Thin and Ultra-Thin Whitetopping (2004)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

13 This chapter includes discussion of the various issues related to the design of TWT and UTW overlays. Information is pre- sented in a fashion suitable as either a lookup reference or as a learning tool. Because many aspects of the design of UTW and TWT overlays are similar to those of a conventional concrete pavement design, the discussion provided will not include sufficient detail to serve as a stand-alone design aid. Instead, it can be used alongside other design tools, especially those cited herein, including the AASHTO Guide for Design of Pave- ment Structures (11) and the PCA Design and Control of Con- crete Mixtures (68). EXISTING PAVEMENT It has been reported that a key to the success of any concrete pavement, including a UTW or TWT overlay, is a uniform and stable support system (69). In such a case, the support is provided by an existing HMA pavement. Therefore, it should be recognized that any contributing factor to the fail- ure of the HMA pavement might similarly lead to a failure of the overlay. A thicker concrete overlay should be considered if the sup- port layers have exhibited poor structural support, by con- tributing to the deformation and/or cracking of the original HMA pavement (70). When designing the UTW or TWT overlay, consideration should be made of the condition of the existing HMA pavement. In some cases, this consideration includes the use of a reduced stiffness value for the HMA. Rational adjustment fac- tors for a decrease in the HMA stiffness owing to damage have been used in other pavement design approaches, and they could be used in whitetopping overlay design as well (2,71). Characterization of the in-place HMA can be found through visual inspection, backcalculation of FWD data, or testing of cores collected from the candidate project. Laboratory testing can include wheel-track testing (e.g., Hamburg) and resilient or dynamic modulus measurements (1,47). The syn- thesis survey revealed that 90% of the respondents relied on some form of visual inspection for characterization of the HMA layer, 62% used pavement management data, 45% used FWD data, and 7% employed laboratory testing. Recent studies have indicated that the susceptibility of the HMA to permanent deformation may be a significant factor in UTW and TWT performance; therefore, testing for this prop- erty may be of benefit, if means are readily available (2,48,49). The survey revealed that 76% of the respondents included rut- depth measurements in their design process. In the near future, consideration could be given to employing the simple perfor- mance test technology currently being developed and eval- uated in NCHRP Reports 465 and 513 (72,73). The devices being considered measure more fundamental properties of the HMA mixture, making extrapolation to a wide variety of loading conditions possible. Although this test is being devel- oped as part of the Superpave system, it shows promise as an improved means to predict the performance of whitetopping overlays (2). In addition to considering the HMA stiffness, another design approach is to adjust the effective thickness of the existing HMA layer. The 1993 AASHTO design guide pro- vides guidance employing this approach (11). Adjustment factors can be derived from a visual condition survey, remaining life analysis, or FWD backcalculation. The stiffness of the pavement system as a whole (includ- ing the HMA layer and support layers) is known to have a significant effect on the performance of the whitetopping overlay. As a result, FWD testing should be carried out if possible. Suggestions for the proper interpretation of the data include the procedures documented in NCHRP Report 372 and elsewhere (69,74). With proper characterization of the pavement support system, the reliability of the overlay design will be improved. PREOVERLAY REPAIR Before overlay, it is recommended that consideration be made to repair the existing pavement, to provide the necessary uni- formity to the support system. Although many existing HMA distresses will not reflect through the whitetopping overlay as they might through an HMA overlay, some types of HMA distress may lead to different types of failures in the overlay. For example, a localized subgrade failure beneath the HMA that leads to alligator cracking may ultimately lead to fault- ing, roughness, or a shattered slab in the whitetopping, if left uncorrected. Of those responding to the survey, 38% do not routinely conduct preoverlay repairs. However, full-depth patching to correct localized subgrade failure was reported as used by 28%, pothole filling by 21%, and crack sealing by 10%. CHAPTER FOUR DESIGN

14 The degree to which preoverlay repairs are conducted will depend on the design of the UTW or TWT, along with the anticipated use of the facility. Usually, if the facility is expected to carry heavy truck traffic, measures should be taken to min- imize weak areas in the existing pavement (1). However, if the whitetopping is being used for aesthetic or other functional purposes, and it is not anticipated to carry large volumes of trucks, the overlay can be placed without significant repair. Cost is a common consideration when determining the degree of preoverlay repair (1). The additional cost of repair may result in a reduced thickness, because the overall project cost may be fixed. Finally, because concrete will be used for the overlay, preoverlay repairs can be made by simply recon- structing areas with full-depth concrete—replacing instead of overlaying the existing HMA. The following sections outline some of the more common existing pavement distresses, plus the recommended preover- lay repairs to mitigate them. Rutting and Shoving Appropriate action should be taken where moderate-to-severe rutting or shoving exists. The ACPA identifies this situation as rutting greater than 50 mm (2 in.) (1). Where only minor rutting or shoving is observed, the distress can usually be ignored. The exception to this situation is when the rutting has occurred within a period of time that was much shorter than normal. In such cases, the HMA (or other support layer) may be highly susceptible to