot mix gap graded, refers to the material specification at the time of construction in April 1997.) and dense graded asphalt concrete ( DGAC) overlays placed on the cracked Goal 1 sections, dry conditions, 20° C ( HVS Sections 514RF, 515RF, 517RF, 518RF) • Goal 3 Rutting, a comparison of rutting performance of ARHM GG and DGAC overlays of previously untrafficked areas of Goal 1 pavements, dry conditions, 40° C or 50° C at 50 mm depth, four different tire/ wheel types ( HVS Sections 504RF, 505RF, 506RF, 507RF, 508RF, 509RF, 510RF, 511RF, 512RF, 513RF) • Goal 5, a comparison of new pavement structures with and without ATPB layer under wet conditions ( water introduced into base layers), moderate temperatures, 20° C ( HVS Sections 543RF, 544RF, 545RF) • Goal 9, initial cracking of asphalt pavement with six replicate sections in preparation for later overlay, new pavement, ambient rainfall, 20° C ( HVS Sections 567RF, 568RF, 569RF, 571RF, 572RF, 573RF) CalME models that the simulations evaluated included: • A stiffness model for asphalt concrete modulus as a function of reduced time based on the model used in NCHRP 1 37A Design Guide, with some adjustments based on field observations; • An asphalt concrete fatigue model that predicts damage, in terms of decrease in modulus, as a function of load repetitions, tensile strain, and stiffness, using parameters from flexural beam testing; • An ability to model partial bonding between asphalt concrete layers; • A model that adjusts the stiffness of unbound layers as a function of the combined bending resistance ( a function of their stiffness and thickness) of the layers above them; • A model that adjusts the stiffness of unbound layers as a function of load level, with an increased load level increasing the moduli for the granular layers and decreasing modulus for the subgrade ( clay); • A permanent deformation model for asphalt concrete as a function of permanent shear strain near the pavement surface beneath the edge of a tire, with permanent shear strain predicted by the calculated elastic shear strain and elastic shear stress; • A permanent deformation model for unbound layers as a function of the vertical strain at the top of each layer; and • A reflection cracking model based on tensile strain calculated using a regression equation developed from a large number of Finite Element analyses and the same damage parameters developed for asphalt concrete fatigue. Stage 5 Distribution vii UCPRC RR 2005 06 Response Models During most of the HVS tests, resilient deflections were measured using the RSD and the MDD. The following figure summarizes the measured deflections with those calculated using CalME damage models for all of the sections in terms of the ratio of the initial deflections before HVS loading to the final deflections at the end of the loading. Assumptions made regarding differences between moduli from different measurement methods, shift factors, slip between layers, and non linear elasticity of unbound layers to obtain reasonably good agreement between measured resilient deflections and those calculated with CalME are discussed in the report. The observed behavior of the aggregate base ( AB) and subbase layers under HVS loading contradicts the commonly accepted wisdom for granular materials, which is based primarily on triaxial testing. The observed behavior is discussed in the report and is modeled in CalME. Using these assumptions, it was possible to model resilient deflections reasonably well for the full history of all of HVS test sections using the layered elastic analysis program ( LEAP) response model. Final/ initial deflection 0.00 1.00 2.00 3.00 4.00 5.00 0.00 1.00 2.00 3.00 4.00 5.00 Measured Calculated RSD MDD Equality Figure ES 1. Ratio of initial to final deflection. Damage of Asphalt Materials Controlled strain fatigue tests conducted on beams were used to derive model parameters for the decrease in modulus for all the asphalt materials — except for the ATPB, where laboratory tests were not available. Working under the assumptions used in the modeling and using a shift factor with these damage models produced the correct changes in resilient deflections during all the HVS tests. Stage 5 Distribution viii UCPRC RR 2005 06 For reflection cracking, a simple model was used to calculate the strains in an overlay caused by existing cracking in the original top layer. Using this model and the laboratory fatigue model, reasonably correct resilient deflections were also predicted. Relating visual cracking to the calculated asphalt damage proved to be difficult, and no single relationship could be derived. Goal 1 and Goal 5 showed differences between the drained and the undrained sections; for Goal 3 Cracking, the increase in visual cracking was quicker than for Goal 1 and Goal 5; and for Goal 9, visual cracking occurred at much less calculated damage than for the other experiments. It is possible that the relationship between visual cracking and calculated damage depends on the thickness of the asphalt layers, and that reflection cracks and cracks in new pavements develop differently. It is also possible that the development of visual cracking depends on factors that the simulations did not consider. No single relationship could be established between the relative increase in deflection and the amount of surface cracking ( shown in the following figure), but it may be noted that visible cracking was not observed until deflection had increased by 50 percent or more. Cracking versus relative deflection 0 2 4 6 8 10 12 1 1.5 2 2.5 3 3.5 4 4.5 Deflection/ initial deflection Cracking m/ m2 500RF 501RF 502CT 503RF 514RF 515RF 517RF 518RF 543RF 544RF 545RF 567RF 568RF 569RF 571RF 572RF 573RF Figure ES 2. Cracking versus increase in deflection. Permanent Deformation of Asphalt The following figure shows the measured and predicted final permanent deformation of the asphalt layers from Goal 1, Goal 3, and Goal 5, where data were available. Permanent deformation was calculated for the upper 100 mm of the asphalt layer( s). Stage 5 Distribution ix UCPRC RR 2005 06 Permanent deformation in AC ( pro rated) 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 Measured, mm Calculated, mm Equality 20 º C at surface 45 or 50 º C at surface Figure ES 3. Measured and predicted final permanent deformation of asphalt. In Figure ES 3, the 20° C outlier point with a measured deformation greater than the calculated one is Section 543RF, which was a wet, drained test where the ATPB stripped and collapsed. The two relatively low values at high temperatures are from the test with a bias ply dual tire and from the test with an aircraft tire. The correlation coefficient between measured and calculated values is 0.82 and the standard error of estimate is 2.2 mm. The parameters for predicting permanent shear strain were based on Repeated Simple Shear Tests at Constant Height ( RSST CH). Permanent Deformation of Granular Layers The permanent deformation of the granular layers for Goal 1, Goal 3, and Goal 5 are shown with average measured values in Figure ES 4. It should be noted that the permanent deformations are rather small except for Section 543RF, the wet drained section, where the permanent deformation includes part of the ATPB. Stage 5 Distribution x UCPRC RR 2005 06 Permanent deformation of granular layers 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 Measured, mm Calculated, mm Goal 1 Goal 3 Goal 5 Equality Figure ES 4. Final permanent deformation of granular layers. Permanent deformation was calculated both at the top of the AB and at the top of the aggregate subbase ( ASB). The two materials are rather similar and might have been treated as a single layer. Permanent Deformation of Subgrade The final permanent deformation of the subgrade is even smaller than that of the granular layers, with a maximum measured value of less than 2 mm. In addition, the data scatter is as large as that of the granular layers, with an average coefficient of variation of 70 percent. This is far from ideal for the calibration of a subgrade permanent deformation model. The mean measured and predicted final deformations are shown in Figure ES 5. The subgrade and granular results indicate that rutting of the unbound layers is probably not a major concern for existing Caltrans pavements that need rehabilitation unless there is poor drainage or significant amounts of water are entering cracks in the asphalt layers. Stage 5 Distribution xi UCPRC RR 2005 06 Permanent deformation of subgrade 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Measured, mm Calculated, mm Figure ES 5. Final permanent deformation of the subgrade. Total Permanent Deformation at Pavement Surface Figure ES 6 shows the final calculated permanent deformation at the pavement surface versus the measured final deformation averaged from profilometer measurements along the HVS test area. Calculated final permanent deformations that underestimated the measured final permanent deformations were worst for the Goal 5 sections in which water was dripped into the base layers, especially for the drained section, 543RF in which the ATPB stripped. The correlation coefficient between measured and calculated deformations is 0.61 and the standard error of estimate is 2.6 mm. Conclusions and Recommendations The overall results from this study indicate that Incremental Recursive models provide reasonable results when predicting the response and performance of pavement under HVS loading. However, now that the models have been shown to match the mechanics of the flexible pavements under HVS loading, additional work remains to be done before these models can be used for pavement design and performance prediction. There are significant differences between HVS testing and field results, and the approach used in this study has limitations because of those differences. These include the effects of age and of seasonal variation that have not been quantified in the simulations because HVS tests are of relatively short duration and are performed, to varying degrees, in controlled environments. Field calibration is required to evaluate the response difference between the field pavement and the Incremental Recursive simulation that should be attributed to aging and seasonal effects. It is likely that the effects of aging can be dealt with using shift factors. Stage 5 Distribution xii UCPRC RR 2005 06 Permanent deformation at pavement surface 0.0 5.0 10.0 15.0 20.0 0.0 5.0 10.0 15.0 20.0 Measured, mm Calculated, mm Goal 1 Goal 3 45/ 55° C Goal 3 20° C Goal 5 Goal 9 Equality Figure ES 6. Final permanent deformation at the pavement surface. The effects of rest periods between loadings and of faster traffic have also not been included in the calibration. It is expected that different shift factors will result because of rest periods and different trafficking patterns. Lastly, moduli from frequency sweep data, triaxial tests, FWD tests, and MDD deflections used in this study are similar but they are not identical. The NCHRP 1 37A Design Guide study proposes relying primarily on triaxial testing to characterize the stiffness of flexible pavement layers and the permanent deformation parameters of asphaltic materials. Recommendations are made in this report for the most practical and economical methods for characterizing materials based on the understanding that the majority of Caltrans’ work over the next several decades will be rehabilitation and reconstruction, with some addition of lane capacity. Recommendations are also made regarding the next steps to develop the CalME models. These include: 1. Perform a sensitivity analysis using “ typical” values for properties and climate in the database established to date, and compare the results from the Classical, Incremental, and Incremental Recursive methods included in CalME to evaluate reasonableness of sensitivity across the three methods. 2. Simulate mainline highway case studies and test track data ( such as WesTrack and NCAT track) using the recommended methods for characterizing flexible pavement materials in conjunction with the Incremental Recursive models in CalME, and compare the simulated and measured results, as was done for the HVS results presented in this report. This step will provide validation for the models Stage 5 Distribution xiii UCPRC RR 2005 06 3. Address the variability of the input parameters ( moduli, thicknesses, traffic loading, etc.) and uncertainty on the damage models. Several approaches should be considered, including the approach used in the NCHRP 1 37A method. 