Category WATER IN ROAD STRUCTURES

In countries that suffer from large amounts of rainfall, the asphaltic wearing sur­faces are often constructed of open graded asphaltic mixtures. The high perme­ability of these wearing surfaces ensures a fast drainage of the water away from

the surface, avoiding hydroplaning and bad visibility conditions due to ‘splash and spray’, Fig. 5.15, and thus improving the overall road safety.

Porous asphalt uses aggregate with a moderate to coarse median particle size and a very steep grading curve – i. e. the majority of the stones in the mixture are of a similar size. This has the effect of developing a mixture with a very high air void volume (20-30%) with stones only adhering to each other by virtue of the films of bitumen at their point of contact. In this way its porosity is very high compared with conventional asphaltic material and water does not easily collect on the surface during rainstorms.

An added benefit not included in the original concept, but now an important driv­ing force for the wider adoption of porous asphalt surfacings is the reduced traffic noise from pavements with these surfacings. The porous nature reduces tyre-surface interaction sounds and acts as a partial absorbent of other vehicle induced noise. Typically they provides a 3-5 dB(A) noise reduction over conventional pavement surfacings. Even greater benefits can be achieved by using two-layer porous asphalt with a finer, filter, layer over a coarser, drainage layer. Noise reduction may then be 8 or 9 dB(A) quieter than conventional asphaltic mixtures and 4 dB(A) quieter than a single-layer porous asphalt.

As the surfacing is so permeable, rain can drain vertically into the porous asphalt layer before being conveyed laterally within the pavement. Typically a porous as­phalt surfacing will have a thickness of between 20 and 100 mm and be placed on top of an impermeable asphaltic base. Hence, water flowing in the surfacing cannot continue to flow vertically but is forced to travel sideways, exiting from the layer at its edge. Unless this edge is free, special attention must be paid as to how the water is to be collected and led away from the pavement. An impermeable edge, such as a conventional kerb, would dam the water within the layer. Consequently special kerbs with inlets and pipe systems have been developed to lead the water into a conventional surface drainage system (Highways Agency, 1997).

Despite the advantages of the material in providing relatively dry surfaces in wet weather, the material and its use pose a number of problems:

1. Lack of durability. Careful mixture design is needed to ensure that there is enough bitumen to coat the stones and ensure longevity of performance – too much and the mixture may rut too readily and the pores become blocked by bitumen (preventing drainage). Too little bitumen and ravelling will be likely (as described in Section 5.5.1), particularly in cold weather when ice could form in the pore space forcing the topmost layers of stone loose. An added issue is that the greater opportunity for bitumen to react with atmospheric oxygen, because of air in the voids, tends to lead to early embrittlement of the bitumen (Herrington et al., 2005). Bitumen film thickness is, thus, of particular importance. Hence, both design and construction practice require careful attention, perhaps more so than for conventional asphaltic mixtures. More detailed coverage of this topic is beyond the scope of this chapter, but interested readers may wish to consult NAPA (2004).

2. Clogging due to ingress of particulates. Small particles and dust, that comes from the environment, blown soil, engine wear, brake wear and from cargoes (see Chapter 6, Section 6.2), tend to get washed into the pore space of the porous asphalt, thereby blocking it. In thin porous surfacings on high-speed roads it ap­pears that reduction in permeability is not of great concern. After some initial de­terioration, further clogging is often not significant, probably because high-speed traffic develops high transient water pulses in the pores of the asphaltic mixture during wet weather, causing a self-cleansing action (Bendtsen et al., 2005). In slower speed roads this action is not evident and clogging is, typically, pro­gressive. These authors monitored an urban test road in Denmark comprising 3 porous asphaltic surfacings and a control surface (Table 5.6).

Using an infiltrometer somewhat like that of Cooley (1999), see Section 5.4.1, Bendtsen et al. (2005) observed, Fig. 5.16, that clogging developed quite rapidly in the finer asphaltic pavements. The two porous pavements with 5 mm aggre­gate, Sections II and III, were effectively clogged after 15-20 months whereas Pavement I with 8 mm aggregate remained in a much better condition. The rea­son for this clogging is believed to be the dirt and fine material from the adjacent dense asphaltic concrete pavement being dragged onto the porous pavements by vehicle tyres, since clogging first appeared at the position nearest to the reference section.