permanent deformation and it should be replaced or otherwise mitigated before overlay. In some instances, the whitetopping concrete can be placed without correction of the rutting or shoving distress (75). Doing so may prove beneficial, because a thicker concrete section will be the result in the wheelpaths, where the traffic- induced stresses are usually the highest. If corrective action is warranted, milling is the most effec- tive for rutting and shoving problems in the existing HMA. This practice also improves the bond between the PCC and HMA, but care should be taken so that adequate HMA thickness remains after milling, if considered essential in the design (17). If the source of severe rutting is found to be the base or subgrade layers, a whitetopping overlay should be used with caution. If there is no replacement of these layers, the per- formance of the overlay may continue to be affected by the poor quality of the layers. Leveling courses have also been used in the past to correct for rutting and shoving, but they are not encouraged for use on UTW and TWT projects unless care is taken in the proper selection of the HMA material. As will be discussed later, the properties of the leveling HMA can affect the PCC–HMA bond as well as the stiffness and stability of the whitetopping support system (48). In addition, a leveling course adds a construction operation, which may adversely affect the user delay costs and project budget. Potholes The ACPA recommends that small potholes that do not have further widespread subsurface failures be repaired before the placement of any whitetopping layer (1). Such distresses can be repaired by filling them with unbound aggregate, cold- or hot-mixed asphalt, or concrete. The surface should be well compacted (consolidated) to reduce further movement by loading that propagates through the whitetopping layer after it has been constructed. The presence of numerous or severe potholes may be an indication of a weakened pavement structure. Therefore, those who design a whitetopping overlay should proceed with cau- tion because the performance of the overlay will be adversely affected by a weak support system (70). Subgrade Failure Localized subgrade failures are often evident by surface fail- ures such as potholes and map or alligator cracking (13). In these cases, the affected area should be removed and replaced before overlay. All of the material affected by the failure or prone to future failures should be removed and the existing HMA structure reconstructed to its original elevation. Exist- ing compaction levels should be matched to avoid differen- tial consolidation of the area after the whitetopping layer has been placed. It is also possible to design full-depth concrete sections for these areas (76). These sections can be placed at the same time that the overlay construction is done. This alter- native would eliminate not only the need for the additional construction process of HMA placement but also the possi- bility of poor support or bond caused by a tender mixture. Cracking The ACPA recommends that cracks in the existing HMA pavement be filled if the crack width is larger than the max- imum size of the aggregate in the whitetopping concrete mix (1). Some researchers have also noted reflective crack- ing from an underlying HMA pavement through a UTW over- lay (50). Therefore, if severe cracking is observed, especially thermal cracking, measures may be taken to repair these cracks before overlay. Because cracking can often be extensive, a cost analysis should be considered to optimize the degree of preoverlay repair (1). For thinner whitetopping overlays constructed over thick and stiff HMA layers, it may be advisable to match the joints with longitudinal cracks that may have developed along

15 construction joints in the HMA. If this is done, it is advisable to seal these cracks to minimize infiltration of water through the PCC and HMA layers and into the underlying support system. It may be possible to use flowable (cementitious) fill for this purpose. PCC–HMA Bond Numerous sources investigated during this synthesis study concurred that a sound bond between the overlay concrete and underlying HMA is key to the success of UTW (1,70,77). The same is true of TWT overlays that have been designed with consideration of a bond (17). The bond at the PCC–HMA interface creates the necessary composite section, lowering the neutral axis so that load- related stresses in the concrete are substantially reduced. Short joint spacings, common on UTW and TWT overlays, further reduce these stresses, because the slabs are not long enough to develop as much bending moment (78). The Iowa Shear Test has been the most commonly reported test for assessing PCC–HMA bond (46). Test results for bond strength are typically in the area of 100 psi for a white- topping considered to have a sound bond (79). However, cau- tion should be used when comparing test results measured at different temperatures and loading rates, because the behavior of the HMA will vary significantly owing to these factors (80). Proper bonding will not only reduce wheel-load stresses but it will also resist curling forces responsible for lifting of the panels’ edges and corners (70,81). The key to a successful bond is proper surface treatment of the HMA layer. Because economic effects are also a consideration, there is commonly a trade-off between assurance of the bond and the associated additional costs. The following sections describe the more common techniques that have been used to develop an inter- face between the existing HMA and the new PCC overlay. No-Mill Option In some cases, a bond has been achieved between the exist- ing HMA and the whitetopping overlay with no mechanical surface preparation, except for cleaning (79). Under direct placement operations, the existing HMA layer is not milled. This method can be used when rutting and/or shoving are not severe enough to require milling for correction of the surface. In this situation, the surface of the HMA should be cleaned before placement of the whitetopping layer. Of the survey respondents, 33% have used sweeping for this purpose, 37% used air blasting, and 30% used water blasting. It should be noted that