4. Make final decisions regarding use of cemented layers in the flexible pavement structure, then calibrate. It is generally recommended that “ semi rigid” pavements, in which asphalt concrete is placed directly on cement treated base ( CTB) or lean concrete base ( LCB), not be used because of the relatively quick reflection of shrinkage cracks. However, because Caltrans has used semi rigid pavements in the past and they remain in the current design method, it is therefore important to have models for the response and performance of these layers. The models in the NCHRP 1 37A Report should be the starting point for such a validation and calibration exercise. Stage 5 Distribution xiv UCPRC RR 2005 06 TABLE OF CONTENTS Executive Summary........................................................................................................................ .................... iv Approaches Included in CalME ...................................................................................................................... iv Use of Heavy Vehicle Simulator Data to Evaluate Models ............................................................................. v Results of HVS Test Simulation Using CalME .............................................................................................. vi Response Models......................................................................................................................... .............. vii Damage of Asphalt Materials ..................................................................................................................... vii Permanent Deformation of Asphalt ........................................................................................................... viii Permanent Deformation of Granular Layers................................................................................................ ix Permanent Deformation of Subgrade ........................................................................................................... x Total Permanent Deformation at Pavement Surface.................................................................................... xi Conclusions and Recommendations................................................................................................................ xi List of Figures........................................................................................................................ .......................... xvii List of Tables ............................................................................................................................... .................... xxv 1.0 Introduction................................................................................................................... ......................... 1 1.1 Models and Approaches Included in CalME .......................................................................................... 1 1.1.1 Validation Using Heavy Vehicle Simulator Data .......................................................................... 2 1.2 HVS tests.......................................................................................................................... ...................... 4 1.2.1 Goal 1 and Goal 3 Tests................................................................................................................. 4 1.2.2 Goal 5 Tests ............................................................................................................................... ... 6 1.2.3 Goal 9 Tests ............................................................................................................................... ... 7 1.3 Response and Damage Models ............................................................................................................... 8 1.3.1 Asphalt Modulus ............................................................................................................................ 8 1.3.2 Fatigue........................................................................................................................ ................. 21 1.4 Weak Bonding........................................................................................................................ .............. 24 1.5 Unbound Layers ............................................................................................................................... .... 25 1.5.1 Triaxial Tests.......................................................................................................................... ..... 25 1.5.2 Influence of Stiffness of Layers above an Unbound Layer.......................................................... 26 1.5.3 Influence of Load Level ............................................................................................................... 39 1.6 Permanent Deformation ........................................................................................................................ 40 1.6.1 Asphalt ............................................................................................................................... ......... 40 1.6.2 Unbound Materials...................................................................................................................... 47 1.7 Reflection Cracking .............................................................................................................................. 48 2.0 Goal 1 Cracking Test Simulations ........................................................................................................ 50 2.1 Goal 1 Resilient Deformations.............................................................................................................. 50 Stage 5 Distribution xv UCPRC RR 2005 06 2.1.1 Section 501RF Resilient Deflections ( Undrained)....................................................................... 50 2.1.2 Section 503RF Resilient Deflections ( Undrained)....................................................................... 56 2.1.3 Section 500RF Resilient Deflections ( Drained)........................................................................... 60 2.1.4 Section 502CT Resilient Deflections ( Drained)........................................................................... 65 2.2 Visual Cracking Versus Damage of the Top Asphalt Layer, Goal 1 .................................................... 69 2.3 Goal 1 Permanent Deformation............................................................................................................. 72 2.3.1 Section 501RF Permanent Deformations..................................................................................... 73 2.3.2 Section 503RF Permanent Deformations..................................................................................... 75 2.3.3 Section 500RF Permanent Deformations..................................................................................... 77 2.3.4 Section 502CT Permanent Deformations..................................................................................... 79 3.0 Goal 3 Reflection cracking tests............................................................................................................ 81 3.1 Resilient Deflections ............................................................................................................................. 81 3.1.1 Section 517RF DGAC on Section 501RF Resilient Deflections ................................................. 81 3.1.2 Section 518RF ARHM on Section 503RF Resilient Deflections................................................. 86 3.1.3 Section 514RF DGAC on Section 500RF Resilient Deflections ................................................. 91 3.1.4 Section 515RF ARHM on Section 502CT Resilient Deflections ................................................ 96 3.2 Visual Cracking Versus Damage of the Overlay, Goal 3, 20 º C.......................................................... 101 3.3 Permanent Deformation Goal 3, 20 º C................................................................................................. 104 3.3.1 Section 517RF 75 mm DGAC Permanent Deformations .......................................................... 105 3.3.2 Section 518RF 38 mm ARHM Permanent Deformation ........................................................... 107 3.3.3 Section 514RF 75 mm DGAC Permanent Deformations .......................................................... 109 3.3.4 Section 515RF 38 mm ARHM Permanent Deformations ......................................................... 112 4.0 Goal 3 Rutting experiments................................................................................................................. 114 4.1 Section 504RF No Overlay, Wide Base Single Tire........................................................................... 115 4.2 Section 505RF DAGC Overlay, Bias Ply Dual Tire........................................................................... 117 4.3 Section 506RF DGAC Overlay, Radial Dual Tire .............................................................................. 119 4.4 Section 507RF DGAC Overlay, Wide Base Single Tire .................................................................... 121 4.5 Section 508RF ARHM Overlay, Wide Base Single Tire.................................................................... 123 4.6 Section 509RF ARHM Overlay, Radial Dual Tire ............................................................................. 125 4.7 Section 510RF ARHM Overlay, Radial Dual Tire ............................................................................. 127 4.8 Section 511RF ARHM Overlay, Wide Base Single Tire.................................................................... 129 4.9 Section 512RF DGAC Overlay, Wide Base Single Tire .................................................................... 131 4.10 Section 513RF DGAC Overlay, Aircraft Tire..................................................................................... 133 5.0 Goal 5 Wet conditions..................................................................................................................... ... 135 5.1 Section 543RF ARHM Overlay, Drained ........................................................................................... 136 5.2 Section 544RF ARHM Overlay, Undrained ....................................................................................... 145 Stage 5 Distribution xvi UCPRC RR 2005 06 5.3 Section 545RF DGAC Overlay, Undrained ........................................................................................ 153 5.4 Visual Cracking versus Damage of the Top Asphalt Layer, Goal 5 ................................................... 160 6.0 Goal 9 Modified Binder ( MB) road, initial tests ................................................................................. 162 6.1 Materials Characterization .................................................................................................................. 162 6.2 Section 567RF MB Road .................................................................................................................... 169 6.3 Section 568RF MB Road .................................................................................................................... 173 6.4 Section 573RF MB Road .................................................................................................................... 176 6.5 Section 571RF MB Road .................................................................................................................... 179 6.6 Section 572RF MB Road .................................................................................................................... 183 6.7 Section 569RF MB Road .................................................................................................................... 186 6.8 Visual Cracking Versus Damage of the Top Asphalt Layer, Goal 9 .................................................. 190 7.0 Summary and recommendations ......................................................................................................... 191 7.1 Shift Factors and Damage Equations Used in Simulations ................................................................. 191 7.2 Response Model.......................................................................................................................... ....... 192 7.3 Damage of Asphalt Materials.............................................................................................................. 197 7.4 Permanent Deformation of Asphalt..................................................................................................... 199 7.5 Permanent Deformation of Granular Layers ....................................................................................... 201 7.6 Permanent Deformation of Subgrade.................................................................................................. 203 7.7 Total Permanent Deformation at Pavement Surface ........................................................................... 205 7.8 Recommendations ............................................................................................................................... 207 8.0 References..................................................................................................................... ..................... 209 9.0 Appendix....................................................................................................................... ..................... 212 9.1 Glossary ............................................................................................................................... .............. 212 9.2 List of Units ............................................................................................................................... ........ 214 9.3 List of Parameters in Equations .......................................................................................................... 214 9.4 Parameter Values Used in Simulations ............................................................................................... 216 9.5 Section 569RF Simulated with a CTB Model from an HVS Nordic Experiment............................... 217 Stage 5 Distribution xvii UCPRC RR 2005 06 LIST OF FIGURES Figure ES 1. Ration of final to initial deflection .........................……………………………………………... vii Figure ES 2. Cracking versus increase in deflection ..........................……………………………………...… viii Figure ES 3. Measured and predicted final permanent deflection of asphalt ........................... .....…………...... ix Figure ES 4. Final permanent deformation of granular layers .............................