70

«Г

I 60

E

50

40

30

о

£ 20

о

E [11]

3. Cold weather clogging by ice and snow. BackstrOm and BergstrOm (2000) evalu­ated the function of porous asphalt in cold climates using a climate room. At the point of freezing point, the infiltration capacity of porous asphalt was approx­imately 50% of the infiltration capacity at +20°C. They simulated conditions of snowmelt by exposing the porous asphalt to alternating melting and freezing over a period of 2 days and found that the infiltration capacity was reduced by approximately 90%.

To overcome clogging, cleaning devices have been developed. These (Figs. 5.17 and 5.18) typically comprise high pressure water jets that aim to disturb and erode fines resting in the pores of the porous asphalt, washing them to the surface from where they may be vacuum-collected by the cleansing machine. It is interesting to note that the results of Bendtsen et al. (Fig. 5.16) show that there is not a significant reduction in the level of clogging due to cleaning (compare ‘before’ (b) to ‘after’ (a) readings in the figure) and that cleaning did not bring back performance close to the original function.

5.3 Conclusions

A range of equipment exits to determine permeability of asphaltic mixtures both by in-situ and laboratory testing. At present the values of permeability collected seem, mostly, to be being used for relative performance assessments and they are not much integrated in whole-pavement water regime modelling. Current advances in computational engineering and mechanical and physio-chemical testing enable the identification of the actual physical processes of water-induced damage in asphaltic mixtures and the evaluation of their effects on the total mixture response. It is hoped that, over a period of time, availability of these resources will enable a gradual tran­sition in mixture design from a design-by-testing approach to an approach in which design is by identification of the optimal choice of individual mixture components on the basis of their physio-chemical and mechanical characteristics and interac­tions. The increasing use of porous asphalt for noise reduction and spray reduction purposes is an important challenge to the pavement engineer. Purposely allowing water into the structure provides the opportunity for much greater and faster ravel­ling. There are also concerns about clogging due to washed-in fines and due to ice formation. Rehabilitating clogged porous asphalt without causing premature dam­age is a challenge.

Acknowledgments Some of the work described in this chapter has been performed with financial, computational or experimental support of Ooms Nederland Holding Bv, the Section of Structural Mechanics of TU-Delft (both in the Netherlands) and the Turner Fairbanks Highway Research Center of the US Federal Highway Administration. The authors wish, therefore, to express their thanks to Dr. A. de Bondt, ir. C. Kasbergen and Dr. J. Youtcheff respectively for this.

Performance characteristics of bound pavement layers are known to be influenced by water-borne pollutants that cause changes in mechanical behaviour, ageing and degradation. With the exception of Portland cement concrete (PCC) pavements, this has not yet received much attention. Asphalt pavements are not seriously affected by inorganic pollutants, but most of the organic chemicals, including gasoline and motor oil, soften up or break down the asphalt binder leaving the asphaltic layer vulnerable to further degradation. Damage of the surface layer, due to ageing (stiff­ening because of ultraviolet light), traffic induced cracks and chemical degradation, opens an ingress route in the pavement system for pollutants from the surface.

In hot climates, salt can be moved by evaporating water to near the pavement surface. This may result in the expansive crystallisation of salt in voids in asphaltic mixtures (or other bound layers) just below the road pavement’s surface leading to blistering of the running surface, to cracking of the pavement and to overall degra­dation. Guidelines to the understanding and treatment of this issue are available (Obika, 2001). In temperate climates, salt (NaCl) seems, usually, not to be a sig­nificant contributor to damage of asphaltic materials. It may, sometimes, accelerate deterioration of poor quality materials, but it appears that it is water damage itself which is the primary cause (see Section 5.5). However, the chemistry of the wa­ter in the pores of asphaltic materials can have an important influence on whether stones and binder adhere efficiently. Greater alkalinity (i. e. higher pH) potentially
results in increased rates of moisture damage, although Calcium Hydroxide (slaked lime) dissolved in the water doesn’t have this effect even though pH rises (Little & Jones, 2003). In cold climates, salt has been implicated in causing damage. Ac­cording to Hudec and Anchampong (1994) certain fine grained aggregates degrade rapidly during wetting and drying cycles and during freeze-thaw cycles especially if deicing salts have been used abun-dantly. The extensive use of chlorides has also been reported to cause accelerated pavement deterioration (Dore et al. 1997, Saarenketo, 2006).

