Category HIGHWAY ENGINEERING HANDBOOK

Perhaps the most important step in designing a pavement is the estimation of the design traffic. Overestimation of the design traffic results in a thicker pavement than necessary with associated higher costs. Underestimation of traffic results in a thin pavement that will fail prematurely, resulting in higher maintenance and user costs. If the proposed pavement will be used to replace an existing pavement, the design traffic could be a projection of the existing traffic. If the proposed pavement is a new loca­tion, the design traffic will have to be estimated on the basis of the proposed use of the pavement. For design purposes, all traffic is equated to an equivalent 18-kip (80-kN) single-axle load, or ESAL. Each vehicle in the expected design traffic volume is converted to an ESAL by an equivalency factor. The equivalency factor is a function of the axle loading, pavement thickness, axle configuration, and terminal serviceability. As dis­cussed in Art. 3.6, the terminal serviceability is an index of the serviceability of a pavement immediately before rehabilitation is needed.

The equivalency factors as given by the AASHTO Pavement Design Guide are pre­sented here for flexible pavements in Tables 3.1 through 3.9, and for rigid pavements in Tables 3.10 through 3.18. For each pavement type, the tables are arranged by axle configuration and terminal serviceability pt. Factors are included for single-axle, tandem-axle, and triple-axle configurations, and for pt values of 2.0, 2.5, and 3.0. In the tables for flexible pavements, the pavement strength is characterized by a pavement structural number (SN), which is defined in Art. 3.7. The use of the tables is illustrated by the following example.

Consider a 30,000-lb (133-kN) transit bus that has a single front axle load of 10,000 lb (44 kN) and a tandem rear axle load of 20,000 lb (89 kN). Before the ESAL can be determined, the pavement thickness or structural number must be known, as well as the terminal serviceability. In an initial design, this necessitates assumptions, and very likely an iteration after the thickness or structural number has initially been determined. In this example, the ESAL is to be determined for a rigid pavement 7 in (178 mm) thick and for a flexible pavement with a pavement structural number of 4. The pt is taken as 2.5. The tables show that, for this case, the equivalency factor for rigid pavement is 0.089 for the front axle (Table 3.13) and 0.220 for the rear axle (Table 3.14). The equivalency factor for flexible pavement is 0.102 for the front axle (Table 3.4) and 0.141 for the rear axle (Table 3.5). Each bus equals 0.089 + 0.220 = 0.309 ESAL for rigid pavement and 0.102 + 0.141 = 0.243 ESAL for flexible pave­ment. A similar analysis would be completed for each vehicle type. A worksheet for making the calculations is provided in Table 3.19, and an example for using the work­sheet is presented in Table 3.20.

The traffic supplied to the designer is usually the total traffic in both directions and all lanes. This traffic needs to be distributed by direction and lane to determine the required pavement thickness. The pavement is first divided by direction by multiplying by the directional factor. In most cases, this factor is equal to 0.5, assuming the loads are distributed equally in both directions. In some cases, the directional factor may be

greater than 0.5. An example would be an industry where material is hauled in by truck and shipped out by rail. In this case, loaded trucks would be going into the plant and empty trucks would be exiting the plant. The next factor is the lane distribution factor. As more lanes are added to a section of road, the traffic will be more distributed among these lanes. However, trucks tend to use the outermost lane, so the distribution of ESALs is not in proportion to the number of lanes added. Many of the state DOTs have developed lane distribution factors for use in pavement design. The AASHTO Pavement Design Guide presents a range of factors used for lane distribution as given below. It should be noted that for the same traffic, the thickness design will be greater for the pavement with the smaller number of lanes.

Abbreviated procedures for determining ESALs have been developed by several states. These procedures usually involve grouping classifications of trucks into several categories and assigning an average equivalency factor to these categories. For example, Ohio groups trucks into two categories, single, or C units, and tractor-trailer, or B combina­tions. The average equivalency factors used by Ohio for these two categories are shown in Table 3.21.