……………………………… x Figure ES 5. Final permanent deformation of the subgrade .............................……………………………...… xi Figure ES 6. Final permanent deformation at the pavement surface............................……………………..… xii Figure 1. Layout of 20° C test sections. Goal 3 rutting sections are distributed in the area between the 20° C test sections. ............................................................................................................................... ....................... 4 Figure 2. Drip watering system for Goal 5 tests. ................................................................................................. 7 Figure 3. Layout of Goal 9 test sections. ............................................................................................................. 9 Figure 4. Example of modulus versus reduced time relationship. ..................................................................... 10 Figure 5. Modulus versus temperature for different viscosity versus temperature relationships....................... 12 Figure 6. Frequency sweep data for Goal 1 and Goal 3 materials compared to models. ................................... 13 Figure 7. Example of input parameters for the modulus versus reduced time relationship for the AC bottom layer. 21 Figure 8. Example of damage versus number of load applications. .................................................................. 23 Figure 9. Example of Equation 6 damage parameters of AC bottom layer ( Goal 1)......................................... 24 Figure 10. Simple Drucker Prager failure condition.......................................................................................... 28 Figure 11. EAC 10,000 MPa, no slip, 40 kN load. .............................................................................................. 29 Figure 12. EAC 10,000 MPa, slip, 40 kN............................................................................................................ 30 Figure 13. EAC 2,000 MPa, slip, 40 kN.............................................................................................................. 31 Figure 14. EAC 2,000 MPa, slip, 100 kN load. ................................................................................................... 33 Figure 15. Displacement field in particulate sample........................................................................................... 34 Figure 16. Displacement field in elastic solid ( FEM). ........................................................................................ 35 Figure 17. Modulus of Layer 2 as a function of the stiffness of the asphalt layers, for the undrained sections. 36 Figure 18. Modulus of Layer 2 as a function of the stiffness of the asphalt layers, for the drained sections. ... 37 Figure 19. Modulus of subgrade as a function of the stiffness of the pavement layers. .................................... 37 Figure 20. Results of RSST CH tests. [ Note: FMFC indicates field mixed field compacted specimen taken by coring the pavement. AV5.5 indicates cores with approximately 5.5 percent air voids. Each title in the legend indicates the RSST CH test temperature ( 40, 50, or 60 ° C) and average shear stress ( MPa).]...... 42 Figure 21. Normalized plastic strain versus number of load repetitions. ( Note: legend is the same as in Figure 20). 44 Figure 22. Average values for Figure 21 curves................................................................................................ 45 Figure 23. Best fitting Gamma function. ( Note: legend is the same as in Figure 20 and Figure 21. ................. 46 Stage 5 Distribution xviii UCPRC RR 2005 06 Figure 24. Input parameters for permanent deformation ( rutting) of subgrade in second column. ................... 48 Figure 25. Comparison of fitted vs. calculated strain for AC on AC overlay, 2D. ........................................... 49 Figure 26. Section 501RF pavement structure................................................................................................... 51 Figure 27. Section 501RF temperatures during testing...................................................................................... 52 Figure 28. Section 501RF 40 kN top modules deflection.................................................................................. 52 Figure 29. Section 501 RF 40 kN resilient compression of pavement layers. ................................................... 53 Figure 30. Section 501RF 40 kN deflection of subgrade................................................................................... 53 Figure 31. Section 501RF 100 kN deflection of top modules............................................................................ 54 Figure 32. Section 501RF 100 kN resilient compression of pavement layers. .................................................. 54 Figure 33. Section 501RF 100 kN deflection of subgrade................................................................................. 55 Figure 34. Section 501RF calculated moduli at 40 kN and actual temperature. ( Note: in this and all other figures showing change in elastic moduli ( E) under loading the lines are plotted for the modulus of each layer, i. e., E1 is the modulus of the first layer, E2 is the modulus of the second layer, etc.)..................... 55 Figure 35. Section 503RF pavement structure................................................................................................... 56 Figure 36. Section 503RF temperatures during testing...................................................................................... 56 Figure 37. Section 503RF 40 kN deflection of top modules.............................................................................. 57 Figure 38. Section 503RF 40 kN compression of pavement layers. .................................................................. 57 Figure 39. Section 503RF 40 kN deflection of subgrade................................................................................... 58 Figure 40. Section 503RF 100 kN deflection of top modules............................................................................ 58 Figure 41. Section 503RF 100 kN resilient compression of pavement layers. .................................................. 59 Figure 42. Section 503RF 100 kN deflection of subgrade................................................................................. 59 Figure 43. Section 503RF calculated layer moduli at 40 kN and actual temperature........................................ 60 Figure 44. Section 500RF pavement structure................................................................................................... 60 Figure 45. Section 500RF temperatures during testing...................................................................................... 61 Figure 46. Section 500RF 40 kN deflection of top modules.............................................................................. 61 Figure 47. Section 500RF 40 kN resilient compression of pavement layers. .................................................... 62 Figure 48. Section 500RF 40 kN deflection of subgrade................................................................................... 62 Figure 49. Section 500RF 100 kN deflection of top modules............................................................................ 63 Figure 50. Section 500RF 100 kN compression of pavement layers. ................................................................ 63 Figure 51. Section 500RF 100 kN deflection of subgrade................................................................................. 64 Figure 52. Section 500RF calculated moduli at 40 kN and actual temperature................................................. 64 Figure 53. Section 502CT pavement structure................................................................................................... 65 Figure 54. Section 502CT 40 kN deflection on top of AC. ............................................................................... 65 Figure 55. Section 502CT 40 kN compression of pavement layers................................................................... 66 Figure 56. Section 502CT 40 kN deflection of subgrade................................................................................... 66 Figure 57. Section 502CT 100 kN deflection at top of AC. .............................................................................. 67 Stage 5 Distribution xix UCPRC RR 2005 06 Figure 58. Section 502CT 100 kN resilient compression of pavement layers. .................................................. 67 Figure 59. Section 502CT 100 kN deflection of subgrade................................................................................. 68 Figure 60. Section 502CT calculated moduli at 40 kN and 20 º C. ..................................................................... 68 Figure 61. Cracking versus relative decrease in modulus of top AC layer for Goal 1....................................... 69 Figure 62. Goal 1, cracking versus increase in deflection. ................................................................................ 70 Figure 63. Permanent compression of AC layers. ............................................................................................. 73 Figure 64. Permanent compression of granular layers....................................................................................... 73 Figure 65. Permanent deformation of subgrade................................................................................................. 74 Figure 66. Permanent deformation at pavement surface.................................................................................... 74 Figure 67. Permanent compression of AC layers. ............................................................................................. 75 Figure 68. Permanent compression of granular layers....................................................................................... 75 Figure 69. Permanent deformation of subgrade................................................................................................. 76 Figure 70. Permanent deformation at pavement surface.................................................................................... 76 Figure 71. Permanent deformation of the AC layers. ........................................................................................ 77 Figure 72. Permanent compression of the granular layers................................................................................. 77 Figure 73. Permanent deformation of the subgrade. .......................................................................................... 78 Figure 74. Permanent deformation at pavement surface.................................................................................... 78 Figure 75. Permanent compression of the AC layers......................................................................................... 79 Figure 76. Permanent deformation of granular layers. ...................................................................................... 79 Figure 77. Permanent deformation of subgrade................................................................................................. 80 Figure 78. Permanent deformation at surface of pavement. .............................................................................. 80 Figure 79. Section 517RF pavement structure................................................................................................... 82 Figure 80. Section 517RF AC temperature during testing................................................................................. 82 Figure 81. Section 517RF 40 kN deflection of top modules.............................................................................. 83 Figure 82. Section 517RF 40 kN compression of pavement layers. .................................................................. 83 Figure 83. Section 517RF deflection of subgrade.............................................................................................. 84 Figure 84. Section 517RF 100 kN top modules deflection................................................................................ 84 Figure 85. Section 517RF 100 kN compression of pavement layers. ................................................................ 85 Figure 86. Section 517RF 100 kN deflection of subgrade................................................................................. 85 Figure 87. Section 517RF calculated moduli at 40 kN and actual temperature................................................. 86 Figure 88. Section 518RF pavement structure................................................................................................... 86 Figure 89. Section 518RF AC temperature during testing................................................................................. 87 Figure 90. Section 518RF 40 kN top modules deflection.................................................................................. 87 Figure 91. Section 518RF 40 kN resilient compression of pavement layers. .................................................... 88 Figure 92. Section 518RF 40 kN deflection of subgrade................................................................................... 88 Figure 93. Section 518RF 100 kN top modules deflection................................................................................ 89 Stage 5 Distribution xx UCPRC RR 2005 06 Figure 94. Section 518RF 100 kN compression of pavement layers. ................................................................ 