In PCC, deterioration is related to complex processes associated with physical and chemical alteration of the cement paste and aggregates. One major chemical degradation mechanism resulting from the long-term application of the popular chemical de-icer sodium chloride (NaCl) is the dissolution of calcium hydroxide (Ca(OH)2). Another common de-icer, CaCl2, is associated with a deleterious chem­ical reaction with PCC. The chemical attack is accompanied by the formation of hydrated calcium oxy-chloride according to the following reaction:

3Ca(OH)2 + CaCl2 + I2H2O ^ 3CaO ■ CaCk ■ I5H2O

This reaction is considered to be disruptive to the concrete matrix because of the expansive pressures generated. Another potential detrimental effect of the ap­plication of chemical de-icing salts is increased alkali – silica reactivity (ASR), which is a distress caused by undesirable chemical reactions between alkalis in the cement paste (Na2O and K2O) and the reactive siliceous components of susceptible aggregates. The product of the reaction is expansive in the presence of moisture, destroying the integrity of the weakened aggregate particle and the surrounding ce­ment paste. When aggregates like dolomitic limestone are used there is a possibility of alkali-carbonate reactivity, where alkalis react with carbonate aggregates. Besides these mentioned processes, there is also the possibility of external and internal sul­phate attack, which can cause deterioration. Other de-icing chemicals (magnesium chloride, calcium magnesium acetate, Ca-acetate, Mg-acetate, urea, etc.) may also have damaging effect on PCC pavement layers (MTTI, 2002).

Hydraulically bound mixtures may be considered as low strength PCC. From this point of view, the effects of the pollutants (mainly de-icers and sulphate) are similar to those on PCC pavements, except that chemical degradation, deterioration and loss of strength of the hydraulically bound layer will be quicker than in PCC layers, leading to higher stress on the layers beneath and faster degradation of the pavement.

The motivation for the following micro-scale finite element simulation, is the on­going discussion about cohesive versus adhesive failure mechanisms in asphaltic mixtures. It is the authors’ belief that, depending on the ability of the individual components and the bond between them, either one of these failure mechanisms may be dominant. It is, therefore, of paramount importance to establish the fun­damental relations between environmental weathering and material strength and stiffness. These relations can be used to assist the designer to optimize the water damage resistance characteristics of the mixture at purchase time on the basis of the response of the individual components.

In Fig. 5.12 details are shown of a micro-mechanical mesh that has been utilized for simulation of the results of pumping action due to traffic loading in a porous mixture (Kringos & Scarpas, 2005a).

Depending on the specified characteristics of the individual mixture components, cohesive Fig. 5.13(a) or adhesive Fig. 5.13(b) failure can occur.

Using the same approach, it is possible to study the combined effect of water – induced damage and traffic loading. Because the time-scale of water-induced dam­age accumulation compared to the time-scale of the mechanically-induced damage differs by several orders of magnitude, mechanical damage can be combined with water-induced damage, at discrete time-intervals. In terms of the finite element sim­ulation, this implies that diffusion studies can be performed until a desired level of water content is attained and the associated water-induced damage in the specimen can be computed. Subsequently, mechanical loading is imposed on the specimen at discrete time intervals and the total damage at a particular time interval can be computed. As can be seen from Fig. 5.14, for the chosen set of material parameters, stripping of the mastic film from the aggregate can be simulated.

Moisture diffusion measurements are still not performed very commonly for as­phaltic materials. There are currently two main test procedures being utilized. The first is an overall measurement of the increase of weight as a sample is ex­posed to a controlled moisture conditioning (Cheng et al., 2003). The second is a slightly more complicated procedure using Fourier transform infrared spectroscopy (Nguyen et al., 1992). It is quite challenging though, to utilize the available test data, since the values can vary greatly. For instance, for the AAD-1 bitumen, values have been published ranging from 4.79 mm2/h (Cheng et al., 2003) to 9.0 x 10-5 mm2/h (Nguyen et al., 1992). Comparing these values with published diffusion coefficients of, rubber, PVC and polyethylene (Abson & Burton, 1979), the lower diffusion value for the AAD-1 asphalt binder (bitumen) seems to be more plausible. Since the mastic matrix in an asphaltic mixture generally consists of asphalt binder as well as sand particles and filler material, a higher diffusion value than for the binder alone can be expected. In the simulations described in this chapter, a diffusion coefficient of 1.0 x 10 3 mm2/h has been utilized for the mastic film.