because this method has only min- imal assurances of bond it is discouraged for UTW. How- ever, it can be used on TWT applications, especially if com- posite action is not a consideration in the design. In cases where low-severity rutting and shoving are not repaired before the whitetopping overlay placement, the excess quantity of PCC needed to fill the surface distortions should be calculated (1). This quantity will be in excess of the amount calculated for the design thickness of the white- topping overlay. Furthermore, the surface irregularities should be considered when recommending the cross-sectional thick- ness (82). If high points in the existing HMA surface are not considered, the result may be thinner concrete sections that may lead to a localized failure. That situation also presents one reason that payments are commonly based on concrete volume for UTW and TWT overlays (83). Milling Milling is commonly used and strongly recommended for UTW overlays. Milling is also used successfully for TWT overlays, but it is not mandatory unless PCC–HMA bond- ing is considered in the design. Of those surveyed, 90% use milling for HMA surface preparation. The benefits of milling include • Smoothing out surface distortions, • Finishing the grade line before placing a whitetopping overlay, • Establishing a cross slope for the new pavement eleva- tion, and • Establishing a good surface to which the whitetopping overlay can develop a good bond. Depending on the equipment used, milling operations com- monly remove between 25 and 75 mm (1 and 3 in.) from the surface of the existing HMA pavement. Figure 7 shows an FIGURE 7 Milled HMA surface before overlay.

16 existing HMA surface that has been milled. The ACPA rec- ommends that care be taken not to remove too much of the HMA surface, because the overlay thickness design some- times relies on a minimum remaining thickness of the exist- ing HMA layer (1). After the milling operation has been completed, the exist- ing HMA surface should be cleaned to remove any particles that had been milled off. A thorough sweeping is most often used for this purpose. Leveling Course Although leveling courses are not recommended to correct the existing HMA surface before UTW or TWT construction, their use has been reported in filling surface distortions up to 50 mm (2 in.) (1). Leveling courses may also be required to correct or modify cross slope or to improve surface drainage. If a leveling course is being considered, an alternate design using a thicker PCC section should also be considered, since the cost of the additional PCC thickness may be less than the cost and time of the HMA laydown. In some reported instances, the new HMA material in the leveling course has been found to further compact and shift under whitetopping surface deflections (17). Such deflec- tions, combined with the instability of the new HMA level- ing course, can cause premature cracking in the overlay owing to voids created beneath the whitetopping surface. Thus, if a leveling course is being considered in the design, appropriate adjustments should be made to the stiffness of the underly- ing HMA and/or to the bond efficiency between the PCC and HMA. Stripping Because stripping of the asphalt binder from the HMA has been reported as a contributing factor to premature deterio- ration of some UTW and TWT overlays, the susceptibility of the existing HMA should be assessed before design (54,84). This action should be coupled with an assessment of the prob- ability that moisture will be present to cause stripping. If the stripping is expected to be unavoidable, the thickness design of the overlay should be adjusted based on the assumption of little or no bond between the PCC and HMA. Furthermore, consideration should be made to reduce the HMA stiffness, to account for the loss of the binder over time. TRAFFIC CHARACTERIZATION Traffic data are an important requirement for the design and analysis of any pavement structure. The number of truck loads or 80-kN (18-kip) ESALs that the pavement must with- stand during its intended design life will likely be the most important factor in determining which of the whitetopping classes (thicknesses) should be used (1). From the survey responses, 59% of states design whitetopping overlays based on ESALs, 38% from a combination of average daily traffic and percent trucks, and 7% from a more general roadway classification. Whereas some design procedures employ these more conventional methods of characterizing traffic, recent trends of traffic characterization include those being adopted in the upcoming mechanistic-empirical design guide (NCHRP Proj- ect 1-37a) (85). The new methods, although not being used by any of the survey respondents at this time, will require additional detail about the anticipated traffic loading, includ- ing the following: • Axle-load spectra—the distribution of the number of axles falling into categories of weight and type of axle. For a given axle type, heavier axles will lead to greater damage in the whitetopping overlay, as they do on other pavement types. • Seasonal distribution of traffic—allowing for a more accurate assessment of damage as affected by seasonal changes in the pavement structure. For example, during the summer months, the underlying HMA will have a lower stiffness that will have an impact on the structural capacity and bond with the PCC. • Growth—traffic level will usually increase with time. Adjustment factors are required to reflect the change in traffic loading over time. • Time of day—an important consideration being the deflected shape of the pavement that will differ from day to night, owing to curling. This behavior can affect the predicted traffic-induced stresses. Furthermore, the temperature of the HMA will have an impact on the stiffness and bond strength. CONCRETE MATERIALS TWT and UTW overlays are intended for existing HMA pavement rehabilitation. If these types of overlays are to be constructed under traffic, it may be desirable to use “fast-track” concrete materials (52% of those surveyed use these methods). Fast-track construction requires a higher early strength than is required of conventional mixes. However, care should be taken when using such mixes, because a greater potential for shrinkage, and thus random cracking, can exist (39,86). Other considerations should also be made, including mix proportioning, admixtures, and particularly the use of fibers. Engineering properties such as strength are also important, because they are commonly used inputs into the design pro- cedures. This section will describe the more relevant facets of concrete-making materials as they relate to UTW and TWT design. Aggregate Properties Because aggregates constitute the largest portion of PCC in both volume and mass, their properties will dominate the