The AASHTO pavement design equations have some variables that are common to both rigid and flexible pavements, including serviceability, traffic loading, reliability, overall standard deviation, and roadbed soil resilient modulus. These parameters are discussed in the following articles. Subsequently, the design procedure is presented for rigid pavements in Art. 3.6 and for flexible pavements in Art. 3.7.

3.3.1 Serviceability

The AASHTO design equations are developed around the concept of serviceability, which serves as the pavement performance parameter by which a pavement’s condition is valued. Present serviceability is defined as the momentary ability of a pavement to serve traffic. The present serviceability rating (PSR) was developed to measure service­ability. PSR is a rating of pavement ride based on a scale of 0, for impassible, to 5, for perfect. For the development of the original AASHO equation, individuals (the raters) would ride the pavements and assign a PSR value. To avoid riding and rating every pavement to determine serviceability, a relationship is usually developed between PSR and measurable pavement attributes. The value determined by this relationship is called the present serviceability index (PSI). At the AASHO Road Test, the PSI was derived to be related to slope variance, cracking, and patching for concrete pavements, and to slope variance, rutting, cracking, and patching for asphalt pavements. The relationship between pavement thickness and serviceability index is defined by the AASHTO pavement design equations.

AASHTO has given interim approval for a new approach to pavement design as described in the AASHTO Interim Mechanistic-Empirical Pavement Design Guide Manual of Practice. Several years in development, this M-E pavement design guide and the accompanying software should provide a significant advancement in pavement performance prediction. As its title implies, mechanistic-empirical models are used to analyze input data for traffic, climate, materials, and the proposed pavement structure, and then to estimate pavement service life damage. The distress prediction models have been calibrated to national averages based on data gathered by the Long-Term Pavement Performance program. However, for the distress models to be fully applica­ble for the particular materials, construction practices, and environmental conditions in a given region, they must be calibrated with local data. The program can best be used by knowledgeable practitioners as application experience is gained. The tradi­tional methods are the focus of this handbook.

Perhaps the most widely used pavement design method in the United States and throughout the world is that presented in the AASHTO Guide for Design of Pavement Structures. A long history of pavement studies has led to the current (1993 with 1998 supplement) edition.

The developments leading to the current AASHTO design procedure began with the Bates Experimental Road, which was constructed in 1922 near Springfield, Illinois. The purpose of the experimental road was to determine what factors affected pavement per­formance. The researchers found that pavement performance could be correlated with truck loading. No further major research was conducted over the ensuing 25 years.

The changes in truck configuration and expansion of the highway network resulting from World War II brought pavement performance to the forefront again. In 1949, the Council of State Governments held a meeting in Columbus, Ohio. At this meeting, highway officials decided there was a “need for more factual data concerning the effects of axle loads of various magnitudes on pavements.” The effort to advance the science of pavement design was led by the American Association of State Highway Officials (AASHO, which later became AASHTO). The regional AASHO associations decided to construct test pavements in each region. The first of these test roads was constructed by the Southeastern AASHO states. Named Road Test One, two test loops were con­structed in 1950 near La Plata, Maryland, each loop containing two 12-ft-wide (3.7-m) pavement lanes. All sections constructed were concrete with a pavement thickness of 7 in (178 mm) thickening to 9 in (229 mm) at the edge of the pavement. Each lane of a loop carried only one loading and axle configuration.

A second regional test road was constructed by the Western Association of State Highway Organizations (WASHO). Named the WASHO Road Test, two test loops were constructed in 1952 near Malad, Idaho, each consisting of two 12-ft (3.7-m) lanes. All pavement was comprised of asphalt concrete on a crushed aggregate base, constructed on subbases from 0 to 16 in (406 mm) thick. Each lane of a loop carried only one load­ing and axle configuration. Because of the limited number of sections constructed in Maryland and Idaho, a rational design procedure could not be developed.