Because water-induced damage influences the dry response of the material, the effects of the physical processes must be coupled with a three dimensional elasto- visco-plastic constitutive model for mastic response (Scarpas et al., 2005). Mastic in asphaltic mixtures is known to be a material whose behaviour, depending on strain rate and/or temperature, exhibits response characteristics varying anywhere between the elasto-plastic and the visco-elastic limits. Constitutive models for such types of materials can be developed by combining the features of purely elasto-plastic and purely visco-elastic materials to create a more general category of constitutive models termed elasto-visco-plastic. This is the approach that was adopted by the Delft researchers, but a detailed description of the formulation is be­yond the scope of this book but is available (Kringos & Scarpas, 2006) for interested readers.

5.5.3.1 Aggregate-Mastic Bond Strength as a Function of Water Content

Clearly, such modelling requires a knowledge of the aggregate-mastic bond as a function of water content. The direct tension test provides the means of assess­ing bond strength and to define a relationship between the mastic-aggregate bond strength and the conditioning time in a water-bath test (see Fig. 5.10). When the pur­pose of the test is to acquire a comparison between particular mastic-stone combina­tions, results of the pull-off test may directly provide useful information, provided that similar geometries and water conditioning are used. To determine, however, the fundamental relationship of the influence of water on the bond strength, the quantity of water at the interface is of paramount importance. Since this type of information cannot be determined from the test, an additional procedure was devel­oped (Copeland et al., 2007) to relate the bond strength to the quantity of water in the bond. By simulating the test specimens with the RoAM software (see above), the relationship between the quantity of water at the mastic-aggregate interface and the soaking time can be found, Fig. 5.10 (left). By combining the results of finite element simulations and the pull-off test, a relationship between the bond strength and the water content is determined, Fig. 5.10 (right).

From test results, using for the mastic a SHRP core bitumen AAD3 (PG4 58-28) with a diabase filler material passing the < 75 ^ m sieve (#200) and for the aggregate substrate a diabase rock, the tensile bond strength, Smd, as a function of volumetric water content, в, was determined as:

Smd = e(ln S0-aVe) (5.2)

Fig. 5.11 Aggregate-mastic bond strength as a function of water (adapted from Copeland et al., 2007) [9] [10]

where ln So = 0.30, So being the dry adhesive strength, and a = 3.76. This can be reformulated into a water-induced damage parameter, d:

d (0) = 1 – e-a^ (5.3)

Figure 5.11 shows the result of the experimental-computational procedure and the regression curve, as expressed by Eq. 5.3.

The existence of a water flow through an asphaltic mixture may cause desorption of parts of the mastic films which are in direct contact with the water flow, Fig. 5.8(a) carrying away elements of the bitumen[8] by advection (see Chapter 6 for a fuller definition of advection). Exposure of an asphaltic mixture to stationary water (i. e. no

water flow) would, therefore, show no advective transport damage. Since practice has shown that exposure of asphaltic mixtures to stationary water does, in time, cause ravelling of the mixture, this process cannot be the only phenomenon causing water damage. In asphaltic mixtures, diffusion of water through the mastic films surrounding the aggregates is a molecular process that may eventually lead to water reaching the interface area between the mastic and the aggregates. Depending on the bond characteristics, the water can then cause an adhesive failure of the mastic – aggregate interface, Fig. 5.8b).

As time increases, the diffusion of water through the mastic weakens the cohesive strength and the stiffness of the mastic and may actually aggravate the desorption. These are modelled using the approaches set out in Chapter 11. More details about these mathematical formulations can be found in Kringos & Scarpas (2004; 2005a & 2006), Kringos (2007) and Kringos et al. (2007). Figure 5.9 shows an example of a mastic desorption simulation with RoAM, in which a coated aggregate is exposed to a fast water flow field.

One of the important realizations is that the problem cannot be solved by mechanical considerations alone. Clearly, water has an effect on the material characteristics of the asphaltic components and their bond, even without mechanical loading. There­fore, both physical and mechanical water damage-inducing processes are included in the model. Another realization is that, in order to acquire a fundamental insight into the processes which cause water damage, the asphaltic mixture needs to be considered at a micro-scale. This implies that the experimental characterization and the computational simulations of the water damage-inducing processes must be dealt with at mixture component level; i. e. the aggregates, the mastic, the bond between the aggregates and the mastic and the (macro) pore space. Each of these

p – Moisture diffusion Ш -»Advective transport

Weakening of mastic Weakening of aggregate-mastic bond

-Mumpirg action – Mechanical damage

Fig. 5.7 Separation of water damage into physical and mechanical processes (Kringos, 2007) contributes to the mechanical performance of the mixture as well as to its moisture susceptibility.

The physical processes that have been identified as important contributors to wa­ter damage are (c. f. Chapter 6, Section 6.3.1):

• the molecular diffusion of water through the mixture components and

• the advective transport, i. e. ‘washing away’, of the mastic due to the moving water flow through the connected macro-pores.