89 Figure 95. Section 518RF 100 kN deflection of subgrade................................................................................. 90 Figure 96. Section 518RF calculated moduli at 40 kN and actual temperature................................................. 90 Figure 97. Section 514RF pavement structure................................................................................................... 91 Figure 98. Section 514RF AC temperature during testing................................................................................. 91 Figure 99. Section 514RF 40 kN top modules deflection.................................................................................. 92 Figure 100. Section 514RF 40 kN compression of pavement layers, MDD1 and MDD2................................. 92 Figure 101. Section 514RF 40 kN compression of pavement layers, MDD3 and MDD4................................. 93 Figure 102. Section 514RF 40 kN deflection of subgrade................................................................................. 93 Figure 103. Section 514RF 100 kN top modules deflection.............................................................................. 94 Figure 104. Section 514RF 100 kN compression of pavement layers, MDD1 and MDD2............................... 94 Figure 105. Section 514RF 100 kN compression of pavement layers, MDD3 and MDD4............................... 95 Figure 106. Section 514RF 100 kN deflection of subgrade............................................................................... 95 Figure 107. Section 514RF calculated moduli at 40 kN and actual temperature............................................... 96 Figure 108. Section 515RF pavement structure................................................................................................. 96 Figure 109. Section 515RF AC temperature during testing............................................................................... 97 Figure 110. Section 515RF 40 kN top modules deflection................................................................................ 97 Figure 111. Section 515RF 40 kN resilient compression of pavement layers. .................................................. 98 Figure 112. Section 515RF 40 kN deflection of subgrade.................................................................................. 98 Figure 113. Section 515RF 100 kN top modules deflection.............................................................................. 99 Figure 114. Section 515RF 100 kN compression of pavement layers. .............................................................. 99 Figure 115. Section 515RF 100 kN deflection of subgrade............................................................................. 100 Figure 116. Section 515RF calculated moduli at 40 kN and actual temperatures. .......................................... 100 Figure 117. Cracking in overlay versus relative decrease in modulus of overlay, Goal 3............................... 101 Figure 118. Goal 3, 20 º C, cracking versus increase in deflection.................................................................... 102 Figure 119. Section 517RF permanent deformation of AC layers................................................................... 105 Figure 120. Section 517RF permanent deformation of granular layers........................................................... 105 Figure 121. Section 517RF permanent deformation of subgrade. ................................................................... 106 Figure 122. Section 517RF permanent deformation at pavement surface. ...................................................... 106 Figure 123. Section 518RF permanent deformation of AC layers................................................................... 107 Figure 124. Section 518RF permanent deformation of granular layers........................................................... 107 Figure 125. Section 518RF permanent deformation of subgrade. ................................................................... 108 Figure 126. Section 518RF permanent deformation at pavement surface. ...................................................... 108 Figure 127. Section 514RF permanent deformation of AC layers................................................................... 109 Figure 128. Section 514RF permanent deformation of granular layers, MDD1 and MDD2........................... 109 Figure 129. Section 514RF permanent deformation of granular layers, MDD3 and MDD4........................... 110 Stage 5 Distribution xxi UCPRC RR 2005 06 Figure 130. Section 514RF permanent deformation of subgrade. ................................................................... 110 Figure 131. Section 514RF permanent deformation at pavement surface. ...................................................... 111 Figure 132. Section 515RF permanent deformation of AC layers................................................................... 112 Figure 133. Section 515RF permanent deformation of granular layers........................................................... 112 Figure 134. Section 515RF permanent deformation of subgrade. ................................................................... 113 Figure 135. Section 515RF permanent deformation of the pavement surface................................................. 113 Figure 136. Section 504RF pavement structure............................................................................................... 115 Figure 137. Section 504RF permanent deformation at pavement surface from profilometer.......................... 115 Figure 138. Section 504RF calculated permanent deformation of pavement layers........................................ 116 Figure 139. Section 505RF pavement structure............................................................................................... 117 Figure 140. Section 505RF temperatures during testing.................................................................................. 117 Figure 141. Section 505RF permanent deformation at pavement surface from profilometer.......................... 118 Figure 142. Section 505RF calculated permanent deformation of pavement layers........................................ 118 Figure 143. Section 506RF pavement structure............................................................................................... 119 Figure 144. Section 506RF temperatures during testing.................................................................................. 119 Figure 145. Section 506RF permanent deformation at pavement surface from profilometer.......................... 120 Figure 146. Section 506RF calculated permanent deformation of pavement layers........................................ 120 Figure 147. Section 507RF pavement structure............................................................................................... 121 Figure 148. Section 507RF temperatures during testing.................................................................................. 121 Figure 149. Section 507RF permanent deformation at pavement surface from profilometer.......................... 122 Figure 150. Section 507RF calculated permanent deformation of pavement layers........................................ 122 Figure 151. Section 508RF pavement structure............................................................................................... 123 Figure 152. Section 508RF temperatures during testing.................................................................................. 123 Figure 153. Section 508RF permanent deformation at pavement surface from profilometer.......................... 124 Figure 154. Section 508RF calculated permanent deformation of pavement layers........................................ 124 Figure 155. Section 509RF pavement structure............................................................................................... 125 Figure 156. Section 509RF temperatures during testing.................................................................................. 125 Figure 157. Section 509RF permanent deformation at pavement surface from profilometer.......................... 126 Figure 158. Section 509RF calculated permanent deformation of pavement layers........................................ 126 Figure 159. Section 510RF pavement structure............................................................................................... 127 Figure 160. Section 510RF temperatures during testing.................................................................................. 127 Figure 161. Section 510RF permanent deformation at pavement surface from profilometer.......................... 128 Figure 162. Section 510RF calculated permanent deformation of pavement layers........................................ 128 Figure 163. Section 511RF pavement structure............................................................................................... 129 Figure 164. Section 511RF temperatures during testing.................................................................................. 129 Figure 165. Section 511RF permanent deformation at the pavement surface. ................................................ 130 Stage 5 Distribution xxii UCPRC RR 2005 06 Figure 166. Section 511RF calculated permanent deformation of the pavement layers.................................. 130 Figure 167. Section 512RF pavement structure............................................................................................... 131 Figure 168. Section 512RF temperatures during testing.................................................................................. 131 Figure 169. Section 512RF permanent deformation at pavement surface from profilometer.......................... 132 Figure 170. Section 512RF calculated permanent deformation of the pavement layers.................................. 132 Figure 171. Section 513RF pavement structure............................................................................................... 133 Figure 172. Section 513RF permanent deformation at pavement surface from profilometer.......................... 134 Figure 173. Section 513RF calculated permanent deformation of pavement layers........................................ 134 Figure 174. Cores from trafficked area of Section 543RF after HVS loading show stripping and disintegration of ATPB, as well as signs of moisture damage between the the three lifts of asphalt concrete ( Bejarano et al. 2003). ............................................................................................................................... .................. 135 Figure 175. Section 543RF pavement structure............................................................................................... 136 Figure 176. Section 543RF temperatures during testing.................................................................................. 137 Figure 177. Section 543RF Road Surface Deflectometer, at 40 kN. ............................................................... 137 Figure 178. Section 543RF Road Surface Deflectometer, at 100 kN. ............................................................. 138 Figure 179. Section 543RF 40 kN top module. ............................................................................................... 138 Figure 180. Section 543RF 40 kN top of aggregate base. ............................................................................... 139 Figure 181. Section 543RF 40 kN top of aggregate subbase........................................................................... 139 Figure 182. Section 543RF 40 kN deflection of subgrade ( 850 mm depth). .................................................... 140 Figure 183. Section 543RF 100 kN top module. ............................................................................................. 140 Figure 184. Section 543RF 100 kN top of aggregate base. ............................................................................. 141 Figure 185. Section 543RF 100 kN top of aggregate subbase......................................................................... 141 Figure 186. Section 543RF 100 kN deflection of subgrade ( 850 mm depth). .................................................. 142 Figure 187. Section 543RF Permanent deformation of asphalt layers............................................................. 142 Figure 188. Section 543RF permanent deformation of granular layers plus top of subgrade.......................... 143 Figure 189. Section 543RF permanent deformation in subgrade ( 850 mm depth)........................................... 143 Figure 190. Section 543RF permanent deformation at pavement surface. ...................................................... 144 Figure 191. Section 543RF calculated layer moduli, at 40 kN and actual temperatures. ................................ 144 Figure 192. Section 544RF pavement structure............................................................................................... 145 Figure 193. Section 544RF temperatures during testing.................................................................................. 145 Figure 194. Section 544RF Road Surface Deflectometer, at 40 kN. ............................................................... 146 Figure 195. Section 544RF Road Surface Deflectometer, at 100 kN. ............................................................. 146 Figure 196. Section 544RF 40 kN top module. ............................................................................................... 