In 1951, support was growing within AASHO for an expanded road test. This led to the construction of the AASHO Road Test near Ottawa, Illinois, which contained six loops with two 12-ft (3.7-m) lanes. The AASHO Road Test contained 468 asphalt sec­tions and 368 concrete sections. Each lane of a loop carried only one loading and axle configuration. A total of 1,114,000 load applications were applied over a 2-year period.

The rigid pavements in the AASHO Road Test were concrete slabs ranging in thickness from 2lA to 121/2 in (63 to 317 mm) thick. The slabs were placed either on a granular subbase or directly on the subgrade. Flexible pavements at the AASHO Road Test consisted of asphalt pavements placed on a base and/or subbase. As confirmed by these tests, rigid pavements carry traffic loads through beam action whereas flexible pavements carry traffic loads by spreading the stress through the underlying layers.

Unpublished preliminary results from the road test were released to the states in 1961 and 1962. The AASHO Interim Guide for Design of Pavement Structures was published in 1972. Chapter 3 of the interim guide was revised in 1981. The first edition of the AASHTO Guide for Design of Pavement Structures (1986) introduced many new con­cepts including the reliability concept. It was published in two volumes, the first giving design procedures and the second providing documentation and explanatory information. The second edition of the guide was published in 1993 and a supplement in 1998.

Rigid pavement constructed with an asphalt overlay is referred to as composite pavement. The advantage of constructing an asphalt overlay on a rigid pavement is solely in the areas of ridability and noise. Rigid pavements are considered by most to create more road noise inside a vehicle than flexible pavements. This phenomenon is largely due to the surface texture specified for rigid pavements to ensure proper skid resistance. By specifying an asphalt overlay with the rigid base, surface texture requirements can be relaxed and noise can be reduced.

There are few documented composite pavement design procedures available to determine the proper thickness ratios between the rigid base thickness and the flexible surface thickness. One way to determine an equivalent composite thickness buildup can be done using elastic layer theory. A convenient computer program called ELSYM5 (public domain) can be used for analysis of different layer combinations, provided the designer is willing to make some assumptions. By accepting the assump­tions, the designer is getting results that are only approximate but relative. ELSYM5 is based on elastic layer theory, which is not entirely appropriate for rigid pavement, since rigid pavement is not continuous, isotropic, and homogeneous in all directions. The procedure involves calculating the required rigid slab thickness for the conditions present where the composite pavement will be constructed. This is done using the AASHTO Pavement Design Guide or another method. The second step is to analyze the required rigid slab using ELSYM5 under the conditions designed for by calculating the deflections, strains, or stresses predicted under the maximum legal loading config­uration. Finally, using a trial-and-error procedure, replace up to 3 in (75 mm) of the rigid slab with enough thickness of asphalt to achieve the same deflections under the same loading scheme. A rule of thumb is to replace the first 1 in (25 mm) of slab thickness with 3 in (75 mm) of asphalt concrete. However, there is not a linear relationship of 1 in (25 mm) of PCC to 3 in (75 mm) asphalt concrete; for additional reductions in the rigid slab thickness, the elastic layer theory is relied upon to calculate equivalent deflections.

Because a composite pavement behaves more like a rigid pavement, special treat­ment is required for the transverse joints. Reflective cracking, cracks that propagate from the rigid pavement joint through the asphalt overlay, can be an intolerable distress which induces a rough riding pavement. The reflective crack allows water to enter, which induces stripping in the asphalt and slow deterioration into a spalled pothole. The suggested treatment to counter the reflective cracking is to saw and seal the asphalt con­crete overlay directly above the concrete joint. The joints should be sawed as soon as the asphalt overlay is placed, and the joint sealant reservoir should be constructed the same way as discussed previously. Figure 3.8 shows a joint in a composite pavement that has been properly sawed and sealed.