17 overall behavior of the mix (87). Some key aggregate thermal properties that affect the behavior and performance of white- topping include the coefficient of thermal expansion (CTE), thermal conductivity, and specific heat. From a stress per- spective, CTE is the most important property. All else being equal, pavements constructed with aggregates of low CTE will perform better than those constructed with aggregates of higher CTE (88). This is because higher slab stresses and wider joint openings can occur when aggregates with higher CTE are used. Another aggregate property to consider is the maximum size of the aggregate. As a rule of thumb, the nominal maxi- mum aggregate size should be no larger than one-third of the slab thickness (89). Optimizing the aggregate gradation and shape should also be considered, because an improvement in strength, durability, and workability can result (90). Both the top size and gradation of the aggregate will also affect aggre- gate interlock at the joint, which is another important con- sideration, because UTW and TWT overlay joints are typi- cally not dowelled. Cement and Supplementary Cementitious Materials As opposed to aggregate properties, cement paste properties change as a function of time, especially during the early-age period (39,91). Cement Types I and III are most commonly used for fast-track mixes (39). Types II and V are occasion- ally used in overlay construction, especially in areas with high sulfates. Each cement type reacts differently when exposed to water, for each releases different quantities of heat at differ- ent rates. Therefore, the type of cement used can affect over- lay behavior and performance. Supplementary cementitious materials (SCMs) such as fly ash and ground-granulated blast furnace slag have been used successfully on UTW and TWT projects nationwide (2). Under hot-weather paving conditions their use is encour- aged (92). Specifications that currently accommodate their use for other PCC paving applications can similarly be adopted for whitetopping. The water-to-cementitious materials ratio is an important indicator. It is defined as the mass of water in the mix divided by the combined mass of the cement and any additional SCMs, including some fly ashes, silica fume, and slag. This ratio is critical in predicting the overall strength and durability of the mix, as well as other mechanical properties including shrink- age and creep (68,91). Admixtures and Fibers At the very least, PCC consists of three basic constituents: aggregate, cement, and water. In addition to these basic com- ponents, admixtures may be used to enhance or otherwise modify a particular characteristic of PCC (68). The use of admixtures can significantly affect the behavior of PCC dur- ing the hydration period, especially in fast-track mixes, which have a greater behavior susceptibility to even small varia- tions in mix design (86). It has been reported that the addition of fibers to concrete enhances the toughness and residual strength of concrete (93–95). Of those respondents surveyed, 64% always use fibers in their whitetopping mixtures, 32% use them some- times, and 4% do not use them at all. Fibers may be of partic- ular benefit to whitetopping overlays, because they reportedly reduce permeability, minimize crack width, reduce surface spalling, and increase wear resistance (94). Fibrillated polypropylene fibers are most commonly used on UTW overlays, and they are sometimes specified on TWT applications (96). The most common dosage rate follows the example of the Louisville, Kentucky, project, which used 1.8 kg/m3 (3 lb/yd3) (10,43). Successful projects have also been constructed employing monofilament fibers (typically polyolefin) and steel fibers (24,97,98). From the survey, 78% of respondents use fibrillated synthetic fibers, 37% use syn- thetic monofilament, and 7% use steel fibers. Tests performed at Gainesville, Florida, indicated no ben- efit to the addition of fibers; however, it was reported that more loading should be applied to determine if there is a ben- efit (70,99). Other sources have stated that the addition of fibers improves the strength and durability of concrete, and it reduces the occurrence of plastic shrinkage cracking (100). Currently, the industry recommends that enough fibers be added to provide a residual strength of 0.6 MPa (80 psi) when measured in accordance with ASTM C1399 (94,101). Concrete Properties The concrete mix for a particular whitetopping project is often selected based on the requirements for opening to traffic. For rapid strength gain under fast-track conditions, a common requirement in urban areas, concrete mix designs sometimes include higher cement contents (39). Strength development is a critical factor in determining the opening of the pave- ment to traffic. The 28-day strength of the mix should also be considered, to satisfy the pavement design, including the long-term effects of fatigue and other traffic-induced dis- tresses (102,103). The impact that including fibers has on the residual strength should also be addressed (104). The designer should consider specifying the use of maturity, because it has been shown to provide better construction control, ensur- ing a higher quality placement (105,106). Maturity is a nonde- structive means to estimate concrete strength by monitoring temperature and elapsed time in the field. Finally, the drying shrinkage potential of the concrete mixture should be a factor during the design, particularly with selection of a proper joint spacing. If high shrinkage cannot be avoided, it could be par- tially compensated for with shorter joint spacing.