A mechanical process that is identified as a contributor to water damage is the oc­currence of intense water pressure fields inside the mixture caused by traffic loads and known as the ‘pumping action’. In the model, these physical material degrada­tion processes interact with a model for mechanical damage to produce the overall water-mechanical damage in the mixture – see Fig. 5.7.

5.5.1 Introduction: The Problem of Water for Road Surfacings

Practice has shown that asphaltic wearing surfaces which are exposed to water generally start losing aggregates prematurely through a damage phenomenon that has become known as asphaltic ‘stripping’ or ‘ravelling’. Stripping is generally at­tributed to water infiltration into the asphaltic mixture, causing a weakening of the mastic, and a weakening of the aggregate-mastic bond. Due to the continuing action of water and traffic loading, progressive dislodgement of the aggregates can occur. This initial stripping rapidly progresses into a more severe ravelling of the wearing surface, and ultimately leads to pothole forming, Fig. 5.4.

Sometimes, open-graded mixtures are deliberately designed and laid. As de­scribed in Section 5.7, this is to help drain pavement surface water. This tends to

allow water to reside more or less permanently within the mixture, contributing to the development of water-induced damage.

An additional challenge in the pavement industry is that there is often a big dif­ference between the asphaltic mixture composition and the material characteristics, which are determined in the laboratory, and the asphaltic mixture which is actually constructed on the road. For instance, with regards to water-induced damage, it is not uncommon for the asphaltic mixture components to be exposed to water, even before construction, Fig. 5.5. Since most aggregates and binders do absorb moisture, when exposed to a wet environment, a binder with a significantly reduced stiffness and an initially damaged mixture would end up on the pavement – see Fig. 5.6.

The only real solution to date, for keeping the asphaltic wearing surfaces at an acceptable performance and safety standard, is frequent closure of the major high­ways for repair and maintenance, implying high costs and frequent road congestion.

Fig. 5.6 (a) water absorption in three SHRP binders (b) reduction of binder stiffness, G*, due to water infiltration (Huber, 2005). Reproduced by permission of N. Kringos

For this reason, it is greatly desired to shift the solution from a repair measure to a preventive measure. This is currently impossible as mixture designers have no prior knowledge of the engineering properties of the mixture at the time of purchase of the bulk materials. Common practice for evaluation of the moisture sensitivity of any particular asphaltic mixture is to perform a set of mechanical tests on dry and moisture-conditioned specimens, giving ‘moisture sensitivity ratios’ for the engi­neering mixture properties. Unfortunately, such ratios can only be used to compare case-specific mixtures under a set of pre-determined conditions, but give no insight into the actual water damage phenomena, nor lead to any fundamental remedies.

For this reason, in recent years, these phenomenological studies are giving way to more fundamental studies in which both experimental and analytical investiga­tions on water-induced damage in asphaltic mixtures are combined. Researchers at Delft University of Technology in the Netherlands have focussed on developing a computational tool which allows a study of the interaction between physical and mechanical water damage inducing processes. The tool developed is named RoAM (Kringos & Scarpas, 2004; Kringos, 2007) and operates as a sub-system of the finite element system developed at TU Delft, CAPA-3D (Scarpas, 2005).

Because cracks play such an important part in allowing water to enter a pavement through the surfacing, laboratory assessments of the permeability of intact asphaltic mixtures are not overly useful. Therefore, a range of techniques have been developed to assess permeability by infiltrating water into the pavement surface from a device which acts over a limited area of the surface (e. g. Ridgeway, 1976; Cooley, 1999; Fwa et al., 2001; Taylor, 2004; Mallick & Daniel, 2006). Used randomly on the surface, the infiltration observed will be likely to relate to the mean value – be­ing a combination of water entering through intact material and via degradation cracks induced by compaction, by the environment or by traffic. Alternatively, the devices may also be used over specific cracks or joints to assess the water that can enter through that crack or joint (e. g. Mallick & Daniel, 2006). Essentially, two approaches have been adopted – to keep the surface of a specific piece of pavement wet and monitor what water supply rate is required to do this, or to provide a falling head arrangement and note the rate of head drop.