Asphalt concrete pavement, also referred to as flexible pavement, is a mixture of sand, aggregate, a filler material, and asphalt cement combined in a controlled process, placed, and compacted. The filler material can range from quarry crushing dust and asphalt-plant baghouse fines to wood fibers (cellulose). There are many additives that can be used in asphalt concrete mixes to encourage thicker cement coatings, more elastic mixes, stiffer mixes, and less temperature-sensitive mixes. Flexible pavements can be of a type constructed on a prepared subgrade, which is called full-depth asphalt concrete pavement (FDACP), or of a type built on an untreated granular base, which is not as carefully identified by the industry but is referred to herein as deep-strength asphalt concrete pavement (DSACP). (See Arts. 1.5.3 and 1.5.5.)

Flexible pavements are designed to bend and rebound with the subgrade. The design concept is to place sufficient layers of base and intermediate courses of pavement so as to control the strains in the subgrade so that no permanent deflections result. Loading of an asphalt pavement requires the stiffest layers to be placed at the surface with successively weaker layers down to the subgrade. The types and thicknesses of subbase materials placed above the subgrade should be selected with consideration of the strength of the subgrade. Very weak subgrades, after compaction, can lose compaction when very stiff aggregate bases are placed above. It is often advantageous to place a granular subbase, which is much weaker than an aggregate base, above weak subgrades to ensure that com­paction is sustained. Most, if not all, flexible pavement design procedures are based on a combination of elastic layer theory and experience. The elastic layer theory is used to cal­culate strains in each of the layers so as to ensure that excessive deflections will not occur. The experience is related to performance parameters that predict the number of times the pavement can bend (loadings) until cracking results.

As the name implies, continuously reinforced concrete (CRC) pavement is a rigid pave­ment constructed with continuous longitudinal reinforcement. No transverse joints are installed. Instead, the pavement is allowed to develop random transverse cracks, and the steel reinforcement holds the cracked sections together. The size and spacing of the cracks are influenced by the percentage of reinforcing steel used. Current practice calls for 0.6 to 0.7 percent of the slab cross-section area. The design of the reinforce­ment is covered in the AASHTO Pavement Design Guide. The thickness of the slab is determined the same way as for other concrete pavements.

>

FIGURE 3.7 Typical preformed joint seal in rigid pavement. Conversion: 1 in = 25.4 mm.

Joint sealing prohibits the infiltration of water into the pavement base and prevents incompressibles from lodging within the joint cavity. The advantages of keeping water out from under a pavement are documented extensively in the AASHTO Pavement

FIGURE 3.5 Layout of joints in rigid pavement at right-angle intersection. Conversions: 1 in = 25.4 mm, 1 ft = 0.305 m.

Design Guide and in various articles in this chapter. With an unsealed joint, contraction under cooler temperatures allows joint cavities to open up and become filled with sand, stone, and other incompressible material. When warmer temperatures try to expand the length of the pavement, the joints are unable to close, compressive stresses develop, and spalling may result.

The purpose of a sealant reservoir (Figs. 3.1, 3.2, and 3.6) is to prevent water and incompressibles from entering the joint cavity. The design criteria for the sealant reservoir ensure that the sealant stays in place. The ability of the sealant to expand and contract with the movement of the joint is a function of the material properties of the sealant (defined by the manufacturer’s specifications) and the expected movement of the joint. Joint movement can be calculated using the following relationship:

AL = CL(A AT + Z) (3.3)

where AL = joint opening created by changes in temperature and loss of moisture during curing (joint movement), ft (mm)

C = constant used to adjust for friction between bottom of slab and the mater­ial that directly supports the pavement (0.65 for granular material, 0.80 for stabilized material)

L = joint spacing, ft (mm)

A = thermal coefficient of concrete, 10 6/°F (10 6/°C), multiplied by the following factors depending on coarse aggregate type: quartz, 6.6 (11.9); sandstone, 6.5 (11.7); gravel 6.0 (10.8); granite, 5.3 (9.5); basalt, 4.8 (8.6); and limestone 3.8 (6.8)

AT = difference in minimum temperature pavement will be subjected to and temperature at which pavement was placed, °F (°C)

Z = drying shrinkage coefficient of the portland cement concrete (PCC) slab, in/in (mm/mm). The coefficient depends on the indirect tensile strength, lb/in2 (MPa) as follows: 0.0008 for 300 (2.1), 0.0006 for 400 (2.8), 0.00045 for 500 (3.4), 0.0003 for 600 (4.1), 0.0002 for 700 (4.8)

(See “AASHTO Design Procedures for New Pavements,” FHWA Report HI-94-023, ERES Consultants, Inc., February 1994; and FHWA Technical Advisory T 5040.30, November 30, 1990.)