CLIMATIC CONSIDERATIONS Along with traffic, climatic conditions can affect the behavior and performance of a whitetopping overlay system. During construction, the most detrimental effects of climatic condi- tions occur at extreme temperature conditions, specifically at air temperatures greater than 32°C (90°F) or less than 4°C (39°F) (92). The design and construction of the whitetopping overlay should consider the possible effects of both lower-than- normal and higher-than-normal curing temperature on concrete properties and pavement behavior at early ages. Developing specifications that properly account for the interaction between the selected concrete mix and the environmental conditions encountered during construction is essential. Distress types that occur at an early age include random cracking, transverse cracking, and even delamination of the PCC from the HMA. These effects can be minimized with proper temperature con- trol provisions. Documents such as the hot-weather and cold- weather placement guidelines from the American Concrete Institute (ACI) should be consulted (93,107). Climatic effects over the life of the overlay should also be considered. For example, environment-related distresses such as D-cracking and freeze–thaw damage can occur if proper precautionary measures are not taken (12,68,108). WHITETOPPING THICKNESS The thickness of the whitetopping overlay is a significant characteristic of the overall pavement system. It will drive not only the performance but also the cost. As a result, the selection of the thickness must balance a number of factors, including the anticipated traffic loading, the strength and stiffness of the existing pavement, the properties of the over- lay concrete, and the degree of load transfer (1,2,17,18,77). In the thickness design, variability of the various design inputs should be accounted for, if possible (1). Reliability is a means by which this variability (as well as variability in the other design inputs) can be rationally considered (11). The level of reliability should be determined based on the impor- tance of the facility. The use of certain reliability levels will ensure that the final design can be evaluated based on the selected confidence levels. Not all design procedures include an explicit means to enter variability or reliability inputs, relying instead on built-in factors of safety (20,77). In any case, pavement designers should always be cautious when selecting “conservative” values for pavement design inputs, unless explicitly instructed. If the designer takes an overly conservative approach, the resulting design may prove more costly than necessary. When the whitetopping thickness is selected, additional checks should be performed, including on the remaining over- head clearance under structures, and matching curb and gut- ter sections. In addition, because whitetopping overlays are 18 commonly preceded by a milling operation, the reduction in HMA thickness should be a factor. Using whitetopping inlays (deep milling followed by place- ment of the overlay) can be considered on very thick HMA sections. This is a commonly used technique, because if design is done properly, there are no drainage and overhead clear- ance problems, the surface is renewed, and additional struc- tural capacity is provided. A more detailed analytical discus- sion of available thickness design methods appears later in this chapter. JOINTS One of the more notable characteristics of UTW and TWT overlays is their shorter joint spacings. Designs using short joint spacing can significantly reduce curling stresses (109). In addition, combined with adequate PCC–HMA bond, the shorter joint spacing reduces the flexure (bending) of the con- crete panels, further lowering the stress (110). The most common rule of thumb for UTW and TWT pan- els is to select a joint spacing that is 12 to 18 times the thick- ness (1,111). This range of size was confirmed from the sur- vey responses. Thinner sections are commonly on the low end of this range (i.e., 12), and thicker TWT sections on the high end (i.e., 18). If a TWT is designed and constructed without a bond to the underlying HMA, recommendations for joint spacing increase to 21 times the thickness. The ratio of the length (longest dimension) to width (shortest dimension) of any given panel is recommended to be no more than 1.5, although some recommend a maximum ratio of 1.25 (1). It should be noted that given the large number of joints typi- cally specified for a UTW or TWT, the cost of the additional sawing might be weighed against the cost of a thicker PCC section, which would require less sawing. In UTW and TWT, load transfer is commonly provided by aggregate interlock, which is enhanced by a small joint opening resulting from the short joint spacing (112). Aggre- gate interlock can be further improved by the selection of a concrete mix having a well-graded aggregate with a larger top size (113). Furthermore, the joint opening can be reduced (enhancing aggregate interlock) by selecting a mix with both low shrinkage characteristics and low CTE (112). Although most UTW and TWT overlays do not need them, dowels and/or tie bars have been used successfully on some TWT projects, and they are recommended if heavy traffic is expected (1). Colorado strongly recommends the use of tie bars to minimize the movement of the longitudinal construc- tion joint (17). Iowa has recommended tying the outside pan- els in cases where little or no support may be provided at the shoulder (114). Recommendations for dowel and tie bar size and spacing can be found in a number of industry publica- tions (1,11,115).