Some of the earliest work on infiltration, in this case through specific joints, involved the fixing of a bottomless wooden box, sealed with clay around its edges, onto an area of pavement containing a measured length of crack crossing the box. Sufficient water is added to the box to maintain a thin layer over the enclosed pave­ment, the rate at which water is added to maintain this condition can be monitored and the quantity of water infiltrating the pavement structure per unit length of crack calculated. The mean infiltration rate generated by Ridgeway (1976) using this ap­proach was approximately 100 cm3/h/cm of crack. Site crack lengths and infiltration rates are given in Table 5.3.

An overall infiltration rate for a large area of pavement can be deduced from that measured as infiltrating through particular cracks. To achieve this, Baldwin, et al. (1997) suggested that maintenance intervention occurs when 10% of the sur­face is cracked which, they said, was equivalent to 0.002 cm of crack/cm2 of pave­ment surface This represents a worst-case value for infiltration through cracks if

they exist in the same magnitude across the whole of a pavement’s surface. Thus the infiltration over such a well cracked pavement would be given by:

IRmax = 0.02 X ic (5.1)

where IRmax is the maximum anticipated infiltration in units of litres/hour/m2 of pavement area and ic is the infiltration measured through one crack in units of

Inner Frame

Tap to which piping is attached

Outflow

Fig. 5.3 Infiltrometer used by Taylor (2004). Reproduced with permission of J. Taylor

cm3/hour/cm of crack. It seems likely that direct use of this equation will over­estimate a pavement’s potential to accept water. This is because water flowing through one crack may spread laterally below the surface whereas water flowing through one of many adjacent cracks at the same time will be constrained by water flowing through its neighbouring cracks.

To overcome this potential over-reading, a double ring infiltrometer can be used (e. g. Fig. 5.3). The level of water in both a central area and an annular ring are main­tained at the same level but only ingress from the central area is monitored as this should flow only vertically through the pavement because of the ‘confinement’ of­fered by the ingress from the outer ring. Data obtained in this way by Taylor (2004) is shown in Table 5.4. Eq. 5.1 has been used to bring the data in the last column to the same units as for the data recorded directly in the penultimate column.

The mean infiltration ability of the pavements studied by Taylor (2004) and of the cracks studied by Ridgeway (1976) are similar if Eq. 5.1 is accepted -3.33 and 2 l/h/m2 respectively, giving confidence of their representativeness.

Another approach, as adopted at the US National Center for Asphalt Technol­ogy (NCAT) (Cooley, 1999) employs a device comprising a series of cylindrical standpipes of reducing cross section stacked on top of each other as a form of a

1 Strategic Highway Research Program (US)

falling head permeameter. The lowest cylinder is sealed to the pavement surface using a metal plate over a rubber disc. When the pavement is most permeable only the lowest standpipe is used while when it is least permeable the tallest, narrowest permeameter is employed. Unlike the devices described above, relatively imperme­able pavements will, therefore, be subjected to unrealistically high surface water pressures. Whereas even a heavy rainstorm will only impose a few millimetres of water on the surface, the NCAT device can deliver as much as 0.5 m of head. With this device permeability values were obtained as shown in Table 5.5. The comparative laboratory data in the Table were obtained from cores taken from the same pavements and tested using a permeameter as described in Section 5.4.1. The much larger permeability values than given in Table 5.2 is evident, illustrating the important, but often overlooked, need to achieve adequate in-situ densities.

For such falling head tests, the coefficient of permeability, K (in m/s), may be estimated using Eq. 3.10. Other procedures for measuring the permeability of porous asphalt have been introduced by Fwa and his co-workers for both laboratory and in-situ evaluations (Tan et al., 1999; Fwa et al., 2002).

In distressed pavements a large proportion of ingress may be through cracks, even if the intact material is relatively impermeable. It has been suggested that there are four factors which influence infiltration rates in cracked asphaltic pavements (Ridgeway, 1976):

• the water-carrying capacity of the crack or joint,

• the amount of cracking present,

• the area that drains to each crack, and

• the intensity and duration of the rainfall.

The first of these is of particular concern in this chapter and is addressed in Sec­tion 5.4.2.

5.2 Measuring Permeability

5.4.1 Laboratory Permeability Determination

Both constant head and falling head laboratory methods are available to determine the permeability of asphaltic cores, often with sides sealed using a membrane and/or a confining pressure to prevent edge-leakage (Cooley, 1999). There is some evidence (Maupin, 2000) that the falling-head device is the better device for testing both cores and moulded cylindrical specimens. There are some standardized test procedures of which the recent standard published by FDOT (2006), is an example. A schematic of their laboratory permeameter with flexible walls is given in Fig. 5.2. There is also a European standard available (CEN, 2004).