147 Figure 197. Section 544RF 40 kN top of aggregate base. ............................................................................... 147 Figure 198. Section 544RF 40 kN top of aggregate subbase........................................................................... 148 Figure 199. Section 544RF 100 kN top module. ............................................................................................. 148 Stage 5 Distribution xxiii UCPRC RR 2005 06 Figure 200. Section 544RF 100 kN top of aggregate base. ............................................................................. 149 Figure 201. Section 544RF 100 kN top of aggregate subbase......................................................................... 149 Figure 202. Section 544RF Permanent deformation of asphalt layers............................................................. 150 Figure 203. Section 544RF permanent deformation of aggregate base. .......................................................... 150 Figure 204. Section 544RF permanent deformation on top of basecourse. ...................................................... 151 Figure 205. Section 544RF permanent deformation at pavement surface. ...................................................... 151 Figure 206. Section 544RF calculated moduli of pavement layers................................................................... 152 Figure 207. Section 545RF pavement structure............................................................................................... 153 Figure 208. Section 545RF temperatures during testing.................................................................................. 153 Figure 209. Section 545RF Road Surface Deflectometer, at 40 kN. ............................................................... 154 Figure 210. Section 545RF Road Surface Deflectometer, 100 kN. ................................................................. 154 Figure 211. Section 545RF 40 kN top module. ............................................................................................... 155 Figure 212. Section 545RF 40 kN top of aggregate base. ............................................................................... 155 Figure 213. Section 545RF 40 kN top of aggregate subbase........................................................................... 156 Figure 214. Section 545RF 100 kN top module. ............................................................................................. 156 Figure 215. Section 545RF 100 kN top of aggregate base. ............................................................................. 157 Figure 216. Section 545RF 100 kN top of aggregate subbase......................................................................... 157 Figure 217. Section 545RF permanent deformation of asphalt layers............................................................. 158 Figure 218. Section 545RF permanent deformation of aggregate base. .......................................................... 158 Figure 219. Section 545RF permanent deformation at pavement surface. ...................................................... 159 Figure 220. Section 545RF calculated moduli of pavement layers.................................................................. 159 Figure 221. Visual cracking versus relative decrease in modulus of layer 1, Goal 5 Wet conditions. ............. 160 Figure 222. Goal 5, cracking versus increase in deflection. ............................................................................ 161 Figure 223. MB road, AC modulus versus reduced time parameters from frequency sweep. ......................... 162 Figure 224. Moduli from FWD compared to frequency sweep tests, Goal 9 ( MB road). ................................ 163 Figure 225. MB road, damage parameters for AC in first column. .................................................................. 165 Figure 226. MB road, backcalculated modulus of AB versus time. ................................................................ 166 Figure 227. MB road, modulus of AB versus stiffness of AC......................................................................... 166 Figure 228. MB road, subgrade modulus versus stiffness of pavement layers. ................................................ 167 Figure 229. Section 567RF pavement structure................................................................................................ 169 Figure 230. Section 567RF load levels. ........................................................................................................... 169 Figure 231. Section 567RF temperatures during testing................................................................................... 170 Figure 232. Section 567RF Road Surface Deflectometer................................................................................. 170 Figure 233. Section 567RF MDDs at 90mm and 330 mm. ............................................................................. 171 Figure 234. Section 567RF permanent deformation of MDDs......................................................................... 171 Figure 235. Section 567RF permanent deformation at pavement surface from profilometer........................... 172 Stage 5 Distribution xxiv UCPRC RR 2005 06 Figure 236. Section 567RF calculated moduli at 40 kN and actual temperature.............................................. 172 Figure 237. Section 568RF pavement structure................................................................................................ 173 Figure 238. Section 568RF temperatures during testing................................................................................... 173 Figure 239. Section 568RF Road Surface Deflectometer................................................................................. 174 Figure 240. Section 568RF permanent deformation at pavement surface from profilometer........................... 174 Figure 241. Section 568RF calculated layer moduli at 40 kN and actual temperature..................................... 175 Figure 242. Section 573RF pavement structure................................................................................................ 176 Figure 243. Section 573RF temperatures during testing................................................................................... 176 Figure 244. Section 573RF Road Surface Deflectometer................................................................................. 177 Figure 245. Section 573RF permanent deformation at pavement surface from profilometer........................... 177 Figure 246. Section 573RF calculated layer moduli at 40 kN and actual temperature..................................... 178 Figure 247. Section 571RF pavement structure................................................................................................ 179 Figure 248. Section 571RF temperatures during testing................................................................................... 179 Figure 249. Section 571RF Road Surface Deflectometer................................................................................. 180 Figure 250. Section 571RF MDDs at 90 mm, 300 mm, and 525 mm. ............................................................. 180 Figure 251. Section 571RF permanent deformation of MDDs......................................................................... 181 Figure 252. Section 571RF permanent deformation at the pavement surface. ................................................. 181 Figure 253. Section 571RF calculated layer moduli at 40 kN and actual temperature..................................... 182 Figure 254. Section 572RF pavement structure................................................................................................ 183 Figure 255. Section 572RF temperatures during testing................................................................................... 183 Figure 256. Section 572RF Road Surface Deflectometer................................................................................. 184 Figure 257. Section 572RF permanent deformation at the pavement surface. ................................................. 184 Figure 258. Section 572RF calculated layer moduli at 40 kN and actual temperature..................................... 185 Figure 259. Section 569RF pavement structure................................................................................................ 186 Figure 260. Section 569RF temperatures during testing................................................................................... 186 Figure 261. Section 569RF Road Surface Deflectometer................................................................................. 187 Figure 262. Section 569RF MDDs at 90 mm, 300 mm, and 525 mm.............................................................. 187 Figure 263. Section 569RF permanent deformation of MDDs......................................................................... 188 Figure 264. Section 569RF permanent deformation at pavement surface from profilometer........................... 188 Figure 265. Section 569RF calculated layer moduli at 40 kN and actual temperature..................................... 189 Figure 266. Cracking versus relative decrease in modulus of AC layer for Goal 9 ( MB road)........................ 190 Figure 267. Goal 9 ( MB road), Cracking versus increase in deflection ( RSD) ................................................ 190 Figure 268. Ratio of final over initial deflection. ............................................................................................. 195 Figure 269. Cracking versus calculated decrease in modulus of top layer. ...................................................... 197 Figure 270. Cracking versus increase in deflection. ......................................................................................... 198 Figure 271. Measured and predicted final permanent deformation of asphalt................................................. 199 Stage 5 Distribution xxv UCPRC RR 2005 06 Figure 272. Final permanent deformation of granular layers. ......................................................................... 201 Figure 273. Final permanent deformation of the subgrade. .............................................................................. 203 Figure 274. Final permanent deformation at the pavement surface ( profile data). ........................................... 205 Figure 275. Modulus of AB layer backcalculated from FWD tests in center line. ........................................... 217 Figure 276. Pavement structure for Section 569RF.......................................................................................... 219 Figure 277. Damage parameters used for DGAC of Section 569RF................................................................ 219 Figure 278. Damage parameters used for CTB of Section 569RF.................................................................... 220 Figure 279. Section 569RF Road Surface Deflectometer................................................................................. 220 Figure 280. Section 569RF MDD resilient deflections..................................................................................... 221 Figure 281. Section 569RF permanent MDD deformations. ............................................................................ 221 Figure 282. Section 569RF average permanent deformation from pavement profile....................................... 222 LIST OF TABLES Table 1. Summary List of HVS Tests ................................................................................................................... 3 Table 2. Design Thicknesses for Goal 1 Sections ................................................................................................. 5 Table 3. As built Thicknesses for Goal 5 Sections ............................................................................................... 6 Table 4. As built Thicknesses of Goal 9 Sections ................................................................................................ 8 Table 5 Influence of Slip Value in LEAP on Calculated Vertical Deflections and Horizontal Strains ............. 25 Table 6. Triaxial Tests on Subgrade.................................................................................................................... 26 Table 7. Triaxial Tests on Aggregate Base ( AB) ................................................................................................ 26 Table 8. Layer Thicknesses Used for FWD Backcalculation.............................................................................. 36 Table 9. Moduli Parameters from FWD ............................................................................................................. 38 Table 10. Moduli Parameters from Calibration to MDD Deflections................................................................. 38 Table 11. Initial Moduli Used in HVS Simulations ( MPa)................................................................................. 39 Table 12. Summary of Moduli ( MPa)................................................................................................................. 39 Table 13. Parameters Used in Equation 20......................................................................................................... 41 Table 14. AC on AC, 2D Structural Parameter Combinations ........................................................................... 