19 75 mm (3 in.). The transition typically occurs over a 2-m (6-ft) extent (111). This technique was reported as being used by 60% of the survey respondents. • Transition slab reinforcement—welded wire mesh or two-way reinforcing bar that can be used to reinforce the first 3 m (10 ft) of PCC from the transition (2). This technique is used by 7% of the survey respondents. • Isolation (expansion) joints, typically constructed 12 to 25 mm (0.5 to 1 in.) wide and filled with a preformed compressible joint filler (13). These joints are used for transitions to an existing pavement. EXISTING DESIGN PROCEDURES AND MODELS This section briefly describes existing procedures for white- topping overlay design. At least four different design proce- dures have been identified in the literature, including those from Colorado (17), New Jersey (18), the Portland Cement Association (PCA) (19), and AASHTO (1). The following two sections describe the PCA UTW and the Colorado TWT design procedures. Although more advanced procedures have recently been developed (2), the procedures discussed in this report are believed to be the most representative of the state of the practice for UTW and TWT overlay design, respec- tively. Information on the other procedures can be found in the references cited. It should also be noted that the formula- tion of the models is such that a spreadsheet application could be developed to assist in the design. PCA UTW Design Procedure This design method is technically an analysis procedure that allows for the prediction of the number of loads to fail- ure for a given UTW configuration (19). The models used are mechanistic–empirical in nature. During their develop- ment, three-dimensional (3-D) finite-element methods (FEM) were employed to consider the unusual geometry of UTW overlays. To verify the 3-D FEM pavement response model, data were collected from three field sites in Missouri and Col- orado. When comparing the predictions with the field obser- vations, it was found that the measured stresses in the UTW slabs were approximately 14% to 34% higher than for the fully bonded 3-D FEM simulation. To account for this dif- ference, the 3-D FEM model was adjusted to simulate the partial bonding condition. The stress multiplier selected for the design procedure is 1.36 (a 36% increase), and it implic- itly incorporates reliability, because it is the sum of the mean stress difference plus one standard deviation. The first step in using the design procedure is to calculate the pavement response under load. To do this, a series of regression equations based on the calibrated FEM model were developed. The final equations are as follows: The designer should also check the expected location of the wheelpath in relation to the joint layout. It has been reported that corner breaks occurred on UTW sections where the wheelpath was located on the longitudinal joints (10,48). If possible, longitudinal joints should be designed away from the wheelpaths, preferably having the wheelpath fall in the center of the panel, directly between two longitudinal joints. If there is an abrupt change in the underlying HMA sec- tion, along a lane–shoulder joint, for example, then consider- ation should be made in matching it with a joint in the thinner UTW overlays. The same is true of any significant transverse transitions—joints on a large patch, for example. In such a case, a transverse joint in the UTW could be aligned to match the patch, even if it falls out of the regular sequence of joint spacing. Because some users have reported reflec- tive cracking from working joints in the HMA, through the UTW overlay (50), this may be useful for mitigation. Sawing of joints on UTW and TWT overlays can be done with conventional sawcutting equipment, but it is more com- monly done with early-entry (green) saws. Because of the thinner sections, higher shrinkage potential, and high restraint caused by the bond between the PCC and HMA layers, the potential for early cracking is greater with UTW and TWT compared with conventional concrete paving (39). Using early- entry saws minimizes the potential for uncontrolled cracking. Both the literature and the survey responses reported that saw- cut depths should be 25% to 33% of the thickness. Shallower cuts—as little as 25 mm (1 in.)—have also been reportedly used with success when early-entry sawing is used (116). Because of the shorter joint spacing, sealing of joints is typ- ically not performed on UTW or on many TWT overlays (9), with only 16% of those surveyed reporting that they seal the joints. However, the user survey also reported that the joints on thicker TWT overlays, commonly containing dowels and/ or tie bars, are sealed. TRANSITIONS AND OTHER DESIGN FEATURES During the design process, detailing of the transition from the whitetopping to the adjoining pavement should be provided. Transitions between the whitetopping overlay and the adjacent pavement are usually more susceptible to damage owing to a number of factors. For example, horizontal movement of the concrete (from shrinkage and/or thermal contraction) can lead to a wide joint at the transition (112). In addition, settlement or shoving of either or both pavement structures can lead to a vertical “bump” or “dip,” which in turn produces additional dynamic loading from vehicle excitation (117). To mitigate those effects, the following design features have been used successfully: • Pavement edge thickening—tapering from the designed whitetopping thickness to the design thickness plus