There are two categories of joint sealants. The field-molded sealant and the pre­formed compression seal are used extensively in rigid pavements. Also, field-molded sealants are gaining acceptance and being used in flexible pavements.

For field-molded sealants, the design is very simple and is controlled by the follow­ing relationship:

W = — (3.4)

where AL = joint opening to be accomodated, in (mm)

W = design width of transverse contraction joint, in (mm)

S = allowable strain specified by sealant manufacturer (typically 25 to 50 percent for field-molded sealants)

To control the strain in field-molded sealants, manufacturers recommend a reservoir shape factor (width to depth), and the use of a backer rod as illustrated in Fig. 3.6. The purpose of the backer rod is to prevent bond at the bottom of the sealant reservoir where the actual crack in the pavement exists. It is at this crack that the greatest strain will occur. Typical joint sealants are either asphalt-based or silicone-based.

For preformed compression seals, the uncompressed width of the compression seal should be chosen according to manufacturer’s specifications, as the material response characteristics are of primary importance. The calculated movement of the joint, normal­ized by the width of the uncompressed seal, should be less than or equal to the allowable movement of the compression seal, as determined by the manufacturer. Figure 3.7 shows a typical preformed seal installed in a pavement joint.

The following basic principles must be observed in developing a correct jointing detail:

1. Never taper concrete down to less than 2 ft (610 mm) in width.

2. Depending upon the amount of transverse reinforcing steel, be careful of the number of lanes that are tied together. In JPCP, tying more than three 12-ft (3.7-m) lanes together may result in uncontrolled longitudinal cracking.

3. Always design the secondary (intersecting) route as independent in movement from the primary route. Thus, as the primary route expands and contracts, no unnecessary forces will be created in the secondary route.

4. Provide for expansion wherever payment is interrupted in its longitudinal direction.

5. Terminate joints at 90° to any intersecting joints, obstructions, or edges of pavement.

6. Where possible, lay out lane widths of the same dimension. This permits the contractor to pave all the lanes without changing the paving machine setup dimensions.

7. Unless unavoidable, all joints should be in a straight line. Curved joints are diffi­cult to saw and generally require additional forming.

8. For plain (nonreinforced) concrete pavement, the slab length/slab width ratio should not exceed 2:1.

Intersection details should always be included in construction plans. A proper jointing layout ensures that cracking occurs at locations where load transfer exists (contraction joints) and away from wheel paths (longitudinal joints). The jointing detail should be a separate detail in the plan to eliminate confusion and allow field personnel to easily lay out the intersection without construction delay. Figures 3.4 and 3.5 show jointing layouts that have been used for typical intersections.

Jointed rigid pavements tend to crack at 13 to 25 ft (4 to 8 m) lengths because of (1) initial shrinkage after placement as excess water evaporates, (2) temperature-induced expansion and contraction resisted by friction with the subgrade, (3) curling and warping caused by temperature and moisture differences between the top and bottom of the slab, and (4) load – induced stresses.