49 Table 15. Initial Moduli MPa ( 1.8 km/ h, 40 kN, Actual Temperature) for Each Section................................... 71 Table 16. Final Moduli MPa ( 1.8 km/ h, 40 kN, Actual Temperature) for Each Section.................................... 71 Table 17. Percentage Decrease in Layer Moduli for Each Section..................................................................... 72 Table 18. Damage Parameter for Asphalt Layers at End of Test for Each Section ............................................ 72 Table 19. Initial Damage Parameters for “ Old” Asphalt Overlay Sections ........................................................ 81 Table 20. Layered Moduli at Start of Test, MPa .............................................................................................. 103 Table 21. Layer Moduli at End of Test, MPa ................................................................................................... 103 Table 22. Percentage Decrease in Moduli ........................................................................................................ 104 Table 23. Initial Permanent Deformations, Goal 3, in mm............................................................................... 104 Stage 5 Distribution xxvi UCPRC RR 2005 06 Table 24. Tire Types and Pressure, MPa .......................................................................................................... 114 Table 25. Estimated Temperatures for Section 513RF ..................................................................................... 133 Table 26. Parameters for FWD Moduli versus Reduced Time, Goal 9 ............................................................ 164 Table 27. MB Road, Stiffness Parameters for AB............................................................................................ 167 Table 28. Summary of Damage Equations and Shift Factors Used in All Simulations.................................... 192 Table 29. Measured and Calculated Road Surface Deflectometer Deflections ( RSD), in mm......................... 193 Table 30. Measured and Calculated Deflections of the Top Multi depth Deflectometer ( MDD), in mm........ 194 Table 31. Final Permanent Deformation of Asphalt, in mm............................................................................. 200 Table 32. Final Permanent Deformation of Granular Layer, in mm................................................................. 202 Table 33. Parameters used for Granular Materials in Equation ( 29) ................................................................ 202 Table 34. Final Permanent Deformation of Subgrade, in mm.......................................................................... 204 Table 35. Parameters Used for Subgrade in Equation ( 29)............................................................................... 204 Table 36. Final Permanent Deformation at the Pavement Surface, in mm....................................................... 206 Stage 5 Distribution 1 UCPRC RR 2005 06 1.0 INTRODUCTION The first step in creating a Mechanistic Empirical ( ME) pavement design or evaluation is to calculate pavement response — in terms of stresses, strains, and/ or displacements — using a mathematical ( or mechanistic) model. In the second step, the calculated response is used as a variable in empirical relationships to predict structural damage ( decrease in moduli or cracking) and functional damage ( rutting and roughness) to the pavement. Both of these steps must be reasonably correct. If the calculated response bears little resemblance to the pavement’s actual response, there is no point in trying to use the calculation to predict future damage to the pavement with the empirical relationship. In other words, only if the calculated response is reasonably correct does it make sense to try to relate the damage to the pavement response. 1.1 Models and Approaches Included in CalME This report presents the modeling of several series of flexible pavement Heavy Vehicle Simulator ( HVS) tests using the set of distress models included in the draft software package, CalME. These models are for the flexible pavement distresses typically observed in California: asphalt fatigue, asphalt rutting, unbound layers rutting and reflection cracking. CalME software provides the user with four approaches for evaluating or designing a flexible pavement structure: • Caltrans current methods, the R value method for new flexible structures, and the deflection reduction method for overlay thickness design for existing flexible structures. • “ Classical” Mechanistic Empirical ( ME) design, largely based on the Asphalt Institute method. This method uses a standard Equivalent Single Axle Load ( ESAL) for the traffic load, one temperature to characterize the entire range of temperatures the asphalt concrete ( AC) layer will experience, and the Asphalt Institute fatigue and unbound layers rutting equations, with an adjustment for air void content and binder content in the asphalt concrete. • An Incremental method, using the typical Miner’s Law approach, permitting damage calculation for the axle load spectrum and expected temperature regimes, but with no updating of materials properties through the life of the project. This is similar to the approach included in the NCHRP 1 37A Design Guide, also referred to as the Mechanistic Empirical Pavement Design Guide ( MEPDG). This type of approach is calibrated against an end failure state, such as 25 percent cracking of the wheelpath, and it assumes a linear accumulation of damage to get to that state. • An Incremental Recursive method in which the materials properties for the pavement are updated in terms of damage as the simulation of the pavement life progresses. The Incremental Recursive approach was used for the simulations included in this report, and is the only approach that can provide an accurate indication of pavement condition at different points during the pavement’s life. The research team proposes that pavement designers should begin their designs by applying either an existing Caltrans method or the Classical method. In CalME both of these options perform a “ design” function, calculating and presenting pavement structures that meet design requirements for a predetermined number of traffic loads. Then, the lowest cost alternatives in the set of candidate pavement structures meeting the design requirements with either of these methods should be checked by the designer with the more comprehensive and precise Incremental Recursive method to be certain that they meet the design requirements. Once a final design has been selected, its Incremental Recursive output can be used to provide a prediction of the pavement’s condition across its entire life. Some distresses and some materials are not considered in either the Caltrans or Classical methods, and can only be evaluated using the Incremental Miner’s Law approach or the Incremental Recursive approach. Stage 5 Distribution 2 UCPRC RR 2005 06 1.1.1 Validation Using Heavy Vehicle Simulator Data The Incremental Recursive models included in CalME were used to predict the performance of all the flexible pavement HVS tests performed to date as part of the Accelerated Pavement Testing ( APT) program operated for the California Department of Transportation ( Caltrans) by the University of California Pavement Research Center ( UCPRC). The HVS test data presented in this report come from tests performed between 1995 and 2004. The HVS response data and the corresponding laboratory test data were extracted from the UCPRC HVS database. HVS tests measure pavement response in terms of deflections, either at the pavement surface, using a Road Surface Deflectometer ( RSD), at multiple depths, using a Multi depth Deflectometer ( MDD), or both. The RSD is very similar to the Benkelman Beam used in the development of the Caltrans new flexible pavement and overlay design methods in the 1950s. In predicting the gradual degradation of the pavement it is important that the response model provides a reasonably accurate prediction of measured deflections. Although a correct prediction of deflections by the response model is no guarantee that it can also correctly predict the stresses and strains in all of the pavement layers, the opposite is true: if the model inaccurately predicts deflections, it will also provide inaccurate predictions of stresses and strains. Therefore, in trying to calibrate the ME models from HVS testing, the research team’s first concern was to make sure that the model predicted resilient deflections reasonably well for the duration of the test and for all load levels. This prediction depended on the moduli ( often referred to as “ stiffnesses” in this report and in the literature) of all of the pavement layers and on the changes to these moduli caused by fatigue damage, slip between asphalt layers, non linear elastic characteristics of the unbound layers, and the effect of confinement on granular layers. Once reasonable agreement was achieved between the measured resilient deflections and the calculated ones, then models of permanent deformation could be calibrated with some confidence. There are different methods for determining moduli and there are often differences in the results from each method ( which should be expected based on the literature.) The methods used in this study included backcalculation of moduli for all layers from Falling Weight Deflectometer ( FWD) and MDD deflection data , and direct measurement of moduli in the laboratory using triaxial tests for unbound materials and flexural frequency sweep tests for asphaltic materials. Differences in measured moduli across the different methods are due to variations in boundary conditions, strain levels, and loading times between the different measurement methods, the effects of which vary among materials. The FWD does not fit under the HVS, so there is no FWD data during an HVS test, there is only FWD data from before the HVS was placed on the pavement and from after the HVS was taken off the trafficked section. The simulations in this report primarily relied on stiffnesses for the asphalt materials taken from flexural frequency sweep data, and stiffnesses for the unbound layers taken from backcalculation of MDD deflection data. In practice, the research team views backcalculation using deflections from the FWD as the primary tool for obtaining the stiffnesses of layers in existing pavements, as opposed to laboratory testing of materials samples taken from the already constructed pavement. FWD deflections and backcalculation take into consideration the stiffness of the layers as they occur in the constructed pavement structure, including the effects of boundary conditions, water and temperature conditions, previous traffic and environmental conditioning, and interaction between layers acting as a system in the in place pavement structure. This is important because most of Caltrans future work will be rehabilitating and reconstructing pavements already in service. The research team sees the flexural beam test as the primary tool for measuring the stiffness and the fatigue characteristics of asphalt overlay materials for new layers. For new pavement construction, the team sees the use of databases of moduli for granular bases and subbases and for subgrades backcalculated from FWD tests on existing pavements, with the materials referenced by characteristics such as the Unified Soil Classification System ( USCS) classification and relative density. The databases should also include some laboratory triaxial tests for these materials, for comparison with any new base, subbase, and subgrade materials Stage 5 Distribution 3 UCPRC RR 2005 06 for which there is no previous FWD testing, and for which laboratory triaxial testing must be used to measure stiffness. The purpose of this study was to compare the overall trends shown by the damage models in simulations of the HVS tests against the actual trends measured in the same HVS tests. Asphalt concrete stiffnesses from flexural frequency sweep data were used in the simulations of the HVS tests, and the stiffnesses of the underlying moduli were adjusted from their initial values as the asphalt concrete stiffness changed with damage so they would match the measured and calculated deflections. The results presented herein show that, overall, the damage trends for deflection and permanent deformation under loading were verified. During HVS testing, deflections often increase markedly, with deflection sometimes rising more than twice as high at the end of the test than they were at the beginning. The flexible pavement design model of the NCHRP 1 37A Design Guide ( NCHRP 2004) does not consider any decrease in the asphalt concrete modulus as a result of fatigue damage ( except for rehabilitation designs). In fact, the NCHRP 1 37A Design Guide includes a model for aging that predicts a continuous increase in the stiffness of asphalt concrete layers across the life of the pavement, resulting in smaller predicted deflections as the pavement is subjected to trafficking. While aging is potentially important, the effect of updating stiffness for aging and not updating it for fatigue damage results in the calculation of unrealistic elastic responses during the pavement life. This may be acceptable for pavements with extremely thick asphalt concrete layers where little fatigue should occur, but it is impossible to use the model to simulate an HVS test and, inversely, to use HVS tests to calibrate the model. The HVS test series in this report were grouped by “ goals,” which are defined as follows: Table 1. Summary List of HVS Tests Goal General Conditions HVS Test