20 (1) (2) (3) (4) (5) (6) where εHMA,18kSAL = HMA bottom strain due to an 18-kip single- axle load (µε); εHMA,36kTAL = HMA bottom strain due to a 36-kip tandem- axle load (µε); σPCC,18kSAL = UTW corner (top) stress due to an 18-kip single-axle load (psi); σPCC,36kTAL = UTW corner (top) stress due to a 36-kip tandem-axle load (psi); ∆εHMA,∆T = additional HMA bottom strain due to temper- ature gradient (µε); ∆σPCC,∆T = additional UTW corner (top) stress due to tem- perature gradient (psi); αPCC = thermal coefficient of expansion of the PCC (ε/°F); ∆T = temperature gradient in UTW (°F); Ladj = adjusted slab length (in.) defined as (7) k = modulus of subgrade reaction (psi/in.); and le = effective radius of relative stiffness for a fully bonded system (in.) defined as: L Ladj = × − +      12 8 24 12 2 ∆ ∆∆σ αPCC T PCC adj e T L l , . . . = − × × + ×   28 037 3 496 18 382 ∆ ∆∆ε αHMA T PCC adj e T L l , . . . = − + × × + ×   28 698 2 131 17 692 log . . log ( ) . log . log . ,10 36 10 10 10 4 898 0 599 1 395 0 963 0 088 σPCC kTAL adj e e adj e k L l l L l ( ) = − × + ×   − × ( ) − ×   log . . log ( ) . log . log ,10 18 10 10 10 5 025 0 465 0 686 1 291 σPCC kSAL adj e e k L l l ( ) = − × + ×   − × ( ) log . . log ( ) . log . ,10 36 10 10 6 070 0 891 0 786 0 028 εHMA kTAL e e k l l ( ) = − × − × ( ) − × log . . log ( ) . log . ,10 18 10 10 5 267 0 927 0 299 0 037 εHMA kSAL adj e e k L l l ( ) = − × + ×   − × (8) where NA = neutral axis from the top of the PCC (in.) defined as (9) EPCC = modulus of elasticity of the UTW PCC (psi); EHMA = modulus of elasticity of the HMA (psi); tPCC = thickness of the UTW PCC (in.); tHMA = thickness of the HMA (in.); and L = actual joint spacing (in.). Once the pavement responses have been calculated through the use of these equations, the next step of the procedure is to calculate the predicted damage as a function of the expected traffic. For this purpose, two modes of failure have been identified in the procedure: (1) fatigue of the PCC at the cor- ner of the UTW and (2) fatigue at the bottom of the HMA. For fatigue of the PCC, the PCA fatigue cracking equa- tions are used (100). For a given stress-to-strength ratio (SR), the number of loads to failure (NPCC) is calculated as: For SR > 0.55 (10) For 0.45 ≤ SR ≤ 0.55 (11) For SR < 0.45 (12) Fatigue damage of the HMA was selected as the second possible failure criterion. For this, the Asphalt Institute method was employed (118). The failure criterion for this method is the number of loads that produce cracking in 20% of the wheelpath area. The equation is a function of the modulus of elasticity of the HMA (EHMA) in psi and the maximum strain at the bottom of the HMA layer (εHMA), as follows: (13)N EHMA HMA HMA = ×   ×  0 0795 1 13 29 0 854 . . . ε NPCC = ∞ N SRPCC = −( )4 25770 43248 3 268. . . log . . 10 0 97187 0 0828N SR PCC( ) = − E t E t t t E t E t PCC PCC HMA HMA PCC HMA PCC PCC HMA HMA ×   + × × +    × + × 2 2 2 E t t NA t k E t t t NA t k PCC PCC PCC PCC PCC HMA HMA HMA PCC HMA HMA × + × −    × −( ) + × + × − +    × −( ) 3 2 2 3 2 2 4 12 2 1 12 2 1 µ µ

21 The accumulated damage can then be calculated by employing Miner’s Hypothesis, which states that failure will occur when (14) This analysis procedure can be conducted by dividing the expected traffic loading into load groups with single and tandem axles of known weights. The number of loads to failure in both the HMA and the PCC can be calculated, and Eq. 14 can be used to determine the fraction of the fatigue life that has been consumed. It should be noted that some of the fatigue life in the HMA may have already been con- sumed by the trafficking before the UTW overlay; there- fore, there may be an initial value of fatigue damage for that failure mode. Benefits Before the publication of this synthesis report, very little guidance existed on the design of UTW pavements. Other design procedures, such as those in the AASHTO design guide, do not consider the unusual geometry and bond inher- ent in a UTW overlay (1). Additional benefits of such a design procedure are • Was developed through the use of a 3-D FEM analysis to more realistically predict the pavement responses, • Recognizes the importance of the bonding characteris- tics of the UTW to the HMA layer, • Was validated through the use of data from a number of field sites, and • Uses multiple failure criteria, recognizing the complex- ity of the UTW system. Limitations Although the approach used in this procedure provides a good foundation for a design method, a number of limitations have subsequently been identified (2). The following items require additional consideration: • To account for the partial bonding between the PCC and HMA layers, a correction term (multiplier) was derived for the corner stress prediction. However, this stress multiplier is based on a single field test location, mak- ing extrapolation prone to error. In addition, the use of a constant correction factor oversimplifies the UTW pavement response under load, especially because a con- stant value for the HMA resilient modulus was used. • The current