As slabs contract as a result of seasonal temperature changes, cracks form and widen, or formed joints widen, allowing incompressible materials into the cracks or joints. Subsequently, expansion is hindered and pressure is built up in the pavement. This pressure can result in pressure spalling or even blowups. To control this, partial depth saw cuts are made at regular intervals which induce concrete to crack at these locations. The timing and depth of these saw cuts are critical to ensure that the pavement cracks at the controlled location. Saw cuts should be made as soon as the pave­ment can support the weight of the saw and operator. The saw cuts should be made at a depth of one-third of the slab thickness for longitudinal joints, and one-fourth of the slab thickness for transverse joints. These saw cuts are then sealed with some type of joint sealer to prevent intrusion of incompressibles. If the saw cut interval (joint spacing) is short enough, intermediate cracks are eliminated. If longer intervals are used, interme­diate cracks will form.

Load transfer is the critical element at joints and cracks. In undowelled, unrein­forced pavements, any load transfer must be provided by aggregate interlock.

Aggregate interlock is lost when slabs contract and the joints or cracks open up. Also, interlock is slowly destroyed by the movement of the concrete as traffic passes over. Given large temperature variations and heavy trucks, aggregate interlock is ineffectual, and faulting is the primary result.

To provide load transfer at the joints, dowels are used which allow for expansion and contraction. Figure 3.1 illustrates a typical doweled joint with saw cut and joint seal. Figure 3.2 shows a similar joint without the dowel to provide load transfer.

Where a long joint spacing is used and intermediate cracks are expected, steel rein­forcement is added to hold the cracks tightly closed (JRCP). This allows the load transfer to be accomplished through aggregate interlock without the associated problems described above. Contraction joints do not provide for expansion of the pavement unless the same amount of contraction has already taken place. This contraction will initially be from shrinkage due to concrete curing. Later changes in the pavement length are due to temperature changes.

Where fixed objects such as structures are placed in the pavement, the use of an expansion joint is warranted. Expansion joints should be used sparingly. The pave­ment will be allowed to creep toward the expansion joint, thus opening the adjacent contraction joints. This can cause movement in the adjacent contraction joints in excess of their design capabilities and result in premature failures. Figure 3.3 shows a detail for a typical expansion joint.

The design of reinforcing steel is a function of seasonal temperature change, subbase friction, and the weight of the slab. Inadequate reinforcing will not be able to hold the

cracks together, and faulting will result. The amount of reinforcing needed to hold cracks together is traditionally calculated using a relationship based on the friction between the subgrade and the bottom of the slab. This relationship assumes that for a crack to open enough to fail the aggregate interlock, the slab will have to slide along the subbase. The current AASHTO recommendation is based on this traditional approach (Guide for Design of Pavement Structures, American Association of State Highway and Transportation Officials, 1993).

The relationship can be expressed as follows:

where As = area of steel required, in2/ft of width (mm2/mm)

W = weight of slab, lb/ft2 (MPa)

F = coefficient of resistance between slab and subgrade (1.5 unless otherwise known)

L = length of slab, ft (mm)

fs = allowable stress in steel, lb/in2 (MPa)

Although these relationships are accepted by most leading authorities and are referred to in almost every reference on the subject, it is important to understand that they make many assumptions about physical quantities that are seldom consistent throughout a length of pavement. For instance, the friction factor can be affected by something as insignificant as a large footprint in the base course prior to paving. Additionally, the environment can play an important role as water and salt erode the steel, thus reducing the sectional area of the steel.

Where reinforcement is not desired, slab lengths must be chosen so that intermediate transverse cracks are eliminated. The most current theory used to determine allowable slab lengths involves a very old concept developed by Dr. H. M. Westergaard. Westergaard defined a constant called the radius of relative stiffness as an algorithm that relates the modulus of subgrade reaction to the flexural stiffness of the slab. Research indicates that cracking can be expected when the ratio between the slab length

and the radius of relative stiffness is greater than 5. The radius can be calculated from the following equation (Federal Highway Administration Technical Advisory T 5040.30, November 30, 1990):

where l = radius of relative stiffness, in (mm)

E = modulus of elasticity of concrete, lb/in2 (MPa) h = pavement thickness, in (mm)

k = modulus of subgrade reaction, (lb/in2)/in (MPa/mm) v = Poisson’s ratio (0.15)