Numbers Original Report References Goal 1: Comparison of structures with and without ATPB layer under dry conditions, moderate temperatures New pavement, dry conditions, 20° C 500RF, 501RF, 502CT, 503RF 14, 15, 16, 17, 18 Goal 3 Cracking: Comparison of reflection cracking performance of ARHM GG and DGAC overlays Overlays of cracked Goal 1 sections, dry conditions, 20° C 514RF, 515RF 517RF, 518RF 8, 11, 13 Goal 3 Rutting: Comparison of rutting performance of ARHM GG and DGAC overlays Overlays of previously untrafficked areas of Goal 1 pavements, dry conditions, 40° C or 50° C at 50 mm depth, four different tire/ wheel types 504RF, 505RF 506RF, 507RF 508RF, 509RF 510RF, 511RF 512RF, 513RF 7, 10 Goal 5: Comparison of structures with and without ATPB layer under wet conditions, moderate temperatures New pavement, wet conditions, 20° C 543RF, 544RF, 545RF 2, 3, 4, 5, 13 Goal 9: Initial cracking of asphalt pavement in preparation for later overlay New pavement, ambient rainfall, 20° C 567RF, 568RF, 569RF, 571RF, 572RF, 573RF 1 ATPB: Asphalt treated permeable base. ARHM GG: Asphalt rubber hot mix gap graded. DGAC: Dense graded asphalt concrete. Stage 5 Distribution 4 UCPRC RR 2005 06 The Goal 1, 3, and 5 tests were performed inside a metal shed built over native subgrade. The shed provided protection from sun, wind, and rain; however, changes in subgrade water content and the depth to the water table were recorded during HVS tests. Goal 9 tests were performed on a road with no cover other than what the HVS and its temperature cabinet provided. The remainder of this chapter presents the general descriptions of the HVS tests and the models used to simulate them. 1.2 HVS tests 1.2.1 Goal 1 and Goal 3 Tests Figure 1 shows the layout of the Goal 1 and Goal 3 cracking sections. Figure 1. Layout of 20° C test sections. Goal 3 rutting sections are distributed in the area between the 20° C test sections. All the sections in Goal 1 had two layers of asphalt concrete ( AC), an aggregate base ( AB), and an aggregate subbase ( ASB). In the two drained sections, part of the AB layer thickness was replaced by an asphalt treated permeable base ( ATPB) at a ratio of 1.4: 1.1. A constant temperature of about 20 º C was maintained during Goal 1. The design layer thicknesses are shown in Table 2. 518RF 517RF 515RF 514RF 503RF 501RF 502CT 500RF Undrained Drained ARHM GG DGAC Goal1 Goal3 Stage 5 Distribution 5 UCPRC RR 2005 06 Table 2. Design Thicknesses for Goal 1 Sections Layer Undrained ( mm) Drained ( mm) AC top lift 61 61 AC bottom lift 76 76 ATPB None 76 AB 274 182 ASB 229 229 The subgrade was clay, with varying plasticity across the pavements. Goal 3 was an overlay study with an aphalt rubber hot mix gap graded ( ARHM GG) ( the acronym ARHM, asphalt rubber hot mix gap graded, refers to the material specification at the time of construction in April 1997.) concrete and a dense graded asphalt concrete ( DGAC). The four Goal 3 Cracking tests were performed at the same temperature as Goal 1 ( 20 º C), with similar wheel types and loads. For these tests the design thickness of the DGAC overlay ( 75 mm) was approximately twice as thick as that of the ARHM ( 40 mm), and the overlays were placed over the four previously cracked Goal 1 sections. The remainder of Goal 3 was done at higher temperatures, with the two overlays placed on previously untested areas of the Goal 1 pavement, and several overlay thicknesses, tire types, and wheel loads used in the testing. ( Table 1 provides the reference numbers for the reports containing the details regarding thickness, tire types, and wheels loads.) All the HVS tests at 20 º C started with a wheel load of 40 kN, which was increased stepwise to 100 kN. Most of the load applications were at 100 kN. Bias ply tires on a dual wheel were used for the Goal 1 tests. The same radial tires on a dual wheel were used for the Goal 3 Cracking tests, and the Goal 5 and Goal 9 tests. Various tires and wheels were used for the Goal 3 Rutting tests, with the load for the entire duration of all but one test fixed at 40 kN. [ De Beer and Fisher ( 1997) describe the details of the tire contact stresses measured.] For all the tests except the Goal 3 Rutting tests, the wheel load was a dual wheel with a centerline distance of 305 mm and an assumed tire pressure of 690 kPa for all load levels. It was assumed that the wheels distributed the load over two circular areas; this assumption was reasonably correct for the low load level ( 40 kN) but not for the high load level ( 100 kN, where the actual load distribution was closer to two rectangles with one side twice the length of the other). All the 20° C sections were instrumented with MDDs. Each section in Goal 1 had two MDDs. Each section had two additional MDDs installed for Goal 3 testing. All MDD anchor depths were assumed to be 3,000 mm. Not all MDD modules functioned for the duration of the tests. The as constructed layer thicknesses given in Harvey et al. ( 1999) were used for the analyses of the Goal 1 and Goal 3 results presented in this report. As built layer thicknesses given in Bejarano et al. ( 2003) and Bejarano et al. ( 2005) were used to analyze the Goal 5 and Goal 9 results respectively. The remaining data was imported from the UC Pavement Research Center database, taken from a subset database named PRCHVS. mdb. Although actual wheel speeds varied, they were assumed to be 7.6 km/ h during HVS testing and 1.8 km/ h during deflection measurement on the MDDs. All tests other than the Goal 3 Rutting tests were performed with bidirectional loading. The Goal 3 Rutting tests were performed with unidirectional loading. The temperature of the test sections was controlled by a “ temperature control box.” The actual temperatures of the asphalt layers were recorded and used in the simulations. Poisson’s ratio was assumed to be 0.35 for all layers. Stage 5 Distribution 6 UCPRC RR 2005 06 1.2.2 Goal 5 Tests The Goal 5 tests were performed on the overlay structures of Goal 3 in locations where the overlay had not been placed on previously trafficked and cracked Goal 1 pavement. The designed structures were: • Section 543RF drained with ATPB and 40 mm ARHM GG wearing surface, • Section 544RF undrained ( no ATPB) and 40 mm ARHM GG wearing surface, and • Section 545RF undrained ( no ATPB) and 75 mm DGAC wearing surface. The structures’ as built thicknesses are shown in Table 3. Table 3. As built Thicknesses for Goal 5 Sections Layer Section 543RF ( mm) Section 544RF ( mm) Section 545RF ( mm) Wearing course 36 51 90 AC ( two lifts combined) 140 149 143 ATPB 64 none none AB 180 272 259 ASB 223 205– 310 206– 280 All of the test sections had a 2 percent transverse gradient and an approximate 0.5 percent longitudinal gradient in all the layers above the subbase. Holes with a diameter of 38 mm were drilled through the asphalt concrete layers on the uphill side of the three HVS test sections, and a drip watering system was installed to continuously put water into the pavement. Holes were drilled into the top of the ATPB layer of Section 543RF so that water entered into that layer. Sections 544RF and 545RF had holes drilled into the top of their aggregate bases, and water entered those layers. The water flow was greater into Section 543RF because of the initial high permeability of its ATPB layer. Considerably less water flowed into the other two sections because of the relative impermeability of their AB layer. Figure 2 shows the drip watering system. Each pavement section had water introduced into it for more than a month prior to HVS loading. This allowed the section to reach an approximate steady state moisture condition. Goal 5 testing used the same dual wheel, radial tire configuration as Goal 3 testing. Similar loading patterns and the same temperature control provided a basis for comparing the results from Goal 3 and Goal 5 tests to the result of Goal 1 ( dry condition) testing. In addition, two MDDs were installed in each pavement section, with depths similar to those used for Goal 1. Stage 5 Distribution 7 UCPRC RR 2005 06 Figure 2. Drip watering system for Goal 5 tests. 1.2.3 Goal 9 Tests Six HVS tests were performed on what were designed to be identical pavement structures. The primary purpose of the six tests was to provide fatigue cracked sections for subsequent placement of different kinds of overlay for HVS reflection cracking tests; these would be similar to the work performed in the Goal 3 cracking tests. The pavements were built so that they aligned with an existing access road. When the existing structure was removed, its subgrade was compacted to state standards. ( Figure 3 shows the layout of the six Stage 5 Distribution 8 UCPRC RR 2005 06 test sections.) The structures’ design thicknesses were: 410 mm of AB with 90 mm of dense graded asphalt concrete ( DGAC). ( Table 4 shows the as built thicknesses.) The dual wheel and radial tire configuration used for the Goal 3 tests was used for Goal 9. Loads were generally lower than those used during Goals 1, 3, and 5 to minimize rutting. Temperatures were maintained close to 20° C, as on the other cracking tests. Two MDDs were placed in each test section. However, most of the MDD modules below the surface either never functioned or failed during the test. Table 4. As built Thicknesses of Goal 9 Sections Layer 567RF 568RF 569RF 571RF 572RF 573RF AC 78 80 81 82 78 76 AB 352 349 337 352 349 337 1.3 Response and Damage Models Response and damage models are presented in this section, and they are discussed in terms of the materials properties common to the Goal 1, 3, and 5 tests. The materials properties of the Goal 9 tests are discussed with the description of the simulations for those tests later in this report. 1.3.1 Asphalt Modulus Asphalt modulus was determined as a function of temperature and loading time, using the NCHRP 1 37A Design Guide model ( NCHRP 2004): ( ) ( ( tr)) E 1 exp log log β γ α δ + + = + ( 1) where E is the modulus in MPa, tr is reduced time in sec, α, β, γ, and δ are constants, and logarithms are to base 10. Reduced time is found from: aTg ref visc visc lt tr ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ = × ( 2) where lt is the loading time ( in sec), viscref is the binder viscosity at the reference temperature, visc is the binder viscosity at the present temperature, and aTg is a constant. Stage 5 Distribution 9 UCPRC RR 2005 06 Figure 3. Layout of Goal 9 test sections. Stage 5 Distribution 10 UCPRC RR 2005 06 100 1000 10000 100000 0.00001 0.0001 0.001 0.01 0.1 1 10 100 1000 10000 Reduced time, sec Modulus, MPa α/ 2 tr ( ) min δ= log E max min log E E ⎛⎜ ⎞⎟ ⎝ ⎠ α = Figure 4. Example of modulus versus reduced time relationship. Stage 5 Distribution 11 UCPRC RR 2005 06 Equation 1 may also be written: ( ) ( ) max min ' min ' / 2 / 2 log log log 1 1 E E E E tr tr tr tr γ γ α α α δ ⎛⎜ ⎞⎟ = + = + ⎝ ⎠ +⎛⎜ ⎞⎟ +⎛⎜ ⎞⎟ ⎝ ⎠ ⎝ ⎠ ( 3) where ( ) ( ( )) / 2 ' ln 10 and tr exp ln 10 α γ γ βγ = = − × trα/ 2 is the reduced time corresponding to log( E) = δ + α/ 2, as indicated in Figure 4. Equation 1 is normally used with frequency sweep data to characterize the master curve. The form of the master curve equation shown in Equation 2 provides some insight, and can be used if Emax and Emin were known. Emax is related to the limiting stiffness of asphalt binder at temperatures below the glass transition temperature. Emin appears to depend on the boundary conditions under which it was measured, with different values coming from backcalculation of in situ pavements and beam fatigue tests. From Equation 2 it can be seen that changing trα/ 2 will shift the curve left or right and changing γ’ will change the curvature. The loading time is determined from the speed of the wheel ( input on the incremental design screen in CalME) and from the depth at which the loading time is desired. Loading time is a rather uncertain notion, as it will vary for different types of responses. For example, the loading time for transverse strain will be much longer than it is for longitudinal strain because of the actual shape of the contact area of the tire on the pavement, which is longer in the longitudinal direction than the transverse direction. The reason for the longer loading time for transverse strain is that the transverse strain is tangential to the load, whereas the longitudinal is radial and therefore has a sign change. The loading time is calculated from ( 200 mm + 2×depth)/( wheel speed in mm/ sec). The reference loading time is 0.015 sec ( 15 msec, roughly corresponding to the loading time of a standard FWD, where loading time refers to a creep test), and the reference temperature is 20 º C. The NCHRP 1 37A Design Guide makes use of an “ effective depth” based on the equivalent thickness of the layers, which results in longer loading times. The guide, however, then converts loading time into frequency, using f = 1/ lt, rather than f = 1/( 2πlt), more than compensating for the longer loading time ( unless loading time is defined differently, i. e., it is not based on a creep test). Viscosity is found from: log( log( visc cPoise)) = A + VTS * log( tK ) ( 4) where tK is the temperature ( in ° K), and A and VTS are constants, and cPoise indicates units of centipoise For all of the asphalt materials in this report a value of A = 9.6307 [ 10.5254 with temperature in º R ( degrees Rankine)] and VTS =  3.5047 were used. These values correspond ( according to the NCHRP 1 37A Design Guide) to an asphalt with a penetration grade of 40– 50. If the minimum modulus, Emin, the maximum modulus, Emax, and the modulus at two different temperatures are known, the viscosity versus temperature relationship ( Equation 4) will have very little influence on the modulus versus temperature relationship. This can be seen in Figure 5, where the resulting modulus versus temperature relationship is shown for asphalts with penetration grades from 40– 50 to 200– 300 and for a PG64 22 grade asphalt. Stage 5 Distribution 12 UCPRC RR 2005 06 Influence of viscosity 0 2000 4000 6000 8000 10000 12000 14000  20 0 20 40 60 Temperature, C Modulus, MPa 40 50 60 70 85 100 120 150 200 300 PG64 22 Figure 5. Modulus versus temperature for different viscosity versus temperature relationships. The constants δ, β, γ, and aTg, and the modulus at the reference temperature ( 20 º C) were derived from flexural frequency sweep tests. The constant a is calculated by the program. The frequency sweep tests were available for the top and bottom asphalt layers of Goal 1 and for the overlays in Goal 3. The fit between frequency sweep data and model data is shown in Figure 6. In fitting the model to the frequency sweep data it was assumed that the minimum modulus ( 10δ) was 200 MPa ( δ = 2.3010). In the frequency sweep test the measured modulus was considerably lower. However, based on FWD testing, it was assumed that an asphalt layer’s modulus, even at very low frequencies and high temperatures, had a minimum value greater than the one measured in the frequency sweep test on a flexural beam. This variance is attributable to the differences in boundary conditions between a laboratory test, such as a flexural beam frequency sweep and the same material when it is part of a layered pavement structure in the field. Specifically, a flexural beam is suspended in space without confinement in a flexural frequency sweep test. In contrast, the same asphalt concrete material, confined below and on its sides when it is part of a pavement layer, has its modulus increased. ( In this confined condition, the aggregate in the asphalt concrete, which has its own relatively unchangeable high modulus, also has a large compressive stress component applied to it.) Figure 7 shows an example of the input parameters for the AC bottom layer. A modulus versusreduced time relationship was assumed for the ATPB, based largely on laboratory triaxial testing. The ATPB had a modulus of 1144 MPa at a temperature of 20 C and a loading time of 0.015 sec. A minimum modulus of 200 MPa was assumed. Stage 5 Distribution 13 UCPRC RR 2005 06 0 2000 4000 6000 8000 10000 12000 14000 16000  4  3  2  1 0 1 2 log reduced time E MPa 28 28 28 28 19 19 19 19 15 15 15 15 AC top 0 2000 4000 6000 8000 10000 12000 14000 16000  4  3  2  1 0 1 2 log reduced time E MPa 28 28 28 28 19 19 19 19 15 15 15 15 AC bottom 0 2000 4000 6000 8000 10000 12000 14000 16000  4  3  2  1 0 1 2 log reduced time E MPa 25 25 10 25 10 25 25 10 25 10 25 10 10 DGAC 0 2000 4000 6000 8000 10000 12000 14000 16000  4  3  2  1 0 1 2 log reduced time E MPa 25 25 25 25 25 25 10 10 10 ARHM Figure 6. Frequency sweep data for Goal 1 and Goal 3 materials compared to models. Stage 5 Distribution 21 UCPRC RR 2005 06 Figure 7. Example of input parameters for the modulus versus reduced time relationship for the AC bottom layer. 1.3.2 Fatigue The modulus of damaged asphalt was calculated as: ( ) ( ) ( ( tr)) E 1 exp log log 1 β γ α ω δ + + × − = + ( 5) where the damage, ω, was calculated from: ( t) MPa E MN A × × ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ × ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ = × × δ με με ω β γ α exp 200 3000 ( 6) where E is the modulus of damaged material, MN is the number of load repetitions in millions ( N/ 106), με is the strain in μstrain, t is the temperature in º C, and α, β, γ, and δ are constants ( these constants are the same as the constants in Equations 1 through 5, and different in Equation 6). Stage 5 Distribution 22 UCPRC RR 2005 06 The initial ( intact) modulus, Ei, corresponds to a damage, ω, of 0, and the minimum modulus, Emin= 10δ, to a damage of 1. Equation 5 leads to: ( ) ( ( ) ) ( ) ⎟⎠ ⎞ ⎜⎝ ⎛ ⎟⎠ ⎞ ⎜⎝ ⎛ = ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ = − = − × − min min log log , log log 1 , E E E E or E E E E E E or i i i i i ω δ δ ω ω ( 7) The NCHRP 1 37A Design Guide calculates the modulus of damaged asphalt, for rehabilitation purposes, using Equation ( 5), but at the same time the Guide defines ω as the relative decrease in modulus ( although this is mistakenly indicated as E/ Ei in the report for the Guide). This definition of ω included in the NCHRP 1 37A Design Guide is inconsistent with Equation 5, as shown in Equation ( 7) The one stage Weibull distribution, Equation ( 8), could be used instead of Equation 6: SR = exp(− α × Nβ ) ( 8) where SR is the stiffness reduction (= E/ Ei), N is the number of load applications, and α and β are constants, different from those used in previous equations. Combining Equations ( 7) and ( 8) one gets: ( ) β α β α ω N E E E E N E E SR E E E E i i i i i × ⎟⎠ ⎞ ⎜⎝ ⎛ = ⎟⎠ ⎞ ⎜⎝ ⎛ − × = ⎟⎠ ⎞ ⎜⎝ ⎛ = ⎟⎠ ⎞ ⎜⎝ ⎛ ⎟⎠ ⎞ ⎜⎝ ⎛ = min min ln min ln min ln ln ln ln ( 9) which has the same format as Equation 6. In the present version of CalME ( September 2006) it is assumed that α and β can be written in the format: ( ( ) ( )) A B t C ( w) A B t C w D t w ln exp ln ln = + × + × = + × + × + × × β β β β α α α α α ( 10) where t is the temperature in º C, w is the internal energy density ( ½ ×ε2×E), and αA, αB, αC, αD, βA, βB, and βC are constants fit from the beam fatigue data. Fatigue parameters were determined for the AC top and AC bottom of Goal 1 and for the DGAC and ARHM overlays of Goal 3, based on four point bending beam tests at 10 Hz under controlled strain. In determining the fatigue parameters it was assumed that for Equation 6 that β was equal to two times γ. This reduces the number of parameters to be determined by one, and it ensures that the damage is a function of the internal energy density ( ½ ×ε2×E). The parameters were determined by minimizing the root mean square ( RMS) difference between the calculated relative modulus ( E/ Ei) and the experimental data for values of E/ Ei > 0.3, the stiffness ratio to which most of the beam fatigue tests were carried out. Stage 5 Distribution 23 UCPRC RR 2005 06 An example of damage versus number of load applications is shown in Figure 8 which shows the damage at a temperature of 20 º C for a constant strain of 400 μstrain. The four materials are surprisingly similar with respect to damage. Damage of asphalt 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 500000 1000000 1500000 2000000 2500000 Number of load applications Damage Goal1Top Goal1Bottom Goal3DGAC Goal3ARHM Figure 8. Example of damage versus number of load applications. For ATPB a damage function ( Equation 6) was chosen based on the damage function for the bottom lift of Goal 1, but with a value of A about fifty times as high. Figure 9 shows an example of the damage parameters for Equation 6 for the AC bottom lift of Goal 1. The parameters of interest here are the values of the first column ( under the heading “ Fatigue, dE/ Ei”). The response type is the horizontal tensile strain at the bottom of the layer ( the minor principal strain) indicated by the Response type “ e”. The parameters A, α, Reference strain ( Respref), β, Reference modulus ( Eref), γ, and δ are given. The parameter Std A is a Standard Deviation factor on A that is used in the Monte Carlo simulations, which are not used in this report. The number of in situ load applications is divided by the Shift Factor given at the bottom of the column to allow for differences between laboratory testing and field conditions. A Shift Factor of 3 was used for all materials, i. e., three HVS loads were assumed to give the same damage as one laboratory load. Stage 5 Distribution 24 UCPRC RR 2005 06 Figure 9. Example of Equation 6 damage parameters of AC bottom layer ( Goal 1). 1.4 Weak Bonding Weak bonding between the top and bottom asphalt lifts of Goals 1, 3, and 5 was found for certain areas of the HVS test sections. It appeared that the top AC layer had moved horizontally with respect to the bottom AC layer. During forensic studies a brown discoloration and scratch marks were observed on the surfaces of the materials at the interface where slip had apparently occured. Cores showed no bonding at the interface. The layered elastic analyis program ( LEAP) option of CalME has a parameter that controls the degree of bonding between two layers, referred to as “ slip” in LEAP if less than full bonding and “ stick” if full bonding. Full bond corresponds to a high value ( 105 is used in CalME), and a value of 0 corresponds to full slip ( i. e., there is no bond between the layers). The logarithm of this parameter is decreased linearly until the final slip is reached, as a function of the number of loads in ESALs. The number of ESALs corresponding to final slip is not known, but the shape of the deflection versus number of loads curve can serve as a guide. The LEAP program treats the pavement structure as a continuum, which means that the two materials above and below the slip interface will be in contact for all points of the interface. This is not a completely realistic assumption, but with the present response models it cannot be changed. [ A three dimensional Finite Element Model ( FEM) would be required to change that assumption.] It is likely that this incorrect modeling of the slip interface will have different effects on deflections, on horizontal strains at the interface, and on the shear stress and strain at a depth of 50 mm. ( This depth is used for calculating the permanent deformation of the AC layer. This is described later). For deflection and shear Stage 5 Distribution 25 UCPRC RR 2005 06 stress, it is likely that friction between the two layers immediately under the wheel will result in only a partial slip occurring because of the high compressive normal force, whereas the maximum horizontal strains may well correspond to a condition closer to full slip. Table 5 Influence of Slip Value in LEAP on Calculated Vertical Deflections and Horizontal Strains Slip Value 0 10 4 10 3 10 2 10 1 1 10 102 103 104 105 d 0 mm* 0.526 0.526 0.525 0.517 0.473 0.396 0.370 0.367 0.367 0.367 d 625 mm* 0.276 0.276 0.276 0.272 0.253 0.228 0.220 0.220 0.219 0.219 Ex top* 230 230 229 222 182 83 14 0  2  2 Ex bottom* 232 232 232 232 226 196 168 162 161 161 * Note: d 0 mm is vertical deflection at surface, d 625 mm is vertical deflection at 625 mm depth, Ex top is horizontal strain at the bottom of the top asphalt layer, Ex bottom is the horizontal strain at the bottom of the bottom asphalt layer. Table 5 shows an example, taken from Section 501RF, of the influence of the slip value on the deflection at the surface ( d 0 mm) and the deflection of the subgrade ( d 625 mm), and the horizontal strain at the bottom of the top AC layer ( Ex top, tensile as positive) and at the bottom of the bottom AC layer ( Ex bottom). A slip value of 0.0001 was chosen for the main simulations. This corresponds to full slip between the layers. At the interface with full slip the shear stress will be zero. The shear stress used for calculating permanent deformation in the asphalt is at a depth of 50 mm, which is only slightly above the interface. When full slip develops, the shear stress at depth 50 mm will therefore decrease considerably. As was mentioned above, this may not be realistic, so a second simulation was carried out with full bonding to determine the permanent deformation of the asphalt. During this simulation the stiffness factors for the unbound layers and the shift factor for asphalt fatigue were adjusted to assure that the pavement deflection history was still correct. These simulations with no slip were only used for determining the permanent deformation of the asphalt. 1.5 Unbound Layers 1.5.1 Triaxial Tests Table 6 shows the results of triaxial tests on the subgrade material ( Harvey et al. 1996). Two specimens were tested, compacted at different density and moisture content ( MC) and either soaked or saturated. The tests were done at a confining stress of 7 kPa, which is close to the static confining stress at the top of the subgrade. The parameters C and n are defined by the equation: n d MPa C E ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ = × 0.1 σ ( 11) where E is the modulus, and σd is the deviator stress. The column “ E( 30 kPa)” indicates the modulus, at a deviator stress of 30 kPa, which is a typical stress at the top of the subgrade under a 40 kN dual wheel load. Stage 5 Distribution 26 UCPRC RR 2005 06 Table 6. Triaxial Tests on Subgrade Density MC% Condition C n E( 30 kPa) 2.06 g/ cm3 22.4 Soaked 36.2  0.34 55 MPa 2.12 g/ cm3 15.8 Soaked 66.5  0.32 98 MPa 2.12 g/ cm3 15.8 Saturated 41.5  0.27 57 MPa Triaxial tests were done on one AB specimen, with the results shown in Table 7. The sample was compacted to a density of 2.47 g/ cm3 at a MC of 5.5 percent. The top of the specimen was exposed for ten days to simulate the effects of exposure to air on the test section. The moisture content dropped from 5.5 percent to 2.9 percent during the ten days. After testing the specimen was saturated and tested again. Table 7. Triaxial Tests on Aggregate Base ( AB) Condition k1 k2 E( 50 kPa) Exposed 481 0.16 430 MPa Saturated 201 0.49 143 MPa The constants k1 and k2 are defined by the equation 2 1 0.1 k MPa k E ⎟ ⎟⎠ ⎞ ⎜ ⎜⎝ ⎛ = × θ ( 12) where E is the modulus, and θ is the bulk stress. The column “ E( 50 kPa)” indicates the modulus at a bulk stress of 50 kPa. According to the theory of elasticity, the bulk stress in the AB a 
PDI.Title  Calibration of incrementalrecursive flexible damage models in CalME using HVS experiments 



B 

C 

I 

S 