design procedure does not account for reli- ability in the design. The only exception is an inherent reliability of the stress multiplier. n N i ii Load Groups   ≥ = ∑ 1 1 • In the current design procedure, two modes of failure govern the predicted service life: fatigue cracking at the corner of the UTW and fatigue cracking of the HMA layer beneath the UTW. Although HMA fatigue may be a contributor to UTW distress, HMA alligator cracking in the wheelpath, as the model used currently predicts, does not apply, owing to the differences in geometry and loading. Furthermore, the developers stated that for UTW pavements, “the asphalt layer is covered by con- crete slabs, and pavement rutting would not be the gov- erning distress.” However, from observations of UTW overlays in service, this mode of failure does appear to be significant, and it should possibly be considered. • The procedure contains the requirements for a number of inputs that may be difficult to define accurately by the typical user including, for example, a “design tem- perature differential.” Colorado TWT Design Procedure In Colorado, TWT sections from 125 to 175 mm (5 to 7 in.) thick, with joint spacings of up to 3.7 m (12 ft), were instru- mented to measure critical stresses and strains as a result of traffic loads and temperature differentials. These results were used to develop a design procedure for TWT overlays (17). In developing the procedure, theoretical design equations for the prediction of critical stresses and strains in the white- topping system were devised first. Correction factors were then calculated to account for the experimental data collected. More specifically, the correction factor adjusts the theoreti- cally predicted fully bonded stresses to the partially bonded stresses seen in the field. It also adjusts for the expected tem- perature gradient. In this procedure, the location of critical stress is hypothe- sized as being centered along a longitudinal free edge joint, such as a PCC curb and gutter. Because the joints loaded by traffic will most likely not be in a free edge condition, the design procedure considers tied longitudinal joints. The rela- tionship between the free edge and tied edge stress is given as (15) where σFE is the longitudinal free joint load-induced stress and σTE is the longitudinal tied joint load-induced stress. Comparison of the theoretical stresses with the measured tied edge loading stresses showed that the measured stresses are greater than the theoretical stresses. The correction fac- tor was reported as (16) where σEX is the measured experimental partially bonded inter- facial stresses and σTH is the theoretical fully bonded interfacial stresses. σ σEX TH= ⋅1 65. σ σFE TE= ⋅1 87.

22 of relative stiffness for the fully bonded slabs. Adjustments are made to the stress equations to account for partial bond and the loss of support as a result of temperature-induced curling. Both PCC and HMA fatigue relations were used as fail- ure criteria in this procedure. The number of allowable load repetitions for the PCC is a function of the flexural stress- to-strength ratio (SR). For the HMA, the number of allow- able loads is a function of the maximum tensile strain in the HMA layer, the HMA modulus of elasticity, and the volume of binder and air voids. The amount of fatigue damage that the HMA has sustained before whitetopping construction is also considered. A subsequent review of this procedure has revealed the following (2): • TWT overlay thickness is substantially dependent on the subgrade modulus of reaction. • TWT thickness is also sensitive to the HMA stiffness and thickness. It was stated in the procedure that the thickness of the HMA layer should not be less than 125 mm (5 in.). • TWT thickness is not sensitive to the number of 80 kN (18 kip) ESALs for greater than 1 million applications. However, when the loads increase, the stiffness of the HMA becomes a more significant factor. For traffic loading greater than 4 million ESALs, TWT should not be specified if the HMA modulus is less than 2.8 GPa (400 ksi). • No guidance is provided on the proper selection of the design temperature gradient. To assess the PCC–HMA interface, the strains were mea- sured in the top of the HMA pavement and also in the bottom of the PCC slab in the field. It was found that the strain in the HMA is less than the strain in the PCC pavement. An equa- tion relating these was given as (17) where εAC is the strain in the HMA surface and εPCC is the strain in the PCC bottom. Because the coefficient in this equation is not 1.0 demon- strates that there is only partial bonding between the PCC and the HMA. The effect of the changing temperature gradient on the load-induced stresses was also assessed. The following equa- tion was given to report the change in the stress as a function of the temperature gradient: (18) where σ% is the percent change in stress from zero tempera- ture gradient and ∆T is the temperature gradient, °F/in. As a result of this analysis, design equations were devel- oped for the prediction of PCC stresses from both 90 kN (20 kip) single-axle loads and 180 kN (40 kip) tandem-axle loads. Similar equations were also developed for the HMA strain. All of these equations depend on the effective radius σ% .= ⋅4 56 ∆T ε εAC PCC= ⋅0 842.